diff --git a/.github/ISSUE_TEMPLATE/bug-report.yml b/.github/ISSUE_TEMPLATE/bug-report.yml new file mode 100644 index 0000000000..a4d5c1fb91 --- /dev/null +++ b/.github/ISSUE_TEMPLATE/bug-report.yml @@ -0,0 +1,52 @@ +name: Bug Report +description: Create a bug report. +labels: ["bug"] +body: +- type: textarea + attributes: + label: Describe the Bug + description: A clear and concise description of what the bug is. You may post screenshots or videos of the bug here. + validations: + required: true + +- type: textarea + attributes: + label: How To Reproduce + description: Steps to reproduce the behavior. You can also post a video of it here. + validations: + required: true + +- type: textarea + attributes: + label: Expected Behavior + description: A clear and concise description of the expected behavior. + placeholder: When I do X, Y should happen. + validations: + required: true + +- type: input + attributes: + label: Operating System + description: Windows version, Linux distribution, etc. + validations: + required: true + +- type: dropdown + attributes: + label: OpenGOAL Version + options: + - Release + - Compiled from source + validations: + required: true + +- type: dropdown + attributes: + label: Game Version + options: + - NTSC 1.0 (black label) + - NTSC Greatest Hits version (red label) + - PAL + - JP + validations: + required: true \ No newline at end of file diff --git a/.github/ISSUE_TEMPLATE/enhancement-request.yml b/.github/ISSUE_TEMPLATE/enhancement-request.yml new file mode 100644 index 0000000000..83afc363f8 --- /dev/null +++ b/.github/ISSUE_TEMPLATE/enhancement-request.yml @@ -0,0 +1,24 @@ +name: Enhancement Request +description: Suggest an improvement for an existing feature. +labels: ["enhancement"] +body: +- type: textarea + attributes: + label: What feature is your idea related to? + description: A clear and concise description of what the problem is. + validations: + required: true + +- type: textarea + attributes: + label: Describe the solution you'd like. + description: A clear and concise description of what you want to happen. + validations: + required: true + +- type: textarea + attributes: + label: Additional context + description: Any other context or screenshots about the enhancement request here. + validations: + required: false \ No newline at end of file diff --git a/.github/ISSUE_TEMPLATE/feature-request.yml b/.github/ISSUE_TEMPLATE/feature-request.yml new file mode 100644 index 0000000000..ddb59ed0cf --- /dev/null +++ b/.github/ISSUE_TEMPLATE/feature-request.yml @@ -0,0 +1,24 @@ +name: Feature Request +description: Suggest an idea for the game +labels: ["game"] +body: +- type: textarea + attributes: + label: Is your feature request related to a problem? + description: A clear and concise description of what the problem is. + validations: + required: true + +- type: textarea + attributes: + label: Describe the solution you'd like. + description: A clear and concise description of what you want to happen. + validations: + required: true + +- type: textarea + attributes: + label: Additional context + description: Add any other context or screenshots about the feature request here. + validations: + required: false \ No newline at end of file diff --git a/.github/scripts/releases/extract_build_linux.sh b/.github/scripts/releases/extract_build_linux.sh index 8fca5cde22..8bd4be176d 100755 --- a/.github/scripts/releases/extract_build_linux.sh +++ b/.github/scripts/releases/extract_build_linux.sh @@ -28,3 +28,4 @@ cp -r $SOURCE/decompiler/config $DEST/data/decompiler/ cp -r $SOURCE/goal_src $DEST/data cp -r $SOURCE/game/assets $DEST/data/game/ cp -r $SOURCE/game/graphics/opengl_renderer/shaders $DEST/data/game/graphics/opengl_renderer +cp -r $SOURCE/custom_levels $DEST/data diff --git a/.github/scripts/releases/extract_build_windows.sh b/.github/scripts/releases/extract_build_windows.sh index 8982346d80..7829505725 100755 --- a/.github/scripts/releases/extract_build_windows.sh +++ b/.github/scripts/releases/extract_build_windows.sh @@ -24,3 +24,4 @@ cp -r $SOURCE/decompiler/config $DEST/data/decompiler/ cp -r $SOURCE/goal_src $DEST/data cp -r $SOURCE/game/assets $DEST/data/game/ cp -r $SOURCE/game/graphics/opengl_renderer/shaders $DEST/data/game/graphics/opengl_renderer +cp -r $SOURCE/custom_levels $DEST/data \ No newline at end of file diff --git a/CMakeLists.txt b/CMakeLists.txt index e8c2e3c65e..a8a81a781a 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -162,6 +162,10 @@ add_subdirectory(third-party/lzokay EXCLUDE_FROM_ALL) # build format library add_subdirectory(third-party/fmt EXCLUDE_FROM_ALL) +add_subdirectory(third-party/stb_image EXCLUDE_FROM_ALL) +add_subdirectory(third-party/tiny_gltf EXCLUDE_FROM_ALL) + + # discord rich presence include_directories(third-party/discord-rpc/include) add_subdirectory(third-party/discord-rpc EXCLUDE_FROM_ALL) diff --git a/FAQ.md b/FAQ.md index 40f4772a0c..29fc2271dc 100644 --- a/FAQ.md +++ b/FAQ.md @@ -37,7 +37,7 @@ The original game had a few issues of varying severity - from progress softlocks In the future yes! We intend to make it as easy as we can to make your own code or import graphics for the game. ### What is this menu? Why am I not dying sometimes? -This might be the result of the game beeing booted in debugging mode. The project is still in development and we rely on the game's built-in debug features (along with some we made ourselves) for testing and development. The final product will have this debug mode disabled. In the meantime, have fun with all the options! Keep in mind some don't work very well. +This might be the result of the game being booted in debugging mode. The project is still in development and we rely on the game's built-in debug features (along with some we made ourselves) for testing and development. The final product will have this debug mode disabled. In the meantime, have fun with all the options! Keep in mind some don't work very well. ### What are the minimum requirements? These are not totally set in stone as the project is not complete. These are the minimum architecture requirements for the project: @@ -49,7 +49,7 @@ These are not totally set in stone as the project is not complete. These are the Probably yes, in the future! ### I'd like to contribute, but don't know where to start! -Great to hear! Any help is good help. If you haven't already, please read the project readme. It has information on what is done, what isn't done, what is being worked on, the project layout and how it all works. Then just join our development discord server and we can help you further. +Great to hear! Any help is good help. If you haven't already, please read the project readme. It has information on what is done, what isn't done, what is being worked on, the project layout and how it all works. Then just join our development Discord server and we can help you further. ### Will you also decompile other games that aren't Jak? No. diff --git a/README.md b/README.md index 00fa072e30..259f103ae0 100644 --- a/README.md +++ b/README.md @@ -46,7 +46,7 @@ This project is to port Jak 1 (NTSC, "black label" version) to PC. Over 98% of t - create tools to repack game assets into a format that our port uses. Our objectives are: -- make the port a "native application" on x86-64, with high performance. It shouldn't emulated, interpreted, or transpiled. +- make the port a "native application" on x86-64, with high performance. It shouldn't be emulated, interpreted, or transpiled. - Our GOAL compiler's performance should be around the same as unoptimized C. - try to match things from the original game and development as possible. For example, the original GOAL compiler supported live modification of code while the game is running, so we do the same, even though it's not required for just porting the game. - support modifications. It should be possible to make edits to the code without everything else breaking. @@ -85,7 +85,7 @@ We don't save any assets from the game - you must bring your own copy of the gam Install packages and init repository: ```sh -sudo apt install gcc make cmake build-essential g++ nasm clang-format libxrandr-dev libxinerama-dev libxcursor-dev libxi-dev python +sudo apt install gcc make cmake build-essential g++ nasm clang-format libxrandr-dev libxinerama-dev libxcursor-dev libpulse-dev libxi-dev python sudo sh -c "$(curl --location https://taskfile.dev/install.sh)" -- -d -b /usr/local/bin ``` @@ -118,10 +118,12 @@ cmake -DCMAKE_SHARED_LINKER_FLAGS="-fuse-ld=lld" -DCMAKE_EXE_LINKER_FLAGS="-fuse Install packages and init repository: ```sh -sudo pacman -S gcc make cmake base-devel g++ nasm python -yay -S taskfile-git +sudo pacman -S cmake libpulse base-devel nasm python +yay -S go-task ``` +For Arch only, replace `task` with `go-task` in the rest of the instructions. + Compile: ```sh @@ -134,6 +136,28 @@ Run tests: ./test.sh ``` +### Fedora + +Install packages and init repository: + +```sh +sudo dnf install cmake lld clang nasm libX11-devel libXrandr-devel libXinerama-devel libXcursor-devel libXi-devel pulseaudio-libs-devel +sudo sh -c "$(curl --location https://taskfile.dev/install.sh)" -- -d -b /usr/local/bin +``` + +Compile with `clang`: + +```sh +cmake -DCMAKE_SHARED_LINKER_FLAGS="-fuse-ld=lld" -DCMAKE_EXE_LINKER_FLAGS="-fuse-ld=lld" -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ -B build +cmake --build build -j$(nproc) +``` + +Run tests: + +```sh +./test.sh +``` + ## Getting Started - Windows ### Required Software @@ -293,7 +317,7 @@ The second component to the project is the decompiler. You must have a copy of t Then run `decomp.sh` (Linux) or `decomp-jak1.bat` (Windows) to run the decompiler. The decompiler will extract assets to the `assets` folder. These assets will be used by the compiler when building the port, and you may want to turn asset extraction off after running it once. The decompiler will output code and other data intended to be inspected by humans in the `decompiler_out` folder. Stuff in this folder will not be used by the compiler. -The third is the game source code, written in OpenGOAL. This is located in `goal_src`. All GOAL and GOOS code should be in this folder. Right now most of this is placeholders or incomplete, but you can take a look at `kernel/gcommon.gc` or `goal-lib.gc` to see some in-progress source code. +The third is the game source code, written in OpenGOAL. This is located in `goal_src`. All GOAL and GOOS code should be in this folder. The final component is the "runtime", located in `game`. This is the part of the game that's written in C++. In the port, that includes: - The "C Kernel", which contains the GOAL linker and some low-level GOAL language features. GOAL has a completely custom dynamically linked object file format so in order to load the first GOAL code, you need a linker written in C++. Some low-level functions for memory allocation, communicating with the I/O Processor, symbol table, strings, and the type system are also implemented in C, as these are required for the linker. It also listens for incoming messages from the compiler and passes them to the running game. This also initializes the game, by initializing the PS2 hardware, allocating the GOAL heaps, loading the GOAL kernel off of the DVD, and executing the kernel dispatcher function. This is in the `game/kernel` folder. This should be as close as possible to the game, and all differences should be noted with a comment. @@ -357,7 +381,7 @@ The final component is the "runtime", located in `game`. This is the part of the - `listener`: The OpenGOAL listener, which connects the compiler to a running GOAL program for the interactive REPL. - `make`: The OpenGOAL build system, builds both code and data files. - `regalloc`: Register allocator. -- `iso_data`: +- `iso_data`: Location of the user-provided DVD contents of the game that the decompiler extracts game assets and code from. - `out`: Outputs from the build process. Only the `iso` subfolder should contain assets used by the game. - `iso`: Final outputs that are used by the game. - `obj`: Object files generated by the compiler. @@ -366,7 +390,8 @@ The final component is the "runtime", located in `game`. This is the part of the - `test`: Unit tests (run on GitHub Actions). - `third-party`: Third party libraries. - CMake Code Coverage. For code coverage statistics on GitHub builds. - - `fmt`. String formatting library. + - `discord-rpc`: Discord Rich Presence library. + - `fmt`: String formatting library. - `googletest`: Test framework. - `inja`: templating library used for generating test code for compiler tests. - `lzokay`: decompression code for Jak 2 and later DGOs. diff --git a/common/CMakeLists.txt b/common/CMakeLists.txt index 45e753162f..6c9656c9fd 100644 --- a/common/CMakeLists.txt +++ b/common/CMakeLists.txt @@ -4,6 +4,7 @@ add_library(common cross_sockets/XSocket.cpp cross_sockets/XSocketServer.cpp cross_sockets/XSocketClient.cpp + custom_data/pack_helpers.cpp custom_data/TFrag3Data.cpp deserialization/subtitles/subtitles.cpp dma/dma.cpp @@ -47,8 +48,7 @@ add_library(common util/os.cpp util/print_float.cpp util/FontUtils.cpp - util/FrameLimiter.cpp - util/image_loading.cpp) + util/FrameLimiter.cpp) target_link_libraries(common fmt lzokay replxx libzstd_static) diff --git a/common/custom_data/TFrag3Data.cpp b/common/custom_data/TFrag3Data.cpp index 70c527eb9a..879ee585da 100644 --- a/common/custom_data/TFrag3Data.cpp +++ b/common/custom_data/TFrag3Data.cpp @@ -1,3 +1,5 @@ +#include +#include #include "Tfrag3Data.h" #include "common/util/Assert.h" @@ -21,6 +23,7 @@ void StripDraw::serialize(Serializer& ser) { ser.from_ptr(&mode); ser.from_ptr(&tree_tex_id); ser.from_pod_vector(&runs); + ser.from_pod_vector(&plain_indices); ser.from_pod_vector(&vis_groups); ser.from_ptr(&num_triangles); } @@ -64,6 +67,7 @@ void TfragTree::serialize(Serializer& ser) { ser.from_pod_vector(&packed_vertices.cluster_origins); ser.from_pod_vector(&colors); bvh.serialize(ser); + ser.from_ptr(&use_strips); } void TieTree::unpack() { @@ -78,7 +82,7 @@ void TieTree::unpack() { vtx.x = proto_vtx.x; vtx.y = proto_vtx.y; vtx.z = proto_vtx.z; - vtx.q = 1.f; + vtx.q_unused = 1.f; vtx.s = proto_vtx.s; vtx.t = proto_vtx.t; i++; @@ -93,7 +97,7 @@ void TieTree::unpack() { vtx.x = temp.x(); vtx.y = temp.y(); vtx.z = temp.z(); - vtx.q = 1.f; + vtx.q_unused = 1.f; vtx.s = proto_vtx.s; vtx.t = proto_vtx.t; i++; @@ -103,6 +107,7 @@ void TieTree::unpack() { for (auto& draw : static_draws) { draw.unpacked.idx_of_first_idx_in_full_buffer = unpacked.indices.size(); + ASSERT(draw.plain_indices.empty()); for (auto& run : draw.runs) { for (u32 ri = 0; ri < run.length; ri++) { unpacked.indices.push_back(run.vertex0 + ri); @@ -152,7 +157,7 @@ void TfragTree::unpack() { o.z = cz + in.zoff * rescale; o.s = in.s / (1024.f); o.t = in.t / (1024.f); - o.q = 1.f; + o.q_unused = 1.f; o.color_index = in.color_index; } @@ -162,8 +167,12 @@ void TfragTree::unpack() { for (u32 ri = 0; ri < run.length; ri++) { unpacked.indices.push_back(run.vertex0 + ri); } - unpacked.indices.push_back(UINT32_MAX); + if (use_strips) { + unpacked.indices.push_back(UINT32_MAX); + } } + unpacked.indices.insert(unpacked.indices.end(), draw.plain_indices.begin(), + draw.plain_indices.end()); } } @@ -288,7 +297,8 @@ void MercModelGroup::serialize(Serializer& ser) { void Level::serialize(Serializer& ser) { ser.from_ptr(&version); if (ser.is_loading() && version != TFRAG3_VERSION) { - ASSERT_MSG(false, fmt::format("version mismatch when loading tfrag3 data. Got {}, expected {}", + ASSERT_MSG(false, fmt::format("version mismatch when loading tfrag3 data. Got {}, expected {}, " + "did you forget to re-decompile?", version, TFRAG3_VERSION)); } @@ -359,6 +369,7 @@ std::array Level::get_memory_usage() c for (const auto& tfrag_tree : tfrag_tree_geoms) { for (const auto& draw : tfrag_tree.draws) { result[TFRAG_INDEX] += draw.runs.size() * sizeof(StripDraw::VertexRun); + result[TFRAG_INDEX] += draw.plain_indices.size() * sizeof(u32); result[TFRAG_VIS] += draw.vis_groups.size() * sizeof(StripDraw::VisGroup); } result[TFRAG_VERTS] += @@ -417,4 +428,54 @@ std::array Level::get_memory_usage() c return result; } +void print_memory_usage(const tfrag3::Level& lev, int uncompressed_data_size) { + int total_accounted = 0; + auto memory_use_by_category = lev.get_memory_usage(); + + std::vector> known_categories = { + {"texture", memory_use_by_category[tfrag3::MemoryUsageCategory::TEXTURE]}, + {"tie-deinst-vis", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_DEINST_VIS]}, + {"tie-deinst-idx", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_DEINST_INDEX]}, + {"tie-inst-vis", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_INST_VIS]}, + {"tie-inst-idx", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_INST_INDEX]}, + {"tie-bvh", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_BVH]}, + {"tie-verts", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_VERTS]}, + {"tie-colors", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_TIME_OF_DAY]}, + {"tie-wind-inst-info", + memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_WIND_INSTANCE_INFO]}, + {"tie-cidx", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_CIDX]}, + {"tie-mats", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_MATRICES]}, + {"tie-grps", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_GRPS]}, + {"tfrag-vis", memory_use_by_category[tfrag3::MemoryUsageCategory::TFRAG_VIS]}, + {"tfrag-idx", memory_use_by_category[tfrag3::MemoryUsageCategory::TFRAG_INDEX]}, + {"tfrag-vert", memory_use_by_category[tfrag3::MemoryUsageCategory::TFRAG_VERTS]}, + {"tfrag-colors", memory_use_by_category[tfrag3::MemoryUsageCategory::TFRAG_TIME_OF_DAY]}, + {"tfrag-cluster", memory_use_by_category[tfrag3::MemoryUsageCategory::TFRAG_CLUSTER]}, + {"tfrag-bvh", memory_use_by_category[tfrag3::MemoryUsageCategory::TFRAG_BVH]}, + {"shrub-colors", memory_use_by_category[tfrag3::MemoryUsageCategory::SHRUB_TIME_OF_DAY]}, + {"shrub-vert", memory_use_by_category[tfrag3::MemoryUsageCategory::SHRUB_VERT]}, + {"shrub-ind", memory_use_by_category[tfrag3::MemoryUsageCategory::SHRUB_IND]}, + {"collision", memory_use_by_category[tfrag3::MemoryUsageCategory::COLLISION]}, + {"merc-vert", memory_use_by_category[tfrag3::MemoryUsageCategory::MERC_VERT]}, + {"merc-idx", memory_use_by_category[tfrag3::MemoryUsageCategory::MERC_INDEX]}}; + for (auto& known : known_categories) { + total_accounted += known.second; + } + + known_categories.push_back({"unknown", uncompressed_data_size - total_accounted}); + + std::sort(known_categories.begin(), known_categories.end(), + [](const auto& a, const auto& b) { return a.second > b.second; }); + + for (const auto& x : known_categories) { + fmt::print("{:30s} : {:6d} kB {:3.1f}%\n", x.first, x.second / 1024, + 100.f * (float)x.second / uncompressed_data_size); + } +} + +std::size_t PreloadedVertex::hash::operator()(const PreloadedVertex& v) const { + return std::hash()(v.x) ^ std::hash()(v.y) ^ std::hash()(v.z) ^ + std::hash()(v.s) ^ std::hash()(v.t) ^ std::hash()(v.color_index); +} + } // namespace tfrag3 diff --git a/common/custom_data/Tfrag3Data.h b/common/custom_data/Tfrag3Data.h index 2354eeac2c..c730190bb4 100644 --- a/common/custom_data/Tfrag3Data.h +++ b/common/custom_data/Tfrag3Data.h @@ -53,17 +53,26 @@ enum MemoryUsageCategory { NUM_CATEGORIES }; -constexpr int TFRAG3_VERSION = 19; +constexpr int TFRAG3_VERSION = 20; // These vertices should be uploaded to the GPU at load time and don't change struct PreloadedVertex { // the vertex position float x, y, z; // texture coordinates - float s, t, q; + float s, t, q_unused; // color table index u16 color_index; u16 pad[3]; + + struct hash { + std::size_t operator()(const PreloadedVertex& x) const; + }; + + bool operator==(const PreloadedVertex& other) const { + return x == other.x && y == other.y && z == other.z && s == other.s && t == other.t && + color_index == other.color_index; + } }; static_assert(sizeof(PreloadedVertex) == 32, "PreloadedVertex size"); @@ -144,12 +153,14 @@ struct StripDraw { u32 idx_of_first_idx_in_full_buffer = 0; } unpacked; + // indices can be specified as lists of runs and plain indices. + // the runs are still drawn with indexed opengl calls, it just uses less space in the file. struct VertexRun { u32 vertex0; u16 length; }; - std::vector runs; + std::vector plain_indices; // to do culling, the above vertex stream is grouped. // by following the visgroups and checking the visibility, you can leave out invisible vertices. @@ -260,6 +271,7 @@ struct TfragTree { PackedTfragVertices packed_vertices; std::vector colors; // vertex colors (pre-interpolation) BVH bvh; // the bvh for frustum culling + bool use_strips = true; struct { std::vector vertices; // mesh vertices @@ -397,4 +409,6 @@ struct Level { std::array get_memory_usage() const; }; +void print_memory_usage(const tfrag3::Level& lev, int uncompressed_data_size); + } // namespace tfrag3 diff --git a/common/custom_data/pack_helpers.cpp b/common/custom_data/pack_helpers.cpp new file mode 100644 index 0000000000..0dd17d700b --- /dev/null +++ b/common/custom_data/pack_helpers.cpp @@ -0,0 +1,71 @@ +#include "pack_helpers.h" +#include + +constexpr float kClusterSize = 4096 * 40; // 100 in-game meters +constexpr float kMasterOffset = 12000 * 4096; + +std::pair position_to_cluster_and_offset(float in) { + in += kMasterOffset; + if (in < 0) { + fmt::print("negative: {}\n", in); + } + ASSERT(in >= 0); + int cluster_cell = (in / kClusterSize); + float leftover = in - (cluster_cell * kClusterSize); + u16 offset = (leftover / kClusterSize) * float(UINT16_MAX); + + float recovered = ((float)cluster_cell + ((float)offset / UINT16_MAX)) * kClusterSize; + float diff = std::fabs(recovered - in); + ASSERT(diff < 7); + ASSERT(cluster_cell >= 0); + ASSERT(cluster_cell < UINT16_MAX); + return {cluster_cell, offset}; +} + +void pack_tfrag_vertices(tfrag3::PackedTfragVertices* result, + const std::vector& vertices) { + u32 next_cluster_idx = 0; + std::map clusters; + + for (auto& vtx : vertices) { + auto x = position_to_cluster_and_offset(vtx.x); + auto y = position_to_cluster_and_offset(vtx.y); + auto z = position_to_cluster_and_offset(vtx.z); + u64 cluster_id = 0; + cluster_id |= x.first; + cluster_id |= (y.first << 16); + cluster_id |= (z.first << 32); + + auto cluster_it = clusters.find(cluster_id); + u32 my_cluster_idx = 0; + if (cluster_it == clusters.end()) { + // first in cluster + clusters[cluster_id] = next_cluster_idx; + my_cluster_idx = next_cluster_idx; + next_cluster_idx++; + } else { + my_cluster_idx = cluster_it->second; + } + + tfrag3::PackedTfragVertices::Vertex out_vtx; + out_vtx.xoff = x.second; + out_vtx.yoff = y.second; + out_vtx.zoff = z.second; + out_vtx.cluster_idx = my_cluster_idx; + // TODO check these + out_vtx.s = vtx.s * 1024; + out_vtx.t = vtx.t * 1024; + out_vtx.color_index = vtx.color_index; + result->vertices.push_back(out_vtx); + } + + result->cluster_origins.resize(next_cluster_idx); + for (auto& cluster : clusters) { + auto& res = result->cluster_origins[cluster.second]; + res.x() = (u16)cluster.first; + res.y() = (u16)(cluster.first >> 16); + res.z() = (u16)(cluster.first >> 32); + } + + ASSERT(next_cluster_idx < UINT16_MAX); +} \ No newline at end of file diff --git a/common/custom_data/pack_helpers.h b/common/custom_data/pack_helpers.h new file mode 100644 index 0000000000..c477a22c40 --- /dev/null +++ b/common/custom_data/pack_helpers.h @@ -0,0 +1,6 @@ +#pragma once + +#include "common/custom_data/Tfrag3Data.h" + +void pack_tfrag_vertices(tfrag3::PackedTfragVertices* result, + const std::vector& vertices); \ No newline at end of file diff --git a/common/math/Vector.h b/common/math/Vector.h index bb0ba31a1c..9351b123ea 100644 --- a/common/math/Vector.h +++ b/common/math/Vector.h @@ -89,6 +89,14 @@ class Vector { return result; } + Vector operator+(const T& other) const { + Vector result; + for (int i = 0; i < Size; i++) { + result[i] = m_data[i] + other; + } + return result; + } + Vector& operator+=(const Vector& other) { for (int i = 0; i < Size; i++) { m_data[i] += other[i]; @@ -103,6 +111,13 @@ class Vector { return *this; } + Vector& operator-=(const T& other) { + for (int i = 0; i < Size; i++) { + m_data[i] -= other; + } + return *this; + } + Vector elementwise_multiply(const Vector& other) const { Vector result; for (int i = 0; i < Size; i++) { @@ -180,6 +195,18 @@ class Vector { void normalize(const T& norm = T(1)) { *this = normalized(norm); } + void max_in_place(const Vector& other) { + for (int i = 0; i < Size; i++) { + m_data[i] = std::max(m_data[i], other[i]); + } + } + + void min_in_place(const Vector& other) { + for (int i = 0; i < Size; i++) { + m_data[i] = std::min(m_data[i], other[i]); + } + } + std::string to_string_aligned() const { std::string result = "["; for (auto x : m_data) { @@ -229,6 +256,8 @@ class Vector { } } + void set_zero() { fill(0); } + private: T m_data[Size]; }; @@ -246,8 +275,24 @@ struct Matrix { return result; } - // const T& operator()(int r, int c) const { return m_data[c + r * Cols]; } - // T& operator()(int r, int c) { return m_data[r + c * Rows]; } + static Matrix identity() { + Matrix result; + for (int c = 0; c < Cols; c++) { + for (int r = 0; r < Rows; r++) { + result(r, c) = r == c ? T(1) : T(0); + } + } + return result; + } + + void set_zero() { + for (auto& x : m_data) { + x = 0; + } + } + + T& operator()(int r, int c) { return m_data[r + c * Rows]; } + const T& operator()(int r, int c) const { return m_data[r + c * Rows]; } Vector col(int c) const { Vector result; @@ -274,6 +319,31 @@ struct Matrix { return result; } + template + Matrix operator*(const Matrix& y) const { + Matrix result; + result.set_zero(); + for (int rx = 0; rx < Rows; rx++) { + for (int cx = 0; cx < Cols; cx++) { + for (int yi = 0; yi < OtherCols; yi++) { + result(rx, yi) += operator()(rx, cx) * y(cx, yi); + } + } + } + return result; + } + + Vector operator*(const Vector& y) const { + Vector result; + result.set_zero(); + for (int rx = 0; rx < Rows; rx++) { + for (int cx = 0; cx < Cols; cx++) { + result[rx] += operator()(rx, cx) * y[cx]; + } + } + return result; + } + private: T m_data[Rows * Cols]; }; diff --git a/common/math/geometry.cpp b/common/math/geometry.cpp index e69de29bb2..59aad2c4c0 100644 --- a/common/math/geometry.cpp +++ b/common/math/geometry.cpp @@ -0,0 +1,40 @@ +#include "geometry.h" + +namespace math { + +Vector4f bsphere_of_triangle(const Vector3f* v) { + Vector4f bsphere; + auto& p1 = v[0]; + auto& p2 = v[1]; + auto& p3 = v[2]; + float A = (p1 - p2).length(); + float B = (p2 - p3).length(); + float C = (p3 - p1).length(); + + const Vector3f *a = &p3, *b = &p1, *c = &p2; + if (B < C) + std::swap(B, C), std::swap(b, c); + if (A < B) + std::swap(A, B), std::swap(a, b); + + float r; + math::Vector3f origin; + if ((B * B) + (C * C) <= (A * A)) { + r = A / 2.f; + origin = (*b + *c) / 2.f; + } else { + float cos_a = (B * B + C * C - A * A) / (B * C * 2); + r = A / (sqrt(1 - cos_a * cos_a) * 2.f); + Vector3f alpha = *a - *c, beta = *b - *c; + origin = (beta * alpha.dot(alpha) - alpha * beta.dot(beta)).cross(alpha.cross(beta)) / + (alpha.cross(beta).dot(alpha.cross(beta)) * 2.f) + + *c; + } + bsphere.x() = origin.x(); + bsphere.y() = origin.y(); + bsphere.z() = origin.z(); + bsphere.w() = r; + return bsphere; +} + +} // namespace math \ No newline at end of file diff --git a/common/math/geometry.h b/common/math/geometry.h index 4c71592bf7..a61988f444 100644 --- a/common/math/geometry.h +++ b/common/math/geometry.h @@ -44,4 +44,10 @@ RaySphereResult ray_sphere_intersect(const Vector3& ray_origin, result.u[1] = minus_b - sqrt_val; return result; } + +math::Vector4f bsphere_of_triangle(const Vector3f* vertices); + +inline bool point_in_bsphere(const Vector4f& sphere, const Vector3f& pt) { + return (sphere.xyz() - pt).squared_length() <= (sphere.w() * sphere.w()); +} } // namespace math \ No newline at end of file diff --git a/common/util/image_loading.cpp b/common/util/image_loading.cpp deleted file mode 100644 index 62971a6551..0000000000 --- a/common/util/image_loading.cpp +++ /dev/null @@ -1,13 +0,0 @@ - -// hides warnings -#ifdef __linux__ -#pragma GCC diagnostic push -#pragma GCC diagnostic ignored "-Wcast-qual" -#endif - -#define STB_IMAGE_IMPLEMENTATION -#include "third-party/stb_image.h" - -#ifdef __linux__ -#pragma GCC diagnostic pop -#endif \ No newline at end of file diff --git a/common/util/image_loading.h b/common/util/image_loading.h deleted file mode 100644 index 36fe77bfaf..0000000000 --- a/common/util/image_loading.h +++ /dev/null @@ -1 +0,0 @@ -#include "third-party/stb_image.h" diff --git a/custom_levels/README.md b/custom_levels/README.md new file mode 100644 index 0000000000..0a6a9aff04 --- /dev/null +++ b/custom_levels/README.md @@ -0,0 +1,36 @@ +# Custom Levels +Disclaimer: custom levels are still in development and are missing most features. + + +The first three steps are already done for "test zone", so this can be used as a starting point. + +# 1: File Setup +To create a custom level, copy the layout of `custom_levels/test-zone`. See `test-zone.jsonc` for information on how to name things. The `.gd` file also contains the level name. + +# 2: Modify the engine +Modify `goal_src/engine/level/level-info.gc` to add level info for each custom level. There is level info for `test-zone` at the bottom that can be used as an example. + +# 3: Modify the build system +Modify `goal_src/game.gp` and add a custom level target: +```lisp +(build-custom-level "test-zone") +;; the DGO file +(custom-level-cgo "TESTZONE.DGO" "test-zone/testzone.gd") +``` + +# 4: Export the GLTF file from blender. +For now, all meshes are displayed and treated as ground collision. This causes buggy collision because walls shouldn't use "floor" mode. + +Blender will create a `.glb` file, which must have the name specified in the `.jsonc` file and should be located in `custom_level/your_level` + +# 5: Rebuild the game +Any time the `.glb` file is changed, you must rebuild the game. Launch the compiler (`goalc`) and run `(mi)` to rebuild everything. It's recommended to leave the compiler open - it will remember files that haven't changed and skip rebuilding them. + +# 6: Go to the custom level +Start the game in debug mode `gk`. + +In the compiler window, run `(lt)` to connect to the game. You must run this again every time you restart the game. If this doesn't work, there could be a firewall issue and you must allow goalc/gk to use the network. They don't make any outside connections. + +In the compiler window, run a command like `(bg-custom 'test-zone-vis)` to load and start at a custom level. + + diff --git a/custom_levels/test-zone/test-zone.jsonc b/custom_levels/test-zone/test-zone.jsonc new file mode 100644 index 0000000000..73c66edf00 --- /dev/null +++ b/custom_levels/test-zone/test-zone.jsonc @@ -0,0 +1,15 @@ +{ + // The "in-game" name of the level. Should be lower case, with dashes (GOAL symbol name) + // the name of this file, and the folder this file is in must have the same name. + "long_name": "test-zone", + // The file name, should be upper case and 8 characters or less. + "iso_name": "TESTZONE", + // The nickname, should be exactly 3 characters + "nickname": "TSZ", // 3 char name, all uppercase + + // Background mesh file. + // Must have vertex colors. Use the blender cycles renderer, bake, diffuse, uncheck color, + // and bake to vertex colors. For now, only the first vertex color group is used, so make sure you + // only have 1. + "gltf_file": "custom_levels/test-zone/test-zone2.glb" +} \ No newline at end of file diff --git a/custom_levels/test-zone/test-zone2.glb b/custom_levels/test-zone/test-zone2.glb new file mode 100644 index 0000000000..2df7f4d595 Binary files /dev/null and b/custom_levels/test-zone/test-zone2.glb differ diff --git a/custom_levels/test-zone/testzone.gd b/custom_levels/test-zone/testzone.gd new file mode 100644 index 0000000000..0ecd61c2ae --- /dev/null +++ b/custom_levels/test-zone/testzone.gd @@ -0,0 +1,8 @@ +;; DGO definition file for Awful Village level +;; We use the convention of having a longer DGO name for levels without precomputed visibility. + +;; the actual file name still needs to be 8.3 +("TSZ.DGO" + ("static-screen.o" "static-screen") + ("test-zone.go" "test-zone") + ) \ No newline at end of file diff --git a/decompiler/CMakeLists.txt b/decompiler/CMakeLists.txt index cb0ab3ebbe..82754934f4 100644 --- a/decompiler/CMakeLists.txt +++ b/decompiler/CMakeLists.txt @@ -83,6 +83,7 @@ target_link_libraries(decomp lzokay common fmt + stb_image ) add_executable(decompiler @@ -92,7 +93,8 @@ target_link_libraries(decompiler decomp common lzokay - fmt) + fmt + stb_image) add_executable(extractor @@ -103,4 +105,5 @@ target_link_libraries(extractor common lzokay fmt - compiler) + compiler + stb_image) diff --git a/decompiler/data/TextureDB.cpp b/decompiler/data/TextureDB.cpp index f145fd1e58..45d46d114c 100644 --- a/decompiler/data/TextureDB.cpp +++ b/decompiler/data/TextureDB.cpp @@ -2,7 +2,7 @@ #include "third-party/fmt/core.h" #include "common/util/Assert.h" -#include "third-party/stb_image.h" +#include "third-party/stb_image/stb_image.h" #include namespace decompiler { diff --git a/decompiler/level_extractor/extract_level.cpp b/decompiler/level_extractor/extract_level.cpp index f63931a31c..ec1a12d34a 100644 --- a/decompiler/level_extractor/extract_level.cpp +++ b/decompiler/level_extractor/extract_level.cpp @@ -53,51 +53,6 @@ bool is_valid_bsp(const decompiler::LinkedObjectFile& file) { return true; } -void print_memory_usage(const tfrag3::Level& lev, int uncompressed_data_size) { - int total_accounted = 0; - auto memory_use_by_category = lev.get_memory_usage(); - - std::vector> known_categories = { - {"texture", memory_use_by_category[tfrag3::MemoryUsageCategory::TEXTURE]}, - {"tie-deinst-vis", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_DEINST_VIS]}, - {"tie-deinst-idx", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_DEINST_INDEX]}, - {"tie-inst-vis", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_INST_VIS]}, - {"tie-inst-idx", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_INST_INDEX]}, - {"tie-bvh", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_BVH]}, - {"tie-verts", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_VERTS]}, - {"tie-colors", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_TIME_OF_DAY]}, - {"tie-wind-inst-info", - memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_WIND_INSTANCE_INFO]}, - {"tie-cidx", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_CIDX]}, - {"tie-mats", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_MATRICES]}, - {"tie-grps", memory_use_by_category[tfrag3::MemoryUsageCategory::TIE_GRPS]}, - {"tfrag-vis", memory_use_by_category[tfrag3::MemoryUsageCategory::TFRAG_VIS]}, - {"tfrag-idx", memory_use_by_category[tfrag3::MemoryUsageCategory::TFRAG_INDEX]}, - {"tfrag-vert", memory_use_by_category[tfrag3::MemoryUsageCategory::TFRAG_VERTS]}, - {"tfrag-colors", memory_use_by_category[tfrag3::MemoryUsageCategory::TFRAG_TIME_OF_DAY]}, - {"tfrag-cluster", memory_use_by_category[tfrag3::MemoryUsageCategory::TFRAG_CLUSTER]}, - {"tfrag-bvh", memory_use_by_category[tfrag3::MemoryUsageCategory::TFRAG_BVH]}, - {"shrub-colors", memory_use_by_category[tfrag3::MemoryUsageCategory::SHRUB_TIME_OF_DAY]}, - {"shrub-vert", memory_use_by_category[tfrag3::MemoryUsageCategory::SHRUB_VERT]}, - {"shrub-ind", memory_use_by_category[tfrag3::MemoryUsageCategory::SHRUB_IND]}, - {"collision", memory_use_by_category[tfrag3::MemoryUsageCategory::COLLISION]}, - {"merc-vert", memory_use_by_category[tfrag3::MemoryUsageCategory::MERC_VERT]}, - {"merc-idx", memory_use_by_category[tfrag3::MemoryUsageCategory::MERC_INDEX]}}; - for (auto& known : known_categories) { - total_accounted += known.second; - } - - known_categories.push_back({"unknown", uncompressed_data_size - total_accounted}); - - std::sort(known_categories.begin(), known_categories.end(), - [](const auto& a, const auto& b) { return a.second > b.second; }); - - for (const auto& x : known_categories) { - fmt::print("{:30s} : {:6d} kB {:3.1f}%\n", x.first, x.second / 1024, - 100.f * (float)x.second / uncompressed_data_size); - } -} - void add_all_textures_from_level(tfrag3::Level& lev, const std::string& level_name, const TextureDB& tex_db) { diff --git a/decompiler/level_extractor/extract_tfrag.cpp b/decompiler/level_extractor/extract_tfrag.cpp index 2af87aa9dc..82a3045d8a 100644 --- a/decompiler/level_extractor/extract_tfrag.cpp +++ b/decompiler/level_extractor/extract_tfrag.cpp @@ -5,6 +5,7 @@ #include "common/util/FileUtil.h" #include "common/dma/gs.h" #include "common/util/Assert.h" +#include "common/custom_data/pack_helpers.h" namespace decompiler { namespace { @@ -2056,9 +2057,9 @@ void make_tfrag3_data(std::map>& draws, vtx.z = vert.pre_cam_trans_pos.z(); vtx.s = vert.stq.x(); vtx.t = vert.stq.y(); - vtx.q = vert.stq.z(); + vtx.q_unused = vert.stq.z(); // if this is true, we can remove a divide in the shader - ASSERT(vtx.q == 1.f); + ASSERT(vtx.q_unused == 1.f); vtx.color_index = vert.rgba / 4; // ASSERT((vert.rgba >> 2) < 1024); spider cave has 2048? ASSERT((vert.rgba & 3) == 0); @@ -2137,85 +2138,6 @@ void merge_groups(std::vector& grps) { } // namespace -constexpr float kClusterSize = 4096 * 40; // 100 in-game meters -constexpr float kMasterOffset = 12000 * 4096; - -std::pair position_to_cluster_and_offset(float in) { - in += kMasterOffset; - if (in < 0) { - fmt::print("negative: {}\n", in); - } - ASSERT(in >= 0); - int cluster_cell = (in / kClusterSize); - float leftover = in - (cluster_cell * kClusterSize); - u16 offset = (leftover / kClusterSize) * float(UINT16_MAX); - - float recovered = ((float)cluster_cell + ((float)offset / UINT16_MAX)) * kClusterSize; - float diff = std::fabs(recovered - in); - ASSERT(diff < 7); - ASSERT(cluster_cell >= 0); - ASSERT(cluster_cell < UINT16_MAX); - return {cluster_cell, offset}; -} - -void pack_vertices(tfrag3::PackedTfragVertices* result, - const std::vector& vertices) { - u32 next_cluster_idx = 0; - std::map clusters; - - for (auto& vtx : vertices) { - auto x = position_to_cluster_and_offset(vtx.x); - auto y = position_to_cluster_and_offset(vtx.y); - auto z = position_to_cluster_and_offset(vtx.z); - u64 cluster_id = 0; - cluster_id |= x.first; - cluster_id |= (y.first << 16); - cluster_id |= (z.first << 32); - - auto cluster_it = clusters.find(cluster_id); - u32 my_cluster_idx = 0; - if (cluster_it == clusters.end()) { - // first in cluster - clusters[cluster_id] = next_cluster_idx; - my_cluster_idx = next_cluster_idx; - next_cluster_idx++; - } else { - my_cluster_idx = cluster_it->second; - } - - tfrag3::PackedTfragVertices::Vertex out_vtx; - out_vtx.xoff = x.second; - out_vtx.yoff = y.second; - out_vtx.zoff = z.second; - out_vtx.cluster_idx = my_cluster_idx; - // TODO check these - out_vtx.s = vtx.s * 1024; - out_vtx.t = vtx.t * 1024; - out_vtx.color_index = vtx.color_index; - result->vertices.push_back(out_vtx); - } - - result->cluster_origins.resize(next_cluster_idx); - for (auto& cluster : clusters) { - auto& res = result->cluster_origins[cluster.second]; - res.x() = (u16)cluster.first; - res.y() = (u16)(cluster.first >> 16); - res.z() = (u16)(cluster.first >> 32); - } - - /* - std::unordered_set - a; - for (auto& v : result->vertices) { - a.insert(v); - } - fmt::print("SIZE: {} vs {} {}\n", a.size(), result->vertices.size(), - (float)a.size() / result->vertices.size()); - */ - - ASSERT(next_cluster_idx < UINT16_MAX); -} - void extract_tfrag(const level_tools::DrawableTreeTfrag* tree, const std::string& debug_name, const std::vector& map, @@ -2282,7 +2204,7 @@ void extract_tfrag(const level_tools::DrawableTreeTfrag* tree, std::vector vertices; emulate_tfrags(geom, as_tfrag_array->tfragments, debug_name, map, out, this_tree, vertices, tex_db, expected_missing_textures, dump_level); - pack_vertices(&this_tree.packed_vertices, vertices); + pack_tfrag_vertices(&this_tree.packed_vertices, vertices); extract_time_of_day(tree, this_tree); for (auto& draw : this_tree.draws) { diff --git a/game/CMakeLists.txt b/game/CMakeLists.txt index c2b4955b67..1a3175406c 100644 --- a/game/CMakeLists.txt +++ b/game/CMakeLists.txt @@ -157,7 +157,7 @@ add_subdirectory(sound) # we build the runtime as a static library. add_library(runtime STATIC ${RUNTIME_SOURCE} "../third-party/glad/src/glad.c") -target_link_libraries(runtime common fmt glfw imgui discord-rpc sound) +target_link_libraries(runtime common fmt glfw imgui discord-rpc sound stb_image) if(WIN32) target_link_libraries(runtime mman) else() diff --git a/game/discord.cpp b/game/discord.cpp index 131df52887..b7f79606c3 100644 --- a/game/discord.cpp +++ b/game/discord.cpp @@ -70,9 +70,9 @@ const char* time_of_day_str(float time) { if (hour >= 0 && hour <= 9) return "green-sun"; - else if (hour >= 10 && hour <= 21) + else if (hour < 22) return "day"; - else if (hour >= 22 && hour <= 24) + else if (hour < 25) return "evening"; else return ""; diff --git a/game/graphics/opengl_renderer/background/Tfrag3.cpp b/game/graphics/opengl_renderer/background/Tfrag3.cpp index 1b6846e0a7..a1ee89a64a 100644 --- a/game/graphics/opengl_renderer/background/Tfrag3.cpp +++ b/game/graphics/opengl_renderer/background/Tfrag3.cpp @@ -78,6 +78,7 @@ void Tfrag3::update_load(const std::vector& tree_kind tree_cache.vis = &tree.bvh; tree_cache.index_data = tree.unpacked.indices.data(); tree_cache.tod_cache = swizzle_time_of_day(tree.colors); + tree_cache.draw_mode = tree.use_strips ? GL_TRIANGLE_STRIP : GL_TRIANGLES; vis_temp_len = std::max(vis_temp_len, tree.bvh.vis_nodes.size()); glBindBuffer(GL_ARRAY_BUFFER, tree_cache.vertex_buffer); // glBufferData(GL_ARRAY_BUFFER, verts * sizeof(tfrag3::PreloadedVertex), @@ -240,11 +241,11 @@ void Tfrag3::render_tree(int geom, prof.add_draw_call(); if (render_state->no_multidraw) { - glDrawElements(GL_TRIANGLE_STRIP, singledraw_indices.second, GL_UNSIGNED_INT, + glDrawElements(tree.draw_mode, singledraw_indices.second, GL_UNSIGNED_INT, (void*)(singledraw_indices.first * sizeof(u32))); } else { - glMultiDrawElements(GL_TRIANGLE_STRIP, - &m_cache.multidraw_count_buffer[multidraw_indices.first], GL_UNSIGNED_INT, + glMultiDrawElements(tree.draw_mode, &m_cache.multidraw_count_buffer[multidraw_indices.first], + GL_UNSIGNED_INT, &m_cache.multidraw_index_offset_buffer[multidraw_indices.first], multidraw_indices.second); } @@ -260,11 +261,11 @@ void Tfrag3::render_tree(int geom, double_draw.aref_second); glDepthMask(GL_FALSE); if (render_state->no_multidraw) { - glDrawElements(GL_TRIANGLE_STRIP, singledraw_indices.second, GL_UNSIGNED_INT, + glDrawElements(tree.draw_mode, singledraw_indices.second, GL_UNSIGNED_INT, (void*)(singledraw_indices.first * sizeof(u32))); } else { glMultiDrawElements( - GL_TRIANGLE_STRIP, &m_cache.multidraw_count_buffer[multidraw_indices.first], + tree.draw_mode, &m_cache.multidraw_count_buffer[multidraw_indices.first], GL_UNSIGNED_INT, &m_cache.multidraw_index_offset_buffer[multidraw_indices.first], multidraw_indices.second); } diff --git a/game/graphics/opengl_renderer/background/Tfrag3.h b/game/graphics/opengl_renderer/background/Tfrag3.h index 24762a087f..295263ee1f 100644 --- a/game/graphics/opengl_renderer/background/Tfrag3.h +++ b/game/graphics/opengl_renderer/background/Tfrag3.h @@ -64,6 +64,7 @@ class Tfrag3 { const tfrag3::BVH* vis = nullptr; const u32* index_data = nullptr; SwizzledTimeOfDay tod_cache; + u64 draw_mode = 0; void reset_stats() { rendered_this_frame = false; diff --git a/game/graphics/opengl_renderer/loader/LoaderStages.cpp b/game/graphics/opengl_renderer/loader/LoaderStages.cpp index e75e76c2a2..618ae46611 100644 --- a/game/graphics/opengl_renderer/loader/LoaderStages.cpp +++ b/game/graphics/opengl_renderer/loader/LoaderStages.cpp @@ -83,6 +83,7 @@ class TfragLoadStage : public LoaderStage { GLuint& tree_out = data.lev_data->tfrag_vertex_data[geo].emplace_back(); glGenBuffers(1, &tree_out); glBindBuffer(GL_ARRAY_BUFFER, tree_out); + glBufferData(GL_ARRAY_BUFFER, in_tree.unpacked.vertices.size() * sizeof(tfrag3::PreloadedVertex), nullptr, GL_STATIC_DRAW); @@ -97,34 +98,38 @@ class TfragLoadStage : public LoaderStage { u32 unique_buffers = 0; while (true) { - const auto& tree = data.lev_data->level->tfrag_trees[m_next_geo][m_next_tree]; - u32 end_vert_in_tree = tree.unpacked.vertices.size(); - // the number of vertices we'd need to finish the tree right now - size_t num_verts_left_in_tree = end_vert_in_tree - m_next_vert; - size_t start_vert_for_chunk; - size_t end_vert_for_chunk; - bool complete_tree; - if (num_verts_left_in_tree > CHUNK_SIZE) { - complete_tree = false; - // should only do partial - start_vert_for_chunk = m_next_vert; - end_vert_for_chunk = start_vert_for_chunk + CHUNK_SIZE; - m_next_vert += CHUNK_SIZE; - } else { - // should do all! - start_vert_for_chunk = m_next_vert; - end_vert_for_chunk = end_vert_in_tree; + if (data.lev_data->level->tfrag_trees[m_next_geo].empty()) { complete_tree = true; - } + } else { + const auto& tree = data.lev_data->level->tfrag_trees[m_next_geo][m_next_tree]; + u32 end_vert_in_tree = tree.unpacked.vertices.size(); + // the number of vertices we'd need to finish the tree right now + size_t num_verts_left_in_tree = end_vert_in_tree - m_next_vert; + size_t start_vert_for_chunk; + size_t end_vert_for_chunk; - glBindBuffer(GL_ARRAY_BUFFER, data.lev_data->tfrag_vertex_data[m_next_geo][m_next_tree]); - u32 upload_size = - (end_vert_for_chunk - start_vert_for_chunk) * sizeof(tfrag3::PreloadedVertex); - glBufferSubData(GL_ARRAY_BUFFER, start_vert_for_chunk * sizeof(tfrag3::PreloadedVertex), - upload_size, tree.unpacked.vertices.data() + start_vert_for_chunk); - uploaded_bytes += upload_size; + if (num_verts_left_in_tree > CHUNK_SIZE) { + complete_tree = false; + // should only do partial + start_vert_for_chunk = m_next_vert; + end_vert_for_chunk = start_vert_for_chunk + CHUNK_SIZE; + m_next_vert += CHUNK_SIZE; + } else { + // should do all! + start_vert_for_chunk = m_next_vert; + end_vert_for_chunk = end_vert_in_tree; + complete_tree = true; + } + + glBindBuffer(GL_ARRAY_BUFFER, data.lev_data->tfrag_vertex_data[m_next_geo][m_next_tree]); + u32 upload_size = + (end_vert_for_chunk - start_vert_for_chunk) * sizeof(tfrag3::PreloadedVertex); + glBufferSubData(GL_ARRAY_BUFFER, start_vert_for_chunk * sizeof(tfrag3::PreloadedVertex), + upload_size, tree.unpacked.vertices.data() + start_vert_for_chunk); + uploaded_bytes += upload_size; + } if (complete_tree) { unique_buffers++; diff --git a/game/graphics/pipelines/opengl.cpp b/game/graphics/pipelines/opengl.cpp index 38e14532ba..04e551236a 100644 --- a/game/graphics/pipelines/opengl.cpp +++ b/game/graphics/pipelines/opengl.cpp @@ -21,7 +21,7 @@ #include "game/system/newpad.h" #include "common/log/log.h" #include "common/goal_constants.h" -#include "common/util/image_loading.h" +#include "third-party/stb_image/stb_image.h" #include "game/runtime.h" #include "common/util/Timer.h" #include "game/graphics/opengl_renderer/debug_gui.h" @@ -192,6 +192,13 @@ static std::shared_ptr gl_make_display(int width, g_gfx_data->debug_gui.m_vsync = false; g_gfx_data->vsync_enabled = false; glfwSwapInterval(false); + if (primary_monitor_video_mode->refreshRate > 100) { + BootVideoMode = VideoMode::FPS150; + g_gfx_data->debug_gui.target_fps = 150; + } else if (primary_monitor_video_mode->refreshRate > 60) { + BootVideoMode = VideoMode::FPS100; + g_gfx_data->debug_gui.target_fps = 100; + } } else { // enable vsync g_gfx_data->debug_gui.framelimiter = false; @@ -228,6 +235,11 @@ static std::shared_ptr gl_make_display(int width, // this does initialization for stuff like the font data ImGui::CreateContext(); + // Init ImGui settings + ImGuiIO& io = ImGui::GetIO(); + io.IniFilename = file_util::get_file_path({"imgui.ini"}).c_str(); + io.LogFilename = file_util::get_file_path({"imgui_log.txt"}).c_str(); + // set up to get inputs for this window ImGui_ImplGlfw_InitForOpenGL(window, true); @@ -400,13 +412,15 @@ GfxDisplayMode GLDisplay::get_fullscreen() { void GLDisplay::get_screen_size(int vmode_idx, s32* w_out, s32* h_out, s32* count_out) { int count = 0; auto vmode = glfwGetVideoMode(glfwGetPrimaryMonitor()); - auto vmodes = glfwGetVideoModes(glfwGetPrimaryMonitor(), &count); - if (vmode_idx >= 0) { - vmode = &vmodes[vmode_idx]; - } else { - for (int i = 0; i < count; ++i) { - if (!vmode || vmode->height < vmodes[i].height) { - vmode = &vmodes[i]; + if (get_fullscreen() == GfxDisplayMode::Fullscreen) { + auto vmodes = glfwGetVideoModes(glfwGetPrimaryMonitor(), &count); + if (vmode_idx >= 0) { + vmode = &vmodes[vmode_idx]; + } else { + for (int i = 0; i < count; ++i) { + if (!vmode || vmode->height < vmodes[i].height) { + vmode = &vmodes[i]; + } } } } diff --git a/game/kernel/kboot.cpp b/game/kernel/kboot.cpp index 6e8aade4a4..7cb2c83bc7 100644 --- a/game/kernel/kboot.cpp +++ b/game/kernel/kboot.cpp @@ -32,6 +32,8 @@ using namespace ee; +VideoMode BootVideoMode; + // Level to load on boot char DebugBootLevel[64]; diff --git a/game/kernel/kboot.h b/game/kernel/kboot.h index 7989048c1c..db29523a71 100644 --- a/game/kernel/kboot.h +++ b/game/kernel/kboot.h @@ -30,6 +30,16 @@ enum class RuntimeExitStatus { RESTART_IN_DEBUG = 3, }; +enum class VideoMode { + NTSC = 0, + PAL = 1, + FPS100 = 2, + FPS150 = 3, +}; + +// Video Mode that's set based on display refresh rate on boot +extern VideoMode BootVideoMode; + // Level to load on boot extern char DebugBootLevel[64]; diff --git a/game/kernel/kscheme.cpp b/game/kernel/kscheme.cpp index ab7a65630a..557c545754 100644 --- a/game/kernel/kscheme.cpp +++ b/game/kernel/kscheme.cpp @@ -2003,7 +2003,7 @@ s32 InitHeapAndSymbol() { method_set_symbol->value = 0; // set *boot-video-mode* - intern_from_c("*boot-video-mode*")->value = 0; + intern_from_c("*boot-video-mode*")->value = (u32)BootVideoMode; lg::info("Initialized GOAL heap in {:.2} ms", heap_init_timer.getMs()); // load the kernel! diff --git a/game/overlord/soundcommon.cpp b/game/overlord/soundcommon.cpp index 40a72785a5..d564fc84ef 100644 --- a/game/overlord/soundcommon.cpp +++ b/game/overlord/soundcommon.cpp @@ -1,6 +1,7 @@ #include "soundcommon.h" #include #include "common/util/Assert.h" +#include // TODO strcpy_toupper // TODO atoi @@ -16,7 +17,9 @@ void ReadBankSoundInfo(SoundBank* bank, SoundBank* unk, s32 unk2) { void PrintBankInfo(SoundBank* bank) { printf("Bank %s\n\n", bank->name); for (u32 i = 0; i < bank->sound_count; i++) { - printf("%d : %16s : min %d max %d curve %d\n", i, bank->sound[i].name, + // Some characters use the full 16 characters (bonelurker-grunt) and dont have a null terminator + std::string name = std::string(bank->sound[i].name, 16); + printf("%d : %16s : min %d max %d curve %d\n", i, name.c_str(), bank->sound[i].fallof_params & 0x3fff, (bank->sound[i].fallof_params >> 14) & 0x3fff, bank->sound[i].fallof_params >> 28); } diff --git a/game/runtime.cpp b/game/runtime.cpp index 659652ed0b..8758933212 100644 --- a/game/runtime.cpp +++ b/game/runtime.cpp @@ -282,8 +282,6 @@ RuntimeExitStatus exec_runtime(int argc, char** argv) { g_argv = argv; g_main_thread_id = std::this_thread::get_id(); - file_util::create_dir_if_needed("game_config/"); - // parse opengoal arguments bool enable_display = true; for (int i = 1; i < argc; i++) { diff --git a/goal_src/engine/ambient/weather-part.gc b/goal_src/engine/ambient/weather-part.gc index 54bcc23c1c..c934004a65 100644 --- a/goal_src/engine/ambient/weather-part.gc +++ b/goal_src/engine/ambient/weather-part.gc @@ -353,7 +353,7 @@ ) (defun update-snow ((arg0 target)) - (let ((gp-0 (-> arg0 control trans))) + (let ((target-position (-> arg0 control trans))) (let ((f0-0 (lerp-scale 0.0 1.0 (vector-length (-> arg0 control transv)) 2048.0 40960.0))) (set! (-> *part-id-table* 34 init-specs 1 initial-valuef) (- 1.0 f0-0)) (set! (-> *part-id-table* 33 init-specs 1 initial-valuef) (* 4.0 f0-0)) @@ -361,22 +361,8 @@ (set! (-> *part-id-table* 33 init-specs 19 initial-valuef) (+ 32768.0 (vector-y-angle (-> arg0 control transv))) ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 34) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 33) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 34) target-position) + (launch-particles (-> *part-id-table* 33) target-position) ) 0 (none) @@ -464,22 +450,8 @@ (let ((gp-0 (new 'stack-no-clear 'vector))) (sp-kill-particle arg0 arg1) (set-vector! gp-0 (-> arg2 x) (-> arg1 user-float) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 39) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 40) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :rate 1.0 (-> *part-id-table* 39) gp-0) + (launch-particles :system *sp-particle-system-3d* :rate 1.0 (-> *part-id-table* 40) gp-0) ) ) (none) @@ -512,22 +484,8 @@ (set! (-> *part-id-table* 38 init-specs 5 initial-valuef) f30-0) (set! (-> *part-id-table* 38 init-specs 5 random-rangef) f30-0) ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 37) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 38) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 37) gp-0) + (launch-particles (-> *part-id-table* 38) gp-0) ) ) 0 diff --git a/goal_src/engine/collide/collide-cache.gc b/goal_src/engine/collide/collide-cache.gc index 031c25d23a..800af98612 100644 --- a/goal_src/engine/collide/collide-cache.gc +++ b/goal_src/engine/collide/collide-cache.gc @@ -384,6 +384,8 @@ (set! (-> v1-5 mesh) (-> obj mesh)) (set! (-> v1-5 inst) #f) ) + ; (add-debug-sphere #t (bucket-id debug) (-> obj bsphere) (-> obj bsphere w) (new 'static 'rgba :g #x80 :a #x80)) + ; (format 0 "~f~%" (-> obj bsphere w)) (+! (-> arg1 num-items) 1) ) (&+! obj 32) diff --git a/goal_src/engine/collide/collide-frag.gc b/goal_src/engine/collide/collide-frag.gc index a9703339dc..7246ac2ab7 100644 --- a/goal_src/engine/collide/collide-frag.gc +++ b/goal_src/engine/collide/collide-frag.gc @@ -83,6 +83,19 @@ (defmethod draw collide-fragment ((obj collide-fragment) (arg0 collide-fragment) (arg1 display-frame)) ;; if we wanted to draw collide-fragment's we'd do it here. + ; (when (< (-> obj bsphere w) (meters 22.)) + ; (format 0 "sp: ~m : ~D~%" (-> obj bsphere w) (-> obj mesh poly-count)) + ; (let ((mins (vector-copy! (new-stack-vector0) (-> obj bsphere))) + ; (maxs (vector-copy! (new-stack-vector0) (-> obj bsphere)))) + ; (dotimes (i 3) + ; (-! (-> mins data i) (-> obj bsphere w)) + ; (+! (-> maxs data i) (-> obj bsphere w)) + ; ) + ; (add-debug-box #t (bucket-id debug) mins maxs (new 'static 'rgba :r #x80 :a #x80) + ; ) + ; ;(add-debug-sphere #t (bucket-id debug) (-> obj bsphere) (-> obj bsphere w) (new 'static 'rgba :r #x80 :a #x80)) + ; ) + ;; (add-debug-point #t (bucket-id debug) (-> obj bsphere)) (none) ) diff --git a/goal_src/engine/draw/drawable.gc b/goal_src/engine/draw/drawable.gc index 9249778496..ccf9200421 100644 --- a/goal_src/engine/draw/drawable.gc +++ b/goal_src/engine/draw/drawable.gc @@ -1335,7 +1335,12 @@ (+! (-> disp base-frame-counter) scaled-seconds) (+! (-> disp part-frame-counter) scaled-seconds) ;; this counts actual frames, not seconds. Will count 2 frames if we lag - (+! (-> disp integral-frame-counter) (the int time-ratio)) + ;; When counting frames while running at > 60fps we need to scale it, but always count at least 1 frame + (+! (-> disp integral-frame-counter) (if (= (-> *setting-control* current video-mode) '150fps) + (max 1 (the int (* 0.4 time-ratio))) + (if (= (-> *setting-control* current video-mode) '100fps) + (max 1 (the int (* 0.6 time-ratio))) + (the int time-ratio)))) ;; this counts actual frames, not doubling for lag. Will count 1 per frame drawn (+! (-> disp actual-frame-counter) 1) ;; game counter will count seconds that we're not in a movie @@ -1640,7 +1645,7 @@ ;; need to call reset-graph with some magic number ;; also stash this parameter so that if things go really wrong and our DMA transfer ;; times out, we can reset-graph to the appropriate video mode - (if (= (-> *setting-control* current video-mode) 'ntsc) + (if (or (= (-> *setting-control* current video-mode) 'ntsc) (= (-> *setting-control* current video-mode) '150fps) (= (-> *setting-control* current video-mode) '100fps)) (set! *video-reset-parm* 2) (set! *video-reset-parm* 3) ) diff --git a/goal_src/engine/draw/process-drawable-h.gc b/goal_src/engine/draw/process-drawable-h.gc index 68b1cda420..1eabf45db9 100644 --- a/goal_src/engine/draw/process-drawable-h.gc +++ b/goal_src/engine/draw/process-drawable-h.gc @@ -148,4 +148,137 @@ (bit-10 10) ) -(define-extern process-entity-status! (function process entity-perm-status symbol int)) \ No newline at end of file +(define-extern process-entity-status! (function process entity-perm-status symbol int)) + +(defmacro ja-group (&key (chan 0)) + "get the frame group for self. default channel is 0, the base channel. returns #f if no frame group." + `(if (> (-> self skel active-channels) ,chan) + (-> self skel root-channel ,chan frame-group)) + ) + +(defmacro ja-group? (group &key (chan 0)) + "is self in this frame group on this channel? default is channel 0, which is the base channel." + `(= (ja-group) ,group) + ) + +(defmacro ja (&key (chan 0) + &key (group! #f) + &key (num! #f) + &key (param0 #f) + &key (param1 #f) + &key (num-func #f) + &key (frame-num #f) + &key (frame-interp #f) + &key (dist #f) + &key (eval? #t) + ) + "set various joint anim parameters for self and eval them. + you can use this for playing animations! + + chan = the channel to modify. defaults to 0 (base channel). this is usually what you want. + group! = when not #f, set this as the new frame-group. defaults to #f + num! = set the frame playback function. this is what determines what frame an animation is at. funcs below. + #f = no func will be set, and there wont be a frame eval. + num-func = sets the num-func field for the channel. this lets you change the function with eval'ing. + param0 = 1st parameter for the playback function. ONLY USE THESE WITH num-func !! + param1 = 2nd parameter for the playback function. ONLY USE THESE WITH num-func !! + frame-num = set the frame-num field. + frame-interp = set the frame-interp field. + dist = set the dist field. + + available num! functions: + - (+!) = advance anim. + - (-!) = reverse anim. + - (identity num) = play 'num' frame. + - (seek! target speed) = animate towards frame target at a speed. + speed is optional and defaults to 1.0 when not provided. + target is optional and defaults to the last frame of the animation. + if you want to set the speed, you therefore must also set the target. + target can be max (no quote), which is just the same as the default value. + - (loop! speed) = loop animation at a speed. default speed is 1.0 when not provided. + - (chan channel) = copy frame from another channel. + - min = the start of the animation. + - max = the end of the animation. + " + + (let* ((num-args (if (pair? num!) (cdr num!) '())) + (num! (if (pair? num!) (car num!) num!)) + (nf (cond + ((or (eq? num! 'identity) + (eq? num! 'min) + (eq? num! 'max) + ) + 'num-func-identity) + ((eq? num! 'none) 'num-func-none) + ((eq? num! '+!) 'num-func-+!) + ((eq? num! '-!) 'num-func--!) + ((eq? num! 'seek!) 'num-func-seek!) + ((eq? num! 'loop!) 'num-func-loop!) + ((eq? num! 'blend-in!) 'num-func-blend-in!) + ((eq? num! 'chan) 'num-func-chan) + )) + (p0 (if param0 param0 + (cond + ((eq? num! 'chan) `(the float ,(car num-args))) + ((eq? num! '+!) (if (null? num-args) 1.0 (car num-args))) + ((eq? num! '-!) (if (null? num-args) 1.0 (car num-args))) + ((eq? num! 'loop!) (if (null? num-args) 1.0 (if (eq? 'max (car num-args)) + (if group! + `(the float (1- (-> (the art-joint-anim ,group!) data 0 length))) + `(the float (1- (-> ja-ch frame-group data 0 length))) + ) + (car num-args)))) + ((eq? num! 'seek!) (if (or (null? num-args) (eq? (car num-args) 'max)) + (if group! + `(the float (1- (-> (the art-joint-anim ,group!) data 0 length))) + `(the float (1- (-> ja-ch frame-group data 0 length))) + ) + (car num-args))) + ))) + (p1 (if param1 param1 + (cond + ((eq? num! 'seek!) (if (or (null? num-args) (null? (cdr num-args))) 1.0 (cadr num-args))) + ))) + (frame-num (if (eq? 'max frame-num) (if group! + `(the float (1- (-> (the art-joint-anim ,group!) data 0 length))) + `(the float (1- (-> ja-ch frame-group data 0 length))) + ) + frame-num)) + (frame-group (if (or p0 p1 frame-num (not nf)) group! #f)) + ) + `(let ((ja-ch (-> self skel root-channel ,chan))) + ,(if frame-interp `(set! (-> ja-ch frame-interp) ,frame-interp) `(none)) + ,(if dist `(set! (-> ja-ch dist) ,dist) `(none)) + ,(if frame-group `(set! (-> ja-ch frame-group) (the art-joint-anim ,frame-group)) `(none)) + ,(if p0 `(set! (-> ja-ch param 0) ,p0) `(none)) + ,(if p1 `(set! (-> ja-ch param 1) ,p1) `(none)) + ,(if num-func `(set! (-> ja-ch num-func) ,num-func) `(none)) + ,(if frame-num `(set! (-> ja-ch frame-num) ,frame-num) `(none)) + ,(if nf + `(,(if eval? 'joint-control-channel-group-eval! 'joint-control-channel-group!) + ja-ch (the art-joint-anim ,group!) ,nf) + `(none)) + ,(cond + ((eq? num! 'min) `(set! (-> ja-ch frame-num) 0.0)) + ((eq? num! 'max) (if group! + `(set! (-> ja-ch frame-num) (the float (1- (-> (the art-joint-anim ,group!) data 0 length)))) + `(set! (-> ja-ch frame-num) (the float (1- (-> ja-ch frame-group data 0 length)))) + )) + ((eq? num! 'identity) `(set! (-> ja-ch frame-num) ,(car num-args))) + (#t `(none)) + ) + )) + ) + +(defmacro ja-no-eval (&key (chan 0) + &key (group! #f) + &key (num! #f) + &key (param0 #f) + &key (param1 #f) + &key (num-func #f) + &key (frame-num #f) + &key (frame-interp #f) + &key (dist #f) + ) + `(ja :eval? #f :chan ,chan :group! ,group! :num! ,num! :param0 ,param0 :param1 ,param1 :num-func ,num-func :frame-num ,frame-num :frame-interp ,frame-interp :dist ,dist) + ) diff --git a/goal_src/engine/draw/process-drawable.gc b/goal_src/engine/draw/process-drawable.gc index da55f294c9..e7edd79d9c 100644 --- a/goal_src/engine/draw/process-drawable.gc +++ b/goal_src/engine/draw/process-drawable.gc @@ -694,18 +694,6 @@ ) ) - -(defmacro ja-group (&key (chan 0)) - "get the frame group for self. default channel is 0, the base channel. returns #f if no frame group." - `(if (> (-> self skel active-channels) ,chan) - (-> self skel root-channel ,chan frame-group)) - ) - -(defmacro ja-group? (group &key (chan 0)) - "is self in this frame group on this channel? default is channel 0, which is the base channel." - `(= (ja-group) ,group) - ) - (defbehavior ja-num-frames process-drawable ((arg0 int)) (+ (-> self skel root-channel arg0 frame-group data 0 length) -1) ) @@ -881,128 +869,6 @@ ) ) -(defmacro ja (&key (chan 0) - &key (group! #f) - &key (num! #f) - &key (param0 #f) - &key (param1 #f) - &key (num-func #f) - &key (frame-num #f) - &key (frame-interp #f) - &key (dist #f) - &key (eval? #t) - ) - "set various joint anim parameters for self and eval them. - you can use this for playing animations! - - chan = the channel to modify. defaults to 0 (base channel). this is usually what you want. - group! = when not #f, set this as the new frame-group. defaults to #f - num! = set the frame playback function. this is what determines what frame an animation is at. funcs below. - #f = no func will be set, and there wont be a frame eval. - num-func = sets the num-func field for the channel. this lets you change the function with eval'ing. - param0 = 1st parameter for the playback function. ONLY USE THESE WITH num-func !! - param1 = 2nd parameter for the playback function. ONLY USE THESE WITH num-func !! - frame-num = set the frame-num field. - frame-interp = set the frame-interp field. - dist = set the dist field. - - available num! functions: - - (+!) = advance anim. - - (-!) = reverse anim. - - (identity num) = play 'num' frame. - - (seek! target speed) = animate towards frame target at a speed. - speed is optional and defaults to 1.0 when not provided. - target is optional and defaults to the last frame of the animation. - if you want to set the speed, you therefore must also set the target. - target can be max (no quote), which is just the same as the default value. - - (loop! speed) = loop animation at a speed. default speed is 1.0 when not provided. - - (chan channel) = copy frame from another channel. - - min = the start of the animation. - - max = the end of the animation. - " - - (let* ((num-args (if (pair? num!) (cdr num!) '())) - (num! (if (pair? num!) (car num!) num!)) - (nf (cond - ((or (eq? num! 'identity) - (eq? num! 'min) - (eq? num! 'max) - ) - 'num-func-identity) - ((eq? num! 'none) 'num-func-none) - ((eq? num! '+!) 'num-func-+!) - ((eq? num! '-!) 'num-func--!) - ((eq? num! 'seek!) 'num-func-seek!) - ((eq? num! 'loop!) 'num-func-loop!) - ((eq? num! 'blend-in!) 'num-func-blend-in!) - ((eq? num! 'chan) 'num-func-chan) - )) - (p0 (if param0 param0 - (cond - ((eq? num! 'chan) `(the float ,(car num-args))) - ((eq? num! '+!) (if (null? num-args) 1.0 (car num-args))) - ((eq? num! '-!) (if (null? num-args) 1.0 (car num-args))) - ((eq? num! 'loop!) (if (null? num-args) 1.0 (if (eq? 'max (car num-args)) - (if group! - `(the float (1- (-> (the art-joint-anim ,group!) data 0 length))) - `(the float (1- (-> ja-ch frame-group data 0 length))) - ) - (car num-args)))) - ((eq? num! 'seek!) (if (or (null? num-args) (eq? (car num-args) 'max)) - (if group! - `(the float (1- (-> (the art-joint-anim ,group!) data 0 length))) - `(the float (1- (-> ja-ch frame-group data 0 length))) - ) - (car num-args))) - ))) - (p1 (if param1 param1 - (cond - ((eq? num! 'seek!) (if (or (null? num-args) (null? (cdr num-args))) 1.0 (cadr num-args))) - ))) - (frame-num (if (eq? 'max frame-num) (if group! - `(the float (1- (-> (the art-joint-anim ,group!) data 0 length))) - `(the float (1- (-> ja-ch frame-group data 0 length))) - ) - frame-num)) - (frame-group (if (or p0 p1 frame-num (not nf)) group! #f)) - ) - `(let ((ja-ch (-> self skel root-channel ,chan))) - ,(if frame-interp `(set! (-> ja-ch frame-interp) ,frame-interp) `(none)) - ,(if dist `(set! (-> ja-ch dist) ,dist) `(none)) - ,(if frame-group `(set! (-> ja-ch frame-group) (the art-joint-anim ,frame-group)) `(none)) - ,(if p0 `(set! (-> ja-ch param 0) ,p0) `(none)) - ,(if p1 `(set! (-> ja-ch param 1) ,p1) `(none)) - ,(if num-func `(set! (-> ja-ch num-func) ,num-func) `(none)) - ,(if frame-num `(set! (-> ja-ch frame-num) ,frame-num) `(none)) - ,(if nf - `(,(if eval? 'joint-control-channel-group-eval! 'joint-control-channel-group!) - ja-ch (the art-joint-anim ,group!) ,nf) - `(none)) - ,(cond - ((eq? num! 'min) `(set! (-> ja-ch frame-num) 0.0)) - ((eq? num! 'max) (if group! - `(set! (-> ja-ch frame-num) (the float (1- (-> (the art-joint-anim ,group!) data 0 length)))) - `(set! (-> ja-ch frame-num) (the float (1- (-> ja-ch frame-group data 0 length)))) - )) - ((eq? num! 'identity) `(set! (-> ja-ch frame-num) ,(car num-args))) - (#t `(none)) - ) - )) - ) - -(defmacro ja-no-eval (&key (chan 0) - &key (group! #f) - &key (num! #f) - &key (param0 #f) - &key (param1 #f) - &key (num-func #f) - &key (frame-num #f) - &key (frame-interp #f) - &key (dist #f) - ) - `(ja :eval? #f :chan ,chan :group! ,group! :num! ,num! :param0 ,param0 :param1 ,param1 :num-func ,num-func :frame-num ,frame-num :frame-interp ,frame-interp :dist ,dist) - ) - (defbehavior ja-eval process-drawable () (let ((gp-0 (-> self skel root-channel 0)) (s5-0 (-> self skel channel (-> self skel active-channels))) diff --git a/goal_src/engine/game/effect-control.gc b/goal_src/engine/game/effect-control.gc index 54bfc59a06..687d7c6add 100644 --- a/goal_src/engine/game/effect-control.gc +++ b/goal_src/engine/game/effect-control.gc @@ -430,14 +430,10 @@ arg1 s5-0 ) - (sp-launch-particles-var - *sp-particle-system-2d* - (the-as sparticle-launcher s3-0) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> obj process node-list data s5-0)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :rate 1.0 + (the-as sparticle-launcher s3-0) + (vector<-cspace! (new 'stack-no-clear 'vector) (-> obj process node-list data s5-0)) + ) ) ((= (-> (the-as basic s3-0) type) sparticle-launch-group) (if *debug-effect-control* diff --git a/goal_src/engine/game/generic-obs.gc b/goal_src/engine/game/generic-obs.gc index ea570ab84b..357b21d8c0 100644 --- a/goal_src/engine/game/generic-obs.gc +++ b/goal_src/engine/game/generic-obs.gc @@ -2364,14 +2364,7 @@ (dotimes (s4-1 3) (quaternion-rotate-local-z! s5-0 s5-0 10922.667) (quaternion-copy! *particle-quat* s5-0) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 2528) - gp-1 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system *sp-particle-system-3d* (-> *part-id-table* 2528) gp-1) ) ) ) diff --git a/goal_src/engine/game/powerups.gc b/goal_src/engine/game/powerups.gc index 77152cc7bd..dd985f9b6d 100644 --- a/goal_src/engine/game/powerups.gc +++ b/goal_src/engine/game/powerups.gc @@ -503,24 +503,10 @@ :init-specs ((sp-flt spt-fade-g 0.0)) ) -(defun eco-blue-glow ((arg0 vector)) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 255) - arg0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) +(defun eco-blue-glow ((origin vector)) + (launch-particles (-> *part-id-table* 255) origin) (if (rand-vu-percent? 0.5) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 257) - arg0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 257) origin) ) 0 (none) @@ -555,14 +541,7 @@ ) (rand-vu-percent? 0.5) ) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 2391) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system *sp-particle-system-3d* (-> *part-id-table* 2391) gp-0) ) ) (let ((gp-1 (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data 70)))) @@ -576,14 +555,7 @@ ) (rand-vu-percent? 0.5) ) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 2391) - gp-1 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system *sp-particle-system-3d* (-> *part-id-table* 2391) gp-1) ) ) ;; PAL patch here @@ -657,29 +629,15 @@ (((pickup-type eco-yellow)) (change-sound! (-> self sound) (static-sound-name "yel-eco-jak")) (let ((s4-0 (rand-vu-int-range 3 (+ (-> self node-list length) -1)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* (if (rand-vu-percent? 0.5) - 269 - 270 - ) - ) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data s4-0)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* (if (rand-vu-percent? 0.5) 269 270)) + (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data s4-0)) + ) ) (dotimes (gp-4 2) (let ((v1-111 (rand-vu-int-range 3 (+ (-> self node-list length) -1)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 271) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data v1-111)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 271) + (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data v1-111)) + ) ) ) ) @@ -695,29 +653,15 @@ (update-transforms! (-> self control)) (change-sound! (-> self sound) (static-sound-name "red-eco-jak")) (let ((s4-2 (rand-vu-int-range 3 (+ (-> self node-list length) -1)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* (if (rand-vu-percent? 0.5) - 273 - 274 - ) - ) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data s4-2)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* (if (rand-vu-percent? 0.5) 273 274)) + (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data s4-2)) + ) ) (dotimes (gp-6 2) (let ((v1-139 (rand-vu-int-range 3 (+ (-> self node-list length) -1)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 275) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data v1-139)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 275) + (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data v1-139)) + ) ) ) ) @@ -736,77 +680,35 @@ (set! (-> *part-id-table* 259 init-specs 4 random-rangef) 16384.0) ) ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 259) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data v1-150)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 259) (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data v1-150)) +) ) (let ((gp-8 (rand-vu-int-range 3 (+ (-> self node-list length) -1)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* (if (rand-vu-percent? 0.5) - 255 - 256 - ) - ) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data gp-8)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* (if (rand-vu-percent? 0.5) 255 256)) + (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data gp-8)) + ) (if (rand-vu-percent? 0.5) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 257) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data gp-8)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 257) + (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data gp-8)) + ) ) ) (let ((v1-168 (rand-vu-int-range 3 (+ (-> self node-list length) -1)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 260) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data v1-168)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 260) + (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data v1-168)) + ) ) (cpad-set-buzz! (-> *cpad-list* cpads 0) 0 1 (seconds 0.1)) ) (((pickup-type eco-green)) (change-sound! (-> self sound) (static-sound-name "green-eco-jak")) (let ((s4-8 (rand-vu-int-range 3 (+ (-> self node-list length) -1)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* (if (rand-vu-percent? 0.5) - 277 - 278 - ) - ) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data s4-8)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* (if (rand-vu-percent? 0.5) 277 278)) + (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data s4-8))) ) (dotimes (gp-11 2) (let ((v1-188 (rand-vu-int-range 3 (+ (-> self node-list length) -1)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 279) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data v1-188)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 279) (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data v1-188))) ) ) ) diff --git a/goal_src/engine/game/settings.gc b/goal_src/engine/game/settings.gc index 8cbc040e31..b1b4964cc0 100644 --- a/goal_src/engine/game/settings.gc +++ b/goal_src/engine/game/settings.gc @@ -445,10 +445,20 @@ (set! (-> gp-0 aspect-ratio) 'aspect4x3) ) ) - (if (zero? *boot-video-mode*) - (set! (-> gp-0 video-mode) 'ntsc) - (set! (-> gp-0 video-mode) 'pal) - ) + (case *boot-video-mode* + ((0) + (set! (-> gp-0 video-mode) 'ntsc) + ) + ((1) + (set! (-> gp-0 video-mode) 'pal) + ) + ((2) + (set! (-> gp-0 video-mode) '100fps) + ) + ((3) + (set! (-> gp-0 video-mode) '150fps) + ) + ) (set! (-> s5-0 sfx-volume) (+ -1.0 (-> gp-0 sfx-volume))) (set! (-> s5-0 music-volume) (+ -1.0 (-> gp-0 music-volume))) (set! (-> s5-0 dialog-volume) (+ -1.0 (-> gp-0 dialog-volume))) diff --git a/goal_src/engine/game/video.gc b/goal_src/engine/game/video.gc index 88ea7b7434..971c3419ee 100644 --- a/goal_src/engine/game/video.gc +++ b/goal_src/engine/game/video.gc @@ -33,6 +33,30 @@ (set! (-> *math-camera* y-clip) 512.0) (set! (-> *shadow-data* texoffset y) 128.5) ) + (('100fps) + (set! (-> *video-parms* screen-sy) 224) + (set! (-> *setting-control* default screenx) 0) + (set! (-> *setting-control* default screeny) 8) + (set! (-> *video-parms* screen-pages-high) 7) + (set! (-> *video-parms* relative-y-scale) 1.0) + (set! *ticks-per-frame* 5859) + (set! (-> *math-camera* isometric vector 1 y) 0.5) + (set! (-> *math-camera* y-pix) 112.0) + (set! (-> *math-camera* y-clip) 448.0) + (set! (-> *shadow-data* texoffset y) 112.5) + ) + (('150fps) + (set! (-> *video-parms* screen-sy) 224) + (set! (-> *setting-control* default screenx) 0) + (set! (-> *setting-control* default screeny) 8) + (set! (-> *video-parms* screen-pages-high) 7) + (set! (-> *video-parms* relative-y-scale) 1.0) + (set! *ticks-per-frame* 3906) + (set! (-> *math-camera* isometric vector 1 y) 0.5) + (set! (-> *math-camera* y-pix) 112.0) + (set! (-> *math-camera* y-clip) 448.0) + (set! (-> *shadow-data* texoffset y) 112.5) + ) ) (set-time-ratios *display* (-> *display* time-ratio)) (set! (-> *video-parms* reset-video-mode) #t) diff --git a/goal_src/engine/gfx/eye.gc b/goal_src/engine/gfx/eye.gc index a331dad64a..bf8a135f23 100644 --- a/goal_src/engine/gfx/eye.gc +++ b/goal_src/engine/gfx/eye.gc @@ -832,7 +832,13 @@ (when (not (paused?)) (cond ((and (>= (-> s5-3 left lid) 0.0) (>= (-> s5-3 right lid) 0.0)) - (set! (-> s5-3 random-time) (the-as uint 60)) + (set! (-> s5-3 random-time) (the int (* (if (or (= (-> *setting-control* current video-mode) '150fps) + (= (-> *setting-control* current video-mode) '100fps)) + (if (= (-> *setting-control* current video-mode) '150fps) + 2.5 + 1.6667) + 1.0) + (the-as uint 60)))) (set! (-> s5-3 blink) 0.0) ) (else @@ -841,7 +847,12 @@ (when (< v1-54 (the-as uint 10)) (set! (-> s5-3 blink) (-> *eye-work* blink-table v1-54)) (if (zero? v1-54) - (set! (-> s5-3 random-time) (the-as uint (the int (rand-vu-float-range 60.0 240.0)))) + (set! (-> s5-3 random-time) (the-as uint (the int (* (if (or (= (-> *setting-control* current video-mode) '150fps) + (= (-> *setting-control* current video-mode) '100fps)) + (if (= (-> *setting-control* current video-mode) '150fps) + 2.5 + 1.6667) + 1.0) (rand-vu-float-range 60.0 240.0))))) ) ) ) diff --git a/goal_src/engine/gfx/hw/display.gc b/goal_src/engine/gfx/hw/display.gc index 42a44e8bb2..1e44cec4e2 100644 --- a/goal_src/engine/gfx/hw/display.gc +++ b/goal_src/engine/gfx/hw/display.gc @@ -45,13 +45,30 @@ ;; 6 "ticks" per frame * 50 fps = 300 ticks per second. (set! (-> obj time-factor) 6.0) ) - (else + (('ntsc) (set! (-> obj time-adjust-ratio) ratio) (set! (-> obj seconds-per-frame) (* 0.016666668 ratio)) (set! (-> obj frames-per-second) (* 60.0 (/ 1.0 ratio))) ;; 5 "ticks" per frame * 60 fps = 300 ticks per second. (set! (-> obj time-factor) 5.0) ) + (('100fps) + (set! (-> obj time-adjust-ratio) (* 0.6 ratio)) + (set! (-> obj seconds-per-frame) (* 0.01 ratio)) + (set! (-> obj frames-per-second) (* 100.0 (/ 1.0 ratio))) + ;; 3 "ticks" per frame * 100 fps = 300 ticks per second. + (set! (-> obj time-factor) 3.0) + ) + (('150fps) + (set! (-> obj time-adjust-ratio) (* 0.4 ratio)) + (set! (-> obj seconds-per-frame) (* 0.006666667 ratio)) + (set! (-> obj frames-per-second) (* 150.0 (/ 1.0 ratio))) + ;; 2 "ticks" per frame * 150 fps = 300 ticks per second. + (set! (-> obj time-factor) 2.0) + ) + (else + (format #t "Warning: Tried to set unsupported video-mode") + ) ) ) (-> obj time-ratio) diff --git a/goal_src/engine/gfx/ocean/ocean.gc b/goal_src/engine/gfx/ocean/ocean.gc index 8c06b24b51..1f4342e766 100644 --- a/goal_src/engine/gfx/ocean/ocean.gc +++ b/goal_src/engine/gfx/ocean/ocean.gc @@ -747,4 +747,4 @@ ) 0 (none) - ) \ No newline at end of file + ) diff --git a/goal_src/engine/gfx/sky/sky-tng.gc b/goal_src/engine/gfx/sky/sky-tng.gc index b0bac528af..69910fb11c 100644 --- a/goal_src/engine/gfx/sky/sky-tng.gc +++ b/goal_src/engine/gfx/sky/sky-tng.gc @@ -168,15 +168,31 @@ (sky-make-sun-data *sky-parms* 1 arg0) (sky-make-moon-data *sky-parms* arg0) (let ((v1-0 *sky-tng-data*)) - (+! (-> v1-0 off-s-0) 16) - (+! (-> v1-0 off-t-0) 32) - (+! (-> v1-0 off-s-1) -21) - (+! (-> v1-0 off-t-1) 42) - (set! (-> v1-0 time) arg0) + (case (-> *setting-control* current video-mode) + (('150fps) + (+! (-> v1-0 off-s-0) 6.4) + (+! (-> v1-0 off-t-0) 12.8) + (+! (-> v1-0 off-s-1) -8.4) + (+! (-> v1-0 off-t-1) 16.8) + ) + (('100fps) + (+! (-> v1-0 off-s-0) 9.6) + (+! (-> v1-0 off-t-0) 19.2) + (+! (-> v1-0 off-s-1) -12.6) + (+! (-> v1-0 off-t-1) 25.2) + ) + (else + (+! (-> v1-0 off-s-0) 16) + (+! (-> v1-0 off-t-0) 32) + (+! (-> v1-0 off-s-1) -21) + (+! (-> v1-0 off-t-1) 42) + ) + ) + (set! (-> v1-0 time) arg0) ) 0 (none) - ) +) #| (defun init-sky-regs () diff --git a/goal_src/engine/gfx/texture.gc b/goal_src/engine/gfx/texture.gc index 8827bde8fc..b7be88434b 100644 --- a/goal_src/engine/gfx/texture.gc +++ b/goal_src/engine/gfx/texture.gc @@ -1371,7 +1371,12 @@ (dotimes (page-idx 9) (set! (-> level texture-page page-idx) #f) ) - + (#when PC_PORT + (when (zero? id-array) + (format #t "ERROR: texture id array is 0, skipping texture login!~%") + (return #f) + ) + ) (if (>= max-page-kind 0) ;; tfrag. ;; login the texture. If the texture isn't there, it will try to load it ;; and allocate with the given allocation function. diff --git a/goal_src/engine/gfx/water/water.gc b/goal_src/engine/gfx/water/water.gc index 614aa8b8b5..d77d7998d0 100644 --- a/goal_src/engine/gfx/water/water.gc +++ b/goal_src/engine/gfx/water/water.gc @@ -48,14 +48,7 @@ (sp-kill-particle arg0 arg1) (set-vector! s5-0 (-> arg2 x) (-> arg1 user-float) (-> arg2 z) 1.0) (sound-play "water-drop" :position (the-as symbol s5-0)) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 108) - s5-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system *sp-particle-system-3d* :rate 1.0 (-> *part-id-table* 108) s5-0) ) ) 0 @@ -728,42 +721,14 @@ (set! (-> *part-id-table* 118 init-specs 1 initial-valuef) (* 0.0000036621095 f28-0)) (set! (-> *part-id-table* 118 init-specs 2 initial-valuef) (* 0.1 f28-0)) (set! (-> *part-id-table* 118 init-specs 13 initial-valuef) 0.7111111) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 118) - s4-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system *sp-particle-system-3d* (-> *part-id-table* 118) s4-0) (set! (-> *part-id-table* 121 init-specs 1 initial-valuef) (* 0.000004150391 f28-0)) (set! (-> *part-id-table* 121 init-specs 18 initial-valuef) f30-1) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 121) - s4-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system *sp-particle-system-3d* (-> *part-id-table* 121) s4-0) (when (< f28-0 4096.0) (set! (-> *part-id-table* 112 init-specs 4 random-rangef) (-> obj ripple-size)) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 112) - s4-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 115) - s4-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system *sp-particle-system-3d* (-> *part-id-table* 112) s4-0) + (launch-particles :system *sp-particle-system-3d* (-> *part-id-table* 115) s4-0) ) ) ) @@ -786,23 +751,9 @@ (* 0.00012207031 (vector-xz-length (-> obj process root transv))) ) ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 110) - s4-1 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 110) s4-1) (set! (-> *part-id-table* 111 init-specs 16 initial-valuef) (-> obj surface-height)) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 111) - s4-1 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 111) s4-1) ) ) (let ((f30-3 (- (+ (-> obj base-height) (-> obj ocean-offset) (-> obj bob-offset) (-> obj align-offset)) @@ -990,14 +941,7 @@ (set! (-> *part-id-table* 145 init-specs 8 initial-valuef) (* 0.05 (- (-> a2-15 x) (-> obj drip-old-pos x)))) (set! (-> *part-id-table* 145 init-specs 9 initial-valuef) (* 0.05 (- (-> a2-15 y) (-> obj drip-old-pos y)))) (set! (-> *part-id-table* 145 init-specs 10 initial-valuef) (* 0.05 (- (-> a2-15 z) (-> obj drip-old-pos z)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 145) - a2-15 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 145) a2-15) ) (set! (-> obj drip-time) (-> *display* base-frame-counter)) (logclear! (-> obj flags) (water-flags wt15)) diff --git a/goal_src/engine/gfx/wind.gc b/goal_src/engine/gfx/wind.gc index 0ee8bfc63c..acef24b799 100644 --- a/goal_src/engine/gfx/wind.gc +++ b/goal_src/engine/gfx/wind.gc @@ -21,9 +21,9 @@ (set! (-> arg0 wind-time) (+ (-> arg0 wind-time) 1)) (let* ((s4-0 (logand (-> arg0 wind-time) 63)) (f0-4 (rand-vu-float-range 0.0 100.0)) - (v1-5 (/ (-> arg0 wind-time) (the-as uint 120))) + (v1-5 (/ (if (or (eq? (-> *setting-control* current video-mode) '150fps) (eq? (-> *setting-control* current video-mode) '100fps)) (the uint (min 1.0 (* (if (eq? (-> *setting-control* current video-mode) '150fps) 0.4 0.6) (-> arg0 wind-time)))) (-> arg0 wind-time)) (the-as uint 120))) (f1-6 (* 0.008333334 - (the float (mod (-> arg0 wind-time) (the-as uint 120))) + (the float (mod (if (or (eq? (-> *setting-control* current video-mode) '150fps) (eq? (-> *setting-control* current video-mode) '100fps)) (the uint (min 1.0 (* (if (eq? (-> *setting-control* current video-mode) '150fps) 0.4 0.6) (-> arg0 wind-time)))) (-> arg0 wind-time)) (the-as uint 120))) ) ) (f2-4 (* 0.0625 diff --git a/goal_src/engine/level/level-info.gc b/goal_src/engine/level/level-info.gc index c8761e7778..5864764e72 100644 --- a/goal_src/engine/level/level-info.gc +++ b/goal_src/engine/level/level-info.gc @@ -1972,3 +1972,44 @@ +;;;;;;;;; CUSTOM LEVELS +(define test-zone (new 'static 'level-load-info + :index 26 + :name 'test-zone + :visname 'test-zone-vis ;; name + -vis + :nickname 'tsz ;; nickname + :packages '() + :sound-banks '() + :music-bank #f + :ambient-sounds '() + :mood '*default-mood* + :mood-func 'update-mood-default + :ocean #f + :sky #t + :continues '((new 'static 'continue-point + :name "test-zone-start" + :level 'test-zone + :trans (new 'static 'vector :x 0.0 :y (meters 10.) :z (meters 10.) :w 1.0) + :quat (new 'static 'quaternion :w 1.0) + :camera-trans (new 'static 'vector :x 0.0 :y 4096.0 :z 0.0 :w 1.0) + :camera-rot (new 'static 'array float 9 1.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0) + :load-commands '() + :vis-nick 'none + :lev0 'test-zone + :disp0 'display + :lev1 'village1 + :disp1 'display + )) + :tasks '() + :priority 100 + :load-commands '() + :alt-load-commands '() + :bsp-mask #xffffffffffffffff + :bsphere (new 'static 'sphere :w 167772160000.0) + :bottom-height (meters -20) + :run-packages '() + :wait-for-load #t + ) + ) + +(cons! *level-load-list* 'test-zone) \ No newline at end of file diff --git a/goal_src/engine/level/level.gc b/goal_src/engine/level/level.gc index 7b3c549f74..371e0fcf8a 100644 --- a/goal_src/engine/level/level.gc +++ b/goal_src/engine/level/level.gc @@ -93,13 +93,13 @@ "Draw a level!" ;; do the draw (draw arg0 arg0 arg3) - + (if (nonzero? *display-strip-lines*) (debug-draw arg0 arg0 arg3) ) (none) ) - + (defmethod print level ((obj level)) "print a level." @@ -115,7 +115,7 @@ (defmethod relocate bsp-header ((obj bsp-header) (dest-heap kheap) (name (pointer uint8))) "Handle a bsp file load." - + ;; we expect that we'll have a loading-level set when we link/login a bsp-header (let ((s5-0 (-> *level* loading-level))) (if s5-0 @@ -175,7 +175,7 @@ (defmethod vis-clear level ((obj level)) "Clear the visibility info for when the level is loading." - + ;; clear vis-infos, so we can't try to look up a vis string. (countdown (v1-0 8) (nop!) ;; the usual. @@ -194,11 +194,11 @@ (defmethod vis-load level ((obj level)) "Start the initial load of a VIS file to the IOP VIS buffer. After this is done, we can use ramdisk-load to load chunks." - - ;; check to see if we have a buffer for loaded vis data. + + ;; check to see if we have a buffer for loaded vis data. (when (zero? (-> obj vis-info (-> obj vis-self-index) ramdisk)) ;; nope, we have no vis data buffer, we need to set it up. - + ;; first, we should see if the other level has loaded vis. if so, kill it. (let ((vis (-> obj other vis-info (-> obj other vis-self-index)))) (when (and vis (nonzero? (-> vis ramdisk))) @@ -207,7 +207,7 @@ 0 ) ) - + ;; set up a ramdisk rpc (fill command, actually load the file from DVD to IOP buffer) (let ((visname (make-file-name (file-kind vis) (the-as string (-> obj nickname)) 0 #f)) (cmd (the-as ramdisk-rpc-fill (add-element *ramdisk-rpc*))) @@ -222,7 +222,7 @@ (set! (-> obj vis-info (-> obj vis-self-index) ramdisk) s5-0) ) ) - + ;; return the ramdisk ID. (-> obj vis-info (-> obj vis-self-index) ramdisk) ) @@ -278,7 +278,7 @@ ) ) ) - + ;; check for up to 6 neighbor level vis info. The last one is always left as null. (dotimes (s5-1 6) (let* ((s3-0 (+ s5-1 1)) @@ -425,7 +425,7 @@ (('alive) (when (and *dproc* (= want-status 'active)) ;; only if we want to do alive -> active - + ;; will set the level to be drawn. (remove-by-param1 *background-draw-engine* (-> obj bsp)) (add-connection *background-draw-engine* *dproc* (the (function object object object object object) add-bsp-drawable) (-> obj bsp) obj #f) @@ -539,10 +539,10 @@ ;; set the level heap. level code logins called from linker may allocate here (set! loading-level (-> obj heap)) - + ;; relocate method of the bsp will look for this (set! (-> *level* loading-level) obj) - + ;; clear out old stuff (set! (-> *level* log-in-level-bsp) #f) (set! (-> obj nickname) #f) @@ -551,12 +551,12 @@ (set! (-> obj ambient) #f) (set! (-> obj linking) #f) (vis-clear obj) - + (set! (-> obj status) 'loading) - + ;; incoming textures should use the level allocator (set! (-> *texture-pool* allocate-func) texture-page-level-allocate) - + ;; build name (if (= (-> obj load-name) (-> obj info visname)) (format (clear *temp-string*) "~S" (-> obj info nickname)) @@ -564,19 +564,19 @@ ) (set! (-> *temp-string* data 8) (the-as uint 0)) (format *temp-string* ".DGO") - + ;; reset temporary allocations on level heap (set! (-> obj heap top) (-> obj heap top-base)) - + ;; allocate DGO loading buffers (let ((s4-0 (kmalloc (-> obj heap) (* 2 1024 1024) (kmalloc-flags align-64 top) "dgo-level-buf-2")) (s5-2 (kmalloc (-> obj heap) (* 2 1024 1024) (kmalloc-flags align-64 top) "dgo-level-buf-2")) ) (load-dbg " DGO buffers at #x~X #x~X~%" s4-0 s5-2) - + ;; we expect to load code first, remember where the heap is now. (set! (-> obj code-memory-start) (-> obj heap current)) - + (format 0 "-----------> begin load ~A [~S]~%" (-> obj load-name) *temp-string*) ;; kick off the load! (dgo-load-begin *temp-string* s5-2 s4-0 (the pointer (align64 (-> obj heap current)))) @@ -586,10 +586,10 @@ (defmethod login-begin level ((obj level)) "Start the login. This is spread over multiple frames." - + ;; done with load, reset the texture page allocator (set! (-> *texture-pool* allocate-func) texture-page-default-allocate) - + (cond ((-> obj bsp) (set! (-> *level* log-in-level-bsp) (-> obj bsp)) @@ -605,7 +605,7 @@ ) ) ) - + ;; set the login state machine at the beginning. (set! (-> *login-state* state) -1) (set! (-> *login-state* pos) (the-as uint 0)) @@ -633,16 +633,16 @@ (sv-16 prototype-bucket-tie) (sv-32 int) ) - + ;; there is some logic for not doing the whole login all at once... ;; for now, we will somewhat ignore that. - - + + (let ((level-drawable-trees (-> loaded-level bsp drawable-trees))) ;;(.mfc0 initial-timer Count) (label cfg-1) ;;(.mfc0 current-timer Count) - + ;; this would quit the login function after some amount of time elapsed. #| (let ((elapsed-timer (- current-timer initial-timer))) @@ -652,16 +652,16 @@ ) ) |# - + (let ((current-login-pos (the-as int (-> level-login-state pos)))) - + ;; Login state -1. ;; in this state, we log in drawables/art-groups that are in referenced in the bsp directly ;; the current-login-pos in the index of the drawable/art to login. - + (when (= (-> level-login-state state) -1) ;;(load-dbg "login state -1~%") - + ;; login some drawables. (when (< current-login-pos (-> level-drawable-trees length)) (let ((current-drawable (-> level-drawable-trees trees (the-as uint current-login-pos)))) @@ -699,7 +699,7 @@ (+! (-> level-login-state pos) 1) (goto cfg-1) ) - + ;; this makes the art groups go at the end. (let ((v1-39 (- (the-as uint current-login-pos) (the-as uint (-> level-drawable-trees length))))) (when (< (the-as int v1-39) (-> loaded-level art-group art-group-array length)) @@ -713,14 +713,14 @@ (goto cfg-1) ) ) - + ;; if we got here, we're done with state -1! (set! (-> level-login-state pos) (the-as uint 0)) (set! (-> level-login-state state) 0) (goto cfg-1) ) - - + + ;; login state 0. ;; we log in children of the drawables from state -1. (when (< (-> level-login-state state) (the-as int (-> level-login-state elts))) @@ -755,7 +755,7 @@ (when (< current-login-pos (-> s1-2 length)) (set! sv-16 (-> s1-2 array-data (the-as uint current-login-pos))) (set! sv-32 0) - + (#when PC_PORT ;; if a TIE uses environment mapping, disable the fade out so it always renderers with ;; the generic renderer. In the port, we just make envmapped things always envmap. @@ -763,7 +763,7 @@ (*! (-> sv-16 envmap-fade-far) 10000.) ) ) - + (while (< sv-32 4) (let ((a0-28 (-> sv-16 geometry sv-32))) ;;(load-dbg " login geom: ~A~%" a0-28) @@ -810,8 +810,8 @@ ) (goto cfg-1) ) - - + + (when (= (-> level-login-state state) (-> level-login-state elts)) (let ((v1-115 (-> loaded-level bsp))) (cond @@ -837,8 +837,7 @@ ) ) ) - - + ;; done! (set! (-> loaded-level nickname) (-> loaded-level bsp nickname)) (if (nonzero? (-> loaded-level bsp nodes)) @@ -857,7 +856,7 @@ (set! (-> *subdivide-settings* close 3) f0-0) (set! (-> *subdivide-settings* far 3) f1-0) ) - + (load-dbg "init-vis~%") (init-vis loaded-level) (load-dbg "package load~%") @@ -905,31 +904,31 @@ (case (-> obj status) (('active 'alive) (format 0 "----------- kill ~A (status ~A)~%" obj (-> obj status)) - + ;; copy data from the level to the game-info storage. This will remember permanent level stuff, like ;; what you collected/completed. (copy-perms-from-level! *game-info* obj) (send-event *camera* 'level-deactivate (-> obj name)) (send-event *target* 'level-deactivate (-> obj name)) - + ;; remove this BSP from the engine. This will stop us from being drawn. (remove-by-param1 *background-draw-engine* (-> obj bsp)) - + ;; track down all the entities and kill them (deactivate-entities (-> obj bsp)) - + ;; kill any remaining particles not associated with a part-tracker (kill-all-particles-in-level obj) - + ;; clean up our level (set! (-> obj inside-sphere?) #f) (set! (-> obj inside-boxes?) #f) (set! (-> obj meta-inside?) #f) (set! (-> obj force-inside?) #f) - + ;; we're still loaded. (set! (-> obj status) 'loaded) - + (set! (-> obj all-visible?) 'loading) ;; clear vis buffers (dotimes (v1-19 128) @@ -955,10 +954,10 @@ (defmethod unload! level ((obj level)) "Unloads the level. This does not free the heap. The level will be made inactive and ready to be loaded some other time." - + (deactivate obj) (when (!= (-> obj status) 'inactive) - + ;; if we linked art group, unlink it. (when (or (= (-> obj status) 'loaded) (= (-> obj status) 'alive) @@ -973,7 +972,7 @@ ) ) ) - + ;; turn some things off (set! (-> obj bsp) #f) (set! (-> obj entity) #f) @@ -981,7 +980,7 @@ (set! (-> obj status) 'inactive) (set! (-> obj art-group string-array length) 0) (set! (-> obj art-group art-group-array length) 0) - + ;; unload texture pages (countdown (s5-1 (-> obj loaded-texture-page-count)) (dotimes (v1-27 32) @@ -993,10 +992,10 @@ ) (set! (-> obj loaded-texture-page-count) 0) (unlink-textures-in-heap! *texture-page-dir* (-> obj heap)) - + ;; unload particle groups that were defined in the level data (unlink-part-group-by-heap (-> obj heap)) - + ;; if there are any in-progress art loads for this level, kill them. (dotimes (s5-2 2) (let ((v1-41 (-> *art-control* buffer s5-2 pending-load-file))) @@ -1019,7 +1018,7 @@ (a0-29 (car s5-3)) ) (while (not (null? s5-3)) - (case (rtype-of a0-29) + (case (rtype-of a0-29) ((symbol) (unload (symbol->string (the-as symbol a0-29))) ) @@ -1033,7 +1032,7 @@ ) (vis-clear obj) - + ;; reset the level heap! (let ((v1-64 (-> obj heap))) (set! (-> v1-64 current) (-> v1-64 base)) @@ -1057,14 +1056,14 @@ ;; note : pc port added option to show every actor regardless (with-pc (if (-> *pc-settings* force-actors?) (return #t))) - + ;; check the vis bits! - (let* (;; lwu v1, 388(a0) + (let* (;; lwu v1, 388(a0) (vis-data (-> obj vis-bits)) ;; sra a0, a1, 3 (byte-idx (sar arg0 3)) ;; daddu v1, a0, v1 - ;; lb v1, 0(v1) + ;; lb v1, 0(v1) (vis-byte (-> (the (pointer int8) vis-data) byte-idx)) ;; andi a0, a1, 7 (bit-idx (logand arg0 #b111)) @@ -1117,7 +1116,7 @@ (defmethod debug-print-splitbox level ((obj level) (arg0 vector) (arg1 string)) "Print the current splitbox, if we're in one." (cond - ((or (not (-> obj bsp)) (zero? (-> obj bsp boxes)) (zero? (-> obj bsp split-box-indices))) + ((or (not (-> obj bsp)) (zero? (-> obj bsp boxes)) (zero? (-> obj bsp split-box-indices))) ;; do nothing! ) (else @@ -1431,6 +1430,11 @@ ) (defun bg ((level-name symbol)) + "Begin game in a given level. + The level name can be the full name (village3), the nickname (vi3), or visname (village3-vis) + If the visname is used (and its a recognized level in level-info), it will use vis mode. + Otherwise, it will use the non-vis DGO name (like VILLAGE3.DGO) which will usually fail. + " (set! *cheat-mode* (if *debug-segment* 'debug #f @@ -1496,6 +1500,62 @@ 0 ) +(defun bg-custom ((level-name symbol)) + "Modified version of bg for the PC Port custom levels." + + ;; lookup info + (format 0 "(bg-custom ~A)%" level-name) + (let ((lev-info (lookup-level-info level-name))) + (when (= lev-info default-level) + (format 0 "Unable to (bg-custom ~A), the level was not found in *level-load-list*~%" level-name) + (return #f) + ) + + ;; kill jak (rip) + (format 0 "doing stop~%") + (stop 'play) + + ;; enable visiblity. the custom level won't use it, but we want it on so other levels can be loaded. + (set! (-> *level* vis?) #t) + + ;; disable border and play mode to prevent loading levels + (set! (-> *level* border?) #f) + (set! (-> *setting-control* default border-mode) #f) + (set! (-> *level* play?) #f) + + (format 0 "doing level load~%") + ;; allocate level. This may start the loading process, but won't finish it. + (let ((lev (level-get-for-use *level* level-name 'active))) + (when (not lev) + (format 0 "Unable to load level, could not level-get-for-use~%") + (return #f) + ) + (format 0 "about to start load loop, game will freeze and hopefully come back soon~%") + + ;; spin in a loop and load it. This will cause the game to freeze during the load, + ;; but this is good enough for now. + (while (or (= (-> lev status) 'loading) + (= (-> lev status) 'loading-bt) + (= (-> lev status) 'login) + ) + (load-continue lev) + ) + + (when (not (-> lev info continues)) + (format 0 "level info has no continues, can't load it.~%") + ) + + (let ((cont (car (-> lev info continues)))) + (start 'play (the continue-point cont)) + ) + + (vis-load lev) + (set! (-> lev all-visible?) #f) + (set! (-> lev force-all-visible?) #t) + ) + ) + ) + (defun play ((use-vis symbol) (init-game symbol)) "The entry point to the actual game! This allocates the level heaps, loads some data, sets some default parameters and sets the startup level." @@ -1817,11 +1877,11 @@ ) ) ) - + ;; load vis info. ;; The load-state's vis-nick is the level we want vis data for. ;; Note that we won't load vis until we are inside the level's boxes. - + ;; this will be the level that is currently being used. (let ((s5-3 #f)) (dotimes (v1-121 (-> *level* length)) @@ -1833,7 +1893,7 @@ ) ) ) - + ;; if we have the wrong vis (when (and (!= s5-3 (-> obj vis-nick)) (-> *level* vis?)) ;; and we want a vis @@ -1859,23 +1919,23 @@ ;; method 16 level-group (debug text stuff) (defmethod level-update level-group ((obj level-group)) - + ;; this does nothing... (camera-pos) (new 'static 'boxed-array :type symbol :length 0 :allocated-length 2) - + ;; compute the settings for this frame (update *setting-control*) - + ;; run the art loading system (update *art-control* #t) (clear-rec *art-control*) - + ;; run level loading! (dotimes (s5-0 2) (load-continue (-> obj level s5-0)) ) - + ;; compute inside for each level (dotimes (s5-1 (-> obj length)) (let ((s4-0 (-> obj level s5-1))) @@ -1889,10 +1949,9 @@ ) ) ) - + ;; update load state machine (the level-border one) (update! *load-state*) - ;; checkpoint assignment (dotimes (s5-2 (-> obj length)) (let ((s4-1 (-> obj level s5-2))) @@ -1927,7 +1986,7 @@ ) ) ) - + ;; determine vis info idx for each level (dotimes (v1-67 (-> obj length)) (let ((a0-26 (-> obj level v1-67))) @@ -1950,7 +2009,7 @@ ) ) ) - + ;; display level vis info (when *display-level-border* (dotimes (s5-3 (-> obj length)) @@ -1978,7 +2037,7 @@ ) ) ) - + ;; if we have vis for level A, but we aren't "in" it, display an error and ;; force us out of the other level. Ideally the boxes and the load boundary system ;; will be consistent and there is no way to set a vis to a level that we aren't in. @@ -2002,7 +2061,6 @@ ) ) ) - ;; if we are outside of the boxes, we consider ourselves "outside of bsp" ;; if we are outside of both levels boxes, then we don't really know what to do ;; for vis, and we can display the classic "outside of bsp" error. @@ -2041,7 +2099,7 @@ ) ) ) - + ;; now, handle setting bit 31 (maybe single vis mode?) (cond ;; special display self mode. @@ -2074,7 +2132,6 @@ ) ) ) - (when (or *display-level-border* *display-texture-download* *display-split-box-info*) (when *display-level-border* (format *stdcon* " want: ~A ~A/~A ~A ~A/~A~%" @@ -2144,13 +2201,13 @@ ) ) ) - + ;; tell PC port about our levels - (__pc-set-levels + (__pc-set-levels (if (= (-> obj level0 status) 'inactive) "none" (symbol->string (-> obj level0 nickname))) (if (= (-> obj level1 status) 'inactive) "none" (symbol->string (-> obj level1 nickname))) ) - + 0 ) diff --git a/goal_src/engine/load/loader.gc b/goal_src/engine/load/loader.gc index 84e99c6e15..68b1c635ee 100644 --- a/goal_src/engine/load/loader.gc +++ b/goal_src/engine/load/loader.gc @@ -892,7 +892,7 @@ (set! sv-56 0) (set! spool-sound (new-sound-id)) (backup-load-state-and-set-cmds *load-state* (-> arg0 command-list)) - (set-setting! 'spooling (the-as symbol (process->ppointer self)) 0.0 0) + (set-setting! 'spooling (process->ppointer self) 0.0 0) (logior! (-> self skel status) (janim-status inited drawn done)) (kill-current-level-hint '() '() 'die) (level-hint-surpress!) @@ -900,16 +900,9 @@ (when (or (handle->process (-> *art-control* spool-lock)) (!= *master-mode* 'game)) (cond (arg1 - (when (!= (if (> (-> self skel active-channels) 0) - (-> self skel root-channel 0 frame-group) - ) - arg1 - ) + (when (!= (ja-group) arg1) (ja-channel-push! 1 (seconds 0.05)) - (let ((s2-0 (-> self skel root-channel 0))) - (joint-control-channel-group-eval! s2-0 arg1 num-func-identity) - (set! (-> s2-0 frame-num) 0.0) - ) + (ja :group! arg1 :num! min) ) ) (else @@ -923,12 +916,9 @@ ) (spool-push *art-control* (-> arg0 name) spool-part self -9.0) (suspend) - (when arg1 - (let ((a0-17 (-> self skel root-channel 0))) - (set! (-> a0-17 param 0) 1.0) - (joint-control-channel-group-eval! a0-17 (the-as art-joint-anim #f) num-func-loop!) + (if arg1 + (ja :num! (loop!)) ) - ) ) ) (let ((v1-46 (process->ppointer self))) @@ -936,7 +926,7 @@ (new 'static 'handle :process v1-46 :pid (-> (the-as process (-> v1-46 0)) pid)) ) ) - (set! sv-48 (-> *display* base-frame-counter)) + (set! sv-48 (the-as int (-> *display* base-frame-counter))) (while (< spool-part (-> arg0 parts)) (spool-push *art-control* (-> arg0 name) spool-part self -20.0) (update *art-control* #f) @@ -944,16 +934,9 @@ (when (!= (file-status *art-control* (-> arg0 name) spool-part) 'active) (cond (arg1 - (when (!= (if (> (-> self skel active-channels) 0) - (-> self skel root-channel 0 frame-group) - ) - arg1 - ) + (when (!= (ja-group) arg1) (ja-channel-set! 1) - (let ((s2-2 (-> self skel root-channel 0))) - (joint-control-channel-group-eval! s2-2 arg1 num-func-identity) - (set! (-> s2-2 frame-num) 0.0) - ) + (ja :group! arg1 :num! min) ) ) (else @@ -967,12 +950,9 @@ (spool-push *art-control* (-> arg0 name) spool-part self -20.0) (format #t "WARNING: ---------------------> loader stall on art ~S ~D~%" (-> arg0 name) spool-part) (suspend) - (when arg1 - (let ((a0-37 (-> self skel root-channel 0))) - (set! (-> a0-37 param 0) 1.0) - (joint-control-channel-group-eval! a0-37 (the-as art-joint-anim #f) num-func-loop!) + (if arg1 + (ja :num! (loop!)) ) - ) ) ) (spool-push *art-control* (-> arg0 name) spool-part self -20.0) @@ -980,28 +960,39 @@ (cond (s2-4 (ja-channel-set! 1) - (let ((a0-42 (-> self skel root-channel 0))) - (set! (-> a0-42 frame-group) s2-4) - (set! (-> a0-42 param 0) (the float (+ (-> s2-4 data 0 length) -1))) - (set! (-> a0-42 param 1) 1.0) - (set! (-> a0-42 frame-num) 0.0) - (joint-control-channel-group! a0-42 s2-4 num-func-seek!) - ) + (ja-no-eval :group! s2-4 :num! (seek!) :frame-num 0.0) (when (zero? spool-part) (str-play-async (-> arg0 name) spool-sound) (set! (-> *art-control* active-stream) (-> arg0 name)) ) - (let* ((f30-0 (* 0.05859375 (-> s2-4 speed))) + ;; When running at higher than 60fps, Jak's idle animations are spedup but no other animations appear to be + ;; We couldn't figure out a simple way to detect these animations, so this is checking against the idle animation names + (let* ((f30-0 (* (if (and (or (= (-> *setting-control* current video-mode) '150fps) (= (-> *setting-control* current video-mode) '100fps)) + (or (string= (-> arg0 name) "eichar-ambient-1") + (string= (-> arg0 name) "eichar-ambient-2") + (string= (-> arg0 name) "eichar-ambient-3") + (string= (-> arg0 name) "eichar-ambient-4") + (string= (-> arg0 name) "fishermans-boat-ride-to-village1-alt") + (string= (-> arg0 name) "fishermans-boat-ride-to-village1") + (string= (-> arg0 name) "fishermans-boat-ride-to-misty") + (string= (-> arg0 name) "gondola-ride-up") + (string= (-> arg0 name) "gondola-ride-down") + ) + ) + (if (= (-> *setting-control* current video-mode) '150fps) 0.4 0.6) + 1.0 + ) + 0.05859375 (-> s2-4 speed))) (f28-0 (+ sv-24 (/ (the float (+ (-> s2-4 data 0 length) -1)) f30-0))) ) (set! sv-72 (current-str-pos spool-sound)) - (set! sv-40 (-> *display* base-frame-counter)) + (set! sv-40 (the-as int (-> *display* base-frame-counter))) (until (>= (the float v0-39) f28-0) (if (= (-> self skel root-channel 0) (-> self skel channel)) (logior! (-> self skel status) (janim-status spool)) ) (if (or (arg3 self) - (and (<= sv-72 0) (>= (- (-> *display* base-frame-counter) sv-40) 1200)) + (and (<= sv-72 0) (>= (- (-> *display* base-frame-counter) sv-40) (seconds 4))) (and (< 300 sv-56) (<= sv-72 0)) ) (goto cfg-88) @@ -1015,22 +1006,17 @@ (cond ((and (< sv-32 sv-72) (= (current-str-id) spool-sound)) (set! sv-56 (+ sv-56 (- (-> *display* base-frame-counter) (-> *display* old-base-frame-counter)))) - (set! sv-40 (-> *display* base-frame-counter)) + (set! sv-40 (the-as int (-> *display* base-frame-counter))) ) (else 0 ) ) (set! sv-32 sv-72) - (set! sv-48 (-> *display* base-frame-counter)) + (set! sv-48 (the-as int (-> *display* base-frame-counter))) (suspend) - (let ((f0-14 (* (- (the float (current-str-pos spool-sound)) sv-24) f30-0)) - (a0-69 (-> self skel root-channel 0)) - ) - (set! (-> a0-69 param 0) (the float (+ (-> a0-69 frame-group data 0 length) -1))) - (set! (-> a0-69 param 1) 1.0) - (set! (-> a0-69 frame-num) f0-14) - (joint-control-channel-group! a0-69 (the-as art-joint-anim #f) num-func-seek!) + (let ((f0-14 (* (- (the float (current-str-pos spool-sound)) sv-24) f30-0))) + (ja-no-eval :num! (seek!) :frame-num f0-14) ) (set! v0-39 (current-str-pos spool-sound)) (set! sv-72 v0-39) @@ -1055,34 +1041,28 @@ (defbehavior ja-abort-spooled-anim process-drawable ((arg0 spool-anim) (arg1 art-joint-anim) (arg2 int)) "Abort a spooled animation." - (restore-load-state-and-cleanup *load-state*) (str-play-stop (-> arg0 name)) (set! (-> *art-control* active-stream) #f) (logclear! (-> self skel status) (janim-status drawn done)) (if (zero? (logand (-> self skel status) (janim-status inited))) - (logclear! (-> self skel status) (janim-status inited)) - ) + (logclear! (-> self skel status) (janim-status inited)) + ) (remove-setting! 'spooling) (cond - ((and arg1 (>= arg2 0)) - (ja-channel-push! 1 (seconds 0.1)) - (set! (-> self skel root-channel 0 frame-group) arg1) - (while (!= (-> self skel root-channel 0) (-> self skel channel)) - (spool-push *art-control* (-> arg0 name) arg2 self -20.0) - (suspend) - ;; TODO macro - (let ((a0-12 (-> self skel root-channel 0))) - (set! (-> a0-12 param 0) (the float (+ (-> a0-12 frame-group data 0 length) -1))) - (set! (-> a0-12 param 1) 1.0) - (joint-control-channel-group-eval! a0-12 (the-as art-joint-anim #f) num-func-seek!) + ((and arg1 (>= arg2 0)) + (ja-channel-push! 1 (seconds 0.1)) + (set! (-> self skel root-channel 0 frame-group) arg1) + (while (!= (-> self skel root-channel 0) (-> self skel channel)) + (spool-push *art-control* (-> arg0 name) arg2 self -20.0) + (suspend) + (ja :num! (seek!)) ) + ) + (else + (ja-channel-set! 0) ) ) - (else - (ja-channel-set! 0) - ) - ) (set! (-> *art-control* spool-lock) (the-as handle #f)) 0 ) diff --git a/goal_src/engine/sparticle/sparticle.gc b/goal_src/engine/sparticle/sparticle.gc index c380ff8a36..3f936a2dc9 100644 --- a/goal_src/engine/sparticle/sparticle.gc +++ b/goal_src/engine/sparticle/sparticle.gc @@ -127,6 +127,12 @@ (new 'global 'sparticle-system 256 0 #t (-> *sprite-array-3d* vec-data) (-> *sprite-array-3d* adgif-data)) ) +(defmacro launch-particles (&key (system *sp-particle-system-2d*) particle origin &key (launch-state (the-as sparticle-launch-state #f)) &key (launch-control (the-as sparticle-launch-control #f)) + &key (rate (if (= (-> *setting-control* current video-mode) '150fps) + (if (= (-> *setting-control* current video-mode) '100fps) 0.6 0.4) + 1.0))) + `(sp-launch-particles-var ,system ,particle ,origin ,launch-state ,launch-control ,rate) + ) ;;;;;;;;;;;;;;;;;;;; @@ -676,46 +682,9 @@ (none) ) -(defun set-particle-frame-time ((arg0 int)) - (cond - ((= arg0 5) - (set-vector! - *sp-frame-time* - 0.00000000000000000000000000000000000001175495 - 5.0 - 1.0 - 1.0 - ) - ) - ((= arg0 6) - (set-vector! - *sp-frame-time* - 0.000000000000000000000000000000000000011754952 - 6.0 - 1.2 - 1.2 - ) - ) - ((= arg0 10) - (set-vector! - *sp-frame-time* - 0.000000000000000000000000000000000000011754958 - 10.0 - 2.0 - 2.0 - ) - ) - ((= arg0 12) - (set-vector! - *sp-frame-time* - 0.00000000000000000000000000000000000001175496 - 12.0 - 2.4 - 2.4 - ) - ) - ) - 0 +(defun set-particle-frame-time ((scaled-seconds int)) + "Adjusts particle frame time based on the frame-rate (scaled-seconds). Note: This used to be a case statement and has been rewritten as a more generic formula" + (set-vector! *sp-frame-time* (the-as float (logior #x800000 scaled-seconds)) (the float scaled-seconds) (* 0.2 scaled-seconds) (* 0.2 scaled-seconds)) (none) ) diff --git a/goal_src/engine/target/target-death.gc b/goal_src/engine/target/target-death.gc index e96242225c..837d1cb9a2 100644 --- a/goal_src/engine/target/target-death.gc +++ b/goal_src/engine/target/target-death.gc @@ -180,9 +180,11 @@ ) ) (let ((s5-2 (level-get *level* (-> arg0 level)))) - (when s5-2 + + ;; vis info check added for PC, don't bother waiting for vis if the level doesn't have it. + (when (and s5-2 (-> s5-2 vis-info 0)) (while (and (= (-> s5-2 all-visible?) 'loading) (-> *level* vis?)) - (suspend) + (suspend) ) ) ) diff --git a/goal_src/engine/target/target-part.gc b/goal_src/engine/target/target-part.gc index b016d29e38..e658c3d855 100644 --- a/goal_src/engine/target/target-part.gc +++ b/goal_src/engine/target/target-part.gc @@ -2172,14 +2172,7 @@ ) ) ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2002) - a2-3 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - (the-as float 1.0) - ) + (launch-particles (-> *part-id-table* 2002) a2-3) ) (suspend) 0 diff --git a/goal_src/engine/target/target.gc b/goal_src/engine/target/target.gc index 09d170fd93..2f051ba5aa 100644 --- a/goal_src/engine/target/target.gc +++ b/goal_src/engine/target/target.gc @@ -809,7 +809,12 @@ ) (until (ja-done? 0) (suspend) - (ja :num! (seek! max (/ (* (fmax 20480.0 (-> self control unknown-float01)) (-> *display* seconds-per-frame)) + ;; This controls the slow walk animation that occurs after landing from a jump. > 60fps was broken here so I think should not be tied to seconds-per-frame + ;; unknown-float01 is the magnitude of xz velocity + (ja :num! (seek! max (/ (* (fmax 20480.0 (-> self control unknown-float01)) (if (or (= (-> *setting-control* current video-mode) '150fps) + (= (-> *setting-control* current video-mode) '100fps)) + 0.016666668 + (-> *display* seconds-per-frame))) (/ (-> *TARGET-bank* run-up-cycle-dist) (-> *TARGET-bank* run-cycle-length)) ) ) @@ -879,7 +884,11 @@ (set! f30-0 (seek f30-0 (fmax 0.0 (fmin 1.0 (* 0.000048828126 (+ -16384.0 (-> self control unknown-float01))))) - (* 2.0 (-> *display* seconds-per-frame)) + ;; Jaks' walk animation felt sped up in > 60fps. I'm thinking it shouldn't be tied to seconds-per-frame but I'm not positive + (* 2.0 (if (or (= (-> *setting-control* current video-mode) '150fps) + (= (-> *setting-control* current video-mode) '100fps)) + 0.016666668 + (-> *display* seconds-per-frame))) ) ) (let ((v1-317 (-> self skel effect))) @@ -2438,7 +2447,13 @@ (seek! (-> self control dynam gravity-length) (-> self control unknown-dynamics00 gravity-length) - (* 245760.0 (-> *display* seconds-per-frame)) + (* 245760.0 (if (or (= (-> *setting-control* current video-mode) '150fps) + (= (-> *setting-control* current video-mode) '100fps) + ) + 0.016666668 + (-> *display* seconds-per-frame) + ) + ) ) ) (when (and (>= (- (-> *display* base-frame-counter) (-> self state-time)) (seconds 0.05)) @@ -2741,10 +2756,14 @@ (or (= v1-6 eichar-flop-down-loop-ja) (= v1-6 eichar-moving-flop-down-ja)) ) ) - (when (and (or (< (target-move-dist (seconds 0.1)) 1638.4) - (and (logtest? (-> self control status) (cshape-moving-flags twall)) (< 0.7 (-> self control poly-angle))) - ) - (zero? (logand (-> self control status) (cshape-moving-flags t-act))) + (when (and (or (< (target-move-dist (if (or (= (-> *setting-control* current video-mode) '150fps) (= (-> *setting-control* current video-mode) '150fps)) + (if (= (-> *setting-control* current video-mode) '150ps) (seconds 0.04) (seconds 0.06)) + (seconds 0.1))) + (if (or (= (-> *setting-control* current video-mode) '150fps) (= (-> *setting-control* current video-mode) '100fps)) (if (= (-> *setting-control* current video-mode) '150fps) 655.36 983.04) 1638.4) + ) + (and (logtest? (-> self control status) (cshape-moving-flags twall)) (< 0.7 (-> self control poly-angle))) + ) + (zero? (logand (-> self control status) (cshape-moving-flags t-act))) (>= (-> self control unknown-uint20) (the-as uint 2)) ) (set! (-> self control unknown-dword36) (-> *display* base-frame-counter)) @@ -2766,9 +2785,15 @@ (lambda ((arg0 target)) (let ((f0-3 (seek (-> arg0 control root-prim local-sphere w) (the-as float 28672.0) - (* 286720.0 (-> *display* seconds-per-frame)) + (* 286720.0 0 (if (or (= (-> *setting-control* current video-mode) '150fps) + (= (-> *setting-control* current video-mode) '100fps) + ) + 0.016666668 + (-> *display* seconds-per-frame) + ) + ) ) - ) + ) ) (set! (-> arg0 control root-prim local-sphere w) f0-3) f0-3 diff --git a/goal_src/engine/target/target2.gc b/goal_src/engine/target/target2.gc index 6e598a2035..23fc4175f8 100644 --- a/goal_src/engine/target/target2.gc +++ b/goal_src/engine/target/target2.gc @@ -224,7 +224,7 @@ ) (case (get-aspect-ratio) (('aspect4x3) - (set! (-> self sides-x-scale) 3.5) + (set! (-> self sides-x-scale) (if (= (-> *pc-settings* aspect-custom-x) 16) 5.0 3.5)) (set! (-> self sides-y-scale) 13.0) (set! (-> self x-offset) 0) 0 diff --git a/goal_src/engine/ui/progress/progress.gc b/goal_src/engine/ui/progress/progress.gc index 177105173e..04daef188a 100644 --- a/goal_src/engine/ui/progress/progress.gc +++ b/goal_src/engine/ui/progress/progress.gc @@ -701,9 +701,9 @@ (let ((f0-1 (* (1/ METER_LENGTH) (the float (-> obj in-out-position))))) (set! (-> obj particles 2 init-pos x) (the float (+ (-> obj right-x-offset) 409 (the int (* 301.5 f0-1))))) (set! (-> obj particles 1 init-pos x) (the float (+ (-> obj left-x-offset) 59))) - (set! (-> obj left-side-x-scale) (meters (+ (/ 3.5 (-> obj sides-x-scale)) (* 10.0 f0-1)))) + (set! (-> obj left-side-x-scale) (meters (+ (/ (if (= (-> *pc-settings* aspect-custom-x) 16) 5.0 3.5) (-> obj sides-x-scale)) (* 10.0 f0-1)))) (set! (-> obj left-side-y-scale) (meters (+ (-> obj sides-y-scale) (* 10.0 f0-1)))) - (set! (-> obj right-side-x-scale) (meters (+ (/ 6.0 (-> obj sides-x-scale)) (* 4.0 f0-1)))) + (set! (-> obj right-side-x-scale) (meters (+ (/ (if (= (-> *pc-settings* aspect-custom-x) 16) 8.5 6.0) (-> obj sides-x-scale)) (* 4.0 f0-1)))) (set! (-> obj right-side-y-scale) (meters (+ (-> obj sides-y-scale) (* 4.0 f0-1)))) ) (dotimes (s5-0 (-> obj nb-of-particles)) diff --git a/goal_src/examples/debug-draw-example.gc b/goal_src/examples/debug-draw-example.gc index e20e9311f4..5ef1ef25e0 100644 --- a/goal_src/examples/debug-draw-example.gc +++ b/goal_src/examples/debug-draw-example.gc @@ -102,23 +102,9 @@ ) (dotimes (i 10) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 37) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 + (launch-particles (-> *part-id-table* 37) gp-0) + (launch-particles (-> *part-id-table* 38) gp-0) ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 38) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) - ) ) ) 0 @@ -368,4 +354,4 @@ (load-state-want-display-level *debug-load-level* #t) ) ) - ) \ No newline at end of file + ) diff --git a/goal_src/examples/display-all-colors.gc b/goal_src/examples/display-all-colors.gc new file mode 100644 index 0000000000..3a068c8a43 --- /dev/null +++ b/goal_src/examples/display-all-colors.gc @@ -0,0 +1,61 @@ +;;-*-Lisp-*- +(in-package goal) + +;; This script creates a simple process that draws text demonstrating +;; all of GOAL's color constants to the on-screen debug output. + +;; Create somewhere for the handle to the process to live. See https://open-goal.github.io/docs/reference/process_and_state +;; as well as kernel/gstate.gc +(define *color-display-handle* (new 'static 'handle)) +(set! *color-display-handle* (the handle #f)) + + +(defun-debug start-display-text-colors () + "Spawn an onscreen string displaying all possible colors" + (if (not (handle->process *color-display-handle*)) + (let ((disp-proc + (process-spawn-function process :name 'display-proc + (lambda :behavior process () + (stack-size-set! (-> self main-thread) 256) + (loop + ;; These constants live in engine/gfx/font-h.gc + (format *stdcon* "~0k~%~% +~0L 0 default ~1L 1 white +~2L 2 gray ~3L 3 orange-red +~4L 4 bright-orange-red ~5L 5 bright-orange-red +~6L 6 bright-green ~7L 7 dark-blue +~8L 8 light-blue ~9L 9 dark-pink +~10L10 lighter-blue ~11L11 dark-light-blue +~12L12 dim-white ~13L13 dim-gray +~14L14 orange-red-2 ~15L15 yellow-green +~16L16 dark-green ~17L17 another-gray +~18L18 dark-dark-green ~19L19 flat-dark-purple +~20L20 flat-yellow ~21L21 blue-white +~22L22 pad-back ~23L23 pad-shine +~24L24 pad-square ~25L25 pad-circle +~26L26 pad-triangle ~27L27 pad-x +~28L28 lighter-lighter-blue ~29L29 yellow-orange +~30L30 yellow-green-2 ~31L31 another-light-blue +~32L32 light-yellow ~33L33 red-orange +~34L34 another-orange-red~0L~% + alternate names + ~3L3 red ~4L4 red2 ~5L5 yellow ~6L6 green ~7L7 blue + ~10L10 cyan ~33L33 red-reverse ~34L34 red-obverse~0L" + ) + (suspend) + ) + ) + ) + )) + (set! *color-display-handle* (ppointer->handle disp-proc)) + ) + ;; else + (format #t "Colors are already being displayed") + ) + ) + + +(defun-debug stop-display-text-colors () + "Kill the example text color display" + (kill-by-name 'display-proc *active-pool*) + ) diff --git a/goal_src/game.gp b/goal_src/game.gp index 40b7e3b588..26e536dff2 100644 --- a/goal_src/game.gp +++ b/goal_src/game.gp @@ -103,6 +103,17 @@ ) ) +(defun custom-level-cgo (output-name desc-file-name) + "Add a CGO with the given output name (in out/iso) and input name (in custom_levels/)" + (let ((out-name (string-append "out/iso/" output-name))) + (defstep :in (string-append "custom_levels/" desc-file-name) + :tool 'dgo + :out `(,out-name) + ) + (set! *all-cgos* (cons out-name *all-cgos*)) + ) + ) + (defun cgo (output-name desc-file-name) "Add a CGO with the given output name (in out/iso) and input name (in goal_src/dgos)" (let ((out-name (string-append "out/iso/" output-name))) @@ -147,6 +158,12 @@ ) ) +(defmacro build-custom-level (name) + (let* ((path (string-append "custom_levels/" name "/" name ".jsonc"))) + `(defstep :in ,path + :tool 'build-level + :out '(,(string-append "out/obj/" name ".go"))))) + (defun get-iso-data-path () (if *use-iso-data-path* (string-append *iso-data* "/") @@ -1553,6 +1570,17 @@ "ndi-volumes-ag" "title-vis") +;;;;;;;;;;;;;;;;;;;;;;;;; +;; Example Custom Level +;;;;;;;;;;;;;;;;;;;;;;;;; + +;; Set up the build system to build the level geometry +;; this path is relative to the custom_levels/ folder +;; it should point to the .jsonc file that specifies the level. +(build-custom-level "test-zone") +;; the DGO file +(custom-level-cgo "TESTZONE.DGO" "test-zone/testzone.gd") + ;;;;;;;;;;;;;;;;;;;;; ;; Game Engine Code ;;;;;;;;;;;;;;;;;;;;; diff --git a/goal_src/levels/beach/beach-part.gc b/goal_src/levels/beach/beach-part.gc index 365d775043..d199592f41 100644 --- a/goal_src/levels/beach/beach-part.gc +++ b/goal_src/levels/beach/beach-part.gc @@ -93,24 +93,10 @@ (f0-0 (vector-vector-distance (-> self root trans) gp-0)) ) (if (or (< (-> gp-0 y) (-> self root trans y)) (< 122880.0 f0-0)) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 666) - (-> self root trans) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 666) (-> self root trans)) ) ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 668) - (-> self root trans) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 668) (-> self root trans)) ) (suspend) ) diff --git a/goal_src/levels/beach/lurkerworm.gc b/goal_src/levels/beach/lurkerworm.gc index 58300d0bc1..b7d16a2ff9 100644 --- a/goal_src/levels/beach/lurkerworm.gc +++ b/goal_src/levels/beach/lurkerworm.gc @@ -241,44 +241,16 @@ (defmethod particle-effect lurkerworm ((obj lurkerworm)) (let ((a2-0 (vector<-cspace! (new 'stack-no-clear 'vector) (-> obj node-list data 5)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 661) - a2-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 661) a2-0) ) (let ((a2-1 (vector<-cspace! (new 'stack-no-clear 'vector) (-> obj node-list data 6)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 661) - a2-1 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 661) a2-1) ) (let ((a2-2 (vector<-cspace! (new 'stack-no-clear 'vector) (-> obj node-list data 7)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 661) - a2-2 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 661) a2-2) ) (let ((a2-3 (vector<-cspace! (new 'stack-no-clear 'vector) (-> obj node-list data 8)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 661) - a2-3 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 661) a2-3) ) 0 (none) diff --git a/goal_src/levels/citadel/citadel-part.gc b/goal_src/levels/citadel/citadel-part.gc index 913d405e81..671cadf8f5 100644 --- a/goal_src/levels/citadel/citadel-part.gc +++ b/goal_src/levels/citadel/citadel-part.gc @@ -63,22 +63,8 @@ (let ((gp-0 (new 'stack-no-clear 'vector))) (sp-kill-particle arg0 arg1) (set-vector! gp-0 (-> arg2 x) (-> arg2 y) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2882) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - (the-as float 1.0) - ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2883) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - (the-as float 1.0) - ) + (launch-particles :rate 1.0 (-> *part-id-table* 2882) gp-0) + (launch-particles :rate 1.0 (-> *part-id-table* 2883) gp-0) ) ) (none) diff --git a/goal_src/levels/citadel/citadel-sages.gc b/goal_src/levels/citadel/citadel-sages.gc index 0c04a53f6e..e8e9215a92 100644 --- a/goal_src/levels/citadel/citadel-sages.gc +++ b/goal_src/levels/citadel/citadel-sages.gc @@ -71,14 +71,7 @@ ) (dotimes (s2-2 12) (vector-matrix*! s5-0 (-> self bar-array s2-2) gp-0) - (sp-launch-particles-var - s3-1 - s4-1 - s5-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system s3-1 s4-1 s5-0) ) ) ) diff --git a/goal_src/levels/citadel/citb-plat.gc b/goal_src/levels/citadel/citb-plat.gc index 897526e5e6..840e4a9cd1 100644 --- a/goal_src/levels/citadel/citb-plat.gc +++ b/goal_src/levels/citadel/citb-plat.gc @@ -887,14 +887,7 @@ (dotimes (s5-0 16) (quaternion-rotate-local-z! gp-0 gp-0 2048.0) (quaternion-copy! *particle-quat* gp-0) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 2541) - (-> self blast-pos) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system *sp-particle-system-3d* (-> *part-id-table* 2541) (-> self blast-pos)) ) ) (spawn (-> self part) (-> self blast-pos)) diff --git a/goal_src/levels/finalboss/light-eco.gc b/goal_src/levels/finalboss/light-eco.gc index ee5e3013a1..2786515d1c 100644 --- a/goal_src/levels/finalboss/light-eco.gc +++ b/goal_src/levels/finalboss/light-eco.gc @@ -188,22 +188,8 @@ (let ((gp-0 (new 'stack-no-clear 'vector))) (sp-kill-particle arg0 arg1) (set-vector! gp-0 (-> arg2 x) (-> arg2 y) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2904) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2905) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :rate 1.0 (-> *part-id-table* 2904) gp-0) + (launch-particles :rate 1.0 (-> *part-id-table* 2905) gp-0) ) ) (none) @@ -321,22 +307,8 @@ (let ((gp-0 (new 'stack-no-clear 'vector))) (sp-kill-particle arg0 arg1) (set-vector! gp-0 (-> arg2 x) (-> arg2 y) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2910) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2911) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :rate 1.0 (-> *part-id-table* 2910) gp-0) + (launch-particles :rate 1.0 (-> *part-id-table* 2911) gp-0) ) ) (none) diff --git a/goal_src/levels/finalboss/sage-finalboss-part.gc b/goal_src/levels/finalboss/sage-finalboss-part.gc index 14f996ec1c..27308eb822 100644 --- a/goal_src/levels/finalboss/sage-finalboss-part.gc +++ b/goal_src/levels/finalboss/sage-finalboss-part.gc @@ -352,22 +352,8 @@ (let ((gp-0 (new 'stack-no-clear 'vector))) (sp-kill-particle arg0 arg1) (set-vector! gp-0 (-> arg2 x) (-> arg2 y) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2933) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2934) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :rate 1.0 (-> *part-id-table* 2933) gp-0) + (launch-particles :rate 1.0 (-> *part-id-table* 2934) gp-0) ) ) (none) @@ -1041,22 +1027,8 @@ (let ((gp-0 (new 'stack-no-clear 'vector))) (sp-kill-particle arg0 arg1) (set-vector! gp-0 (-> arg2 x) (-> arg2 y) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2962) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2963) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :rate 1.0 (-> *part-id-table* 2962) gp-0) + (launch-particles :rate 1.0 (-> *part-id-table* 2963) gp-0) ) ) (none) diff --git a/goal_src/levels/jungle/darkvine.gc b/goal_src/levels/jungle/darkvine.gc index 291f34c129..c00e933300 100644 --- a/goal_src/levels/jungle/darkvine.gc +++ b/goal_src/levels/jungle/darkvine.gc @@ -218,14 +218,7 @@ (ja-channel-push! 1 (seconds 0.15)) (ja-no-eval :group! darkvine-retreat-ja :num! (seek!) :frame-num 0.0) (until (ja-done? 0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 800) - (-> self root-override trans) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 800) (-> self root-override trans)) (suspend) (ja :num! (seek!)) ) diff --git a/goal_src/levels/jungle/fisher.gc b/goal_src/levels/jungle/fisher.gc index da79c80233..f68997ad64 100644 --- a/goal_src/levels/jungle/fisher.gc +++ b/goal_src/levels/jungle/fisher.gc @@ -853,14 +853,7 @@ (set! (-> *part-id-table* 118 init-specs 1 initial-valuef) 0.05) (set! (-> *part-id-table* 118 init-specs 2 initial-valuef) 0.0) (set! (-> *part-id-table* 118 init-specs 13 initial-valuef) 0.35555556) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 118) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system *sp-particle-system-3d* (-> *part-id-table* 118) gp-0) ) 0 (none) @@ -1738,42 +1731,14 @@ (when (-> self training) (let ((gp-0 (new-stack-vector0))) (vector<-cspace! gp-0 (-> self node-list data 74)) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2001) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 2001) gp-0) (fisher-fish-water gp-0 (+ 32768.0 (vector-y-angle (-> self node-list data 75 bone transform vector 1)))) (vector<-cspace! gp-0 (-> self node-list data 77)) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2001) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 2001) gp-0) (fisher-fish-water gp-0 (+ 32768.0 (vector-y-angle (-> self node-list data 78 bone transform vector 1)))) (vector<-cspace! gp-0 (-> self node-list data 80)) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 828) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2013) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 828) gp-0) + (launch-particles (-> *part-id-table* 2013) gp-0) (fisher-fish-water gp-0 (+ 32768.0 (vector-y-angle (-> self node-list data 80 bone transform vector 1)))) ) ) diff --git a/goal_src/levels/jungle/jungle-mirrors.gc b/goal_src/levels/jungle/jungle-mirrors.gc index 61c740d3d5..714d768efd 100644 --- a/goal_src/levels/jungle/jungle-mirrors.gc +++ b/goal_src/levels/jungle/jungle-mirrors.gc @@ -998,14 +998,7 @@ (when (periscope-has-power-input?) (update! (-> self sound)) (if (logtest? (-> self draw status) (draw-status was-drawn)) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 825) - (-> self reflector-trans) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 825) (-> self reflector-trans)) ) ) 0 @@ -1828,14 +1821,7 @@ (draw-power-beam gp-0 (-> self beam-end)) (update! (-> self sound)) (when (logtest? (-> self draw status) (draw-status was-drawn)) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 825) - (-> self beam-end) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 825) (-> self beam-end)) (when (and *target* (>= 24576.0 (vector-vector-distance (-> self root-override trans) (-> *target* control trans)))) (start-hint-timer (game-text-id jungle-mirrors-break-the-mirror-jak)) (level-hint-spawn diff --git a/goal_src/levels/maincave/maincave-obs.gc b/goal_src/levels/maincave/maincave-obs.gc index 56aa06a64c..9a96531ac2 100644 --- a/goal_src/levels/maincave/maincave-obs.gc +++ b/goal_src/levels/maincave/maincave-obs.gc @@ -483,22 +483,8 @@ (cond ((< gp-0 a0-1) (when (sphere-in-view-frustum? (the-as sphere (-> self root-override root-prim prim-core))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 704) - (the-as vector (-> self launch-pos)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 705) - (the-as vector (&-> self stack 112)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 704) (the-as vector (-> self launch-pos))) + (launch-particles (-> *part-id-table* 705) (the-as vector (&-> self stack 112))) ) (when (-> self should-play-sound?) (set! (-> self should-play-sound?) #f) diff --git a/goal_src/levels/maincave/maincave-part.gc b/goal_src/levels/maincave/maincave-part.gc index 696762375c..9a85d60567 100644 --- a/goal_src/levels/maincave/maincave-part.gc +++ b/goal_src/levels/maincave/maincave-part.gc @@ -293,22 +293,8 @@ (sound-play "water-drop") ) (set-vector! gp-0 (-> arg2 x) (-> arg1 user-float) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2231) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 2232) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :rate 1.0 (-> *part-id-table* 2231) gp-0) + (launch-particles :system *sp-particle-system-3d* :rate 1.0 (-> *part-id-table* 2232) gp-0) ) ) (none) diff --git a/goal_src/levels/misty/misty-obs.gc b/goal_src/levels/misty/misty-obs.gc index 056470a7c0..2657d82112 100644 --- a/goal_src/levels/misty/misty-obs.gc +++ b/goal_src/levels/misty/misty-obs.gc @@ -1394,14 +1394,7 @@ (defstate breakaway-about-to-fall (breakaway) :code (behavior () (sound-play "falling-bones") - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 281) - (-> self root-override trans) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 281) (-> self root-override trans)) (let ((gp-1 #f) (s5-1 (-> *display* base-frame-counter)) ) diff --git a/goal_src/levels/misty/muse.gc b/goal_src/levels/misty/muse.gc index 1a994dae59..be7ddcf094 100644 --- a/goal_src/levels/misty/muse.gc +++ b/goal_src/levels/misty/muse.gc @@ -150,14 +150,7 @@ (defmethod dummy-51 muse ((obj muse)) (dotimes (s5-0 2) (let ((v1-2 (rand-vu-int-range 3 (+ (-> obj node-list length) -1)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 271) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> obj node-list data v1-2)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 271) (vector<-cspace! (new 'stack-no-clear 'vector) (-> obj node-list data v1-2))) ) ) 0 diff --git a/goal_src/levels/racer_common/racer-states.gc b/goal_src/levels/racer_common/racer-states.gc index 89e8fdd180..6582b1b6be 100644 --- a/goal_src/levels/racer_common/racer-states.gc +++ b/goal_src/levels/racer_common/racer-states.gc @@ -407,14 +407,7 @@ ) (when (>= (-> self control unknown-float01) 40960.0) (set! (-> *part-id-table* 2225 init-specs 1 initial-valuef) 100.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2225) - (-> self control trans) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 2225) (-> self control trans)) ) (target-land-effect) (when (and (>= (-> self control ground-impact-vel) 61440.0) (zero? (-> self racer bounce))) diff --git a/goal_src/levels/racer_common/target-racer.gc b/goal_src/levels/racer_common/target-racer.gc index 63ecaf5c54..b882c77432 100644 --- a/goal_src/levels/racer_common/target-racer.gc +++ b/goal_src/levels/racer_common/target-racer.gc @@ -650,14 +650,10 @@ (when (< 0.0 (-> *part-id-table* 2212 init-specs 11 initial-valuef)) (set! (-> *part-id-table* 2212 init-specs 14 initial-valuef) (the-as float gp-0)) (set! (-> *part-id-table* 2212 init-specs 3 initial-valuef) 15155.2) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 2212) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> self manipy 0 node-list data 4)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system *sp-particle-system-3d* + (-> *part-id-table* 2212) + (vector<-cspace! (new 'stack-no-clear 'vector) (-> self manipy 0 node-list data 4)) + ) ) ) (let ((gp-2 (vector-y-quaternion! (new 'stack-no-clear 'vector) (-> self control unknown-quaternion00)))) @@ -670,14 +666,10 @@ (when (< 0.0 (-> *part-id-table* 2212 init-specs 11 initial-valuef)) (set! (-> *part-id-table* 2212 init-specs 14 initial-valuef) (the-as float gp-2)) (set! (-> *part-id-table* 2212 init-specs 3 initial-valuef) 10240.0) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 2212) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> self manipy 0 node-list data 10)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system *sp-particle-system-3d* + (-> *part-id-table* 2212) + (vector<-cspace! (new 'stack-no-clear 'vector) (-> self manipy 0 node-list data 10)) + ) ) ) (when (and (racer-on-ground?) (< (-> self control unknown-float01) 90112.0)) @@ -696,22 +688,8 @@ (set! (-> *part-id-table* 2275 init-specs 19 initial-valuef) (+ 49152.0 f1-3)) (set! (-> *part-id-table* 2275 init-specs 1 initial-valuef) (* 0.0000036621095 f0-17)) (set! (-> *part-id-table* 2275 init-specs 2 initial-valuef) (* 0.1 f0-17)) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 2275) - s4-2 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 2276) - s4-2 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system *sp-particle-system-3d* (-> *part-id-table* 2275) s4-2) + (launch-particles :system *sp-particle-system-3d* (-> *part-id-table* 2276) s4-2) ) ) (-> *part-id-table* 2208) @@ -753,17 +731,10 @@ ) (set! (-> s5-2 y) (-> self control shadow-pos y)) (if (nonzero? a1-13) - (sp-launch-particles-var - (if gp-4 - *sp-particle-system-3d* - *sp-particle-system-2d* - ) - a1-13 - s5-2 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system (if gp-4 *sp-particle-system-3d* *sp-particle-system-2d*) + a1-13 + s5-2 + ) ) ) (let* ((gp-5 #f) @@ -790,17 +761,10 @@ ) (set! (-> a2-7 y) (-> self control shadow-pos y)) (if (nonzero? s5-3) - (sp-launch-particles-var - (if gp-5 - *sp-particle-system-3d* - *sp-particle-system-2d* - ) - (the-as sparticle-launcher s5-3) - a2-7 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system (if gp-5 *sp-particle-system-3d* *sp-particle-system-2d*) + (the-as sparticle-launcher s5-3) + a2-7 + ) ) ) ) @@ -916,14 +880,9 @@ (when (!= (-> self racer boost-output) 0.0) (dotimes (gp-7 8) (let ((v1-258 (rand-vu-int-range 3 (+ (-> self node-list length) -1)))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2229) - (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data v1-258)) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 2229) + (vector<-cspace! (new 'stack-no-clear 'vector) (-> self node-list data v1-258)) + ) ) ) (cpad-set-buzz! (-> *cpad-list* cpads 0) 0 1 (seconds 0.1)) @@ -948,14 +907,7 @@ ) ) ) - (sp-launch-particles-var - *sp-particle-system-2d* - a1-54 - gp-8 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles a1-54 gp-8) ) (cpad-set-buzz! (-> *cpad-list* cpads 0) 0 51 (seconds 0.05)) (sound-play-by-name @@ -986,14 +938,7 @@ ) ) ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2227) - a2-28 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 2227) a2-28) ) ) 0 diff --git a/goal_src/levels/rolling/rolling-lightning-mole.gc b/goal_src/levels/rolling/rolling-lightning-mole.gc index c692629c1e..37ecb572cb 100644 --- a/goal_src/levels/rolling/rolling-lightning-mole.gc +++ b/goal_src/levels/rolling/rolling-lightning-mole.gc @@ -1157,14 +1157,7 @@ (let ((gp-0 (new 'stack-no-clear 'vector))) (sp-kill-particle arg0 arg1) (set-vector! gp-0 (-> arg2 x) (-> arg1 user-float) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 1771) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :rate 1.0 (-> *part-id-table* 1771) gp-0) ) ) (none) @@ -1178,14 +1171,7 @@ (sound-play "land-grass") ) (set-vector! gp-0 (-> arg2 x) (-> arg1 user-float) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 1772) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :rate 1.0 (-> *part-id-table* 1772) gp-0) ) ) (none) diff --git a/goal_src/levels/sunken/orbit-plat.gc b/goal_src/levels/sunken/orbit-plat.gc index bdb3c610de..e3cd038d55 100644 --- a/goal_src/levels/sunken/orbit-plat.gc +++ b/goal_src/levels/sunken/orbit-plat.gc @@ -152,14 +152,7 @@ ) (set! (-> *part-id-table* 1717 init-specs 1 random-rangef) (* 0.000009494358 f30-0)) ) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 1717) - arg0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 1717) arg0) 0 (none) ) diff --git a/goal_src/levels/sunken/sun-exit-chamber.gc b/goal_src/levels/sunken/sun-exit-chamber.gc index e3ec051794..b6041d69d3 100644 --- a/goal_src/levels/sunken/sun-exit-chamber.gc +++ b/goal_src/levels/sunken/sun-exit-chamber.gc @@ -571,14 +571,7 @@ (vector-matrix*! gp-0 gp-0 s4-0) (vector-float*! gp-0 gp-0 (/ 1.0 (-> gp-0 w))) (set! (-> *part-id-table* 2515 init-specs 13 initial-valuef) (+ 24576.0 (-> gp-0 y))) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 2515) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 2515) gp-0) ) (none) ) diff --git a/goal_src/levels/sunken/sunken-water.gc b/goal_src/levels/sunken/sunken-water.gc index 01cdde025c..8bb8022a5c 100644 --- a/goal_src/levels/sunken/sunken-water.gc +++ b/goal_src/levels/sunken/sunken-water.gc @@ -80,14 +80,7 @@ ) (set! (-> gp-0 vertex-skip) 128) (dotimes (s5-0 (-> gp-0 vertex-count)) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 1736) - (-> gp-0 data s5-0) - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles (-> *part-id-table* 1736) (-> gp-0 data s5-0)) ) ) #f diff --git a/goal_src/levels/training/training-part.gc b/goal_src/levels/training/training-part.gc index 264cc74f21..b999b2d384 100644 --- a/goal_src/levels/training/training-part.gc +++ b/goal_src/levels/training/training-part.gc @@ -293,22 +293,8 @@ (sound-play "water-drop") ) (set-vector! gp-0 (-> arg2 x) (-> arg1 user-float) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 763) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 764) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :rate 1.0 (-> *part-id-table* 763) gp-0) + (launch-particles :system *sp-particle-system-3d* :rate 1.0 (-> *part-id-table* 764) gp-0) ) ) (none) diff --git a/goal_src/levels/village1/assistant.gc b/goal_src/levels/village1/assistant.gc index b05217c073..0ffdeff23b 100644 --- a/goal_src/levels/village1/assistant.gc +++ b/goal_src/levels/village1/assistant.gc @@ -359,14 +359,7 @@ (sound-play "water-drop") ) (set-vector! gp-0 (-> arg2 x) (-> arg1 user-float) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 367) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :rate 1.0 (-> *part-id-table* 367) gp-0) ) ) (none) diff --git a/goal_src/levels/village1/fishermans-boat.gc b/goal_src/levels/village1/fishermans-boat.gc index a221ff40d3..ed5a3c28a5 100644 --- a/goal_src/levels/village1/fishermans-boat.gc +++ b/goal_src/levels/village1/fishermans-boat.gc @@ -688,14 +688,7 @@ (set! (-> *part-id-table* 2896 init-specs 19 initial-valuef) (+ 49152.0 arg1)) (set! (-> *part-id-table* 2896 init-specs 1 initial-valuef) (* 0.0000036621095 arg2)) (set! (-> *part-id-table* 2896 init-specs 2 initial-valuef) (* 0.1 arg2)) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 2896) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :system *sp-particle-system-3d* (-> *part-id-table* 2896) gp-0) ) 0 (none) diff --git a/goal_src/levels/village1/village1-part.gc b/goal_src/levels/village1/village1-part.gc index c72df280b5..bda4286fd3 100644 --- a/goal_src/levels/village1/village1-part.gc +++ b/goal_src/levels/village1/village1-part.gc @@ -1209,22 +1209,8 @@ (let ((gp-0 (new 'stack-no-clear 'vector))) (sp-kill-particle arg0 arg1) (set-vector! gp-0 (-> arg2 x) (-> arg1 user-float) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 434) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - (the-as float 1.0) - ) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 435) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - (the-as float 1.0) - ) + (launch-particles :rate 1.0 (-> *part-id-table* 434) gp-0) + (launch-particles :system *sp-particle-system-3d* :rate 1.0 (-> *part-id-table* 435) gp-0) ) ) (none) diff --git a/goal_src/levels/village1/village1-part2.gc b/goal_src/levels/village1/village1-part2.gc index 9604e20dee..c83094d2f0 100644 --- a/goal_src/levels/village1/village1-part2.gc +++ b/goal_src/levels/village1/village1-part2.gc @@ -1613,22 +1613,8 @@ (sound-play "water-drop") ) (set-vector! gp-0 (-> arg2 x) (-> arg1 user-float) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 502) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - (the-as float 1.0) - ) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 503) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - (the-as float 1.0) - ) + (launch-particles :rate 1.0 (-> *part-id-table* 502) gp-0) + (launch-particles :system *sp-particle-system-3d* :rate 1.0 (-> *part-id-table* 503) gp-0) ) ) (none) diff --git a/goal_src/levels/village2/assistant-village2.gc b/goal_src/levels/village2/assistant-village2.gc index f11bb6a5e0..7161b86d80 100644 --- a/goal_src/levels/village2/assistant-village2.gc +++ b/goal_src/levels/village2/assistant-village2.gc @@ -687,14 +687,7 @@ (sound-play "water-drop") ) (set-vector! gp-0 (-> arg2 x) (-> arg1 user-float) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 1324) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :rate 1.0 (-> *part-id-table* 1324) gp-0) ) ) (none) diff --git a/goal_src/levels/village2/swamp-blimp.gc b/goal_src/levels/village2/swamp-blimp.gc index bfc94caec3..96aceefecb 100644 --- a/goal_src/levels/village2/swamp-blimp.gc +++ b/goal_src/levels/village2/swamp-blimp.gc @@ -1030,14 +1030,7 @@ (let ((a2-1 (new 'static 'vector))) (set! (-> a2-1 quad) (-> self root-override trans quad)) (set! (-> a2-1 y) 0.0) - (sp-launch-particles-var - *sp-particle-system-3d* - (-> *part-id-table* 2017) - a2-1 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - 1.0 - ) + (launch-particles :rate 1.0 (-> *part-id-table* 2017) a2-1) ) ) (none) diff --git a/goal_src/levels/village2/village2-part.gc b/goal_src/levels/village2/village2-part.gc index 2eb4111ae2..d05f9afcd2 100644 --- a/goal_src/levels/village2/village2-part.gc +++ b/goal_src/levels/village2/village2-part.gc @@ -1856,14 +1856,7 @@ (sound-play "water-drop") ) (set-vector! gp-0 (-> arg2 x) (-> arg1 user-float) (-> arg2 z) 1.0) - (sp-launch-particles-var - *sp-particle-system-2d* - (-> *part-id-table* 1207) - gp-0 - (the-as sparticle-launch-state #f) - (the-as sparticle-launch-control #f) - (the-as float 1.0) - ) + (launch-particles :rate 1.0 (-> *part-id-table* 1207) gp-0) ) ) (none) diff --git a/goal_src/pc_debug/pc-pad-utils.gc b/goal_src/pc_debug/pc-pad-utils.gc index 5ae03dee3b..fd00295898 100644 --- a/goal_src/pc_debug/pc-pad-utils.gc +++ b/goal_src/pc_debug/pc-pad-utils.gc @@ -104,13 +104,30 @@ (format *stdcon* "~%") ) (dotimes (ii 2) - (format *stdcon* " ~3Lcpad ~D~0L~%" ii) + (format *stdcon* " ~0k~3Lcpad ~D~0L~%" ii) (dotimes (j 8) (dotimes (i 2) - (let ( + (format *stdcon* " ") + (let* ( (btn-idx (+ i(* j 2))) + (btn-name (-> *pc-pad-button-names* btn-idx)) + (keycode (pc-pad-get-mapped-button ii btn-idx)) ) - (format *stdcon* " ~8L~S: ~0L~D " (-> *pc-pad-button-names* btn-idx) (pc-pad-get-mapped-button ii btn-idx)) + (format *stdcon* "~8L~S: " btn-name) + ;; longest button string is TRIANGLE, which is 8 characters in length + ;; but only shows up in the left column. CIRCLE and SQUARE are longest in the right with 6 + (dotimes (_ (- + (cond ((= i 0) 8) ((= i 1) 6)) + (length btn-name))) + (format *stdcon* " ") + ) + (format *stdcon* "~0L~D" keycode) + (when (< keycode 100) + (format *stdcon* " ") + (if (< keycode 10) + (format *stdcon* " ") + ) + ) ) ) (format *stdcon* "~%") diff --git a/goalc/CMakeLists.txt b/goalc/CMakeLists.txt index 24b84263bb..40aa875e4b 100644 --- a/goalc/CMakeLists.txt +++ b/goalc/CMakeLists.txt @@ -5,6 +5,16 @@ add_library(compiler emitter/ObjectGenerator.cpp emitter/Register.cpp debugger/disassemble.cpp + build_level/build_level.cpp + build_level/collide_bvh.cpp + build_level/collide_drawable.cpp + build_level/collide_pack.cpp + build_level/color_quantization.cpp + build_level/FileInfo.cpp + build_level/gltf_mesh_extract.cpp + build_level/LevelFile.cpp + build_level/ResLump.cpp + build_level/Tfrag.cpp compiler/Compiler.cpp compiler/Env.cpp compiler/Val.cpp @@ -44,7 +54,7 @@ add_library(compiler regalloc/Allocator_v2.cpp ) -target_link_libraries(compiler common Zydis) +target_link_libraries(compiler common Zydis tiny_gltf) if (WIN32) target_link_libraries(compiler mman) diff --git a/goalc/build_level/FileInfo.cpp b/goalc/build_level/FileInfo.cpp new file mode 100644 index 0000000000..1c5367143e --- /dev/null +++ b/goalc/build_level/FileInfo.cpp @@ -0,0 +1,30 @@ +#include "FileInfo.h" +#include "goalc/data_compiler/DataObjectGenerator.h" +#include "common/versions.h" + +size_t FileInfo::add_to_object_file(DataObjectGenerator& gen) const { + gen.align_to_basic(); + gen.add_type_tag("file-info"); + size_t offset = gen.current_offset_bytes(); + gen.add_type_tag(file_type); + gen.add_ref_to_string_in_pool(file_name); + gen.add_word(major_version); + gen.add_word(minor_version); + gen.add_ref_to_string_in_pool(maya_file_name); + gen.add_ref_to_string_in_pool(tool_debug); + gen.add_ref_to_string_in_pool(tool_debug); + + return offset; +} + +FileInfo make_file_info_for_level(const std::string& file_name) { + FileInfo info; + info.file_type = "bsp-header"; + info.file_name = file_name; + info.major_version = versions::jak1::LEVEL_FILE_VERSION; + info.minor_version = 0; + info.maya_file_name = "Unknown"; + info.tool_debug = "Created by OpenGOAL buildlevel"; + info.mdb_file_name = "Unknown"; + return info; +} \ No newline at end of file diff --git a/goalc/build_level/FileInfo.h b/goalc/build_level/FileInfo.h new file mode 100644 index 0000000000..7932018ca1 --- /dev/null +++ b/goalc/build_level/FileInfo.h @@ -0,0 +1,28 @@ +#pragma once + +#include + +#include "common/common_types.h" + +class DataObjectGenerator; + +struct FileInfo { + // (file-type symbol :offset-assert 4) + std::string file_type; + // (file-name basic :offset-assert 8) + std::string file_name; + // (major-version uint32 :offset-assert 12) + u32 major_version = 0; + // (minor-version uint32 :offset-assert 16) + u32 minor_version = 0; + // (maya-file-name basic :offset-assert 20) + std::string maya_file_name; + // (tool-debug basic :offset-assert 24) + std::string tool_debug; + // (mdb-file-name basic :offset-assert 28) + std::string mdb_file_name; + + size_t add_to_object_file(DataObjectGenerator& gen) const; +}; + +FileInfo make_file_info_for_level(const std::string& file_name); \ No newline at end of file diff --git a/goalc/build_level/LevelFile.cpp b/goalc/build_level/LevelFile.cpp new file mode 100644 index 0000000000..b5902eb988 --- /dev/null +++ b/goalc/build_level/LevelFile.cpp @@ -0,0 +1,105 @@ +#include "LevelFile.h" +#include "goalc/data_compiler/DataObjectGenerator.h" + +size_t DrawableTreeArray::add_to_object_file(DataObjectGenerator& gen) const { + /* + (deftype drawable-tree-array (drawable-group) + ((trees drawable-tree 1 :offset 32 :score 100)) + :flag-assert #x1200000024 + ) + (deftype drawable-group (drawable) + ((length int16 :offset 6) + (data drawable 1 :offset-assert 32) + ) + (:methods + (new (symbol type int) _type_) + ) + :flag-assert #x1200000024 + ) + */ + gen.align_to_basic(); + gen.add_type_tag("drawable-tree-array"); + size_t result = gen.current_offset_bytes(); + int num_trees = 0; + num_trees += tfrags.size(); + num_trees += collides.size(); + gen.add_word(num_trees << 16); + gen.add_word(0); + gen.add_word(0); + + gen.add_word(0); + gen.add_word(0); + gen.add_word(0); + gen.add_word(0); + + // todo add trees... + + if (num_trees == 0) { + gen.add_word(0); // the one at the end. + } else { + int tree_word = (int)gen.current_offset_bytes() / 4; + for (int i = 0; i < num_trees; i++) { + gen.add_word(0); + } + + for (auto& tfrag : tfrags) { + // gen.set_word(tree_word++, tfrag.add_to_object_file(gen)); + gen.link_word_to_byte(tree_word++, tfrag.add_to_object_file(gen)); + } + + for (auto& collide : collides) { + gen.link_word_to_byte(tree_word++, collide.add_to_object_file(gen)); + } + } + + return result; +} + +std::vector LevelFile::save_object_file() const { + DataObjectGenerator gen; + gen.add_type_tag("bsp-header"); + + // add blank space for the bsp-header + while (gen.words() < 100) { + gen.add_word(0); + } + + //(info file-info :offset 4) + auto file_info_slot = info.add_to_object_file(gen); + gen.link_word_to_byte(1, file_info_slot); + + //(bsphere vector :inline :offset-assert 16) + //(all-visible-list (pointer uint16) :offset-assert 32) + //(visible-list-length int32 :offset-assert 36) + //(drawable-trees drawable-tree-array :offset-assert 40) + gen.link_word_to_byte(40 / 4, drawable_trees.add_to_object_file(gen)); + //(pat pointer :offset-assert 44) + //(pat-length int32 :offset-assert 48) + //(texture-remap-table (pointer uint64) :offset-assert 52) + //(texture-remap-table-len int32 :offset-assert 56) + //(texture-ids (pointer texture-id) :offset-assert 60) + //(texture-page-count int32 :offset-assert 64) + //(unk-zero-0 basic :offset-assert 68) + //(name symbol :offset-assert 72) + gen.link_word_to_symbol(name, 72 / 4); + //(nickname symbol :offset-assert 76) + gen.link_word_to_symbol(nickname, 76 / 4); + //(vis-info level-vis-info 8 :offset-assert 80) + //(actors drawable-inline-array-actor :offset-assert 112) + //(cameras (array entity-camera) :offset-assert 116) + //(nodes (inline-array bsp-node) :offset-assert 120) + //(level level :offset-assert 124) + //(current-leaf-idx uint16 :offset-assert 128) + //(unk-data-2 uint16 9 :offset-assert 130) + //(boxes box8s-array :offset-assert 148) + //(current-bsp-back-flags uint32 :offset-assert 152) + //(ambients drawable-inline-array-ambient :offset-assert 156) + //(unk-data-4 float :offset-assert 160) + //(unk-data-5 float :offset-assert 164) + //(adgifs adgif-shader-array :offset-assert 168) + //(actor-birth-order (pointer uint32) :offset-assert 172) + //(split-box-indices (pointer uint16) :offset-assert 176) + //(unk-data-8 uint32 55 :offset-assert 180) + + return gen.generate_v2(); +} \ No newline at end of file diff --git a/goalc/build_level/LevelFile.h b/goalc/build_level/LevelFile.h new file mode 100644 index 0000000000..29f53dd4c0 --- /dev/null +++ b/goalc/build_level/LevelFile.h @@ -0,0 +1,140 @@ +#pragma once + +#include +#include +#include + +#include "common/common_types.h" +#include "goalc/build_level/FileInfo.h" +#include "goalc/build_level/Tfrag.h" +#include "goalc/build_level/collide_pack.h" +#include "goalc/build_level/collide_common.h" +#include "goalc/build_level/collide_bvh.h" +#include "goalc/build_level/collide_drawable.h" + +struct VisibilityString { + std::vector bytes; +}; + +struct DrawableTreeInstanceTie {}; + +struct DrawableTreeActor {}; + +struct DrawableTreeAmbient {}; + +struct DrawableTreeInstanceShrub {}; + +struct DrawableTreeArray { + std::vector tfrags; + std::vector ties; + std::vector actors; // unused? + std::vector collides; + std::vector ambients; + std::vector shrubs; + size_t add_to_object_file(DataObjectGenerator& gen) const; +}; + +struct TextureRemap {}; + +struct TextureId {}; + +struct VisInfo {}; + +struct DrawableActor {}; + +struct DrawableInlineArrayActor {}; + +struct EntityCamera {}; + +struct BspNode {}; + +struct Box8s {}; + +struct DrawableAmbient {}; + +struct DrawableInlineArrayAmbient {}; + +struct AdgifShaderArray {}; + +// This is a place to collect all the data that should go into the bsp-header file. +struct LevelFile { + // (info file-info :offset 4) + FileInfo info; + + // (all-visible-list (pointer uint16) :offset-assert 32) + // (visible-list-length int32 :offset-assert 36) + VisibilityString all_visibile_list; + + // (drawable-trees drawable-tree-array :offset-assert 40) + DrawableTreeArray drawable_trees; + + // (pat pointer :offset-assert 44) + // (pat-length int32 :offset-assert 48) + std::vector pat; + + // (texture-remap-table (pointer uint64) :offset-assert 52) + // (texture-remap-table-len int32 :offset-assert 56) + std::vector texture_remap_table; + + // (texture-ids (pointer texture-id) :offset-assert 60) + // (texture-page-count int32 :offset-assert 64) + std::vector texture_ids; + + // (unk-zero-0 basic :offset-assert 68) + // "misc", seems like it can be zero and is unused. + + // (name symbol :offset-assert 72) + std::string name; // full name + + // (nickname symbol :offset-assert 76) + std::string nickname; // 3 char name + + // (vis-info level-vis-info 8 :offset-assert 80) ;; note: 0 when + std::array vis_infos; + + // (actors drawable-inline-array-actor :offset-assert 112) + DrawableInlineArrayActor actors; + + // (cameras (array entity-camera) :offset-assert 116) + std::vector cameras; + + // (nodes (inline-array bsp-node) :offset-assert 120) + std::vector nodes; + + // (level level :offset-assert 124) + // zero + + // (current-leaf-idx uint16 :offset-assert 128) + // zero + + // (unk-data-2 uint16 9 :offset-assert 130) + // looks like padding plus 4 floats? unused + + // (boxes box8s-array :offset-assert 148) + std::vector boxes; + + // (current-bsp-back-flags uint32 :offset-assert 152) + // zero + + // (ambients drawable-inline-array-ambient :offset-assert 156) + DrawableInlineArrayAmbient ambients; + + // (unk-data-4 float :offset-assert 160) + float close_subdiv = 0; + + // (unk-data-5 float :offset-assert 164) + float far_subdiv = 0; + + // (adgifs adgif-shader-array :offset-assert 168) + AdgifShaderArray adgifs; + + // (actor-birth-order (pointer uint32) :offset-assert 172) + std::vector actor_birth_order; + + // (split-box-indices (pointer uint16) :offset-assert 176) + std::vector split_box_indices; + + // (unk-data-8 uint32 55 :offset-assert 180) + + std::vector save_object_file() const; +}; \ No newline at end of file diff --git a/goalc/build_level/ResLump.cpp b/goalc/build_level/ResLump.cpp new file mode 100644 index 0000000000..f70a09ee2b --- /dev/null +++ b/goalc/build_level/ResLump.cpp @@ -0,0 +1,4 @@ +#include "ResLump.h" + +#include "third-party/fmt/core.h" +#include "goalc/data_compiler/DataObjectGenerator.h" diff --git a/goalc/build_level/ResLump.h b/goalc/build_level/ResLump.h new file mode 100644 index 0000000000..19b6ee6453 --- /dev/null +++ b/goalc/build_level/ResLump.h @@ -0,0 +1,7 @@ +#pragma once + +#include +#include +#include + +#include "common/common_types.h" diff --git a/goalc/build_level/TexturePool.h b/goalc/build_level/TexturePool.h new file mode 100644 index 0000000000..2cef160246 --- /dev/null +++ b/goalc/build_level/TexturePool.h @@ -0,0 +1,12 @@ +#pragma once + +#include +#include +#include + +#include "common/custom_data/Tfrag3Data.h" + +struct TexturePool { + std::unordered_map textures_by_name; + std::vector textures_by_idx; +}; \ No newline at end of file diff --git a/goalc/build_level/Tfrag.cpp b/goalc/build_level/Tfrag.cpp new file mode 100644 index 0000000000..449cdbb22f --- /dev/null +++ b/goalc/build_level/Tfrag.cpp @@ -0,0 +1,99 @@ +#include + +#include "Tfrag.h" +#include "common/custom_data/pack_helpers.h" +#include "goalc/data_compiler/DataObjectGenerator.h" +#include "goalc/build_level/gltf_mesh_extract.h" + +void tfrag_from_gltf(const gltf_mesh_extract::TfragOutput& mesh_extract_out, + DrawableTreeTfrag& out, + tfrag3::TfragTree& out_pc) { + out_pc.kind = tfrag3::TFragmentTreeKind::NORMAL; // todo more types? + out_pc.draws = std::move(mesh_extract_out.strip_draws); + fmt::print("have {} draws\n", out_pc.draws.size()); + pack_tfrag_vertices(&out_pc.packed_vertices, mesh_extract_out.vertices); + fmt::print("have {} vertices\n", out_pc.packed_vertices.vertices.size()); + + for (auto& col : mesh_extract_out.color_palette) { + tfrag3::TimeOfDayColor todc; + for (auto& rgba : todc.rgba) { + rgba = col; + } + out_pc.colors.push_back(todc); + } + out_pc.use_strips = false; +} + +/* + +(deftype drawable-group (drawable) + ((length int16 :offset 6) + (data drawable 1 :offset-assert 32) + ) + (:methods + (new (symbol type int) _type_) + ) + :flag-assert #x1200000024 + ) + + (deftype drawable-tree (drawable-group) + () + :flag-assert #x1200000024 + ) + +(deftype drawable-inline-array (drawable) + ((length int16 :offset 6) ;; this is kinda weird. + ) + :method-count-assert 18 + :size-assert #x20 + :flag-assert #x1200000020 + ) + +(deftype drawable-inline-array-tfrag (drawable-inline-array) + ((data tfragment 1 :inline :offset-assert 32) + (pad uint32)) + :method-count-assert 18 + :size-assert #x64 + :flag-assert #x1200000064 + ) + +(deftype drawable-tree-tfrag (drawable-tree) + ((time-of-day-pal time-of-day-palette :offset 12) + (arrays drawable-inline-array 1 :offset 32 :score 100) ;; either drawable-inline-array-node +or drawable-inline-array-tfrag + ) + :method-count-assert #x12 + :size-assert #x24 + :flag-assert #x1200000024 + ) + + + */ + +size_t add_empty_dia(const std::string& name, DataObjectGenerator& gen, int total_size) { + gen.align_to_basic(); + gen.add_type_tag(name); + size_t result = gen.current_offset_bytes(); + total_size -= 4; + while (total_size > 0) { + gen.add_word(0); + total_size -= 4; + } + + return result; +} + +size_t DrawableTreeTfrag::add_to_object_file(DataObjectGenerator& gen) const { + gen.align_to_basic(); + gen.add_type_tag("drawable-tree-tfrag"); + size_t result = gen.current_offset_bytes(); + gen.add_word(1 << 16); + for (int i = 0; i < 6; i++) { + gen.add_word(0); + } + size_t slot = gen.add_word(0); + ASSERT(slot * 4 - result == 28); + gen.link_word_to_byte(slot, add_empty_dia("drawable-inline-array-tfrag", gen, 0x64)); + + return result; +} \ No newline at end of file diff --git a/goalc/build_level/Tfrag.h b/goalc/build_level/Tfrag.h new file mode 100644 index 0000000000..f9c137a473 --- /dev/null +++ b/goalc/build_level/Tfrag.h @@ -0,0 +1,17 @@ +#pragma once + +#include + +#include "common/custom_data/Tfrag3Data.h" +#include "goalc/build_level/TexturePool.h" +#include "goalc/build_level/gltf_mesh_extract.h" + +class DataObjectGenerator; + +struct DrawableTreeTfrag { + size_t add_to_object_file(DataObjectGenerator& gen) const; +}; + +void tfrag_from_gltf(const gltf_mesh_extract::TfragOutput& mesh_extract_out, + DrawableTreeTfrag& out, + tfrag3::TfragTree& out_pc); \ No newline at end of file diff --git a/goalc/build_level/build_level.cpp b/goalc/build_level/build_level.cpp new file mode 100644 index 0000000000..9edb3586ef --- /dev/null +++ b/goalc/build_level/build_level.cpp @@ -0,0 +1,102 @@ +#include "third-party/fmt/core.h" +#include "common/util/json_util.h" +#include "common/util/FileUtil.h" +#include "common/log/log.h" +#include "goalc/build_level/LevelFile.h" +#include "goalc/build_level/FileInfo.h" +#include "goalc/build_level/Tfrag.h" +#include "goalc/build_level/gltf_mesh_extract.h" +#include "goalc/build_level/collide_bvh.h" +#include "goalc/build_level/collide_pack.h" + +#include "common/custom_data/Tfrag3Data.h" +#include "common/util/compress.h" + +void save_pc_data(const std::string& nickname, tfrag3::Level& data) { + Serializer ser; + data.serialize(ser); + auto compressed = + compression::compress_zstd(ser.get_save_result().first, ser.get_save_result().second); + fmt::print("stats for {}\n", data.level_name); + print_memory_usage(data, ser.get_save_result().second); + fmt::print("compressed: {} -> {} ({:.2f}%)\n", ser.get_save_result().second, compressed.size(), + 100.f * compressed.size() / ser.get_save_result().second); + file_util::write_binary_file(file_util::get_file_path({fmt::format("assets/{}.fr3", nickname)}), + compressed.data(), compressed.size()); +} + +std::vector get_build_level_deps(const std::string& input_file) { + auto level_json = parse_commented_json( + file_util::read_text_file(file_util::get_file_path({input_file})), input_file); + return {level_json.at("gltf_file").get()}; +} + +bool run_build_level(const std::string& input_file, const std::string& output_file) { + auto level_json = parse_commented_json( + file_util::read_text_file(file_util::get_file_path({input_file})), input_file); + LevelFile file; // GOAL level file + tfrag3::Level pc_level; // PC level file + TexturePool tex_pool; // pc level texture pool + + // process input mesh from blender + gltf_mesh_extract::Input mesh_extract_in; + mesh_extract_in.filename = + file_util::get_file_path({level_json.at("gltf_file").get()}); + mesh_extract_in.tex_pool = &tex_pool; + gltf_mesh_extract::Output mesh_extract_out; + gltf_mesh_extract::extract(mesh_extract_in, mesh_extract_out); + + // add stuff to the GOAL level structure + file.info = make_file_info_for_level(std::filesystem::path(input_file).filename().string()); + // all vis + // drawable trees + // pat + // texture remap + // texture ids + // unk zero + // name + file.name = level_json.at("long_name").get(); + // nick + file.nickname = level_json.at("nickname").get(); + // vis infos + // actors + // cameras + // nodes + // boxes + // ambients + // subdivs + // adgifs + // actor birht + // split box + + // add stuff to the PC level structure + pc_level.level_name = file.name; + + // TFRAG + auto& tfrag_drawable_tree = file.drawable_trees.tfrags.emplace_back(); + tfrag_from_gltf(mesh_extract_out.tfrag, tfrag_drawable_tree, + pc_level.tfrag_trees[0].emplace_back()); + pc_level.textures = std::move(tex_pool.textures_by_idx); + + // COLLIDE + if (mesh_extract_out.collide.faces.empty()) { + lg::error("No collision geometry was found"); + } else { + auto& collide_drawable_tree = file.drawable_trees.collides.emplace_back(); + collide_drawable_tree.bvh = collide::construct_collide_bvh(mesh_extract_out.collide.faces); + collide_drawable_tree.packed_frags = pack_collide_frags(collide_drawable_tree.bvh.frags.frags); + } + + // Save the GOAL level + auto result = file.save_object_file(); + fmt::print("Level bsp file size {} bytes\n", result.size()); + auto save_path = file_util::get_file_path({output_file}); + file_util::create_dir_if_needed_for_file(save_path); + fmt::print("Saving to {}\n", save_path); + file_util::write_binary_file(save_path, result.data(), result.size()); + + // Save the PC level + save_pc_data(file.nickname, pc_level); + + return true; +} diff --git a/goalc/build_level/build_level.h b/goalc/build_level/build_level.h new file mode 100644 index 0000000000..afe685daac --- /dev/null +++ b/goalc/build_level/build_level.h @@ -0,0 +1,7 @@ +#pragma once + +#include +#include + +bool run_build_level(const std::string& input_file, const std::string& output_file); +std::vector get_build_level_deps(const std::string& input_file); \ No newline at end of file diff --git a/goalc/build_level/collide_bvh.cpp b/goalc/build_level/collide_bvh.cpp new file mode 100644 index 0000000000..8d410b2a3c --- /dev/null +++ b/goalc/build_level/collide_bvh.cpp @@ -0,0 +1,284 @@ +#include + +#include "collide_bvh.h" +#include "common/util/Assert.h" +#include "common/log/log.h" +#include "common/util/Timer.h" + +// Collision BVH algorithm +// We start with all the points in a single node, then recursively split nodes in 8 until no nodes +// have too many faces. +// The splitting is done by doing median cuts along the x, y, or z axis. + +// The bspheres are built at the end. + +namespace collide { + +namespace { + +constexpr int MAX_FACES_IN_FRAG = 100; + +/*! + * The Collide node. + * Has either children collide node or children faces, but not both + * The size of child_nodes is either 0 or 8 at all times. + */ +struct CNode { + std::vector child_nodes; + std::vector faces; + math::Vector4f bsphere; +}; + +struct BBox { + math::Vector3f mins, maxs; + std::string sz_to_string() const { + return fmt::format("({})", ((maxs - mins) / 4096.f).to_string_aligned()); + } +}; + +/*! + * Make the bounding box hold this node and all its children. + */ +void add_to_bbox_recursive(const CNode& node, BBox& bbox) { + if (node.faces.empty()) { + ASSERT(node.child_nodes.size() == 8); + for (auto& child : node.child_nodes) { + add_to_bbox_recursive(child, bbox); + } + } else { + for (auto& face : node.faces) { + for (auto& vert : face.v) { + bbox.mins.min_in_place(vert); + bbox.maxs.max_in_place(vert); + } + } + } +} + +BBox bbox_of_node(const CNode& node) { + BBox bbox; + bbox.mins.fill(std::numeric_limits::max()); + bbox.maxs.fill(-std::numeric_limits::max()); + add_to_bbox_recursive(node, bbox); + return bbox; +} + +/*! + * Make the bsphere hold this node and all its children. + */ +void update_bsphere_recursive(const CNode& node, const math::Vector3f& origin, float& r_squared) { + if (node.faces.empty()) { + ASSERT(node.child_nodes.size() == 8); + for (auto& child : node.child_nodes) { + update_bsphere_recursive(child, origin, r_squared); + } + } else { + for (auto& face : node.faces) { + for (auto& vert : face.v) { + r_squared = std::max(r_squared, (vert - origin).squared_length()); + } + } + } +} + +/*! + * Compute the bsphere of a single node. + */ +void compute_my_bsphere(CNode& node) { + // first compute bbox. + BBox bbox = bbox_of_node(node); + float r = 0; + math::Vector3f origin = (bbox.maxs + bbox.mins) * 0.5; + update_bsphere_recursive(node, origin, r); + node.bsphere.x() = origin.x(); + node.bsphere.y() = origin.y(); + node.bsphere.z() = origin.z(); + node.bsphere.w() = std::sqrt(r); +} + +/*! + * Split faces in two along a coordinate plane. + * Will clear the input faces + */ +void split_along_dim(std::vector& faces, + int dim, + std::vector* out0, + std::vector* out1) { + std::sort(faces.begin(), faces.end(), [=](const CollideFace& a, const CollideFace& b) { + return a.bsphere[dim] < b.bsphere[dim]; + }); + size_t split_idx = faces.size() / 2; + out0->insert(out0->end(), faces.begin(), faces.begin() + split_idx); + out1->insert(out1->end(), faces.begin() + split_idx, faces.end()); +} + +/*! + * Split a node into two nodes. The outputs should be uninitialized nodes + */ +void split_node_once(CNode& node, CNode* out0, CNode* out1) { + CNode temps[6]; + // split_along_dim(node.faces, pick_dim_for_split(node.faces), &out0->faces, &out1->faces); + split_along_dim(node.faces, 0, &temps[0].faces, &temps[1].faces); + split_along_dim(node.faces, 1, &temps[2].faces, &temps[3].faces); + split_along_dim(node.faces, 2, &temps[4].faces, &temps[5].faces); + node.faces.clear(); + for (auto& t : temps) { + compute_my_bsphere(t); + } + + float max_bspheres[3] = {0, 0, 0}; + for (int i = 0; i < 3; i++) { + max_bspheres[i] = std::max(temps[i * 2].bsphere.w(), temps[i * 2 + 1].bsphere.w()); + } + + int best_dim = 0; + float best_w = max_bspheres[0]; + for (int i = 0; i < 3; i++) { + if (max_bspheres[i] < best_w) { + best_dim = i; + best_w = max_bspheres[i]; + } + } + + *out0 = temps[best_dim * 2]; + *out1 = temps[best_dim * 2 + 1]; +} + +/*! + * Split a node into 8 children and store these in the given node. + */ +void split_node_to_8_children(CNode& node) { + ASSERT(node.child_nodes.empty()); + node.child_nodes.resize(8); + // level 0 + CNode level0[2]; + split_node_once(node, &level0[0], &level0[1]); + // level 1 + CNode level1[4]; + split_node_once(level0[0], &level1[0], &level1[1]); + split_node_once(level0[1], &level1[2], &level1[3]); + // level 2 + split_node_once(level1[0], &node.child_nodes[0], &node.child_nodes[1]); + split_node_once(level1[1], &node.child_nodes[2], &node.child_nodes[3]); + split_node_once(level1[2], &node.child_nodes[4], &node.child_nodes[5]); + split_node_once(level1[3], &node.child_nodes[6], &node.child_nodes[7]); +} + +/*! + * Split all leaf nodes. Returns the number of faces in the leaf with the most faces after + * splitting. + * This slightly unusual recursion pattern is to make sure we split everything to same depth, + * which we believe might be a requirement of the collision system. + */ +size_t split_all_leaves(CNode& node) { + size_t worst_leaf_face_count = 0; + if (node.child_nodes.empty()) { + // we're a leaf! + // split us: + split_node_to_8_children(node); + for (auto& child : node.child_nodes) { + worst_leaf_face_count = std::max(worst_leaf_face_count, child.faces.size()); + } + return worst_leaf_face_count; + } else { + // not a leaf, recurse + for (auto& child : node.child_nodes) { + worst_leaf_face_count = std::max(worst_leaf_face_count, split_all_leaves(child)); + } + return worst_leaf_face_count; + } +} + +/*! + * Main BVH construction function. Splits leaves until it is no longer needed. + */ +void split_as_needed(CNode& root) { + int initial_tri_count = root.faces.size(); + int num_leaves = 1; + bool need_to_split = true; + while (need_to_split) { + int faces_in_worst = split_all_leaves(root); + num_leaves *= 8; + lg::info("after splitting, the worst leaf has {} tris", faces_in_worst); + if (faces_in_worst < MAX_FACES_IN_FRAG) { + need_to_split = false; + } + } + lg::info("average triangles per leaf: {}", initial_tri_count / num_leaves); + lg::info("leaf count: {}", num_leaves); +} + +/*! + * Recursively compute bspheres of all children + * (note that we don't do bspheres of bspheres... I think this is better?) + */ +void bsphere_recursive(CNode& node) { + compute_my_bsphere(node); + for (auto& child : node.child_nodes) { + bsphere_recursive(child); + } +} + +void drawable_layout_helper(CNode& node, int depth, CollideTree& tree_out, size_t my_idx_check) { + if (node.child_nodes.empty()) { + // we're a leaf! add us to the frags + auto& frag = tree_out.frags.frags.emplace_back(); + frag.bsphere = node.bsphere; + frag.faces = node.faces; + } else { + // not a leaf + if ((int)tree_out.node_arrays.size() <= depth) { + tree_out.node_arrays.resize(depth + 1); + } + ASSERT(my_idx_check == tree_out.node_arrays.at(depth).nodes.size()); + auto& draw_node = tree_out.node_arrays[depth].nodes.emplace_back(); + draw_node.bsphere = node.bsphere; + for (int i = 0; i < 8; i++) { + draw_node.children[i] = my_idx_check * 8 + i; + drawable_layout_helper(node.child_nodes.at(i), depth + 1, tree_out, draw_node.children[i]); + } + } +} + +CollideTree build_collide_tree(CNode& root) { + CollideTree tree; + drawable_layout_helper(root, 0, tree, 0); + return tree; +} + +void debug_stats(const CollideTree& tree) { + lg::info("Tree build: {} draw node layers", tree.node_arrays.size()); + float sum_w = 0, max_w = 0; + for (auto& frag : tree.frags.frags) { + sum_w += frag.bsphere.w(); + max_w = std::max(frag.bsphere.w(), max_w); + } + lg::info("Max bsphere radius: {:.2f}m, average {:.2f} (aiming for around 20-30m avg)", + max_w / 4096, sum_w / (4096 * tree.frags.frags.size())); +} + +} // namespace + +CollideTree construct_collide_bvh(const std::vector& tris) { + // part 1: build the tree + Timer bvh_timer; + lg::info("Building collide bvh from {} triangles", tris.size()); + CNode root; + root.faces = tris; + split_as_needed(root); + lg::info("BVH tree constructed in {:.2f} ms", bvh_timer.getMs()); + + // part 2: compute bspheres + bvh_timer.start(); + bsphere_recursive(root); + lg::info("Found bspheres in {:.2f} ms", bvh_timer.getMs()); + + // part 3: layout tree + bvh_timer.start(); + auto tree = build_collide_tree(root); + debug_stats(tree); + lg::info("Tree layout done in {:.2f} ms", bvh_timer.getMs()); + return tree; +} + +} // namespace collide \ No newline at end of file diff --git a/goalc/build_level/collide_bvh.h b/goalc/build_level/collide_bvh.h new file mode 100644 index 0000000000..11b4f4bdaa --- /dev/null +++ b/goalc/build_level/collide_bvh.h @@ -0,0 +1,37 @@ +#pragma once + +#include +#include "goalc/build_level/collide_common.h" + +// requirements: +// max depth of 3 (maybe?) +// max face per frag = 90 +// max vert per frag = 110 +// branching factor of 8 everywhere. +namespace collide { + +struct DrawNode { + s32 children[8] = {-1, -1, -1, -1, -1, -1, -1, -1}; + math::Vector4f bsphere; +}; + +struct CollideFrag { + math::Vector4f bsphere; + std::vector faces; +}; + +struct DrawableInlineArrayNode { + std::vector nodes; +}; + +struct DrawableInlineArrayCollideFrag { + std::vector frags; +}; + +struct CollideTree { + std::vector node_arrays; + DrawableInlineArrayCollideFrag frags; +}; + +CollideTree construct_collide_bvh(const std::vector& tris); +} // namespace collide diff --git a/goalc/build_level/collide_common.h b/goalc/build_level/collide_common.h new file mode 100644 index 0000000000..3ef5146167 --- /dev/null +++ b/goalc/build_level/collide_common.h @@ -0,0 +1,110 @@ +#pragma once +#include "common/common_types.h" +#include "common/math/Vector.h" + +struct PatSurface { + enum class Mode { GROUND = 0, WALL = 1, OBSTACLE = 2 }; + enum class Material { + STONE = 0, + ICE = 1, + QUICKSAND = 2, + WATERBOTTOM = 3, + TAR = 4, + SAND = 5, + WOOD = 6, + GRASS = 7, + PCMETAL = 8, + SNOW = 9, + DEEPSNOW = 10, + HOTCOALS = 11, + LAVA = 12, + CRWOOD = 13, + GRAVEL = 14, + DIRT = 15, + METAL = 16, + STRAW = 17, + TUBE = 18, + SWAMP = 19, + STOPPROJ = 20, + ROTATE = 21, + NEUTRAL = 22, + }; + + enum class Event { + NONE = 0, + DEADLY = 1, + ENDLESSFALL = 2, + BURN = 3, + DEADLYUP = 4, + BURNUP = 5, + MELT = 6, + }; + + void set_noentity(bool x) { + if (x) { + val |= (1 << 0); + } else { + val &= ~(1 << 0); + } + } + bool get_noentity() const { return val & (1 << 0); } + + void set_nocamera(bool x) { + if (x) { + val |= (1 << 1); + } else { + val &= ~(1 << 1); + } + } + bool get_nocamera() const { return val & (1 << 1); } + + void set_noedge(bool x) { + if (x) { + val |= (1 << 2); + } else { + val &= ~(1 << 2); + } + } + bool get_noedge() const { return val & (1 << 2); } + + void set_mode(Mode mode) { + val &= ~(0b111 << 3); + val |= ((u32)mode << 3); + } + Mode get_mode() const { return (Mode)(0b111 & (val >> 3)); } + + void set_material(Material mat) { + val &= ~(0b111111 << 6); + val |= ((u32)mat << 6); + } + Material get_material() const { return (Material)(0b111111 & (val >> 6)); } + + void set_nolineofsight(bool x) { + if (x) { + val |= (1 << 12); + } else { + val &= ~(1 << 12); + } + } + bool get_nolineofsight() const { return val & (1 << 12); } + + void set_event(Event ev) { + val &= ~(0b111111 << 14); + val |= ((u32)ev << 6); + } + Event get_event() const { return (Event)(0b111111 & (val >> 14)); } + + bool operator==(const PatSurface& other) const { return val == other.val; } + // bits 13, [15-31] are unused, or have unknown purpose. + u32 val = 0; +}; + +struct CollideVertex { + float x, y, z; +}; + +struct CollideFace { + math::Vector4f bsphere; + math::Vector3f v[3]; + PatSurface pat; +}; diff --git a/goalc/build_level/collide_drawable.cpp b/goalc/build_level/collide_drawable.cpp new file mode 100644 index 0000000000..4198eadfbd --- /dev/null +++ b/goalc/build_level/collide_drawable.cpp @@ -0,0 +1,268 @@ +#include "collide_drawable.h" +#include "goalc/data_compiler/DataObjectGenerator.h" +#include "common/util/Assert.h" + +/* +(deftype drawable (basic) + ((id int16 :offset-assert 4) + (bsphere vector :inline :offset-assert 16) + ) + +(deftype drawable-group (drawable) + ((length int16 :offset 6) + (data drawable 1 :offset-assert 32) + ) + +(deftype drawable-tree (drawable-group) + () + +(deftype drawable-inline-array (drawable) + ((length int16 :offset 6) ;; this is kinda weird. + ) + +(deftype drawable-tree-collide-fragment (drawable-tree) + ((data-override drawable-inline-array 1 :offset 32)) ;; should be 1 there + +(deftype drawable-inline-array-collide-fragment (drawable-inline-array) + ((data collide-fragment 1 :inline :offset-assert 32) + + (deftype collide-fragment (drawable) + ((mesh collide-frag-mesh :offset 8) + ) + :method-count-assert 18 + :size-assert #x20 + +(deftype collide-frag-mesh (basic) + ((packed-data uint32 :offset-assert 4) + (pat-array uint32 :offset-assert 8) + (strip-data-len uint16 :offset-assert 12) + (poly-count uint16 :offset-assert 14) + (base-trans vector :inline :offset-assert 16) + ;; these go in the w of the vector above. + (vertex-count uint8 :offset 28) + (vertex-data-qwc uint8 :offset 29) + (total-qwc uint8 :offset 30) + (unused uint8 :offset 31) + ) + :method-count-assert 9 + :size-assert #x20 + +(deftype draw-node (drawable) + ((child-count uint8 :offset 6) + (flags uint8 :offset 7) + (child drawable :offset 8) + (distance float :offset 12) + ) + :size-assert #x20 + + (deftype drawable-inline-array-node (drawable-inline-array) + ((data draw-node 1 :inline) + (pad uint32) + ) + :method-count-assert 18 + :size-assert #x44 + */ + +size_t generate_pat_array(DataObjectGenerator& gen, const std::vector& pats) { + gen.align_to_basic(); + size_t result = gen.current_offset_bytes(); + for (auto& pat : pats) { + gen.add_word(pat.val); + } + return result; +} + +size_t generate_packed_collide_data(DataObjectGenerator& gen, const std::vector& data) { + gen.align_to_basic(); + size_t result = gen.current_offset_bytes(); + ASSERT((data.size() % 4) == 0); + for (size_t i = 0; i < data.size(); i += 4) { + u32 word; + memcpy(&word, data.data() + i, 4); + gen.add_word(word); + } + return result; +} + +size_t generate_collide_frag_mesh(DataObjectGenerator& gen, + const CollideFragMeshData& mesh, + size_t packed_data_loc, + size_t pat_array_loc) { + gen.align_to_basic(); + gen.add_type_tag("collide-frag-mesh"); // 0 + size_t result = gen.current_offset_bytes(); + gen.link_word_to_byte(gen.add_word(0), packed_data_loc); // 4 + gen.link_word_to_byte(gen.add_word(0), pat_array_loc); // 8 + gen.add_word(mesh.strip_data_len | (mesh.poly_count << 16)); // 12 + gen.add_word(mesh.base_trans_xyz_s32.x()); // 16 + gen.add_word(mesh.base_trans_xyz_s32.y()); // 20 + gen.add_word(mesh.base_trans_xyz_s32.z()); // 24 + u32 packed = 0; + packed |= mesh.vertex_count; + packed |= ((u32)mesh.vertex_data_qwc) << 8; + packed |= ((u32)mesh.total_qwc) << 16; + gen.add_word(packed); // 28 + return result; +} + +size_t generate_collide_fragment(DataObjectGenerator& gen, + const CollideFragMeshData& mesh, + size_t frag_mesh_loc) { + /* + .type collide-fragment + .word 0x10000 + .word L705 + .word 0x0 + .word 0x46bf480a + .word 0x43dc730b + .word 0xb71ed4fe + .word 0x46c42e44 + */ + gen.align_to_basic(); + gen.add_type_tag("collide-fragment"); + size_t result = gen.current_offset_bytes(); + gen.add_word(0x10000); // ??? + gen.link_word_to_byte(gen.add_word(0), frag_mesh_loc); + gen.add_word(0); + for (int i = 0; i < 4; i++) { + gen.add_word_float(mesh.bsphere[i]); + } + + return result; +} + +size_t generate_collide_fragment_array(DataObjectGenerator& gen, + const std::vector& meshes, + const std::vector& frag_mesh_locs, + std::vector& parent_ref_out) { + gen.align_to_basic(); + gen.add_type_tag("drawable-inline-array-collide-fragment"); // 0 + size_t result = gen.current_offset_bytes(); + ASSERT(meshes.size() < UINT16_MAX); + gen.add_word(meshes.size() << 16); // 4, 6 + gen.add_word(0); // 8 + gen.add_word(0); // 12 + gen.add_word(0); // 16 + gen.add_word(0); // 20 + gen.add_word(0); // 24 + gen.add_word(0); // 28 + + fmt::print("have: {}\n", meshes.size()); + + ASSERT(meshes.size() == frag_mesh_locs.size()); + for (size_t i = 0; i < meshes.size(); i++) { + auto& mesh = meshes[i]; + // should be 8 words here: + gen.add_type_tag("collide-fragment"); // 1 + size_t me = gen.current_offset_bytes(); + gen.add_word(0x10000); // ??? + gen.link_word_to_byte(gen.add_word(0), frag_mesh_locs[i]); + gen.add_word(0); + for (int j = 0; j < 4; j++) { + gen.add_word_float(mesh.bsphere[j]); + } + if ((i % 8) == 0) { + parent_ref_out.push_back(me); + } + } + + return result; +} + +size_t generate_collide_draw_node_array(DataObjectGenerator& gen, + const std::vector& nodes, + u32 flag, + const std::vector& children, + std::vector& parent_ref_out) { + gen.align_to_basic(); + gen.add_type_tag("drawable-inline-array-node"); // 0 + size_t result = gen.current_offset_bytes(); + gen.add_word(nodes.size() << 16); // 4, 6 + gen.add_word(0); // 8 + gen.add_word(0); // 12 + gen.add_word(0); // 16 + gen.add_word(0); // 20 + gen.add_word(0); // 24 + gen.add_word(0); // 28 + + ASSERT(nodes.size() == children.size()); + for (size_t i = 0; i < nodes.size(); i++) { + auto& node = nodes[i]; + // should be 8 words here: + gen.add_type_tag("draw-node"); // 1 + size_t me = gen.current_offset_bytes(); + u32 packed_flags = 0; + packed_flags |= (8 << 16); // TODO hard-coded size here + packed_flags |= (flag << 24); + gen.add_word(packed_flags); // 2 + gen.link_word_to_byte(gen.add_word(0), children[i]); // 3 + gen.add_word(0); // 4 + if ((i % 8) == 0) { + parent_ref_out.push_back(me); + } + gen.add_word_float(node.bsphere.x()); // 5 + gen.add_word_float(node.bsphere.y()); // 6 + gen.add_word_float(node.bsphere.z()); // 7 + gen.add_word_float(node.bsphere.w()); // 8 + } + + return result; +} + +size_t DrawableTreeCollideFragment::add_to_object_file(DataObjectGenerator& gen) const { + for (auto& lev : bvh.node_arrays) { + fmt::print("lev: {}\n", lev.nodes.size()); + } + // generate pat array + size_t pat_array_loc = generate_pat_array(gen, packed_frags.pats); + + // generated packed data + std::vector packed_data_locs; + for (auto& mesh : packed_frags.packed_frag_data) { + packed_data_locs.push_back(generate_packed_collide_data(gen, mesh.packed_data)); + } + + // generate collide frag meshes + std::vector collide_frag_meshes; + for (size_t i = 0; i < packed_data_locs.size(); i++) { + collide_frag_meshes.push_back(generate_collide_frag_mesh(gen, packed_frags.packed_frag_data[i], + packed_data_locs[i], pat_array_loc)); + } + + std::vector array_locs; + array_locs.resize(bvh.node_arrays.size() + 1); // plus one for the frags. + int array_slot = bvh.node_arrays.size(); + + std::vector children_refs; + array_locs[array_slot--] = generate_collide_fragment_array(gen, packed_frags.packed_frag_data, + collide_frag_meshes, children_refs); + u32 flag = 0; + while (array_slot >= 0) { + fmt::print("sizes: {} {}\n", children_refs.size(), bvh.node_arrays.at(array_slot).nodes.size()); + ASSERT(children_refs.size() == bvh.node_arrays.at(array_slot).nodes.size()); + std::vector next_children; + + array_locs[array_slot] = generate_collide_draw_node_array( + gen, bvh.node_arrays.at(array_slot).nodes, flag, children_refs, next_children); + + children_refs = std::move(next_children); + array_slot--; + flag = 1; + } + + { + gen.align_to_basic(); + gen.add_type_tag("drawable-tree-collide-fragment"); + size_t result = gen.current_offset_bytes(); + gen.add_word((array_locs.size() - 1) << 16); // todo the minus one here?? + for (int i = 0; i < 6; i++) { + gen.add_word(0); + } + + for (size_t i = 1; i < array_locs.size(); i++) { // todo the offset here? + gen.link_word_to_byte(gen.add_word(0), array_locs[i]); + } + + return result; + } +} diff --git a/goalc/build_level/collide_drawable.h b/goalc/build_level/collide_drawable.h new file mode 100644 index 0000000000..5804dbb5c1 --- /dev/null +++ b/goalc/build_level/collide_drawable.h @@ -0,0 +1,12 @@ +#pragma once + +#include "goalc/build_level/collide_bvh.h" +#include "goalc/build_level/collide_pack.h" + +class DataObjectGenerator; + +struct DrawableTreeCollideFragment { + CollideFragMeshDataArray packed_frags; + collide::CollideTree bvh; + size_t add_to_object_file(DataObjectGenerator& gen) const; +}; diff --git a/goalc/build_level/collide_pack.cpp b/goalc/build_level/collide_pack.cpp new file mode 100644 index 0000000000..595ad6c89c --- /dev/null +++ b/goalc/build_level/collide_pack.cpp @@ -0,0 +1,263 @@ +#include + +#include "collide_pack.h" +#include "common/util/Assert.h" +#include "common/log/log.h" +#include "common/util/Timer.h" + +struct PackedU16Verts { + math::Vector base; + std::vector> vertex; +}; + +/*! + * Assert that the given float can be converted to a packed u16, and return that u16. + */ +u16 magic_float_to_u16(float in) { + u16 u16s[2]; + memcpy(u16s, &in, 4); + ASSERT_MSG(u16s[1] == 0x4d00, + "Unable to pack collision vertex data to u16's. Try to make smaller triangles and " + "avoid skinny triangles. (or there is a bug in the packer)"); + return u16s[0]; +} + +namespace { +float u32_to_float(u32 in) { + float out; + memcpy(&out, &in, 4); + return out; +} +} // namespace + +/*! + * Pack vertices to base + u16. + * The format is quite strange to allow for fast unpacking. + */ +PackedU16Verts pack_verts_to_u16(const std::vector& input) { + PackedU16Verts result; + ASSERT(!input.empty()); + + // this "magic" offset is a large float where a ulp is 16.f, or 1/256th of a meter. + // this means that float -> int can be done as a single addition. + // (or, in some cases, we can avoid float->int entirely) + math::Vector3f magic_offset; + magic_offset.fill(u32_to_float(0x4d000000)); + + // we'll be treating everything as an offset from this minimum vertex: + math::Vector3f min_vtx = input[0]; + for (auto& vtx : input) { + min_vtx.min_in_place(vtx); + } + // give us a tiny bit of extra room to avoid rounding problems + min_vtx -= 16.f; + + // to round down to nearest integer + result.base = min_vtx.cast(); + auto base = result.base.cast(); + + // compute offset relative to base. + for (auto& vtx : input) { + // add the "magic offset" to make this a 0x4dXXXXXX float. + // subtract the base to make this a 0x4d00XXXX float. + auto vertex_magic = vtx + magic_offset - base; + // and if we did it right, we should be able to pack to u16's here + result.vertex.emplace_back(magic_float_to_u16(vertex_magic[0]), + magic_float_to_u16(vertex_magic[1]), + magic_float_to_u16(vertex_magic[2])); + } + + // verify + /* + for (size_t i = 0; i < input.size(); i++) { + math::Vector3f v; + v[0] = u32_to_float(0x4d000000 + result.vertex[i][0]); + v[1] = u32_to_float(0x4d000000 + result.vertex[i][1]); + v[2] = u32_to_float(0x4d000000 + result.vertex[i][2]); + float base_offset = u32_to_float(0x4d000000); + math::Vector3f vf13_combo_offset(base_offset, base_offset, base_offset); + math::Vector3f vf14_base_trans_float(result.base[0], result.base[1], result.base[2]); + vf13_combo_offset -= vf14_base_trans_float; + v -= vf13_combo_offset; + fmt::print("error {}\n", (v - input[i]).to_string_aligned());; + } + */ + + return result; +} + +struct PatSurfaceHash { + size_t operator()(const PatSurface& in) const { return std::hash()(in.val); } +}; + +/*! + * pat -> pat index mapping. + * There's a pat "palette" with up 255 unique pats. + */ +struct PatMap { + std::unordered_map map; + std::vector pats; + + u32 add_pat(PatSurface pat) { + const auto& lookup = map.find(pat); + if (lookup == map.end()) { + u32 new_idx = pats.size(); + if (new_idx > UINT8_MAX) { + lg::die("Too many pats. Use fewer. Or improve the pat code to use multiple pat arrays."); + } + map[pat] = new_idx; + pats.push_back(pat); + return new_idx; + } else { + return lookup->second; + } + } +}; + +/*! + * A face, represented as indices + */ +struct IndexFace { + math::Vector vertex_indices; // per vertex, winding order matters here + u8 pat_idx; // pat for the whole face +}; + +/*! + * All the faces in a frag + */ +struct IndexedFaces { + std::vector faces; // index in to the two vertex array below: + std::vector vertices_float; + PackedU16Verts vertices_u16; +}; + +struct Vector3fHash { + size_t operator()(const math::Vector3f& in) const { + return std::hash()(in.x()) ^ std::hash()(in.y()) ^ std::hash()(in.z()); + } +}; + +/*! + * Deduplicate vertices, converted to indexed, add to pat palette, pack to u16s. + */ +IndexedFaces dedup_frag_mesh(const collide::CollideFrag& frag, PatMap* pat_map) { + IndexedFaces result; + std::unordered_map vertex_map; + // avoid confusion with 0 in strip table. (todo, can probably remove) + result.vertices_float.push_back(math::Vector3f::zero()); + + for (auto& face_in : frag.faces) { + auto& face_out = result.faces.emplace_back(); + // pat: + face_out.pat_idx = pat_map->add_pat(face_in.pat); + // vertices + for (int i = 0; i < 3; i++) { + const auto& lookup = vertex_map.find(face_in.v[i]); + if (lookup == vertex_map.end()) { + u32 idx = result.vertices_float.size(); + result.vertices_float.push_back(face_in.v[i]); + face_out.vertex_indices[i] = idx; + vertex_map[face_in.v[i]] = idx; + } else { + face_out.vertex_indices[i] = lookup->second; + } + } + } + // fmt::print("{} -> {}\n", frag.faces.size() * 3, result.vertices_float.size()); + result.vertices_u16 = pack_verts_to_u16(result.vertices_float); + return result; +} + +/*! + * make strip table that doesn't do any stripping. It will be quite long, which might cause problem + */ +std::vector make_dumb_strip_table(const IndexedFaces& faces) { + std::vector out; + ASSERT_MSG( + faces.vertices_float.size() < UINT8_MAX, + "somehow have UINT8_MAX deduped vertices in a single fragment, likely a bug somewhere."); + for (auto& face : faces.faces) { + out.push_back(face.vertex_indices[0]); + out.push_back(face.vertex_indices[1]); + out.push_back(face.vertex_indices[2]); + out.push_back(0); + } + out.push_back(-1); + + return out; +} + +CollideFragMeshDataArray pack_collide_frags(const std::vector& frag_data) { + Timer pack_timer; + CollideFragMeshDataArray result; + PatMap pat_map; + + size_t total_pack_bytes = 0; + lg::info("Packing {} fragments", frag_data.size()); + + for (auto& frag_in : frag_data) { + auto& frag_out = result.packed_frag_data.emplace_back(); + auto indexed = dedup_frag_mesh(frag_in, &pat_map); + // first part of packed_data is the u16 vertex data: + frag_out.vertex_count = indexed.vertices_u16.vertex.size(); + frag_out.packed_data.resize(sizeof(u16) * frag_out.vertex_count * 3); + memcpy(frag_out.packed_data.data(), indexed.vertices_u16.vertex.data(), + frag_out.packed_data.size()); + // align to 16-bytes + while (frag_out.packed_data.size() & 0xf) { + frag_out.packed_data.push_back(0); + } + // remember where + frag_out.vertex_data_qwc = frag_out.packed_data.size() / 16; + + // up next, the strip table + auto strip = make_dumb_strip_table(indexed); + frag_out.packed_data.insert(frag_out.packed_data.end(), strip.begin(), strip.end()); + frag_out.strip_data_len = strip.size(); + ASSERT(frag_out.strip_data_len < UINT16_MAX); // probably in big trouble in here. + + // pat table + for (auto& face : indexed.faces) { + frag_out.packed_data.push_back(face.pat_idx); + } + + // align to 16-bytes so total_qwc works. + while (frag_out.packed_data.size() & 0xf) { + frag_out.packed_data.push_back(0); + } + // gonna guess here: + frag_out.poly_count = indexed.faces.size(); + frag_out.total_qwc = frag_out.packed_data.size() / 16; + frag_out.base_trans_xyz_s32 = indexed.vertices_u16.base; + frag_out.bsphere = frag_in.bsphere; + total_pack_bytes += frag_out.packed_data.size(); + } + + result.pats = pat_map.pats; + lg::info("Total packed data size: {} kB, took {:.2f} ms", total_pack_bytes / 1024, + pack_timer.getMs()); + return result; +} + +/* +(deftype collide-frag-mesh (basic) + ((packed-data uint32 :offset-assert 4) <- ptr + (pat-array uint32 :offset-assert 8) <- ptr + (strip-data-len uint16 :offset-assert 12) + (poly-count uint16 :offset-assert 14) + (base-trans vector :inline :offset-assert 16) + ;; these go in the w of the vector above. + (vertex-count uint8 :offset 28) // done! + (vertex-data-qwc uint8 :offset 29) // done! + (total-qwc uint8 :offset 30) + (unused uint8 :offset 31) + ) + :method-count-assert 9 + :size-assert #x20 + :flag-assert #x900000020 + ) + */ + +// packed_data: +// (u16x3) per vertex, packed float vtx format. +// diff --git a/goalc/build_level/collide_pack.h b/goalc/build_level/collide_pack.h new file mode 100644 index 0000000000..316b56e9f4 --- /dev/null +++ b/goalc/build_level/collide_pack.h @@ -0,0 +1,21 @@ +#pragma once + +#include "goalc/build_level/collide_bvh.h" + +struct CollideFragMeshData { + math::Vector4f bsphere; // not part of the collide frag, but is part of the drawable wrapping it + std::vector packed_data; + u32 strip_data_len; + u32 poly_count; + math::Vector base_trans_xyz_s32; + u8 vertex_count; + u8 vertex_data_qwc; + u8 total_qwc; +}; + +struct CollideFragMeshDataArray { + std::vector packed_frag_data; + std::vector pats; +}; + +CollideFragMeshDataArray pack_collide_frags(const std::vector& frag_data); \ No newline at end of file diff --git a/goalc/build_level/color_quantization.cpp b/goalc/build_level/color_quantization.cpp new file mode 100644 index 0000000000..91784dc117 --- /dev/null +++ b/goalc/build_level/color_quantization.cpp @@ -0,0 +1,220 @@ +#include +#include "color_quantization.h" +#include "common/util/Assert.h" +#include "common/log/log.h" + +/*! + * Just removes duplicate colors, which can work if there are only a few unique colors. + */ +QuantizedColors quantize_colors_dumb(const std::vector>& in) { + QuantizedColors result; + std::unordered_map color_to_slot; + for (auto& vtx : in) { + u64 key; + memcpy(&key, vtx.data(), sizeof(u64)); + const auto& existing = color_to_slot.find(key); + if (existing == color_to_slot.end()) { + auto cidx = result.final_colors.size(); + result.vtx_to_color.push_back(cidx); + color_to_slot[key] = cidx; + result.final_colors.push_back(vtx); + } else { + result.vtx_to_color.push_back(existing->second); + } + } + fmt::print("quantize_colors_dumb: {} -> {}\n", in.size(), result.final_colors.size()); + ASSERT(result.final_colors.size() < 8192); + return result; +} + +namespace { + +using Color = math::Vector; + +// An octree node. +// Represents a color in the output if rgb_sum_count > 0. +// Otherwise, just organizational. +struct Node { + u32 r_sum = 0; + u32 g_sum = 0; + u32 b_sum = 0; + u32 rgb_sum_count = 0; + u8 depth = 0xff; // 0 for root, 7 deepest + + // children stuff + u32 leaves_under_me = 0; + std::vector children; + Node* parent = nullptr; + + u32 final_idx = UINT32_MAX; +}; + +u8 child_index(Color color, u8 depth) { + u8 r_bit = (color.x() >> (7 - depth)) & 1; + u8 g_bit = (color.y() >> (7 - depth)) & 1; + u8 b_bit = (color.z() >> (7 - depth)) & 1; + return (r_bit) + (g_bit * 2) + (b_bit * 4); +} + +void insert(Node& root, Color color, u8 current_depth) { + if (current_depth == 7) { + root.r_sum += color.x(); + root.g_sum += color.y(); + root.b_sum += color.z(); + if (root.rgb_sum_count == 0) { + for (auto* up = root.parent; up; up = up->parent) { + up->leaves_under_me++; + } + } + root.rgb_sum_count++; + } else { + if (root.children.empty()) { + root.children.resize(8); + } + auto& next_node = root.children[child_index(color, current_depth)]; + if (next_node.depth == 0xff) { + next_node.depth = current_depth + 1; + next_node.parent = &root; + } + insert(next_node, color, current_depth + 1); + } +} + +template +void for_each_node(Node& root, T&& func) { + func(root); + for (auto& child : root.children) { + for_each_node(child, func); + } +} + +u32 count_leaves(Node& root) { + u32 result = 0; + for_each_node(root, [&](Node& n) { + if (n.rgb_sum_count) { + ASSERT(n.children.empty()); + result++; + } + }); + return result; +} + +void collapse1(Node& root) { + ASSERT(!root.children.empty()); + u32 total_children_removed = 0; + u32 total_rgb_sum_moved_up = 0; + bool started_as_leaf = root.rgb_sum_count; + for (auto& child : root.children) { + if (child.depth != 0xff) { + ASSERT(child.children.empty()); + ASSERT(child.rgb_sum_count); + total_children_removed++; + root.r_sum += child.r_sum; + root.g_sum += child.g_sum; + root.b_sum += child.b_sum; + root.rgb_sum_count += child.rgb_sum_count; + total_rgb_sum_moved_up += child.rgb_sum_count; + } + } + ASSERT(total_children_removed == root.leaves_under_me); + + if (!started_as_leaf && root.rgb_sum_count) { + total_children_removed--; + } + root.children.clear(); + root.leaves_under_me = 0; + if (total_children_removed) { + for (auto* up = root.parent; up; up = up->parent) { + up->leaves_under_me -= total_children_removed; + } + } + + ASSERT(count_leaves(root) == 1); +} + +void find_nodes_at_level(Node& n, std::vector& out, u8 level) { + if (n.depth == level) { + out.push_back(&n); + } else if (n.depth < level) { + for (auto& child : n.children) { + find_nodes_at_level(child, out, level); + } + } +} + +void collapse_at_level(Node& root, u8 level, u32 target_leaf_count) { + std::vector nodes_at_level; + find_nodes_at_level(root, nodes_at_level, level); + std::stable_sort(nodes_at_level.begin(), nodes_at_level.end(), + [](Node* a, Node* b) { return a->leaves_under_me < b->leaves_under_me; }); + + size_t at_level_to_try = 0; + while (root.leaves_under_me > target_leaf_count && at_level_to_try < nodes_at_level.size()) { + collapse1(*nodes_at_level[at_level_to_try++]); + } +} + +void collapse_as_needed(Node& root, u32 target_leaf_count) { + u32 level_to_reduce = 6; + while (root.leaves_under_me > target_leaf_count) { + collapse_at_level(root, level_to_reduce--, target_leaf_count); + } +} + +void assign_colors(Node& root, std::vector& palette_out) { + u32 idx = 0; + for_each_node(root, [&](Node& n) { + if (n.rgb_sum_count) { + n.final_idx = idx++; + palette_out.emplace_back(n.r_sum / n.rgb_sum_count, n.g_sum / n.rgb_sum_count, + n.b_sum / n.rgb_sum_count); + } + }); +} + +u32 lookup_node_for_color(Node& root, Color c, u8 depth) { + if (root.children.empty()) { + return root.final_idx; + } else { + return lookup_node_for_color(root.children[child_index(c, depth)], c, depth + 1); + } +} + +} // namespace + +/*! + * Quantize colors using an octree for clustering. + */ +QuantizedColors quantize_colors_octree(const std::vector>& in, + u32 target_count) { + Node root; + root.depth = 0; + for (auto& color : in) { + insert(root, color, 0); + } + + collapse_as_needed(root, target_count); + + QuantizedColors out; + assign_colors(root, out.final_colors); + for (auto& color : in) { + out.vtx_to_color.push_back(lookup_node_for_color(root, color, 0)); + } + + float total_error[3] = {0, 0, 0}; + for (size_t i = 0; i < in.size(); i++) { + // fmt::print(" {} -> {}\n", in[i].to_string_hex_byte(), + // out.final_colors[out.vtx_to_color[i]].to_string_hex_byte()); + auto diff = in[i].cast() - out.final_colors[out.vtx_to_color[i]].cast(); + + for (int j = 0; j < 3; j++) { + total_error[j] += std::abs(diff[j]); + } + } + + lg::info("Octree quantize average error (as 8-bit ints): r: {}, g: {} b: {}", + total_error[0] / in.size(), total_error[1] / in.size(), total_error[2] / in.size()); + lg::info("Final palette size: {}", out.final_colors.size()); + + return out; +} \ No newline at end of file diff --git a/goalc/build_level/color_quantization.h b/goalc/build_level/color_quantization.h new file mode 100644 index 0000000000..44b5ba68cf --- /dev/null +++ b/goalc/build_level/color_quantization.h @@ -0,0 +1,20 @@ +#pragma once + +#include +#include "common/common_types.h" +#include "common/math/Vector.h" + +// TODO: come up with something for time of day colors. +// time of day colors make the colors effectively 24-channel instead of just 3. +// the octree approach doesn't work too well here (we'd need a 16777216-tree) +// but a k-d tree seems like the right approach. + +struct QuantizedColors { + std::vector> final_colors; + std::vector vtx_to_color; +}; + +QuantizedColors quantize_colors_dumb(const std::vector>& in); + +QuantizedColors quantize_colors_octree(const std::vector>& in, + u32 target_count); \ No newline at end of file diff --git a/goalc/build_level/gltf_mesh_extract.cpp b/goalc/build_level/gltf_mesh_extract.cpp new file mode 100644 index 0000000000..a81463c052 --- /dev/null +++ b/goalc/build_level/gltf_mesh_extract.cpp @@ -0,0 +1,688 @@ +/*! + * Mesh extraction for GLTF meshes. + */ + +#include "gltf_mesh_extract.h" +#include "goalc/build_level/color_quantization.h" +#include "third-party/tiny_gltf/tiny_gltf.h" +#include "common/log/log.h" +#include "common/util/Timer.h" +#include "common/math/geometry.h" + +namespace gltf_mesh_extract { + +namespace { + +/*! + * Convert a GLTF index buffer to std::vector + */ +template +std::vector index_list_to_u32(const u8* data, u32 num_verts, u32 offset, u32 stride) { + std::vector result; + result.reserve(num_verts); + for (u32 i = 0; i < num_verts; i++) { + T val; + memcpy(&val, data, sizeof(T)); + result.push_back(offset + val); + data += stride; + } + return result; +} + +/*! + * Convert a GLTF position buffer or similar to std::vector + */ +std::vector extract_vec3f(const u8* data, u32 count, u32 stride) { + std::vector result; + result.reserve(count); + for (u32 i = 0; i < count; i++) { + memcpy(&result.emplace_back(), data, sizeof(math::Vector3f)); + data += stride; + } + return result; +} + +std::vector extract_vec2f(const u8* data, u32 count, u32 stride) { + std::vector result; + result.reserve(count); + for (u32 i = 0; i < count; i++) { + memcpy(&result.emplace_back(), data, sizeof(math::Vector2f)); + data += stride; + } + return result; +} + +/*! + * Convert a GLTF color buffer to u8 colors. + */ +std::vector> extract_color_from_vec4_u16(const u8* data, + u32 count, + u32 stride) { + std::vector> result; + result.reserve(count); + for (u32 i = 0; i < count; i++) { + math::Vector temp; + memcpy(&temp, data, sizeof(math::Vector)); + data += stride; + result.emplace_back(temp.x() >> 8, temp.y() >> 8, temp.z() >> 8, temp.w() >> 8); + } + return result; +} + +/*! + * Convert a GLTF index buffer + */ +std::vector gltf_index_buffer(const tinygltf::Model& model, + int indices_idx, + u32 index_offset) { + const auto& indices_accessor = model.accessors[indices_idx]; + const auto& buffer_view = model.bufferViews[indices_accessor.bufferView]; + const auto& buffer = model.buffers[buffer_view.buffer]; + const auto data_ptr = buffer.data.data() + buffer_view.byteOffset + indices_accessor.byteOffset; + const auto stride = indices_accessor.ByteStride(buffer_view); + const auto count = indices_accessor.count; + + switch (indices_accessor.componentType) { + case TINYGLTF_COMPONENT_TYPE_BYTE: + return index_list_to_u32(data_ptr, count, index_offset, stride); + case TINYGLTF_COMPONENT_TYPE_UNSIGNED_BYTE: + return index_list_to_u32(data_ptr, count, index_offset, stride); + case TINYGLTF_COMPONENT_TYPE_SHORT: + return index_list_to_u32(data_ptr, count, index_offset, stride); + case TINYGLTF_COMPONENT_TYPE_UNSIGNED_SHORT: + return index_list_to_u32(data_ptr, count, index_offset, stride); + case TINYGLTF_COMPONENT_TYPE_INT: + return index_list_to_u32(data_ptr, count, index_offset, stride); + case TINYGLTF_COMPONENT_TYPE_UNSIGNED_INT: + return index_list_to_u32(data_ptr, count, index_offset, stride); + default: + ASSERT_MSG(false, "unsupported component type"); + } +} + +struct ExtractedVertices { + std::vector vtx; + std::vector> vtx_colors; + std::vector normals; +}; + +ExtractedVertices gltf_vertices(const tinygltf::Model& model, + const std::map& attributes, + const math::Matrix4f& w_T_local, + bool get_colors, + bool get_normals, + const std::string& debug_name) { + std::vector result; + std::vector> vtx_colors; + + { + const auto& position_attrib = attributes.find("POSITION"); + ASSERT_MSG(position_attrib != attributes.end(), "Did not find position attribute."); + + const auto attrib_accessor = model.accessors[position_attrib->second]; + const auto& buffer_view = model.bufferViews[attrib_accessor.bufferView]; + const auto& buffer = model.buffers[buffer_view.buffer]; + const auto data_ptr = buffer.data.data() + buffer_view.byteOffset + attrib_accessor.byteOffset; + const auto byte_stride = attrib_accessor.ByteStride(buffer_view); + const auto count = attrib_accessor.count; + + ASSERT_MSG(attrib_accessor.type == TINYGLTF_TYPE_VEC3, "POSITION wasn't vec3"); + ASSERT_MSG(attrib_accessor.componentType == TINYGLTF_COMPONENT_TYPE_FLOAT, + "POSITION wasn't float"); + // for (auto& attrib : attributes) { + // fmt::print("attrib: {}\n", attrib.first); + //} + auto mesh_verts = extract_vec3f(data_ptr, count, byte_stride); + result.reserve(mesh_verts.size()); + for (auto& vert : mesh_verts) { + auto& new_vert = result.emplace_back(); + math::Vector4f v_in(vert.x(), vert.y(), vert.z(), 1); + math::Vector4f v_w = w_T_local * v_in; + new_vert.x = v_w.x() * 4096; + new_vert.y = v_w.y() * 4096; + new_vert.z = v_w.z() * 4096; + } + } + + if (get_colors) { + const auto& color_attrib = attributes.find("COLOR_0"); + if (color_attrib == attributes.end()) { + lg::error("Mesh {} didn't have any colors, using white", debug_name); + for (size_t i = 0; i < result.size(); i++) { + vtx_colors.emplace_back(0x80, 0x80, 0x80, 0xff); + } + } else { + const auto attrib_accessor = model.accessors[color_attrib->second]; + const auto& buffer_view = model.bufferViews[attrib_accessor.bufferView]; + const auto& buffer = model.buffers[buffer_view.buffer]; + const auto data_ptr = + buffer.data.data() + buffer_view.byteOffset + attrib_accessor.byteOffset; + const auto byte_stride = attrib_accessor.ByteStride(buffer_view); + const auto count = attrib_accessor.count; + + ASSERT_MSG(attrib_accessor.type == TINYGLTF_TYPE_VEC4, "COLOR_0 wasn't vec4"); + ASSERT_MSG( + attrib_accessor.componentType == TINYGLTF_COMPONENT_TYPE_UNSIGNED_SHORT, + fmt::format("COLOR_0 wasn't float, got {} instead", attrib_accessor.componentType)); + auto colors = extract_color_from_vec4_u16(data_ptr, count, byte_stride); + vtx_colors.insert(vtx_colors.end(), colors.begin(), colors.end()); + } + + // ASSERT_MSG(color_attrib != attributes.end(), "Did not find color attribute."); + } + + bool got_texture = false; + { + const auto& texcoord_attrib = attributes.find("TEXCOORD_0"); + if (texcoord_attrib != attributes.end()) { + const auto attrib_accessor = model.accessors[texcoord_attrib->second]; + const auto& buffer_view = model.bufferViews[attrib_accessor.bufferView]; + const auto& buffer = model.buffers[buffer_view.buffer]; + const auto data_ptr = + buffer.data.data() + buffer_view.byteOffset + attrib_accessor.byteOffset; + const auto byte_stride = attrib_accessor.ByteStride(buffer_view); + const auto count = attrib_accessor.count; + + ASSERT_MSG(attrib_accessor.type == TINYGLTF_TYPE_VEC2, "TEXCOORD wasn't vec2"); + ASSERT_MSG(attrib_accessor.componentType == TINYGLTF_COMPONENT_TYPE_FLOAT, + "TEXCOORD wasn't float"); + auto mesh_verts = extract_vec2f(data_ptr, count, byte_stride); + ASSERT(mesh_verts.size() == result.size()); + got_texture = true; + for (size_t i = 0; i < mesh_verts.size(); i++) { + result[i].s = mesh_verts[i].x(); + result[i].t = mesh_verts[i].y(); + } + } else { + if (!get_normals) { + // don't warn if we're just getting collision + lg::warn("No texcoord attribute for mesh: {}", debug_name); + } + } + } + + std::vector normals; + if (get_normals) { + const auto& normal_attrib = attributes.find("NORMAL"); + if (normal_attrib != attributes.end()) { + const auto attrib_accessor = model.accessors[normal_attrib->second]; + const auto& buffer_view = model.bufferViews[attrib_accessor.bufferView]; + const auto& buffer = model.buffers[buffer_view.buffer]; + const auto data_ptr = + buffer.data.data() + buffer_view.byteOffset + attrib_accessor.byteOffset; + const auto byte_stride = attrib_accessor.ByteStride(buffer_view); + const auto count = attrib_accessor.count; + + ASSERT_MSG(attrib_accessor.type == TINYGLTF_TYPE_VEC3, "NORMAL wasn't vec3"); + ASSERT_MSG(attrib_accessor.componentType == TINYGLTF_COMPONENT_TYPE_FLOAT, + "NORMAL wasn't float"); + normals = extract_vec3f(data_ptr, count, byte_stride); + ASSERT(normals.size() == result.size()); + } else { + lg::error("No NORMAL attribute for mesh: {}", debug_name); + } + } + + for (auto& v : result) { + v.color_index = 0; + if (!got_texture) { + v.s = 0; + v.t = 0; + } + + v.q_unused = 0; + v.pad[0] = 0; + v.pad[1] = 0; + v.pad[2] = 0; + } + // TODO: other properties + return {result, vtx_colors, normals}; +} + +DrawMode make_default_draw_mode() { + DrawMode mode; + mode.set_depth_write_enable(true); + mode.set_depth_test(GsTest::ZTest::GEQUAL); + mode.set_alpha_blend(DrawMode::AlphaBlend::DISABLED); + mode.set_aref(0); + mode.set_alpha_fail(GsTest::AlphaFail::KEEP); + mode.set_clamp_s_enable(true); + mode.set_clamp_t_enable(true); + mode.disable_filt(); // for checkerboard... + mode.enable_tcc(); // ? + mode.disable_at(); + mode.enable_zt(); + mode.disable_ab(); + mode.disable_decal(); + mode.enable_fog(); + return mode; +} + +int texture_pool_debug_checker(TexturePool* pool) { + const auto& existing = pool->textures_by_name.find("DEBUG_CHECKERBOARD"); + if (existing == pool->textures_by_name.end()) { + size_t idx = pool->textures_by_idx.size(); + pool->textures_by_name["DEBUG_CHECKERBOARD"] = idx; + auto& tex = pool->textures_by_idx.emplace_back(); + tex.w = 16; + tex.h = 16; + tex.debug_name = "DEBUG_CHECKERBOARD"; + tex.debug_tpage_name = "DEBUG"; + tex.load_to_pool = false; + tex.combo_id = 0; // doesn't matter, not a pool tex + tex.data.resize(16 * 16); + u32 c0 = 0xa0303030; + u32 c1 = 0xa0e0e0e0; + for (int i = 0; i < 16; i++) { + for (int j = 0; j < 16; j++) { + tex.data[i * 16 + j] = (((i / 4) & 1) ^ ((j / 4) & 1)) ? c1 : c0; + } + } + return idx; + } else { + return existing->second; + } +} + +int texture_pool_add_texture(TexturePool* pool, const tinygltf::Image& tex) { + const auto& existing = pool->textures_by_name.find(tex.name); + if (existing != pool->textures_by_name.end()) { + lg::info("Reusing image: {}", tex.name); + return existing->second; + } + + ASSERT(tex.bits == 8); + ASSERT(tex.component == 4); + ASSERT(tex.pixel_type == TINYGLTF_TEXTURE_TYPE_UNSIGNED_BYTE); + + size_t idx = pool->textures_by_idx.size(); + pool->textures_by_name[tex.name] = idx; + auto& tt = pool->textures_by_idx.emplace_back(); + tt.w = tex.width; + tt.h = tex.height; + tt.debug_name = tex.name; + tt.debug_tpage_name = "custom-level"; + tt.load_to_pool = false; + tt.combo_id = 0; // doesn't matter, not a pool tex + tt.data.resize(tt.w * tt.h * 4); + ASSERT(tex.image.size() >= tt.data.size()); + memcpy(tt.data.data(), tex.image.data(), tt.data.size()); + return idx; +} +} // namespace + +math::Matrix4f affine_translation(const math::Vector3f& translation) { + math::Matrix4f result = math::Matrix4f::identity(); + result(0, 3) = translation[0]; + result(1, 3) = translation[1]; + result(2, 3) = translation[2]; + result(3, 3) = 1; + return result; +} + +math::Matrix4f affine_scale(const math::Vector3f& scale) { + math::Matrix4f result = math::Matrix4f::zero(); + result(0, 0) = scale[0]; + result(1, 1) = scale[1]; + result(2, 2) = scale[2]; + result(3, 3) = 1; + return result; +} + +math::Matrix4f affine_rot_qxyzw(const math::Vector4f& quat) { + math::Matrix4f result = math::Matrix4f::zero(); + result(3, 3) = 1; + result(0, 0) = 1.0 - 2.0 * (quat.y() * quat.y() + quat.z() * quat.z()); + result(0, 1) = 2.0 * (quat.x() * quat.y() - quat.z() * quat.w()); + result(0, 2) = 2.0 * (quat.x() * quat.z() + quat.y() * quat.w()); + result(1, 0) = 2.0 * (quat.x() * quat.y() + quat.z() * quat.w()); + result(1, 1) = 1.0 - 2.0 * (quat.x() * quat.x() + quat.z() * quat.z()); + result(1, 2) = 2.0 * (quat.y() * quat.z() - quat.x() * quat.w()); + result(2, 0) = 2.0 * (quat.x() * quat.z() - quat.y() * quat.w()); + result(2, 1) = 2.0 * (quat.y() * quat.z() + quat.x() * quat.w()); + result(2, 2) = 1.0 - 2.0 * (quat.x() * quat.x() + quat.y() * quat.y()); + return result; +} + +math::Vector3f vector3f_from_gltf(const std::vector& in) { + ASSERT(in.size() == 3); + return math::Vector3f{in[0], in[1], in[2]}; +} + +math::Vector4f vector4f_from_gltf(const std::vector& in) { + ASSERT(in.size() == 4); + return math::Vector4f{in[0], in[1], in[2], in[3]}; +} + +math::Matrix4f matrix_from_node(const tinygltf::Node& node) { + if (!node.matrix.empty()) { + math::Matrix4f result; + for (int i = 0; i < 16; i++) { + result.data()[i] = node.matrix[i]; + } + return result; + } else { + // from trs + math::Matrix4f t, r, s; + if (!node.translation.empty()) { + t = affine_translation(vector3f_from_gltf(node.translation)); + } else { + t = math::Matrix4f::identity(); + } + + if (!node.rotation.empty()) { + r = affine_rot_qxyzw(vector4f_from_gltf(node.rotation)); + } else { + r = math::Matrix4f::identity(); + } + + if (!node.scale.empty()) { + s = affine_scale(vector3f_from_gltf(node.scale)); + } else { + s = math::Matrix4f::identity(); + } + + return t * r * s; + } +} + +struct NodeWithTransform { + int node_idx; + math::Matrix4f w_T_node; +}; + +/*! + * Recursively walk the tree of nodes, flatten, and compute w_T_node for each. + */ +void node_find_helper(const tinygltf::Model& model, + const math::Matrix4f& w_T_parent, + int node_idx, + std::vector* out) { + const auto& node = model.nodes.at(node_idx); + math::Matrix4f w_T_node = w_T_parent * matrix_from_node(node); + out->push_back({node_idx, w_T_node}); + for (auto& child : node.children) { + node_find_helper(model, w_T_node, child, out); + } +} + +std::vector flatten_nodes_from_all_scenes(const tinygltf::Model& model) { + std::vector out; + for (auto& scene : model.scenes) { + for (auto& nidx : scene.nodes) { + math::Matrix4f identity = math::Matrix4f::identity(); + node_find_helper(model, identity, nidx, &out); + } + } + return out; +} + +void dedup_vertices(const std::vector& vertices_in, + std::vector& vertices_out, + std::vector& old_to_new_out) { + ASSERT(vertices_out.empty()); + ASSERT(old_to_new_out.empty()); + old_to_new_out.resize(vertices_in.size(), -1); + + std::unordered_map vtx_to_new; + + for (size_t in_idx = 0; in_idx < vertices_in.size(); in_idx++) { + auto& vtx = vertices_in[in_idx]; + const auto& lookup = vtx_to_new.find(vtx); + if (lookup == vtx_to_new.end()) { + // first time seeing this one + size_t new_idx = vertices_out.size(); + vertices_out.push_back(vtx); + old_to_new_out[in_idx] = new_idx; + vtx_to_new[vtx] = new_idx; + } else { + old_to_new_out[in_idx] = lookup->second; + } + } +} + +void dedup_vertices(TfragOutput& data) { + Timer timer; + size_t original_size = data.vertices.size(); + std::vector new_verts; + std::vector old_to_new; + + dedup_vertices(data.vertices, new_verts, old_to_new); + data.vertices = std::move(new_verts); + + for (auto& draw : data.strip_draws) { + ASSERT(draw.runs.empty()); // not supported yet + for (auto& idx : draw.plain_indices) { + idx = old_to_new.at(idx); + } + } + + lg::info("Deduplication took {:.2f} ms, {} -> {} ({:.2f} %)", timer.getMs(), original_size, + data.vertices.size(), 100.f * data.vertices.size() / original_size); +} + +DrawMode draw_mode_from_sampler(const tinygltf::Sampler& sampler) { + DrawMode mode = make_default_draw_mode(); + if (sampler.magFilter == TINYGLTF_TEXTURE_FILTER_NEAREST) { + ASSERT(sampler.minFilter == TINYGLTF_TEXTURE_FILTER_NEAREST); + mode.set_filt_enable(false); + } else { + ASSERT(sampler.minFilter != TINYGLTF_TEXTURE_FILTER_NEAREST); + mode.set_filt_enable(true); + } + + switch (sampler.wrapS) { + case TINYGLTF_TEXTURE_WRAP_CLAMP_TO_EDGE: + mode.set_clamp_s_enable(true); + break; + case TINYGLTF_TEXTURE_WRAP_REPEAT: + mode.set_clamp_s_enable(false); + break; + default: + ASSERT(false); + } + + switch (sampler.wrapT) { + case TINYGLTF_TEXTURE_WRAP_CLAMP_TO_EDGE: + mode.set_clamp_t_enable(true); + break; + case TINYGLTF_TEXTURE_WRAP_REPEAT: + mode.set_clamp_t_enable(false); + break; + default: + ASSERT(false); + } + + return mode; +} + +void extract(const Input& in, + TfragOutput& out, + const tinygltf::Model& model, + const std::vector& all_nodes) { + std::vector> all_vtx_colors; + ASSERT(out.vertices.empty()); + std::map draw_by_material; + int mesh_count = 0; + int prim_count = 0; + + for (const auto& n : all_nodes) { + const auto& node = model.nodes[n.node_idx]; + if (node.mesh >= 0) { + const auto& mesh = model.meshes[node.mesh]; + if (!mesh.extras.Has("tfrag")) { + // fmt::print("skip tfrag: {}\n", mesh.name); + // continue; + } + mesh_count++; + for (const auto& prim : mesh.primitives) { + prim_count++; + // extract index buffer + std::vector prim_indices = gltf_index_buffer(model, prim.indices, out.vertices.size()); + ASSERT_MSG(prim.mode == TINYGLTF_MODE_TRIANGLES, "Unsupported triangle mode"); + // extract vertices + auto verts = + gltf_vertices(model, prim.attributes, n.w_T_node, in.get_colors, false, mesh.name); + out.vertices.insert(out.vertices.end(), verts.vtx.begin(), verts.vtx.end()); + if (in.get_colors) { + all_vtx_colors.insert(all_vtx_colors.end(), verts.vtx_colors.begin(), + verts.vtx_colors.end()); + } + + // TODO: just putting it all in one material + auto& draw = draw_by_material[prim.material]; + draw.mode = make_default_draw_mode(); // todo rm + draw.tree_tex_id = texture_pool_debug_checker(in.tex_pool); // todo rm + draw.num_triangles += prim_indices.size() / 3; + if (draw.vis_groups.empty()) { + auto& grp = draw.vis_groups.emplace_back(); + grp.num_inds += prim_indices.size(); + grp.num_tris += draw.num_triangles; + grp.vis_idx_in_pc_bvh = UINT32_MAX; + } else { + auto& grp = draw.vis_groups.back(); + grp.num_inds += prim_indices.size(); + grp.num_tris += draw.num_triangles; + grp.vis_idx_in_pc_bvh = UINT32_MAX; + } + + draw.plain_indices.insert(draw.plain_indices.end(), prim_indices.begin(), + prim_indices.end()); + } + } + } + + for (const auto& [mat_idx, d_] : draw_by_material) { + out.strip_draws.push_back(d_); + auto& draw = out.strip_draws.back(); + draw.mode = make_default_draw_mode(); + + if (mat_idx == -1) { + lg::warn("Draw had a material index of -1, using default texture."); + draw.tree_tex_id = texture_pool_debug_checker(in.tex_pool); + continue; + } + const auto& mat = model.materials[mat_idx]; + int tex_idx = mat.pbrMetallicRoughness.baseColorTexture.index; + if (tex_idx == -1) { + lg::warn("Material {} has no texture, using default texture.", mat.name); + draw.tree_tex_id = texture_pool_debug_checker(in.tex_pool); + continue; + } + + const auto& tex = model.textures[tex_idx]; + ASSERT(tex.sampler >= 0); + ASSERT(tex.source >= 0); + draw.mode = draw_mode_from_sampler(model.samplers.at(tex.sampler)); + + const auto& img = model.images[tex.source]; + draw.tree_tex_id = texture_pool_add_texture(in.tex_pool, img); + } + lg::info("total of {} unique materials", out.strip_draws.size()); + + lg::info("Merged {} meshes and {} prims into {} vertices", mesh_count, prim_count, + out.vertices.size()); + + if (in.get_colors) { + Timer quantize_timer; + auto quantized = quantize_colors_octree(all_vtx_colors, 1024); + for (size_t i = 0; i < out.vertices.size(); i++) { + out.vertices[i].color_index = quantized.vtx_to_color[i]; + } + out.color_palette = std::move(quantized.final_colors); + lg::info("Color palette generation took {:.2f} ms", quantize_timer.getMs()); + } + + dedup_vertices(out); +} + +void extract(const Input& in, + CollideOutput& out, + const tinygltf::Model& model, + const std::vector& all_nodes) { + int mesh_count = 0; + int prim_count = 0; + int suspicious_faces = 0; + + for (const auto& n : all_nodes) { + const auto& node = model.nodes[n.node_idx]; + fmt::print("node: {} {}\n", node.name, node.mesh); + if (node.mesh >= 0) { + const auto& mesh = model.meshes[node.mesh]; + if (!mesh.extras.Has("collide")) { + // fmt::print("skip collide: {}\n", mesh.name); + // continue; + } + mesh_count++; + for (const auto& prim : mesh.primitives) { + prim_count++; + // extract index buffer + std::vector prim_indices = gltf_index_buffer(model, prim.indices, 0); + ASSERT_MSG(prim.mode == TINYGLTF_MODE_TRIANGLES, "Unsupported triangle mode"); + // extract vertices + auto verts = gltf_vertices(model, prim.attributes, n.w_T_node, false, true, mesh.name); + + for (size_t iidx = 0; iidx < prim_indices.size(); iidx += 3) { + CollideFace face; + + // get the positions + for (int j = 0; j < 3; j++) { + auto& vtx = verts.vtx.at(prim_indices.at(iidx + j)); + face.v[j].x() = vtx.x; + face.v[j].y() = vtx.y; + face.v[j].z() = vtx.z; + } + + // now face normal + math::Vector3f face_normal = + (face.v[2] - face.v[0]).cross(face.v[1] - face.v[0]).normalized(); + + float dots[3]; + for (int j = 0; j < 3; j++) { + dots[j] = face_normal.dot(verts.normals.at(prim_indices.at(iidx + j)).normalized()); + } + + if (dots[0] > 1e-3 && dots[1] > 1e-3 && dots[2] > 1e-3) { + suspicious_faces++; + std::swap(face.v[2], face.v[1]); + } + + face.bsphere = math::bsphere_of_triangle(face.v); + face.bsphere.w() += 1e-1; + for (int j = 0; j < 3; j++) { + float output_dist = face.bsphere.w() - (face.bsphere.xyz() - face.v[j]).length(); + if (output_dist < 0) { + fmt::print("{}\n", output_dist); + fmt::print("BAD:\n{}\n{}\n{}\n", face.v[0].to_string_aligned(), + face.v[1].to_string_aligned(), face.v[2].to_string_aligned()); + fmt::print("bsphere: {}\n", face.bsphere.to_string_aligned()); + } + } + + out.faces.push_back(face); + } + } + } + } + + lg::info("{} out of {} faces were suspicious (a small number is ok)", suspicious_faces, + out.faces.size()); + // lg::info("Collision extract{} {}", mesh_count, prim_count); +} + +void extract(const Input& in, Output& out) { + lg::info("Reading gltf mesh: {}", in.filename); + Timer read_timer; + tinygltf::TinyGLTF loader; + tinygltf::Model model; + std::string err, warn; + bool res = loader.LoadBinaryFromFile(&model, &err, &warn, in.filename); + ASSERT_MSG(warn.empty(), warn.c_str()); + ASSERT_MSG(err.empty(), err.c_str()); + ASSERT_MSG(res, "Failed to load GLTF file!"); + auto all_nodes = flatten_nodes_from_all_scenes(model); + extract(in, out.tfrag, model, all_nodes); + extract(in, out.collide, model, all_nodes); + lg::info("GLTF total took {:.2f} ms", read_timer.getMs()); +} +} // namespace gltf_mesh_extract diff --git a/goalc/build_level/gltf_mesh_extract.h b/goalc/build_level/gltf_mesh_extract.h new file mode 100644 index 0000000000..5bc0e47606 --- /dev/null +++ b/goalc/build_level/gltf_mesh_extract.h @@ -0,0 +1,34 @@ +#pragma once + +#include + +#include "common/custom_data/Tfrag3Data.h" +#include "goalc/build_level/TexturePool.h" +#include "goalc/build_level/collide_common.h" + +namespace gltf_mesh_extract { + +struct Input { + std::string filename; + TexturePool* tex_pool = nullptr; + bool get_colors = true; +}; + +struct TfragOutput { + std::vector strip_draws; + std::vector vertices; + std::vector> color_palette; +}; + +struct CollideOutput { + std::vector faces; +}; + +struct Output { + TfragOutput tfrag; + CollideOutput collide; +}; + +void extract(const Input& in, Output& out); + +} // namespace gltf_mesh_extract \ No newline at end of file diff --git a/goalc/data_compiler/DataObjectGenerator.cpp b/goalc/data_compiler/DataObjectGenerator.cpp index 12fa9767d7..6864bfa2f6 100644 --- a/goalc/data_compiler/DataObjectGenerator.cpp +++ b/goalc/data_compiler/DataObjectGenerator.cpp @@ -51,6 +51,17 @@ int DataObjectGenerator::add_word(u32 word) { return result; } +int DataObjectGenerator::add_word_float(float f) { + auto result = int(m_words.size()); + m_words.push_back(0); + memcpy(&m_words.back(), &f, sizeof(float)); + return result; +} + +void DataObjectGenerator::set_word(u32 word_idx, u32 val) { + m_words.at(word_idx) = val; +} + void DataObjectGenerator::link_word_to_word(int source, int target, int offset) { link_word_to_byte(source, target * 4 + offset); } @@ -68,6 +79,10 @@ int DataObjectGenerator::add_ref_to_string_in_pool(const std::string& str) { return result; } +void DataObjectGenerator::link_word_to_string_in_pool(const std::string& str, int word_idx) { + m_string_pool[str].push_back(word_idx); +} + int DataObjectGenerator::add_type_tag(const std::string& str) { auto result = int(m_words.size()); m_words.push_back(0); @@ -82,12 +97,20 @@ int DataObjectGenerator::add_symbol_link(const std::string& str) { return result; } +void DataObjectGenerator::link_word_to_symbol(const std::string& str, int word_idx) { + m_symbol_links[str].push_back(word_idx); +} + void DataObjectGenerator::align(int alignment_words) { while (m_words.size() % alignment_words) { m_words.push_back(0); } } +void DataObjectGenerator::align_to_basic() { + align(4); +} + int DataObjectGenerator::words() const { return int(m_words.size()); } diff --git a/goalc/data_compiler/DataObjectGenerator.h b/goalc/data_compiler/DataObjectGenerator.h index 3522b9a3b7..97e5b32dd1 100644 --- a/goalc/data_compiler/DataObjectGenerator.h +++ b/goalc/data_compiler/DataObjectGenerator.h @@ -8,15 +8,22 @@ class DataObjectGenerator { public: int add_word(u32 word); + int add_word_float(float f); + void set_word(u32 word_idx, u32 val); void link_word_to_word(int source, int target, int offset = 0); void link_word_to_byte(int source_word, int target_byte); int add_ref_to_string_in_pool(const std::string& str); + void link_word_to_string_in_pool(const std::string& str, int word_idx); int add_type_tag(const std::string& str); int add_symbol_link(const std::string& str); + void link_word_to_symbol(const std::string& str, int word_idx); std::vector generate_v2(); std::vector generate_v4(); void align(int alignment_words); + void align_to_basic(); int words() const; + size_t current_offset_bytes() const { return m_words.size() * sizeof(u32); } + u8* data() { return (u8*)m_words.data(); } private: void add_strings(); diff --git a/goalc/make/MakeSystem.cpp b/goalc/make/MakeSystem.cpp index a59154d44b..21f9c1b8d8 100644 --- a/goalc/make/MakeSystem.cpp +++ b/goalc/make/MakeSystem.cpp @@ -66,6 +66,7 @@ MakeSystem::MakeSystem() { add_tool(); add_tool(); add_tool(); + add_tool(); } /*! diff --git a/goalc/make/Tools.cpp b/goalc/make/Tools.cpp index d387bb75bf..01194a4ceb 100644 --- a/goalc/make/Tools.cpp +++ b/goalc/make/Tools.cpp @@ -1,5 +1,3 @@ - - #include "Tools.h" #include @@ -11,6 +9,7 @@ #include "goalc/data_compiler/dir_tpages.h" #include "goalc/data_compiler/game_count.h" #include "goalc/data_compiler/game_text_common.h" +#include "goalc/build_level/build_level.h" CompilerTool::CompilerTool(Compiler* compiler) : Tool("goalc"), m_compiler(compiler) {} @@ -186,3 +185,20 @@ bool SubtitleTool::run(const ToolInput& task) { } return true; } + +BuildLevelTool::BuildLevelTool() : Tool("build-level") {} + +bool BuildLevelTool::needs_run(const ToolInput& task) { + if (task.input.size() != 1) { + throw std::runtime_error(fmt::format("Invalid amount of inputs to {} tool", name())); + } + auto deps = get_build_level_deps(task.input.at(0)); + return Tool::needs_run({task.input, deps, task.output, task.arg}); +} + +bool BuildLevelTool::run(const ToolInput& task) { + if (task.input.size() != 1) { + throw std::runtime_error(fmt::format("Invalid amount of inputs to {} tool", name())); + } + return run_build_level(task.input.at(0), task.output.at(0)); +} diff --git a/goalc/make/Tools.h b/goalc/make/Tools.h index e6d61b2b04..966a1820f2 100644 --- a/goalc/make/Tools.h +++ b/goalc/make/Tools.h @@ -62,3 +62,10 @@ class SubtitleTool : public Tool { bool run(const ToolInput& task) override; bool needs_run(const ToolInput& task) override; }; + +class BuildLevelTool : public Tool { + public: + BuildLevelTool(); + bool run(const ToolInput& task) override; + bool needs_run(const ToolInput& task) override; +}; diff --git a/third-party/stb_image.h b/third-party/stb_image.h deleted file mode 100644 index 9c88b32170..0000000000 --- a/third-party/stb_image.h +++ /dev/null @@ -1,8709 +0,0 @@ -/* stb_image - v2.27 - public domain image loader - http://nothings.org/stb - no warranty implied; use at your own risk - - Do this: - #define STB_IMAGE_IMPLEMENTATION - before you include this file in *one* C or C++ file to create the implementation. - - // i.e. it should look like this: - #include ... - #include ... - #include ... - #define STB_IMAGE_IMPLEMENTATION - #include "stb_image.h" - - You can #define STBI_ASSERT(x) before the #include to avoid using assert.h. - And #define STBI_MALLOC, STBI_REALLOC, and STBI_FREE to avoid using malloc,realloc,free - - - QUICK NOTES: - Primarily of interest to game developers and other people who can - avoid problematic images and only need the trivial interface - - JPEG baseline & progressive (12 bpc/arithmetic not supported, same as stock IJG lib) - PNG 1/2/4/8/16-bit-per-channel - - TGA (not sure what subset, if a subset) - BMP non-1bpp, non-RLE - PSD (composited view only, no extra channels, 8/16 bit-per-channel) - - GIF (*comp always reports as 4-channel) - HDR (radiance rgbE format) - PIC (Softimage PIC) - PNM (PPM and PGM binary only) - - Animated GIF still needs a proper API, but here's one way to do it: - http://gist.github.com/urraka/685d9a6340b26b830d49 - - - decode from memory or through FILE (define STBI_NO_STDIO to remove code) - - decode from arbitrary I/O callbacks - - SIMD acceleration on x86/x64 (SSE2) and ARM (NEON) - - Full documentation under "DOCUMENTATION" below. - - -LICENSE - - See end of file for license information. - -RECENT REVISION HISTORY: - - 2.27 (2021-07-11) document stbi_info better, 16-bit PNM support, bug fixes - 2.26 (2020-07-13) many minor fixes - 2.25 (2020-02-02) fix warnings - 2.24 (2020-02-02) fix warnings; thread-local failure_reason and flip_vertically - 2.23 (2019-08-11) fix clang static analysis warning - 2.22 (2019-03-04) gif fixes, fix warnings - 2.21 (2019-02-25) fix typo in comment - 2.20 (2019-02-07) support utf8 filenames in Windows; fix warnings and platform ifdefs - 2.19 (2018-02-11) fix warning - 2.18 (2018-01-30) fix warnings - 2.17 (2018-01-29) bugfix, 1-bit BMP, 16-bitness query, fix warnings - 2.16 (2017-07-23) all functions have 16-bit variants; optimizations; bugfixes - 2.15 (2017-03-18) fix png-1,2,4; all Imagenet JPGs; no runtime SSE detection on GCC - 2.14 (2017-03-03) remove deprecated STBI_JPEG_OLD; fixes for Imagenet JPGs - 2.13 (2016-12-04) experimental 16-bit API, only for PNG so far; fixes - 2.12 (2016-04-02) fix typo in 2.11 PSD fix that caused crashes - 2.11 (2016-04-02) 16-bit PNGS; enable SSE2 in non-gcc x64 - RGB-format JPEG; remove white matting in PSD; - allocate large structures on the stack; - correct channel count for PNG & BMP - 2.10 (2016-01-22) avoid warning introduced in 2.09 - 2.09 (2016-01-16) 16-bit TGA; comments in PNM files; STBI_REALLOC_SIZED - - See end of file for full revision history. - - - ============================ Contributors ========================= - - Image formats Extensions, features - Sean Barrett (jpeg, png, bmp) Jetro Lauha (stbi_info) - Nicolas Schulz (hdr, psd) Martin "SpartanJ" Golini (stbi_info) - Jonathan Dummer (tga) James "moose2000" Brown (iPhone PNG) - Jean-Marc Lienher (gif) Ben "Disch" Wenger (io callbacks) - Tom Seddon (pic) Omar Cornut (1/2/4-bit PNG) - Thatcher Ulrich (psd) Nicolas Guillemot (vertical flip) - Ken Miller (pgm, ppm) Richard Mitton (16-bit PSD) - github:urraka (animated gif) Junggon Kim (PNM comments) - Christopher Forseth (animated gif) Daniel Gibson (16-bit TGA) - socks-the-fox (16-bit PNG) - Jeremy Sawicki (handle all ImageNet JPGs) - Optimizations & bugfixes Mikhail Morozov (1-bit BMP) - Fabian "ryg" Giesen Anael Seghezzi (is-16-bit query) - Arseny Kapoulkine Simon Breuss (16-bit PNM) - John-Mark Allen - Carmelo J Fdez-Aguera - - Bug & warning fixes - Marc LeBlanc David Woo Guillaume George Martins Mozeiko - Christpher Lloyd Jerry Jansson Joseph Thomson Blazej Dariusz Roszkowski - Phil Jordan Dave Moore Roy Eltham - Hayaki Saito Nathan Reed Won Chun - Luke Graham Johan Duparc Nick Verigakis the Horde3D community - Thomas Ruf Ronny Chevalier github:rlyeh - Janez Zemva John Bartholomew Michal Cichon github:romigrou - Jonathan Blow Ken Hamada Tero Hanninen github:svdijk - Eugene Golushkov Laurent Gomila Cort Stratton github:snagar - Aruelien Pocheville Sergio Gonzalez Thibault Reuille github:Zelex - Cass Everitt Ryamond Barbiero github:grim210 - Paul Du Bois Engin Manap Aldo Culquicondor github:sammyhw - Philipp Wiesemann Dale Weiler Oriol Ferrer Mesia github:phprus - Josh Tobin Matthew Gregan github:poppolopoppo - Julian Raschke Gregory Mullen Christian Floisand github:darealshinji - Baldur Karlsson Kevin Schmidt JR Smith github:Michaelangel007 - Brad Weinberger Matvey Cherevko github:mosra - Luca Sas Alexander Veselov Zack Middleton [reserved] - Ryan C. Gordon [reserved] [reserved] - DO NOT ADD YOUR NAME HERE - - Jacko Dirks - - To add your name to the credits, pick a random blank space in the middle and fill it. - 80% of merge conflicts on stb PRs are due to people adding their name at the end - of the credits. -*/ - -#ifndef STBI_INCLUDE_STB_IMAGE_H -#define STBI_INCLUDE_STB_IMAGE_H - -// DOCUMENTATION -// -// Limitations: -// - no 12-bit-per-channel JPEG -// - no JPEGs with arithmetic coding -// - GIF always returns *comp=4 -// -// Basic usage (see HDR discussion below for HDR usage): -// int x,y,n; -// unsigned char *data = stbi_load(filename, &x, &y, &n, 0); -// // ... process data if not NULL ... -// // ... x = width, y = height, n = # 8-bit components per pixel ... -// // ... replace '0' with '1'..'4' to force that many components per pixel -// // ... but 'n' will always be the number that it would have been if you said 0 -// stbi_image_free(data) -// -// Standard parameters: -// int *x -- outputs image width in pixels -// int *y -- outputs image height in pixels -// int *channels_in_file -- outputs # of image components in image file -// int desired_channels -- if non-zero, # of image components requested in result -// -// The return value from an image loader is an 'unsigned char *' which points -// to the pixel data, or NULL on an allocation failure or if the image is -// corrupt or invalid. The pixel data consists of *y scanlines of *x pixels, -// with each pixel consisting of N interleaved 8-bit components; the first -// pixel pointed to is top-left-most in the image. There is no padding between -// image scanlines or between pixels, regardless of format. The number of -// components N is 'desired_channels' if desired_channels is non-zero, or -// *channels_in_file otherwise. If desired_channels is non-zero, -// *channels_in_file has the number of components that _would_ have been -// output otherwise. E.g. if you set desired_channels to 4, you will always -// get RGBA output, but you can check *channels_in_file to see if it's trivially -// opaque because e.g. there were only 3 channels in the source image. -// -// An output image with N components has the following components interleaved -// in this order in each pixel: -// -// N=#comp components -// 1 grey -// 2 grey, alpha -// 3 red, green, blue -// 4 red, green, blue, alpha -// -// If image loading fails for any reason, the return value will be NULL, -// and *x, *y, *channels_in_file will be unchanged. The function -// stbi_failure_reason() can be queried for an extremely brief, end-user -// unfriendly explanation of why the load failed. Define STBI_NO_FAILURE_STRINGS -// to avoid compiling these strings at all, and STBI_FAILURE_USERMSG to get slightly -// more user-friendly ones. -// -// Paletted PNG, BMP, GIF, and PIC images are automatically depalettized. -// -// To query the width, height and component count of an image without having to -// decode the full file, you can use the stbi_info family of functions: -// -// int x,y,n,ok; -// ok = stbi_info(filename, &x, &y, &n); -// // returns ok=1 and sets x, y, n if image is a supported format, -// // 0 otherwise. -// -// Note that stb_image pervasively uses ints in its public API for sizes, -// including sizes of memory buffers. This is now part of the API and thus -// hard to change without causing breakage. As a result, the various image -// loaders all have certain limits on image size; these differ somewhat -// by format but generally boil down to either just under 2GB or just under -// 1GB. When the decoded image would be larger than this, stb_image decoding -// will fail. -// -// Additionally, stb_image will reject image files that have any of their -// dimensions set to a larger value than the configurable STBI_MAX_DIMENSIONS, -// which defaults to 2**24 = 16777216 pixels. Due to the above memory limit, -// the only way to have an image with such dimensions load correctly -// is for it to have a rather extreme aspect ratio. Either way, the -// assumption here is that such larger images are likely to be malformed -// or malicious. If you do need to load an image with individual dimensions -// larger than that, and it still fits in the overall size limit, you can -// #define STBI_MAX_DIMENSIONS on your own to be something larger. -// -// =========================================================================== -// -// UNICODE: -// -// If compiling for Windows and you wish to use Unicode filenames, compile -// with -// #define STBI_WINDOWS_UTF8 -// and pass utf8-encoded filenames. Call stbi_convert_wchar_to_utf8 to convert -// Windows wchar_t filenames to utf8. -// -// =========================================================================== -// -// Philosophy -// -// stb libraries are designed with the following priorities: -// -// 1. easy to use -// 2. easy to maintain -// 3. good performance -// -// Sometimes I let "good performance" creep up in priority over "easy to maintain", -// and for best performance I may provide less-easy-to-use APIs that give higher -// performance, in addition to the easy-to-use ones. Nevertheless, it's important -// to keep in mind that from the standpoint of you, a client of this library, -// all you care about is #1 and #3, and stb libraries DO NOT emphasize #3 above all. -// -// Some secondary priorities arise directly from the first two, some of which -// provide more explicit reasons why performance can't be emphasized. -// -// - Portable ("ease of use") -// - Small source code footprint ("easy to maintain") -// - No dependencies ("ease of use") -// -// =========================================================================== -// -// I/O callbacks -// -// I/O callbacks allow you to read from arbitrary sources, like packaged -// files or some other source. Data read from callbacks are processed -// through a small internal buffer (currently 128 bytes) to try to reduce -// overhead. -// -// The three functions you must define are "read" (reads some bytes of data), -// "skip" (skips some bytes of data), "eof" (reports if the stream is at the end). -// -// =========================================================================== -// -// SIMD support -// -// The JPEG decoder will try to automatically use SIMD kernels on x86 when -// supported by the compiler. For ARM Neon support, you must explicitly -// request it. -// -// (The old do-it-yourself SIMD API is no longer supported in the current -// code.) -// -// On x86, SSE2 will automatically be used when available based on a run-time -// test; if not, the generic C versions are used as a fall-back. On ARM targets, -// the typical path is to have separate builds for NEON and non-NEON devices -// (at least this is true for iOS and Android). Therefore, the NEON support is -// toggled by a build flag: define STBI_NEON to get NEON loops. -// -// If for some reason you do not want to use any of SIMD code, or if -// you have issues compiling it, you can disable it entirely by -// defining STBI_NO_SIMD. -// -// =========================================================================== -// -// HDR image support (disable by defining STBI_NO_HDR) -// -// stb_image supports loading HDR images in general, and currently the Radiance -// .HDR file format specifically. You can still load any file through the existing -// interface; if you attempt to load an HDR file, it will be automatically remapped -// to LDR, assuming gamma 2.2 and an arbitrary scale factor defaulting to 1; -// both of these constants can be reconfigured through this interface: -// -// stbi_hdr_to_ldr_gamma(2.2f); -// stbi_hdr_to_ldr_scale(1.0f); -// -// (note, do not use _inverse_ constants; stbi_image will invert them -// appropriately). -// -// Additionally, there is a new, parallel interface for loading files as -// (linear) floats to preserve the full dynamic range: -// -// float *data = stbi_loadf(filename, &x, &y, &n, 0); -// -// If you load LDR images through this interface, those images will -// be promoted to floating point values, run through the inverse of -// constants corresponding to the above: -// -// stbi_ldr_to_hdr_scale(1.0f); -// stbi_ldr_to_hdr_gamma(2.2f); -// -// Finally, given a filename (or an open file or memory block--see header -// file for details) containing image data, you can query for the "most -// appropriate" interface to use (that is, whether the image is HDR or -// not), using: -// -// stbi_is_hdr(char *filename); -// -// =========================================================================== -// -// iPhone PNG support: -// -// We optionally support converting iPhone-formatted PNGs (which store -// premultiplied BGRA) back to RGB, even though they're internally encoded -// differently. To enable this conversion, call -// stbi_convert_iphone_png_to_rgb(1). -// -// Call stbi_set_unpremultiply_on_load(1) as well to force a divide per -// pixel to remove any premultiplied alpha *only* if the image file explicitly -// says there's premultiplied data (currently only happens in iPhone images, -// and only if iPhone convert-to-rgb processing is on). -// -// =========================================================================== -// -// ADDITIONAL CONFIGURATION -// -// - You can suppress implementation of any of the decoders to reduce -// your code footprint by #defining one or more of the following -// symbols before creating the implementation. -// -// STBI_NO_JPEG -// STBI_NO_PNG -// STBI_NO_BMP -// STBI_NO_PSD -// STBI_NO_TGA -// STBI_NO_GIF -// STBI_NO_HDR -// STBI_NO_PIC -// STBI_NO_PNM (.ppm and .pgm) -// -// - You can request *only* certain decoders and suppress all other ones -// (this will be more forward-compatible, as addition of new decoders -// doesn't require you to disable them explicitly): -// -// STBI_ONLY_JPEG -// STBI_ONLY_PNG -// STBI_ONLY_BMP -// STBI_ONLY_PSD -// STBI_ONLY_TGA -// STBI_ONLY_GIF -// STBI_ONLY_HDR -// STBI_ONLY_PIC -// STBI_ONLY_PNM (.ppm and .pgm) -// -// - If you use STBI_NO_PNG (or _ONLY_ without PNG), and you still -// want the zlib decoder to be available, #define STBI_SUPPORT_ZLIB -// -// - If you define STBI_MAX_DIMENSIONS, stb_image will reject images greater -// than that size (in either width or height) without further processing. -// This is to let programs in the wild set an upper bound to prevent -// denial-of-service attacks on untrusted data, as one could generate a -// valid image of gigantic dimensions and force stb_image to allocate a -// huge block of memory and spend disproportionate time decoding it. By -// default this is set to (1 << 24), which is 16777216, but that's still -// very big. - -#ifndef STBI_NO_STDIO -#include -#endif // STBI_NO_STDIO - -#define STBI_VERSION 1 - -enum { - STBI_default = 0, // only used for desired_channels - - STBI_grey = 1, - STBI_grey_alpha = 2, - STBI_rgb = 3, - STBI_rgb_alpha = 4 -}; - -#include -typedef unsigned char stbi_uc; -typedef unsigned short stbi_us; - -#ifdef __cplusplus -extern "C" { -#endif - -#ifndef STBIDEF -#ifdef STB_IMAGE_STATIC -#define STBIDEF static -#else -#define STBIDEF extern -#endif -#endif - -////////////////////////////////////////////////////////////////////////////// -// -// PRIMARY API - works on images of any type -// - -// -// load image by filename, open file, or memory buffer -// - -typedef struct { - int (*read)(void* user, - char* data, - int size); // fill 'data' with 'size' bytes. return number of bytes actually read - void (*skip)(void* user, - int n); // skip the next 'n' bytes, or 'unget' the last -n bytes if negative - int (*eof)(void* user); // returns nonzero if we are at end of file/data -} stbi_io_callbacks; - -//////////////////////////////////// -// -// 8-bits-per-channel interface -// - -STBIDEF stbi_uc* stbi_load_from_memory(stbi_uc const* buffer, - int len, - int* x, - int* y, - int* channels_in_file, - int desired_channels); -STBIDEF stbi_uc* stbi_load_from_callbacks(stbi_io_callbacks const* clbk, - void* user, - int* x, - int* y, - int* channels_in_file, - int desired_channels); - -#ifndef STBI_NO_STDIO -STBIDEF stbi_uc* stbi_load(char const* filename, - int* x, - int* y, - int* channels_in_file, - int desired_channels); -STBIDEF stbi_uc* stbi_load_from_file(FILE* f, - int* x, - int* y, - int* channels_in_file, - int desired_channels); -// for stbi_load_from_file, file pointer is left pointing immediately after image -#endif - -#ifndef STBI_NO_GIF -STBIDEF stbi_uc* stbi_load_gif_from_memory(stbi_uc const* buffer, - int len, - int** delays, - int* x, - int* y, - int* z, - int* comp, - int req_comp); -#endif - -#ifdef STBI_WINDOWS_UTF8 -STBIDEF int stbi_convert_wchar_to_utf8(char* buffer, size_t bufferlen, const wchar_t* input); -#endif - -//////////////////////////////////// -// -// 16-bits-per-channel interface -// - -STBIDEF stbi_us* stbi_load_16_from_memory(stbi_uc const* buffer, - int len, - int* x, - int* y, - int* channels_in_file, - int desired_channels); -STBIDEF stbi_us* stbi_load_16_from_callbacks(stbi_io_callbacks const* clbk, - void* user, - int* x, - int* y, - int* channels_in_file, - int desired_channels); - -#ifndef STBI_NO_STDIO -STBIDEF stbi_us* stbi_load_16(char const* filename, - int* x, - int* y, - int* channels_in_file, - int desired_channels); -STBIDEF stbi_us* stbi_load_from_file_16(FILE* f, - int* x, - int* y, - int* channels_in_file, - int desired_channels); -#endif - -//////////////////////////////////// -// -// float-per-channel interface -// -#ifndef STBI_NO_LINEAR -STBIDEF float* stbi_loadf_from_memory(stbi_uc const* buffer, - int len, - int* x, - int* y, - int* channels_in_file, - int desired_channels); -STBIDEF float* stbi_loadf_from_callbacks(stbi_io_callbacks const* clbk, - void* user, - int* x, - int* y, - int* channels_in_file, - int desired_channels); - -#ifndef STBI_NO_STDIO -STBIDEF float* stbi_loadf(char const* filename, - int* x, - int* y, - int* channels_in_file, - int desired_channels); -STBIDEF float* stbi_loadf_from_file(FILE* f, - int* x, - int* y, - int* channels_in_file, - int desired_channels); -#endif -#endif - -#ifndef STBI_NO_HDR -STBIDEF void stbi_hdr_to_ldr_gamma(float gamma); -STBIDEF void stbi_hdr_to_ldr_scale(float scale); -#endif // STBI_NO_HDR - -#ifndef STBI_NO_LINEAR -STBIDEF void stbi_ldr_to_hdr_gamma(float gamma); -STBIDEF void stbi_ldr_to_hdr_scale(float scale); -#endif // STBI_NO_LINEAR - -// stbi_is_hdr is always defined, but always returns false if STBI_NO_HDR -STBIDEF int stbi_is_hdr_from_callbacks(stbi_io_callbacks const* clbk, void* user); -STBIDEF int stbi_is_hdr_from_memory(stbi_uc const* buffer, int len); -#ifndef STBI_NO_STDIO -STBIDEF int stbi_is_hdr(char const* filename); -STBIDEF int stbi_is_hdr_from_file(FILE* f); -#endif // STBI_NO_STDIO - -// get a VERY brief reason for failure -// on most compilers (and ALL modern mainstream compilers) this is threadsafe -STBIDEF const char* stbi_failure_reason(void); - -// free the loaded image -- this is just free() -STBIDEF void stbi_image_free(void* retval_from_stbi_load); - -// get image dimensions & components without fully decoding -STBIDEF int stbi_info_from_memory(stbi_uc const* buffer, int len, int* x, int* y, int* comp); -STBIDEF int stbi_info_from_callbacks(stbi_io_callbacks const* clbk, - void* user, - int* x, - int* y, - int* comp); -STBIDEF int stbi_is_16_bit_from_memory(stbi_uc const* buffer, int len); -STBIDEF int stbi_is_16_bit_from_callbacks(stbi_io_callbacks const* clbk, void* user); - -#ifndef STBI_NO_STDIO -STBIDEF int stbi_info(char const* filename, int* x, int* y, int* comp); -STBIDEF int stbi_info_from_file(FILE* f, int* x, int* y, int* comp); -STBIDEF int stbi_is_16_bit(char const* filename); -STBIDEF int stbi_is_16_bit_from_file(FILE* f); -#endif - -// for image formats that explicitly notate that they have premultiplied alpha, -// we just return the colors as stored in the file. set this flag to force -// unpremultiplication. results are undefined if the unpremultiply overflow. -STBIDEF void stbi_set_unpremultiply_on_load(int flag_true_if_should_unpremultiply); - -// indicate whether we should process iphone images back to canonical format, -// or just pass them through "as-is" -STBIDEF void stbi_convert_iphone_png_to_rgb(int flag_true_if_should_convert); - -// flip the image vertically, so the first pixel in the output array is the bottom left -STBIDEF void stbi_set_flip_vertically_on_load(int flag_true_if_should_flip); - -// as above, but only applies to images loaded on the thread that calls the function -// this function is only available if your compiler supports thread-local variables; -// calling it will fail to link if your compiler doesn't -STBIDEF void stbi_set_unpremultiply_on_load_thread(int flag_true_if_should_unpremultiply); -STBIDEF void stbi_convert_iphone_png_to_rgb_thread(int flag_true_if_should_convert); -STBIDEF void stbi_set_flip_vertically_on_load_thread(int flag_true_if_should_flip); - -// ZLIB client - used by PNG, available for other purposes - -STBIDEF char* stbi_zlib_decode_malloc_guesssize(const char* buffer, - int len, - int initial_size, - int* outlen); -STBIDEF char* stbi_zlib_decode_malloc_guesssize_headerflag(const char* buffer, - int len, - int initial_size, - int* outlen, - int parse_header); -STBIDEF char* stbi_zlib_decode_malloc(const char* buffer, int len, int* outlen); -STBIDEF int stbi_zlib_decode_buffer(char* obuffer, int olen, const char* ibuffer, int ilen); - -STBIDEF char* stbi_zlib_decode_noheader_malloc(const char* buffer, int len, int* outlen); -STBIDEF int stbi_zlib_decode_noheader_buffer(char* obuffer, - int olen, - const char* ibuffer, - int ilen); - -#ifdef __cplusplus -} -#endif - -// -// -//// end header file ///////////////////////////////////////////////////// -#endif // STBI_INCLUDE_STB_IMAGE_H - -#ifdef STB_IMAGE_IMPLEMENTATION - -#if defined(STBI_ONLY_JPEG) || defined(STBI_ONLY_PNG) || defined(STBI_ONLY_BMP) || \ - defined(STBI_ONLY_TGA) || defined(STBI_ONLY_GIF) || defined(STBI_ONLY_PSD) || \ - defined(STBI_ONLY_HDR) || defined(STBI_ONLY_PIC) || defined(STBI_ONLY_PNM) || \ - defined(STBI_ONLY_ZLIB) -#ifndef STBI_ONLY_JPEG -#define STBI_NO_JPEG -#endif -#ifndef STBI_ONLY_PNG -#define STBI_NO_PNG -#endif -#ifndef STBI_ONLY_BMP -#define STBI_NO_BMP -#endif -#ifndef STBI_ONLY_PSD -#define STBI_NO_PSD -#endif -#ifndef STBI_ONLY_TGA -#define STBI_NO_TGA -#endif -#ifndef STBI_ONLY_GIF -#define STBI_NO_GIF -#endif -#ifndef STBI_ONLY_HDR -#define STBI_NO_HDR -#endif -#ifndef STBI_ONLY_PIC -#define STBI_NO_PIC -#endif -#ifndef STBI_ONLY_PNM -#define STBI_NO_PNM -#endif -#endif - -#if defined(STBI_NO_PNG) && !defined(STBI_SUPPORT_ZLIB) && !defined(STBI_NO_ZLIB) -#define STBI_NO_ZLIB -#endif - -#include -#include // ptrdiff_t on osx -#include -#include -#include - -#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) -#include // ldexp, pow -#endif - -#ifndef STBI_NO_STDIO -#include -#endif - -#ifndef STBI_ASSERT -#include -#define STBI_ASSERT(x) assert(x) -#endif - -#ifdef __cplusplus -#define STBI_EXTERN extern "C" -#else -#define STBI_EXTERN extern -#endif - -#ifndef _MSC_VER -#ifdef __cplusplus -#define stbi_inline inline -#else -#define stbi_inline -#endif -#else -#define stbi_inline __forceinline -#endif - -#ifndef STBI_NO_THREAD_LOCALS -#if defined(__cplusplus) && __cplusplus >= 201103L -#define STBI_THREAD_LOCAL thread_local -#elif defined(__GNUC__) && __GNUC__ < 5 -#define STBI_THREAD_LOCAL __thread -#elif defined(_MSC_VER) -#define STBI_THREAD_LOCAL __declspec(thread) -#elif defined(__STDC_VERSION__) && __STDC_VERSION__ >= 201112L && !defined(__STDC_NO_THREADS__) -#define STBI_THREAD_LOCAL _Thread_local -#endif - -#ifndef STBI_THREAD_LOCAL -#if defined(__GNUC__) -#define STBI_THREAD_LOCAL __thread -#endif -#endif -#endif - -#ifdef _MSC_VER -typedef unsigned short stbi__uint16; -typedef signed short stbi__int16; -typedef unsigned int stbi__uint32; -typedef signed int stbi__int32; -#else -#include -typedef uint16_t stbi__uint16; -typedef int16_t stbi__int16; -typedef uint32_t stbi__uint32; -typedef int32_t stbi__int32; -#endif - -// should produce compiler error if size is wrong -typedef unsigned char validate_uint32[sizeof(stbi__uint32) == 4 ? 1 : -1]; - -#ifdef _MSC_VER -#define STBI_NOTUSED(v) (void)(v) -#else -#define STBI_NOTUSED(v) (void)sizeof(v) -#endif - -#ifdef _MSC_VER -#define STBI_HAS_LROTL -#endif - -#ifdef STBI_HAS_LROTL -#define stbi_lrot(x, y) _lrotl(x, y) -#else -#define stbi_lrot(x, y) (((x) << (y)) | ((x) >> (-(y)&31))) -#endif - -#if defined(STBI_MALLOC) && defined(STBI_FREE) && \ - (defined(STBI_REALLOC) || defined(STBI_REALLOC_SIZED)) -// ok -#elif !defined(STBI_MALLOC) && !defined(STBI_FREE) && !defined(STBI_REALLOC) && \ - !defined(STBI_REALLOC_SIZED) -// ok -#else -#error \ - "Must define all or none of STBI_MALLOC, STBI_FREE, and STBI_REALLOC (or STBI_REALLOC_SIZED)." -#endif - -#ifndef STBI_MALLOC -#define STBI_MALLOC(sz) malloc(sz) -#define STBI_REALLOC(p, newsz) realloc(p, newsz) -#define STBI_FREE(p) free(p) -#endif - -#ifndef STBI_REALLOC_SIZED -#define STBI_REALLOC_SIZED(p, oldsz, newsz) STBI_REALLOC(p, newsz) -#endif - -// x86/x64 detection -#if defined(__x86_64__) || defined(_M_X64) -#define STBI__X64_TARGET -#elif defined(__i386) || defined(_M_IX86) -#define STBI__X86_TARGET -#endif - -#if defined(__GNUC__) && defined(STBI__X86_TARGET) && !defined(__SSE2__) && !defined(STBI_NO_SIMD) -// gcc doesn't support sse2 intrinsics unless you compile with -msse2, -// which in turn means it gets to use SSE2 everywhere. This is unfortunate, -// but previous attempts to provide the SSE2 functions with runtime -// detection caused numerous issues. The way architecture extensions are -// exposed in GCC/Clang is, sadly, not really suited for one-file libs. -// New behavior: if compiled with -msse2, we use SSE2 without any -// detection; if not, we don't use it at all. -#define STBI_NO_SIMD -#endif - -#if defined(__MINGW32__) && defined(STBI__X86_TARGET) && !defined(STBI_MINGW_ENABLE_SSE2) && \ - !defined(STBI_NO_SIMD) -// Note that __MINGW32__ doesn't actually mean 32-bit, so we have to avoid STBI__X64_TARGET -// -// 32-bit MinGW wants ESP to be 16-byte aligned, but this is not in the -// Windows ABI and VC++ as well as Windows DLLs don't maintain that invariant. -// As a result, enabling SSE2 on 32-bit MinGW is dangerous when not -// simultaneously enabling "-mstackrealign". -// -// See https://github.com/nothings/stb/issues/81 for more information. -// -// So default to no SSE2 on 32-bit MinGW. If you've read this far and added -// -mstackrealign to your build settings, feel free to #define STBI_MINGW_ENABLE_SSE2. -#define STBI_NO_SIMD -#endif - -#if !defined(STBI_NO_SIMD) && (defined(STBI__X86_TARGET) || defined(STBI__X64_TARGET)) -#define STBI_SSE2 -#include - -#ifdef _MSC_VER - -#if _MSC_VER >= 1400 // not VC6 -#include // __cpuid -static int stbi__cpuid3(void) { - int info[4]; - __cpuid(info, 1); - return info[3]; -} -#else -static int stbi__cpuid3(void) { - int res; - __asm { - mov eax,1 - cpuid - mov res,edx - } - return res; -} -#endif - -#define STBI_SIMD_ALIGN(type, name) __declspec(align(16)) type name - -#if !defined(STBI_NO_JPEG) && defined(STBI_SSE2) -static int stbi__sse2_available(void) { - int info3 = stbi__cpuid3(); - return ((info3 >> 26) & 1) != 0; -} -#endif - -#else // assume GCC-style if not VC++ -#define STBI_SIMD_ALIGN(type, name) type name __attribute__((aligned(16))) - -#if !defined(STBI_NO_JPEG) && defined(STBI_SSE2) -static int stbi__sse2_available(void) { - // If we're even attempting to compile this on GCC/Clang, that means - // -msse2 is on, which means the compiler is allowed to use SSE2 - // instructions at will, and so are we. - return 1; -} -#endif - -#endif -#endif - -// ARM NEON -#if defined(STBI_NO_SIMD) && defined(STBI_NEON) -#undef STBI_NEON -#endif - -#ifdef STBI_NEON -#include -#ifdef _MSC_VER -#define STBI_SIMD_ALIGN(type, name) __declspec(align(16)) type name -#else -#define STBI_SIMD_ALIGN(type, name) type name __attribute__((aligned(16))) -#endif -#endif - -#ifndef STBI_SIMD_ALIGN -#define STBI_SIMD_ALIGN(type, name) type name -#endif - -#ifndef STBI_MAX_DIMENSIONS -#define STBI_MAX_DIMENSIONS (1 << 24) -#endif - -/////////////////////////////////////////////// -// -// stbi__context struct and start_xxx functions - -// stbi__context structure is our basic context used by all images, so it -// contains all the IO context, plus some basic image information -typedef struct { - stbi__uint32 img_x, img_y; - int img_n, img_out_n; - - stbi_io_callbacks io; - void* io_user_data; - - int read_from_callbacks; - int buflen; - stbi_uc buffer_start[128]; - int callback_already_read; - - stbi_uc *img_buffer, *img_buffer_end; - stbi_uc *img_buffer_original, *img_buffer_original_end; -} stbi__context; - -static void stbi__refill_buffer(stbi__context* s); - -// initialize a memory-decode context -static void stbi__start_mem(stbi__context* s, stbi_uc const* buffer, int len) { - s->io.read = NULL; - s->read_from_callbacks = 0; - s->callback_already_read = 0; - s->img_buffer = s->img_buffer_original = (stbi_uc*)buffer; - s->img_buffer_end = s->img_buffer_original_end = (stbi_uc*)buffer + len; -} - -// initialize a callback-based context -static void stbi__start_callbacks(stbi__context* s, stbi_io_callbacks* c, void* user) { - s->io = *c; - s->io_user_data = user; - s->buflen = sizeof(s->buffer_start); - s->read_from_callbacks = 1; - s->callback_already_read = 0; - s->img_buffer = s->img_buffer_original = s->buffer_start; - stbi__refill_buffer(s); - s->img_buffer_original_end = s->img_buffer_end; -} - -#ifndef STBI_NO_STDIO - -static int stbi__stdio_read(void* user, char* data, int size) { - return (int)fread(data, 1, size, (FILE*)user); -} - -static void stbi__stdio_skip(void* user, int n) { - int ch; - fseek((FILE*)user, n, SEEK_CUR); - ch = fgetc((FILE*)user); /* have to read a byte to reset feof()'s flag */ - if (ch != EOF) { - ungetc(ch, (FILE*)user); /* push byte back onto stream if valid. */ - } -} - -static int stbi__stdio_eof(void* user) { - return feof((FILE*)user) || ferror((FILE*)user); -} - -static stbi_io_callbacks stbi__stdio_callbacks = { - stbi__stdio_read, - stbi__stdio_skip, - stbi__stdio_eof, -}; - -static void stbi__start_file(stbi__context* s, FILE* f) { - stbi__start_callbacks(s, &stbi__stdio_callbacks, (void*)f); -} - -// static void stop_file(stbi__context *s) { } - -#endif // !STBI_NO_STDIO - -static void stbi__rewind(stbi__context* s) { - // conceptually rewind SHOULD rewind to the beginning of the stream, - // but we just rewind to the beginning of the initial buffer, because - // we only use it after doing 'test', which only ever looks at at most 92 bytes - s->img_buffer = s->img_buffer_original; - s->img_buffer_end = s->img_buffer_original_end; -} - -enum { STBI_ORDER_RGB, STBI_ORDER_BGR }; - -typedef struct { - int bits_per_channel; - int num_channels; - int channel_order; -} stbi__result_info; - -#ifndef STBI_NO_JPEG -static int stbi__jpeg_test(stbi__context* s); -static void* stbi__jpeg_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri); -static int stbi__jpeg_info(stbi__context* s, int* x, int* y, int* comp); -#endif - -#ifndef STBI_NO_PNG -static int stbi__png_test(stbi__context* s); -static void* stbi__png_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri); -static int stbi__png_info(stbi__context* s, int* x, int* y, int* comp); -static int stbi__png_is16(stbi__context* s); -#endif - -#ifndef STBI_NO_BMP -static int stbi__bmp_test(stbi__context* s); -static void* stbi__bmp_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri); -static int stbi__bmp_info(stbi__context* s, int* x, int* y, int* comp); -#endif - -#ifndef STBI_NO_TGA -static int stbi__tga_test(stbi__context* s); -static void* stbi__tga_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri); -static int stbi__tga_info(stbi__context* s, int* x, int* y, int* comp); -#endif - -#ifndef STBI_NO_PSD -static int stbi__psd_test(stbi__context* s); -static void* stbi__psd_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri, - int bpc); -static int stbi__psd_info(stbi__context* s, int* x, int* y, int* comp); -static int stbi__psd_is16(stbi__context* s); -#endif - -#ifndef STBI_NO_HDR -static int stbi__hdr_test(stbi__context* s); -static float* stbi__hdr_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri); -static int stbi__hdr_info(stbi__context* s, int* x, int* y, int* comp); -#endif - -#ifndef STBI_NO_PIC -static int stbi__pic_test(stbi__context* s); -static void* stbi__pic_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri); -static int stbi__pic_info(stbi__context* s, int* x, int* y, int* comp); -#endif - -#ifndef STBI_NO_GIF -static int stbi__gif_test(stbi__context* s); -static void* stbi__gif_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri); -static void* stbi__load_gif_main(stbi__context* s, - int** delays, - int* x, - int* y, - int* z, - int* comp, - int req_comp); -static int stbi__gif_info(stbi__context* s, int* x, int* y, int* comp); -#endif - -#ifndef STBI_NO_PNM -static int stbi__pnm_test(stbi__context* s); -static void* stbi__pnm_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri); -static int stbi__pnm_info(stbi__context* s, int* x, int* y, int* comp); -static int stbi__pnm_is16(stbi__context* s); -#endif - -static -#ifdef STBI_THREAD_LOCAL - STBI_THREAD_LOCAL -#endif - const char* stbi__g_failure_reason; - -STBIDEF const char* stbi_failure_reason(void) { - return stbi__g_failure_reason; -} - -#ifndef STBI_NO_FAILURE_STRINGS -static int stbi__err(const char* str) { - stbi__g_failure_reason = str; - return 0; -} -#endif - -static void* stbi__malloc(size_t size) { - return STBI_MALLOC(size); -} - -// stb_image uses ints pervasively, including for offset calculations. -// therefore the largest decoded image size we can support with the -// current code, even on 64-bit targets, is INT_MAX. this is not a -// significant limitation for the intended use case. -// -// we do, however, need to make sure our size calculations don't -// overflow. hence a few helper functions for size calculations that -// multiply integers together, making sure that they're non-negative -// and no overflow occurs. - -// return 1 if the sum is valid, 0 on overflow. -// negative terms are considered invalid. -static int stbi__addsizes_valid(int a, int b) { - if (b < 0) - return 0; - // now 0 <= b <= INT_MAX, hence also - // 0 <= INT_MAX - b <= INTMAX. - // And "a + b <= INT_MAX" (which might overflow) is the - // same as a <= INT_MAX - b (no overflow) - return a <= INT_MAX - b; -} - -// returns 1 if the product is valid, 0 on overflow. -// negative factors are considered invalid. -static int stbi__mul2sizes_valid(int a, int b) { - if (a < 0 || b < 0) - return 0; - if (b == 0) - return 1; // mul-by-0 is always safe - // portable way to check for no overflows in a*b - return a <= INT_MAX / b; -} - -#if !defined(STBI_NO_JPEG) || !defined(STBI_NO_PNG) || !defined(STBI_NO_TGA) || \ - !defined(STBI_NO_HDR) -// returns 1 if "a*b + add" has no negative terms/factors and doesn't overflow -static int stbi__mad2sizes_valid(int a, int b, int add) { - return stbi__mul2sizes_valid(a, b) && stbi__addsizes_valid(a * b, add); -} -#endif - -// returns 1 if "a*b*c + add" has no negative terms/factors and doesn't overflow -static int stbi__mad3sizes_valid(int a, int b, int c, int add) { - return stbi__mul2sizes_valid(a, b) && stbi__mul2sizes_valid(a * b, c) && - stbi__addsizes_valid(a * b * c, add); -} - -// returns 1 if "a*b*c*d + add" has no negative terms/factors and doesn't overflow -#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) || !defined(STBI_NO_PNM) -static int stbi__mad4sizes_valid(int a, int b, int c, int d, int add) { - return stbi__mul2sizes_valid(a, b) && stbi__mul2sizes_valid(a * b, c) && - stbi__mul2sizes_valid(a * b * c, d) && stbi__addsizes_valid(a * b * c * d, add); -} -#endif - -#if !defined(STBI_NO_JPEG) || !defined(STBI_NO_PNG) || !defined(STBI_NO_TGA) || \ - !defined(STBI_NO_HDR) -// mallocs with size overflow checking -static void* stbi__malloc_mad2(int a, int b, int add) { - if (!stbi__mad2sizes_valid(a, b, add)) - return NULL; - return stbi__malloc(a * b + add); -} -#endif - -static void* stbi__malloc_mad3(int a, int b, int c, int add) { - if (!stbi__mad3sizes_valid(a, b, c, add)) - return NULL; - return stbi__malloc(a * b * c + add); -} - -#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) || !defined(STBI_NO_PNM) -static void* stbi__malloc_mad4(int a, int b, int c, int d, int add) { - if (!stbi__mad4sizes_valid(a, b, c, d, add)) - return NULL; - return stbi__malloc(a * b * c * d + add); -} -#endif - -// stbi__err - error -// stbi__errpf - error returning pointer to float -// stbi__errpuc - error returning pointer to unsigned char - -#ifdef STBI_NO_FAILURE_STRINGS -#define stbi__err(x, y) 0 -#elif defined(STBI_FAILURE_USERMSG) -#define stbi__err(x, y) stbi__err(y) -#else -#define stbi__err(x, y) stbi__err(x) -#endif - -#define stbi__errpf(x, y) ((float*)(size_t)(stbi__err(x, y) ? NULL : NULL)) -#define stbi__errpuc(x, y) ((unsigned char*)(size_t)(stbi__err(x, y) ? NULL : NULL)) - -STBIDEF void stbi_image_free(void* retval_from_stbi_load) { - STBI_FREE(retval_from_stbi_load); -} - -#ifndef STBI_NO_LINEAR -static float* stbi__ldr_to_hdr(stbi_uc* data, int x, int y, int comp); -#endif - -#ifndef STBI_NO_HDR -static stbi_uc* stbi__hdr_to_ldr(float* data, int x, int y, int comp); -#endif - -static int stbi__vertically_flip_on_load_global = 0; - -STBIDEF void stbi_set_flip_vertically_on_load(int flag_true_if_should_flip) { - stbi__vertically_flip_on_load_global = flag_true_if_should_flip; -} - -#ifndef STBI_THREAD_LOCAL -#define stbi__vertically_flip_on_load stbi__vertically_flip_on_load_global -#else -static STBI_THREAD_LOCAL int stbi__vertically_flip_on_load_local, stbi__vertically_flip_on_load_set; - -STBIDEF void stbi_set_flip_vertically_on_load_thread(int flag_true_if_should_flip) { - stbi__vertically_flip_on_load_local = flag_true_if_should_flip; - stbi__vertically_flip_on_load_set = 1; -} - -#define stbi__vertically_flip_on_load \ - (stbi__vertically_flip_on_load_set ? stbi__vertically_flip_on_load_local \ - : stbi__vertically_flip_on_load_global) -#endif // STBI_THREAD_LOCAL - -static void* stbi__load_main(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri, - int bpc) { - memset(ri, 0, sizeof(*ri)); // make sure it's initialized if we add new fields - ri->bits_per_channel = 8; // default is 8 so most paths don't have to be changed - ri->channel_order = STBI_ORDER_RGB; // all current input & output are this, but this is here so - // we can add BGR order - ri->num_channels = 0; - -// test the formats with a very explicit header first (at least a FOURCC -// or distinctive magic number first) -#ifndef STBI_NO_PNG - if (stbi__png_test(s)) - return stbi__png_load(s, x, y, comp, req_comp, ri); -#endif -#ifndef STBI_NO_BMP - if (stbi__bmp_test(s)) - return stbi__bmp_load(s, x, y, comp, req_comp, ri); -#endif -#ifndef STBI_NO_GIF - if (stbi__gif_test(s)) - return stbi__gif_load(s, x, y, comp, req_comp, ri); -#endif -#ifndef STBI_NO_PSD - if (stbi__psd_test(s)) - return stbi__psd_load(s, x, y, comp, req_comp, ri, bpc); -#else - STBI_NOTUSED(bpc); -#endif -#ifndef STBI_NO_PIC - if (stbi__pic_test(s)) - return stbi__pic_load(s, x, y, comp, req_comp, ri); -#endif - -// then the formats that can end up attempting to load with just 1 or 2 -// bytes matching expectations; these are prone to false positives, so -// try them later -#ifndef STBI_NO_JPEG - if (stbi__jpeg_test(s)) - return stbi__jpeg_load(s, x, y, comp, req_comp, ri); -#endif -#ifndef STBI_NO_PNM - if (stbi__pnm_test(s)) - return stbi__pnm_load(s, x, y, comp, req_comp, ri); -#endif - -#ifndef STBI_NO_HDR - if (stbi__hdr_test(s)) { - float* hdr = stbi__hdr_load(s, x, y, comp, req_comp, ri); - return stbi__hdr_to_ldr(hdr, *x, *y, req_comp ? req_comp : *comp); - } -#endif - -#ifndef STBI_NO_TGA - // test tga last because it's a crappy test! - if (stbi__tga_test(s)) - return stbi__tga_load(s, x, y, comp, req_comp, ri); -#endif - - return stbi__errpuc("unknown image type", "Image not of any known type, or corrupt"); -} - -static stbi_uc* stbi__convert_16_to_8(stbi__uint16* orig, int w, int h, int channels) { - int i; - int img_len = w * h * channels; - stbi_uc* reduced; - - reduced = (stbi_uc*)stbi__malloc(img_len); - if (reduced == NULL) - return stbi__errpuc("outofmem", "Out of memory"); - - for (i = 0; i < img_len; ++i) - reduced[i] = - (stbi_uc)((orig[i] >> 8) & - 0xFF); // top half of each byte is sufficient approx of 16->8 bit scaling - - STBI_FREE(orig); - return reduced; -} - -static stbi__uint16* stbi__convert_8_to_16(stbi_uc* orig, int w, int h, int channels) { - int i; - int img_len = w * h * channels; - stbi__uint16* enlarged; - - enlarged = (stbi__uint16*)stbi__malloc(img_len * 2); - if (enlarged == NULL) - return (stbi__uint16*)stbi__errpuc("outofmem", "Out of memory"); - - for (i = 0; i < img_len; ++i) - enlarged[i] = - (stbi__uint16)((orig[i] << 8) + - orig[i]); // replicate to high and low byte, maps 0->0, 255->0xffff - - STBI_FREE(orig); - return enlarged; -} - -static void stbi__vertical_flip(void* image, int w, int h, int bytes_per_pixel) { - int row; - size_t bytes_per_row = (size_t)w * bytes_per_pixel; - stbi_uc temp[2048]; - stbi_uc* bytes = (stbi_uc*)image; - - for (row = 0; row < (h >> 1); row++) { - stbi_uc* row0 = bytes + row * bytes_per_row; - stbi_uc* row1 = bytes + (h - row - 1) * bytes_per_row; - // swap row0 with row1 - size_t bytes_left = bytes_per_row; - while (bytes_left) { - size_t bytes_copy = (bytes_left < sizeof(temp)) ? bytes_left : sizeof(temp); - memcpy(temp, row0, bytes_copy); - memcpy(row0, row1, bytes_copy); - memcpy(row1, temp, bytes_copy); - row0 += bytes_copy; - row1 += bytes_copy; - bytes_left -= bytes_copy; - } - } -} - -#ifndef STBI_NO_GIF -static void stbi__vertical_flip_slices(void* image, int w, int h, int z, int bytes_per_pixel) { - int slice; - int slice_size = w * h * bytes_per_pixel; - - stbi_uc* bytes = (stbi_uc*)image; - for (slice = 0; slice < z; ++slice) { - stbi__vertical_flip(bytes, w, h, bytes_per_pixel); - bytes += slice_size; - } -} -#endif - -static unsigned char* stbi__load_and_postprocess_8bit(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp) { - stbi__result_info ri; - void* result = stbi__load_main(s, x, y, comp, req_comp, &ri, 8); - - if (result == NULL) - return NULL; - - // it is the responsibility of the loaders to make sure we get either 8 or 16 bit. - STBI_ASSERT(ri.bits_per_channel == 8 || ri.bits_per_channel == 16); - - if (ri.bits_per_channel != 8) { - result = stbi__convert_16_to_8((stbi__uint16*)result, *x, *y, req_comp == 0 ? *comp : req_comp); - ri.bits_per_channel = 8; - } - - // @TODO: move stbi__convert_format to here - - if (stbi__vertically_flip_on_load) { - int channels = req_comp ? req_comp : *comp; - stbi__vertical_flip(result, *x, *y, channels * sizeof(stbi_uc)); - } - - return (unsigned char*)result; -} - -static stbi__uint16* stbi__load_and_postprocess_16bit(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp) { - stbi__result_info ri; - void* result = stbi__load_main(s, x, y, comp, req_comp, &ri, 16); - - if (result == NULL) - return NULL; - - // it is the responsibility of the loaders to make sure we get either 8 or 16 bit. - STBI_ASSERT(ri.bits_per_channel == 8 || ri.bits_per_channel == 16); - - if (ri.bits_per_channel != 16) { - result = stbi__convert_8_to_16((stbi_uc*)result, *x, *y, req_comp == 0 ? *comp : req_comp); - ri.bits_per_channel = 16; - } - - // @TODO: move stbi__convert_format16 to here - // @TODO: special case RGB-to-Y (and RGBA-to-YA) for 8-bit-to-16-bit case to keep more precision - - if (stbi__vertically_flip_on_load) { - int channels = req_comp ? req_comp : *comp; - stbi__vertical_flip(result, *x, *y, channels * sizeof(stbi__uint16)); - } - - return (stbi__uint16*)result; -} - -#if !defined(STBI_NO_HDR) && !defined(STBI_NO_LINEAR) -static void stbi__float_postprocess(float* result, int* x, int* y, int* comp, int req_comp) { - if (stbi__vertically_flip_on_load && result != NULL) { - int channels = req_comp ? req_comp : *comp; - stbi__vertical_flip(result, *x, *y, channels * sizeof(float)); - } -} -#endif - -#ifndef STBI_NO_STDIO - -#if defined(_WIN32) && defined(STBI_WINDOWS_UTF8) -STBI_EXTERN __declspec(dllimport) int __stdcall MultiByteToWideChar(unsigned int cp, - unsigned long flags, - const char* str, - int cbmb, - wchar_t* widestr, - int cchwide); -STBI_EXTERN __declspec(dllimport) int __stdcall WideCharToMultiByte(unsigned int cp, - unsigned long flags, - const wchar_t* widestr, - int cchwide, - char* str, - int cbmb, - const char* defchar, - int* used_default); -#endif - -#if defined(_WIN32) && defined(STBI_WINDOWS_UTF8) -STBIDEF int stbi_convert_wchar_to_utf8(char* buffer, size_t bufferlen, const wchar_t* input) { - return WideCharToMultiByte(65001 /* UTF8 */, 0, input, -1, buffer, (int)bufferlen, NULL, NULL); -} -#endif - -static FILE* stbi__fopen(char const* filename, char const* mode) { - FILE* f; -#if defined(_WIN32) && defined(STBI_WINDOWS_UTF8) - wchar_t wMode[64]; - wchar_t wFilename[1024]; - if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, filename, -1, wFilename, - sizeof(wFilename) / sizeof(*wFilename))) - return 0; - - if (0 == - MultiByteToWideChar(65001 /* UTF8 */, 0, mode, -1, wMode, sizeof(wMode) / sizeof(*wMode))) - return 0; - -#if defined(_MSC_VER) && _MSC_VER >= 1400 - if (0 != _wfopen_s(&f, wFilename, wMode)) - f = 0; -#else - f = _wfopen(wFilename, wMode); -#endif - -#elif defined(_MSC_VER) && _MSC_VER >= 1400 - if (0 != fopen_s(&f, filename, mode)) - f = 0; -#else - f = fopen(filename, mode); -#endif - return f; -} - -STBIDEF stbi_uc* stbi_load(char const* filename, int* x, int* y, int* comp, int req_comp) { - FILE* f = stbi__fopen(filename, "rb"); - unsigned char* result; - if (!f) - return stbi__errpuc("can't fopen", "Unable to open file"); - result = stbi_load_from_file(f, x, y, comp, req_comp); - fclose(f); - return result; -} - -STBIDEF stbi_uc* stbi_load_from_file(FILE* f, int* x, int* y, int* comp, int req_comp) { - unsigned char* result; - stbi__context s; - stbi__start_file(&s, f); - result = stbi__load_and_postprocess_8bit(&s, x, y, comp, req_comp); - if (result) { - // need to 'unget' all the characters in the IO buffer - fseek(f, -(int)(s.img_buffer_end - s.img_buffer), SEEK_CUR); - } - return result; -} - -STBIDEF stbi__uint16* stbi_load_from_file_16(FILE* f, int* x, int* y, int* comp, int req_comp) { - stbi__uint16* result; - stbi__context s; - stbi__start_file(&s, f); - result = stbi__load_and_postprocess_16bit(&s, x, y, comp, req_comp); - if (result) { - // need to 'unget' all the characters in the IO buffer - fseek(f, -(int)(s.img_buffer_end - s.img_buffer), SEEK_CUR); - } - return result; -} - -STBIDEF stbi_us* stbi_load_16(char const* filename, int* x, int* y, int* comp, int req_comp) { - FILE* f = stbi__fopen(filename, "rb"); - stbi__uint16* result; - if (!f) - return (stbi_us*)stbi__errpuc("can't fopen", "Unable to open file"); - result = stbi_load_from_file_16(f, x, y, comp, req_comp); - fclose(f); - return result; -} - -#endif //! STBI_NO_STDIO - -STBIDEF stbi_us* stbi_load_16_from_memory(stbi_uc const* buffer, - int len, - int* x, - int* y, - int* channels_in_file, - int desired_channels) { - stbi__context s; - stbi__start_mem(&s, buffer, len); - return stbi__load_and_postprocess_16bit(&s, x, y, channels_in_file, desired_channels); -} - -STBIDEF stbi_us* stbi_load_16_from_callbacks(stbi_io_callbacks const* clbk, - void* user, - int* x, - int* y, - int* channels_in_file, - int desired_channels) { - stbi__context s; - stbi__start_callbacks(&s, (stbi_io_callbacks*)clbk, user); - return stbi__load_and_postprocess_16bit(&s, x, y, channels_in_file, desired_channels); -} - -STBIDEF stbi_uc* stbi_load_from_memory(stbi_uc const* buffer, - int len, - int* x, - int* y, - int* comp, - int req_comp) { - stbi__context s; - stbi__start_mem(&s, buffer, len); - return stbi__load_and_postprocess_8bit(&s, x, y, comp, req_comp); -} - -STBIDEF stbi_uc* stbi_load_from_callbacks(stbi_io_callbacks const* clbk, - void* user, - int* x, - int* y, - int* comp, - int req_comp) { - stbi__context s; - stbi__start_callbacks(&s, (stbi_io_callbacks*)clbk, user); - return stbi__load_and_postprocess_8bit(&s, x, y, comp, req_comp); -} - -#ifndef STBI_NO_GIF -STBIDEF stbi_uc* stbi_load_gif_from_memory(stbi_uc const* buffer, - int len, - int** delays, - int* x, - int* y, - int* z, - int* comp, - int req_comp) { - unsigned char* result; - stbi__context s; - stbi__start_mem(&s, buffer, len); - - result = (unsigned char*)stbi__load_gif_main(&s, delays, x, y, z, comp, req_comp); - if (stbi__vertically_flip_on_load) { - stbi__vertical_flip_slices(result, *x, *y, *z, *comp); - } - - return result; -} -#endif - -#ifndef STBI_NO_LINEAR -static float* stbi__loadf_main(stbi__context* s, int* x, int* y, int* comp, int req_comp) { - unsigned char* data; -#ifndef STBI_NO_HDR - if (stbi__hdr_test(s)) { - stbi__result_info ri; - float* hdr_data = stbi__hdr_load(s, x, y, comp, req_comp, &ri); - if (hdr_data) - stbi__float_postprocess(hdr_data, x, y, comp, req_comp); - return hdr_data; - } -#endif - data = stbi__load_and_postprocess_8bit(s, x, y, comp, req_comp); - if (data) - return stbi__ldr_to_hdr(data, *x, *y, req_comp ? req_comp : *comp); - return stbi__errpf("unknown image type", "Image not of any known type, or corrupt"); -} - -STBIDEF float* stbi_loadf_from_memory(stbi_uc const* buffer, - int len, - int* x, - int* y, - int* comp, - int req_comp) { - stbi__context s; - stbi__start_mem(&s, buffer, len); - return stbi__loadf_main(&s, x, y, comp, req_comp); -} - -STBIDEF float* stbi_loadf_from_callbacks(stbi_io_callbacks const* clbk, - void* user, - int* x, - int* y, - int* comp, - int req_comp) { - stbi__context s; - stbi__start_callbacks(&s, (stbi_io_callbacks*)clbk, user); - return stbi__loadf_main(&s, x, y, comp, req_comp); -} - -#ifndef STBI_NO_STDIO -STBIDEF float* stbi_loadf(char const* filename, int* x, int* y, int* comp, int req_comp) { - float* result; - FILE* f = stbi__fopen(filename, "rb"); - if (!f) - return stbi__errpf("can't fopen", "Unable to open file"); - result = stbi_loadf_from_file(f, x, y, comp, req_comp); - fclose(f); - return result; -} - -STBIDEF float* stbi_loadf_from_file(FILE* f, int* x, int* y, int* comp, int req_comp) { - stbi__context s; - stbi__start_file(&s, f); - return stbi__loadf_main(&s, x, y, comp, req_comp); -} -#endif // !STBI_NO_STDIO - -#endif // !STBI_NO_LINEAR - -// these is-hdr-or-not is defined independent of whether STBI_NO_LINEAR is -// defined, for API simplicity; if STBI_NO_LINEAR is defined, it always -// reports false! - -STBIDEF int stbi_is_hdr_from_memory(stbi_uc const* buffer, int len) { -#ifndef STBI_NO_HDR - stbi__context s; - stbi__start_mem(&s, buffer, len); - return stbi__hdr_test(&s); -#else - STBI_NOTUSED(buffer); - STBI_NOTUSED(len); - return 0; -#endif -} - -#ifndef STBI_NO_STDIO -STBIDEF int stbi_is_hdr(char const* filename) { - FILE* f = stbi__fopen(filename, "rb"); - int result = 0; - if (f) { - result = stbi_is_hdr_from_file(f); - fclose(f); - } - return result; -} - -STBIDEF int stbi_is_hdr_from_file(FILE* f) { -#ifndef STBI_NO_HDR - long pos = ftell(f); - int res; - stbi__context s; - stbi__start_file(&s, f); - res = stbi__hdr_test(&s); - fseek(f, pos, SEEK_SET); - return res; -#else - STBI_NOTUSED(f); - return 0; -#endif -} -#endif // !STBI_NO_STDIO - -STBIDEF int stbi_is_hdr_from_callbacks(stbi_io_callbacks const* clbk, void* user) { -#ifndef STBI_NO_HDR - stbi__context s; - stbi__start_callbacks(&s, (stbi_io_callbacks*)clbk, user); - return stbi__hdr_test(&s); -#else - STBI_NOTUSED(clbk); - STBI_NOTUSED(user); - return 0; -#endif -} - -#ifndef STBI_NO_LINEAR -static float stbi__l2h_gamma = 2.2f, stbi__l2h_scale = 1.0f; - -STBIDEF void stbi_ldr_to_hdr_gamma(float gamma) { - stbi__l2h_gamma = gamma; -} -STBIDEF void stbi_ldr_to_hdr_scale(float scale) { - stbi__l2h_scale = scale; -} -#endif - -static float stbi__h2l_gamma_i = 1.0f / 2.2f, stbi__h2l_scale_i = 1.0f; - -STBIDEF void stbi_hdr_to_ldr_gamma(float gamma) { - stbi__h2l_gamma_i = 1 / gamma; -} -STBIDEF void stbi_hdr_to_ldr_scale(float scale) { - stbi__h2l_scale_i = 1 / scale; -} - -////////////////////////////////////////////////////////////////////////////// -// -// Common code used by all image loaders -// - -enum { STBI__SCAN_load = 0, STBI__SCAN_type, STBI__SCAN_header }; - -static void stbi__refill_buffer(stbi__context* s) { - int n = (s->io.read)(s->io_user_data, (char*)s->buffer_start, s->buflen); - s->callback_already_read += (int)(s->img_buffer - s->img_buffer_original); - if (n == 0) { - // at end of file, treat same as if from memory, but need to handle case - // where s->img_buffer isn't pointing to safe memory, e.g. 0-byte file - s->read_from_callbacks = 0; - s->img_buffer = s->buffer_start; - s->img_buffer_end = s->buffer_start + 1; - *s->img_buffer = 0; - } else { - s->img_buffer = s->buffer_start; - s->img_buffer_end = s->buffer_start + n; - } -} - -stbi_inline static stbi_uc stbi__get8(stbi__context* s) { - if (s->img_buffer < s->img_buffer_end) - return *s->img_buffer++; - if (s->read_from_callbacks) { - stbi__refill_buffer(s); - return *s->img_buffer++; - } - return 0; -} - -#if defined(STBI_NO_JPEG) && defined(STBI_NO_HDR) && defined(STBI_NO_PIC) && defined(STBI_NO_PNM) -// nothing -#else -stbi_inline static int stbi__at_eof(stbi__context* s) { - if (s->io.read) { - if (!(s->io.eof)(s->io_user_data)) - return 0; - // if feof() is true, check if buffer = end - // special case: we've only got the special 0 character at the end - if (s->read_from_callbacks == 0) - return 1; - } - - return s->img_buffer >= s->img_buffer_end; -} -#endif - -#if defined(STBI_NO_JPEG) && defined(STBI_NO_PNG) && defined(STBI_NO_BMP) && \ - defined(STBI_NO_PSD) && defined(STBI_NO_TGA) && defined(STBI_NO_GIF) && defined(STBI_NO_PIC) -// nothing -#else -static void stbi__skip(stbi__context* s, int n) { - if (n == 0) - return; // already there! - if (n < 0) { - s->img_buffer = s->img_buffer_end; - return; - } - if (s->io.read) { - int blen = (int)(s->img_buffer_end - s->img_buffer); - if (blen < n) { - s->img_buffer = s->img_buffer_end; - (s->io.skip)(s->io_user_data, n - blen); - return; - } - } - s->img_buffer += n; -} -#endif - -#if defined(STBI_NO_PNG) && defined(STBI_NO_TGA) && defined(STBI_NO_HDR) && defined(STBI_NO_PNM) -// nothing -#else -static int stbi__getn(stbi__context* s, stbi_uc* buffer, int n) { - if (s->io.read) { - int blen = (int)(s->img_buffer_end - s->img_buffer); - if (blen < n) { - int res, count; - - memcpy(buffer, s->img_buffer, blen); - - count = (s->io.read)(s->io_user_data, (char*)buffer + blen, n - blen); - res = (count == (n - blen)); - s->img_buffer = s->img_buffer_end; - return res; - } - } - - if (s->img_buffer + n <= s->img_buffer_end) { - memcpy(buffer, s->img_buffer, n); - s->img_buffer += n; - return 1; - } else - return 0; -} -#endif - -#if defined(STBI_NO_JPEG) && defined(STBI_NO_PNG) && defined(STBI_NO_PSD) && defined(STBI_NO_PIC) -// nothing -#else -static int stbi__get16be(stbi__context* s) { - int z = stbi__get8(s); - return (z << 8) + stbi__get8(s); -} -#endif - -#if defined(STBI_NO_PNG) && defined(STBI_NO_PSD) && defined(STBI_NO_PIC) -// nothing -#else -static stbi__uint32 stbi__get32be(stbi__context* s) { - stbi__uint32 z = stbi__get16be(s); - return (z << 16) + stbi__get16be(s); -} -#endif - -#if defined(STBI_NO_BMP) && defined(STBI_NO_TGA) && defined(STBI_NO_GIF) -// nothing -#else -static int stbi__get16le(stbi__context* s) { - int z = stbi__get8(s); - return z + (stbi__get8(s) << 8); -} -#endif - -#ifndef STBI_NO_BMP -static stbi__uint32 stbi__get32le(stbi__context* s) { - stbi__uint32 z = stbi__get16le(s); - z += (stbi__uint32)stbi__get16le(s) << 16; - return z; -} -#endif - -#define STBI__BYTECAST(x) ((stbi_uc)((x)&255)) // truncate int to byte without warnings - -#if defined(STBI_NO_JPEG) && defined(STBI_NO_PNG) && defined(STBI_NO_BMP) && \ - defined(STBI_NO_PSD) && defined(STBI_NO_TGA) && defined(STBI_NO_GIF) && \ - defined(STBI_NO_PIC) && defined(STBI_NO_PNM) -// nothing -#else -////////////////////////////////////////////////////////////////////////////// -// -// generic converter from built-in img_n to req_comp -// individual types do this automatically as much as possible (e.g. jpeg -// does all cases internally since it needs to colorspace convert anyway, -// and it never has alpha, so very few cases ). png can automatically -// interleave an alpha=255 channel, but falls back to this for other cases -// -// assume data buffer is malloced, so malloc a new one and free that one -// only failure mode is malloc failing - -static stbi_uc stbi__compute_y(int r, int g, int b) { - return (stbi_uc)(((r * 77) + (g * 150) + (29 * b)) >> 8); -} -#endif - -#if defined(STBI_NO_PNG) && defined(STBI_NO_BMP) && defined(STBI_NO_PSD) && \ - defined(STBI_NO_TGA) && defined(STBI_NO_GIF) && defined(STBI_NO_PIC) && defined(STBI_NO_PNM) -// nothing -#else -static unsigned char* stbi__convert_format(unsigned char* data, - int img_n, - int req_comp, - unsigned int x, - unsigned int y) { - int i, j; - unsigned char* good; - - if (req_comp == img_n) - return data; - STBI_ASSERT(req_comp >= 1 && req_comp <= 4); - - good = (unsigned char*)stbi__malloc_mad3(req_comp, x, y, 0); - if (good == NULL) { - STBI_FREE(data); - return stbi__errpuc("outofmem", "Out of memory"); - } - - for (j = 0; j < (int)y; ++j) { - unsigned char* src = data + j * x * img_n; - unsigned char* dest = good + j * x * req_comp; - -#define STBI__COMBO(a, b) ((a)*8 + (b)) -#define STBI__CASE(a, b) \ - case STBI__COMBO(a, b): \ - for (i = x - 1; i >= 0; --i, src += a, dest += b) - // convert source image with img_n components to one with req_comp components; - // avoid switch per pixel, so use switch per scanline and massive macros - switch (STBI__COMBO(img_n, req_comp)) { - STBI__CASE(1, 2) { - dest[0] = src[0]; - dest[1] = 255; - } - break; - STBI__CASE(1, 3) { dest[0] = dest[1] = dest[2] = src[0]; } - break; - STBI__CASE(1, 4) { - dest[0] = dest[1] = dest[2] = src[0]; - dest[3] = 255; - } - break; - STBI__CASE(2, 1) { dest[0] = src[0]; } - break; - STBI__CASE(2, 3) { dest[0] = dest[1] = dest[2] = src[0]; } - break; - STBI__CASE(2, 4) { - dest[0] = dest[1] = dest[2] = src[0]; - dest[3] = src[1]; - } - break; - STBI__CASE(3, 4) { - dest[0] = src[0]; - dest[1] = src[1]; - dest[2] = src[2]; - dest[3] = 255; - } - break; - STBI__CASE(3, 1) { dest[0] = stbi__compute_y(src[0], src[1], src[2]); } - break; - STBI__CASE(3, 2) { - dest[0] = stbi__compute_y(src[0], src[1], src[2]); - dest[1] = 255; - } - break; - STBI__CASE(4, 1) { dest[0] = stbi__compute_y(src[0], src[1], src[2]); } - break; - STBI__CASE(4, 2) { - dest[0] = stbi__compute_y(src[0], src[1], src[2]); - dest[1] = src[3]; - } - break; - STBI__CASE(4, 3) { - dest[0] = src[0]; - dest[1] = src[1]; - dest[2] = src[2]; - } - break; - default: - STBI_ASSERT(0); - STBI_FREE(data); - STBI_FREE(good); - return stbi__errpuc("unsupported", "Unsupported format conversion"); - } -#undef STBI__CASE - } - - STBI_FREE(data); - return good; -} -#endif - -#if defined(STBI_NO_PNG) && defined(STBI_NO_PSD) -// nothing -#else -static stbi__uint16 stbi__compute_y_16(int r, int g, int b) { - return (stbi__uint16)(((r * 77) + (g * 150) + (29 * b)) >> 8); -} -#endif - -#if defined(STBI_NO_PNG) && defined(STBI_NO_PSD) -// nothing -#else -static stbi__uint16* stbi__convert_format16(stbi__uint16* data, - int img_n, - int req_comp, - unsigned int x, - unsigned int y) { - int i, j; - stbi__uint16* good; - - if (req_comp == img_n) - return data; - STBI_ASSERT(req_comp >= 1 && req_comp <= 4); - - good = (stbi__uint16*)stbi__malloc(req_comp * x * y * 2); - if (good == NULL) { - STBI_FREE(data); - return (stbi__uint16*)stbi__errpuc("outofmem", "Out of memory"); - } - - for (j = 0; j < (int)y; ++j) { - stbi__uint16* src = data + j * x * img_n; - stbi__uint16* dest = good + j * x * req_comp; - -#define STBI__COMBO(a, b) ((a)*8 + (b)) -#define STBI__CASE(a, b) \ - case STBI__COMBO(a, b): \ - for (i = x - 1; i >= 0; --i, src += a, dest += b) - // convert source image with img_n components to one with req_comp components; - // avoid switch per pixel, so use switch per scanline and massive macros - switch (STBI__COMBO(img_n, req_comp)) { - STBI__CASE(1, 2) { - dest[0] = src[0]; - dest[1] = 0xffff; - } - break; - STBI__CASE(1, 3) { dest[0] = dest[1] = dest[2] = src[0]; } - break; - STBI__CASE(1, 4) { - dest[0] = dest[1] = dest[2] = src[0]; - dest[3] = 0xffff; - } - break; - STBI__CASE(2, 1) { dest[0] = src[0]; } - break; - STBI__CASE(2, 3) { dest[0] = dest[1] = dest[2] = src[0]; } - break; - STBI__CASE(2, 4) { - dest[0] = dest[1] = dest[2] = src[0]; - dest[3] = src[1]; - } - break; - STBI__CASE(3, 4) { - dest[0] = src[0]; - dest[1] = src[1]; - dest[2] = src[2]; - dest[3] = 0xffff; - } - break; - STBI__CASE(3, 1) { dest[0] = stbi__compute_y_16(src[0], src[1], src[2]); } - break; - STBI__CASE(3, 2) { - dest[0] = stbi__compute_y_16(src[0], src[1], src[2]); - dest[1] = 0xffff; - } - break; - STBI__CASE(4, 1) { dest[0] = stbi__compute_y_16(src[0], src[1], src[2]); } - break; - STBI__CASE(4, 2) { - dest[0] = stbi__compute_y_16(src[0], src[1], src[2]); - dest[1] = src[3]; - } - break; - STBI__CASE(4, 3) { - dest[0] = src[0]; - dest[1] = src[1]; - dest[2] = src[2]; - } - break; - default: - STBI_ASSERT(0); - STBI_FREE(data); - STBI_FREE(good); - return (stbi__uint16*)stbi__errpuc("unsupported", "Unsupported format conversion"); - } -#undef STBI__CASE - } - - STBI_FREE(data); - return good; -} -#endif - -#ifndef STBI_NO_LINEAR -static float* stbi__ldr_to_hdr(stbi_uc* data, int x, int y, int comp) { - int i, k, n; - float* output; - if (!data) - return NULL; - output = (float*)stbi__malloc_mad4(x, y, comp, sizeof(float), 0); - if (output == NULL) { - STBI_FREE(data); - return stbi__errpf("outofmem", "Out of memory"); - } - // compute number of non-alpha components - if (comp & 1) - n = comp; - else - n = comp - 1; - for (i = 0; i < x * y; ++i) { - for (k = 0; k < n; ++k) { - output[i * comp + k] = - (float)(pow(data[i * comp + k] / 255.0f, stbi__l2h_gamma) * stbi__l2h_scale); - } - } - if (n < comp) { - for (i = 0; i < x * y; ++i) { - output[i * comp + n] = data[i * comp + n] / 255.0f; - } - } - STBI_FREE(data); - return output; -} -#endif - -#ifndef STBI_NO_HDR -#define stbi__float2int(x) ((int)(x)) -static stbi_uc* stbi__hdr_to_ldr(float* data, int x, int y, int comp) { - int i, k, n; - stbi_uc* output; - if (!data) - return NULL; - output = (stbi_uc*)stbi__malloc_mad3(x, y, comp, 0); - if (output == NULL) { - STBI_FREE(data); - return stbi__errpuc("outofmem", "Out of memory"); - } - // compute number of non-alpha components - if (comp & 1) - n = comp; - else - n = comp - 1; - for (i = 0; i < x * y; ++i) { - for (k = 0; k < n; ++k) { - float z = (float)pow(data[i * comp + k] * stbi__h2l_scale_i, stbi__h2l_gamma_i) * 255 + 0.5f; - if (z < 0) - z = 0; - if (z > 255) - z = 255; - output[i * comp + k] = (stbi_uc)stbi__float2int(z); - } - if (k < comp) { - float z = data[i * comp + k] * 255 + 0.5f; - if (z < 0) - z = 0; - if (z > 255) - z = 255; - output[i * comp + k] = (stbi_uc)stbi__float2int(z); - } - } - STBI_FREE(data); - return output; -} -#endif - -////////////////////////////////////////////////////////////////////////////// -// -// "baseline" JPEG/JFIF decoder -// -// simple implementation -// - doesn't support delayed output of y-dimension -// - simple interface (only one output format: 8-bit interleaved RGB) -// - doesn't try to recover corrupt jpegs -// - doesn't allow partial loading, loading multiple at once -// - still fast on x86 (copying globals into locals doesn't help x86) -// - allocates lots of intermediate memory (full size of all components) -// - non-interleaved case requires this anyway -// - allows good upsampling (see next) -// high-quality -// - upsampled channels are bilinearly interpolated, even across blocks -// - quality integer IDCT derived from IJG's 'slow' -// performance -// - fast huffman; reasonable integer IDCT -// - some SIMD kernels for common paths on targets with SSE2/NEON -// - uses a lot of intermediate memory, could cache poorly - -#ifndef STBI_NO_JPEG - -// huffman decoding acceleration -#define FAST_BITS 9 // larger handles more cases; smaller stomps less cache - -typedef struct { - stbi_uc fast[1 << FAST_BITS]; - // weirdly, repacking this into AoS is a 10% speed loss, instead of a win - stbi__uint16 code[256]; - stbi_uc values[256]; - stbi_uc size[257]; - unsigned int maxcode[18]; - int delta[17]; // old 'firstsymbol' - old 'firstcode' -} stbi__huffman; - -typedef struct { - stbi__context* s; - stbi__huffman huff_dc[4]; - stbi__huffman huff_ac[4]; - stbi__uint16 dequant[4][64]; - stbi__int16 fast_ac[4][1 << FAST_BITS]; - - // sizes for components, interleaved MCUs - int img_h_max, img_v_max; - int img_mcu_x, img_mcu_y; - int img_mcu_w, img_mcu_h; - - // definition of jpeg image component - struct { - int id; - int h, v; - int tq; - int hd, ha; - int dc_pred; - - int x, y, w2, h2; - stbi_uc* data; - void *raw_data, *raw_coeff; - stbi_uc* linebuf; - short* coeff; // progressive only - int coeff_w, coeff_h; // number of 8x8 coefficient blocks - } img_comp[4]; - - stbi__uint32 code_buffer; // jpeg entropy-coded buffer - int code_bits; // number of valid bits - unsigned char marker; // marker seen while filling entropy buffer - int nomore; // flag if we saw a marker so must stop - - int progressive; - int spec_start; - int spec_end; - int succ_high; - int succ_low; - int eob_run; - int jfif; - int app14_color_transform; // Adobe APP14 tag - int rgb; - - int scan_n, order[4]; - int restart_interval, todo; - - // kernels - void (*idct_block_kernel)(stbi_uc* out, int out_stride, short data[64]); - void (*YCbCr_to_RGB_kernel)(stbi_uc* out, - const stbi_uc* y, - const stbi_uc* pcb, - const stbi_uc* pcr, - int count, - int step); - stbi_uc* ( - *resample_row_hv_2_kernel)(stbi_uc* out, stbi_uc* in_near, stbi_uc* in_far, int w, int hs); -} stbi__jpeg; - -static int stbi__build_huffman(stbi__huffman* h, int* count) { - int i, j, k = 0; - unsigned int code; - // build size list for each symbol (from JPEG spec) - for (i = 0; i < 16; ++i) - for (j = 0; j < count[i]; ++j) - h->size[k++] = (stbi_uc)(i + 1); - h->size[k] = 0; - - // compute actual symbols (from jpeg spec) - code = 0; - k = 0; - for (j = 1; j <= 16; ++j) { - // compute delta to add to code to compute symbol id - h->delta[j] = k - code; - if (h->size[k] == j) { - while (h->size[k] == j) - h->code[k++] = (stbi__uint16)(code++); - if (code - 1 >= (1u << j)) - return stbi__err("bad code lengths", "Corrupt JPEG"); - } - // compute largest code + 1 for this size, preshifted as needed later - h->maxcode[j] = code << (16 - j); - code <<= 1; - } - h->maxcode[j] = 0xffffffff; - - // build non-spec acceleration table; 255 is flag for not-accelerated - memset(h->fast, 255, 1 << FAST_BITS); - for (i = 0; i < k; ++i) { - int s = h->size[i]; - if (s <= FAST_BITS) { - int c = h->code[i] << (FAST_BITS - s); - int m = 1 << (FAST_BITS - s); - for (j = 0; j < m; ++j) { - h->fast[c + j] = (stbi_uc)i; - } - } - } - return 1; -} - -// build a table that decodes both magnitude and value of small ACs in -// one go. -static void stbi__build_fast_ac(stbi__int16* fast_ac, stbi__huffman* h) { - int i; - for (i = 0; i < (1 << FAST_BITS); ++i) { - stbi_uc fast = h->fast[i]; - fast_ac[i] = 0; - if (fast < 255) { - int rs = h->values[fast]; - int run = (rs >> 4) & 15; - int magbits = rs & 15; - int len = h->size[fast]; - - if (magbits && len + magbits <= FAST_BITS) { - // magnitude code followed by receive_extend code - int k = ((i << len) & ((1 << FAST_BITS) - 1)) >> (FAST_BITS - magbits); - int m = 1 << (magbits - 1); - if (k < m) - k += (~0U << magbits) + 1; - // if the result is small enough, we can fit it in fast_ac table - if (k >= -128 && k <= 127) - fast_ac[i] = (stbi__int16)((k * 256) + (run * 16) + (len + magbits)); - } - } - } -} - -static void stbi__grow_buffer_unsafe(stbi__jpeg* j) { - do { - unsigned int b = j->nomore ? 0 : stbi__get8(j->s); - if (b == 0xff) { - int c = stbi__get8(j->s); - while (c == 0xff) - c = stbi__get8(j->s); // consume fill bytes - if (c != 0) { - j->marker = (unsigned char)c; - j->nomore = 1; - return; - } - } - j->code_buffer |= b << (24 - j->code_bits); - j->code_bits += 8; - } while (j->code_bits <= 24); -} - -// (1 << n) - 1 -static const stbi__uint32 stbi__bmask[17] = {0, 1, 3, 7, 15, 31, 63, 127, 255, - 511, 1023, 2047, 4095, 8191, 16383, 32767, 65535}; - -// decode a jpeg huffman value from the bitstream -stbi_inline static int stbi__jpeg_huff_decode(stbi__jpeg* j, stbi__huffman* h) { - unsigned int temp; - int c, k; - - if (j->code_bits < 16) - stbi__grow_buffer_unsafe(j); - - // look at the top FAST_BITS and determine what symbol ID it is, - // if the code is <= FAST_BITS - c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS) - 1); - k = h->fast[c]; - if (k < 255) { - int s = h->size[k]; - if (s > j->code_bits) - return -1; - j->code_buffer <<= s; - j->code_bits -= s; - return h->values[k]; - } - - // naive test is to shift the code_buffer down so k bits are - // valid, then test against maxcode. To speed this up, we've - // preshifted maxcode left so that it has (16-k) 0s at the - // end; in other words, regardless of the number of bits, it - // wants to be compared against something shifted to have 16; - // that way we don't need to shift inside the loop. - temp = j->code_buffer >> 16; - for (k = FAST_BITS + 1;; ++k) - if (temp < h->maxcode[k]) - break; - if (k == 17) { - // error! code not found - j->code_bits -= 16; - return -1; - } - - if (k > j->code_bits) - return -1; - - // convert the huffman code to the symbol id - c = ((j->code_buffer >> (32 - k)) & stbi__bmask[k]) + h->delta[k]; - STBI_ASSERT((((j->code_buffer) >> (32 - h->size[c])) & stbi__bmask[h->size[c]]) == h->code[c]); - - // convert the id to a symbol - j->code_bits -= k; - j->code_buffer <<= k; - return h->values[c]; -} - -// bias[n] = (-1<code_bits < n) - stbi__grow_buffer_unsafe(j); - - sgn = j->code_buffer >> - 31; // sign bit always in MSB; 0 if MSB clear (positive), 1 if MSB set (negative) - k = stbi_lrot(j->code_buffer, n); - j->code_buffer = k & ~stbi__bmask[n]; - k &= stbi__bmask[n]; - j->code_bits -= n; - return k + (stbi__jbias[n] & (sgn - 1)); -} - -// get some unsigned bits -stbi_inline static int stbi__jpeg_get_bits(stbi__jpeg* j, int n) { - unsigned int k; - if (j->code_bits < n) - stbi__grow_buffer_unsafe(j); - k = stbi_lrot(j->code_buffer, n); - j->code_buffer = k & ~stbi__bmask[n]; - k &= stbi__bmask[n]; - j->code_bits -= n; - return k; -} - -stbi_inline static int stbi__jpeg_get_bit(stbi__jpeg* j) { - unsigned int k; - if (j->code_bits < 1) - stbi__grow_buffer_unsafe(j); - k = j->code_buffer; - j->code_buffer <<= 1; - --j->code_bits; - return k & 0x80000000; -} - -// given a value that's at position X in the zigzag stream, -// where does it appear in the 8x8 matrix coded as row-major? -static const stbi_uc stbi__jpeg_dezigzag[64 + 15] = { - 0, 1, 8, 16, 9, 2, 3, 10, 17, 24, 32, 25, 18, 11, 4, 5, 12, 19, 26, 33, 40, 48, 41, 34, 27, 20, - 13, 6, 7, 14, 21, 28, 35, 42, 49, 56, 57, 50, 43, 36, 29, 22, 15, 23, 30, 37, 44, 51, 58, 59, - 52, 45, 38, 31, 39, 46, 53, 60, 61, 54, 47, 55, 62, 63, - // let corrupt input sample past end - 63, 63, 63, 63, 63, 63, 63, 63, 63, 63, 63, 63, 63, 63, 63}; - -// decode one 64-entry block-- -static int stbi__jpeg_decode_block(stbi__jpeg* j, - short data[64], - stbi__huffman* hdc, - stbi__huffman* hac, - stbi__int16* fac, - int b, - stbi__uint16* dequant) { - int diff, dc, k; - int t; - - if (j->code_bits < 16) - stbi__grow_buffer_unsafe(j); - t = stbi__jpeg_huff_decode(j, hdc); - if (t < 0 || t > 15) - return stbi__err("bad huffman code", "Corrupt JPEG"); - - // 0 all the ac values now so we can do it 32-bits at a time - memset(data, 0, 64 * sizeof(data[0])); - - diff = t ? stbi__extend_receive(j, t) : 0; - dc = j->img_comp[b].dc_pred + diff; - j->img_comp[b].dc_pred = dc; - data[0] = (short)(dc * dequant[0]); - - // decode AC components, see JPEG spec - k = 1; - do { - unsigned int zig; - int c, r, s; - if (j->code_bits < 16) - stbi__grow_buffer_unsafe(j); - c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS) - 1); - r = fac[c]; - if (r) { // fast-AC path - k += (r >> 4) & 15; // run - s = r & 15; // combined length - j->code_buffer <<= s; - j->code_bits -= s; - // decode into unzigzag'd location - zig = stbi__jpeg_dezigzag[k++]; - data[zig] = (short)((r >> 8) * dequant[zig]); - } else { - int rs = stbi__jpeg_huff_decode(j, hac); - if (rs < 0) - return stbi__err("bad huffman code", "Corrupt JPEG"); - s = rs & 15; - r = rs >> 4; - if (s == 0) { - if (rs != 0xf0) - break; // end block - k += 16; - } else { - k += r; - // decode into unzigzag'd location - zig = stbi__jpeg_dezigzag[k++]; - data[zig] = (short)(stbi__extend_receive(j, s) * dequant[zig]); - } - } - } while (k < 64); - return 1; -} - -static int stbi__jpeg_decode_block_prog_dc(stbi__jpeg* j, - short data[64], - stbi__huffman* hdc, - int b) { - int diff, dc; - int t; - if (j->spec_end != 0) - return stbi__err("can't merge dc and ac", "Corrupt JPEG"); - - if (j->code_bits < 16) - stbi__grow_buffer_unsafe(j); - - if (j->succ_high == 0) { - // first scan for DC coefficient, must be first - memset(data, 0, 64 * sizeof(data[0])); // 0 all the ac values now - t = stbi__jpeg_huff_decode(j, hdc); - if (t < 0 || t > 15) - return stbi__err("can't merge dc and ac", "Corrupt JPEG"); - diff = t ? stbi__extend_receive(j, t) : 0; - - dc = j->img_comp[b].dc_pred + diff; - j->img_comp[b].dc_pred = dc; - data[0] = (short)(dc * (1 << j->succ_low)); - } else { - // refinement scan for DC coefficient - if (stbi__jpeg_get_bit(j)) - data[0] += (short)(1 << j->succ_low); - } - return 1; -} - -// @OPTIMIZE: store non-zigzagged during the decode passes, -// and only de-zigzag when dequantizing -static int stbi__jpeg_decode_block_prog_ac(stbi__jpeg* j, - short data[64], - stbi__huffman* hac, - stbi__int16* fac) { - int k; - if (j->spec_start == 0) - return stbi__err("can't merge dc and ac", "Corrupt JPEG"); - - if (j->succ_high == 0) { - int shift = j->succ_low; - - if (j->eob_run) { - --j->eob_run; - return 1; - } - - k = j->spec_start; - do { - unsigned int zig; - int c, r, s; - if (j->code_bits < 16) - stbi__grow_buffer_unsafe(j); - c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS) - 1); - r = fac[c]; - if (r) { // fast-AC path - k += (r >> 4) & 15; // run - s = r & 15; // combined length - j->code_buffer <<= s; - j->code_bits -= s; - zig = stbi__jpeg_dezigzag[k++]; - data[zig] = (short)((r >> 8) * (1 << shift)); - } else { - int rs = stbi__jpeg_huff_decode(j, hac); - if (rs < 0) - return stbi__err("bad huffman code", "Corrupt JPEG"); - s = rs & 15; - r = rs >> 4; - if (s == 0) { - if (r < 15) { - j->eob_run = (1 << r); - if (r) - j->eob_run += stbi__jpeg_get_bits(j, r); - --j->eob_run; - break; - } - k += 16; - } else { - k += r; - zig = stbi__jpeg_dezigzag[k++]; - data[zig] = (short)(stbi__extend_receive(j, s) * (1 << shift)); - } - } - } while (k <= j->spec_end); - } else { - // refinement scan for these AC coefficients - - short bit = (short)(1 << j->succ_low); - - if (j->eob_run) { - --j->eob_run; - for (k = j->spec_start; k <= j->spec_end; ++k) { - short* p = &data[stbi__jpeg_dezigzag[k]]; - if (*p != 0) - if (stbi__jpeg_get_bit(j)) - if ((*p & bit) == 0) { - if (*p > 0) - *p += bit; - else - *p -= bit; - } - } - } else { - k = j->spec_start; - do { - int r, s; - int rs = stbi__jpeg_huff_decode( - j, hac); // @OPTIMIZE see if we can use the fast path here, advance-by-r is so slow, eh - if (rs < 0) - return stbi__err("bad huffman code", "Corrupt JPEG"); - s = rs & 15; - r = rs >> 4; - if (s == 0) { - if (r < 15) { - j->eob_run = (1 << r) - 1; - if (r) - j->eob_run += stbi__jpeg_get_bits(j, r); - r = 64; // force end of block - } else { - // r=15 s=0 should write 16 0s, so we just do - // a run of 15 0s and then write s (which is 0), - // so we don't have to do anything special here - } - } else { - if (s != 1) - return stbi__err("bad huffman code", "Corrupt JPEG"); - // sign bit - if (stbi__jpeg_get_bit(j)) - s = bit; - else - s = -bit; - } - - // advance by r - while (k <= j->spec_end) { - short* p = &data[stbi__jpeg_dezigzag[k++]]; - if (*p != 0) { - if (stbi__jpeg_get_bit(j)) - if ((*p & bit) == 0) { - if (*p > 0) - *p += bit; - else - *p -= bit; - } - } else { - if (r == 0) { - *p = (short)s; - break; - } - --r; - } - } - } while (k <= j->spec_end); - } - } - return 1; -} - -// take a -128..127 value and stbi__clamp it and convert to 0..255 -stbi_inline static stbi_uc stbi__clamp(int x) { - // trick to use a single test to catch both cases - if ((unsigned int)x > 255) { - if (x < 0) - return 0; - if (x > 255) - return 255; - } - return (stbi_uc)x; -} - -#define stbi__f2f(x) ((int)(((x)*4096 + 0.5))) -#define stbi__fsh(x) ((x)*4096) - -// derived from jidctint -- DCT_ISLOW -#define STBI__IDCT_1D(s0, s1, s2, s3, s4, s5, s6, s7) \ - int t0, t1, t2, t3, p1, p2, p3, p4, p5, x0, x1, x2, x3; \ - p2 = s2; \ - p3 = s6; \ - p1 = (p2 + p3) * stbi__f2f(0.5411961f); \ - t2 = p1 + p3 * stbi__f2f(-1.847759065f); \ - t3 = p1 + p2 * stbi__f2f(0.765366865f); \ - p2 = s0; \ - p3 = s4; \ - t0 = stbi__fsh(p2 + p3); \ - t1 = stbi__fsh(p2 - p3); \ - x0 = t0 + t3; \ - x3 = t0 - t3; \ - x1 = t1 + t2; \ - x2 = t1 - t2; \ - t0 = s7; \ - t1 = s5; \ - t2 = s3; \ - t3 = s1; \ - p3 = t0 + t2; \ - p4 = t1 + t3; \ - p1 = t0 + t3; \ - p2 = t1 + t2; \ - p5 = (p3 + p4) * stbi__f2f(1.175875602f); \ - t0 = t0 * stbi__f2f(0.298631336f); \ - t1 = t1 * stbi__f2f(2.053119869f); \ - t2 = t2 * stbi__f2f(3.072711026f); \ - t3 = t3 * stbi__f2f(1.501321110f); \ - p1 = p5 + p1 * stbi__f2f(-0.899976223f); \ - p2 = p5 + p2 * stbi__f2f(-2.562915447f); \ - p3 = p3 * stbi__f2f(-1.961570560f); \ - p4 = p4 * stbi__f2f(-0.390180644f); \ - t3 += p1 + p4; \ - t2 += p2 + p3; \ - t1 += p2 + p4; \ - t0 += p1 + p3; - -static void stbi__idct_block(stbi_uc* out, int out_stride, short data[64]) { - int i, val[64], *v = val; - stbi_uc* o; - short* d = data; - - // columns - for (i = 0; i < 8; ++i, ++d, ++v) { - // if all zeroes, shortcut -- this avoids dequantizing 0s and IDCTing - if (d[8] == 0 && d[16] == 0 && d[24] == 0 && d[32] == 0 && d[40] == 0 && d[48] == 0 && - d[56] == 0) { - // no shortcut 0 seconds - // (1|2|3|4|5|6|7)==0 0 seconds - // all separate -0.047 seconds - // 1 && 2|3 && 4|5 && 6|7: -0.047 seconds - int dcterm = d[0] * 4; - v[0] = v[8] = v[16] = v[24] = v[32] = v[40] = v[48] = v[56] = dcterm; - } else { - STBI__IDCT_1D(d[0], d[8], d[16], d[24], d[32], d[40], d[48], d[56]) - // constants scaled things up by 1<<12; let's bring them back - // down, but keep 2 extra bits of precision - x0 += 512; - x1 += 512; - x2 += 512; - x3 += 512; - v[0] = (x0 + t3) >> 10; - v[56] = (x0 - t3) >> 10; - v[8] = (x1 + t2) >> 10; - v[48] = (x1 - t2) >> 10; - v[16] = (x2 + t1) >> 10; - v[40] = (x2 - t1) >> 10; - v[24] = (x3 + t0) >> 10; - v[32] = (x3 - t0) >> 10; - } - } - - for (i = 0, v = val, o = out; i < 8; ++i, v += 8, o += out_stride) { - // no fast case since the first 1D IDCT spread components out - STBI__IDCT_1D(v[0], v[1], v[2], v[3], v[4], v[5], v[6], v[7]) - // constants scaled things up by 1<<12, plus we had 1<<2 from first - // loop, plus horizontal and vertical each scale by sqrt(8) so together - // we've got an extra 1<<3, so 1<<17 total we need to remove. - // so we want to round that, which means adding 0.5 * 1<<17, - // aka 65536. Also, we'll end up with -128 to 127 that we want - // to encode as 0..255 by adding 128, so we'll add that before the shift - x0 += 65536 + (128 << 17); - x1 += 65536 + (128 << 17); - x2 += 65536 + (128 << 17); - x3 += 65536 + (128 << 17); - // tried computing the shifts into temps, or'ing the temps to see - // if any were out of range, but that was slower - o[0] = stbi__clamp((x0 + t3) >> 17); - o[7] = stbi__clamp((x0 - t3) >> 17); - o[1] = stbi__clamp((x1 + t2) >> 17); - o[6] = stbi__clamp((x1 - t2) >> 17); - o[2] = stbi__clamp((x2 + t1) >> 17); - o[5] = stbi__clamp((x2 - t1) >> 17); - o[3] = stbi__clamp((x3 + t0) >> 17); - o[4] = stbi__clamp((x3 - t0) >> 17); - } -} - -#ifdef STBI_SSE2 -// sse2 integer IDCT. not the fastest possible implementation but it -// produces bit-identical results to the generic C version so it's -// fully "transparent". -static void stbi__idct_simd(stbi_uc* out, int out_stride, short data[64]) { - // This is constructed to match our regular (generic) integer IDCT exactly. - __m128i row0, row1, row2, row3, row4, row5, row6, row7; - __m128i tmp; - -// dot product constant: even elems=x, odd elems=y -#define dct_const(x, y) _mm_setr_epi16((x), (y), (x), (y), (x), (y), (x), (y)) - -// out(0) = c0[even]*x + c0[odd]*y (c0, x, y 16-bit, out 32-bit) -// out(1) = c1[even]*x + c1[odd]*y -#define dct_rot(out0, out1, x, y, c0, c1) \ - __m128i c0##lo = _mm_unpacklo_epi16((x), (y)); \ - __m128i c0##hi = _mm_unpackhi_epi16((x), (y)); \ - __m128i out0##_l = _mm_madd_epi16(c0##lo, c0); \ - __m128i out0##_h = _mm_madd_epi16(c0##hi, c0); \ - __m128i out1##_l = _mm_madd_epi16(c0##lo, c1); \ - __m128i out1##_h = _mm_madd_epi16(c0##hi, c1) - -// out = in << 12 (in 16-bit, out 32-bit) -#define dct_widen(out, in) \ - __m128i out##_l = _mm_srai_epi32(_mm_unpacklo_epi16(_mm_setzero_si128(), (in)), 4); \ - __m128i out##_h = _mm_srai_epi32(_mm_unpackhi_epi16(_mm_setzero_si128(), (in)), 4) - -// wide add -#define dct_wadd(out, a, b) \ - __m128i out##_l = _mm_add_epi32(a##_l, b##_l); \ - __m128i out##_h = _mm_add_epi32(a##_h, b##_h) - -// wide sub -#define dct_wsub(out, a, b) \ - __m128i out##_l = _mm_sub_epi32(a##_l, b##_l); \ - __m128i out##_h = _mm_sub_epi32(a##_h, b##_h) - -// butterfly a/b, add bias, then shift by "s" and pack -#define dct_bfly32o(out0, out1, a, b, bias, s) \ - { \ - __m128i abiased_l = _mm_add_epi32(a##_l, bias); \ - __m128i abiased_h = _mm_add_epi32(a##_h, bias); \ - dct_wadd(sum, abiased, b); \ - dct_wsub(dif, abiased, b); \ - out0 = _mm_packs_epi32(_mm_srai_epi32(sum_l, s), _mm_srai_epi32(sum_h, s)); \ - out1 = _mm_packs_epi32(_mm_srai_epi32(dif_l, s), _mm_srai_epi32(dif_h, s)); \ - } - -// 8-bit interleave step (for transposes) -#define dct_interleave8(a, b) \ - tmp = a; \ - a = _mm_unpacklo_epi8(a, b); \ - b = _mm_unpackhi_epi8(tmp, b) - -// 16-bit interleave step (for transposes) -#define dct_interleave16(a, b) \ - tmp = a; \ - a = _mm_unpacklo_epi16(a, b); \ - b = _mm_unpackhi_epi16(tmp, b) - -#define dct_pass(bias, shift) \ - { \ - /* even part */ \ - dct_rot(t2e, t3e, row2, row6, rot0_0, rot0_1); \ - __m128i sum04 = _mm_add_epi16(row0, row4); \ - __m128i dif04 = _mm_sub_epi16(row0, row4); \ - dct_widen(t0e, sum04); \ - dct_widen(t1e, dif04); \ - dct_wadd(x0, t0e, t3e); \ - dct_wsub(x3, t0e, t3e); \ - dct_wadd(x1, t1e, t2e); \ - dct_wsub(x2, t1e, t2e); \ - /* odd part */ \ - dct_rot(y0o, y2o, row7, row3, rot2_0, rot2_1); \ - dct_rot(y1o, y3o, row5, row1, rot3_0, rot3_1); \ - __m128i sum17 = _mm_add_epi16(row1, row7); \ - __m128i sum35 = _mm_add_epi16(row3, row5); \ - dct_rot(y4o, y5o, sum17, sum35, rot1_0, rot1_1); \ - dct_wadd(x4, y0o, y4o); \ - dct_wadd(x5, y1o, y5o); \ - dct_wadd(x6, y2o, y5o); \ - dct_wadd(x7, y3o, y4o); \ - dct_bfly32o(row0, row7, x0, x7, bias, shift); \ - dct_bfly32o(row1, row6, x1, x6, bias, shift); \ - dct_bfly32o(row2, row5, x2, x5, bias, shift); \ - dct_bfly32o(row3, row4, x3, x4, bias, shift); \ - } - - __m128i rot0_0 = - dct_const(stbi__f2f(0.5411961f), stbi__f2f(0.5411961f) + stbi__f2f(-1.847759065f)); - __m128i rot0_1 = - dct_const(stbi__f2f(0.5411961f) + stbi__f2f(0.765366865f), stbi__f2f(0.5411961f)); - __m128i rot1_0 = - dct_const(stbi__f2f(1.175875602f) + stbi__f2f(-0.899976223f), stbi__f2f(1.175875602f)); - __m128i rot1_1 = - dct_const(stbi__f2f(1.175875602f), stbi__f2f(1.175875602f) + stbi__f2f(-2.562915447f)); - __m128i rot2_0 = - dct_const(stbi__f2f(-1.961570560f) + stbi__f2f(0.298631336f), stbi__f2f(-1.961570560f)); - __m128i rot2_1 = - dct_const(stbi__f2f(-1.961570560f), stbi__f2f(-1.961570560f) + stbi__f2f(3.072711026f)); - __m128i rot3_0 = - dct_const(stbi__f2f(-0.390180644f) + stbi__f2f(2.053119869f), stbi__f2f(-0.390180644f)); - __m128i rot3_1 = - dct_const(stbi__f2f(-0.390180644f), stbi__f2f(-0.390180644f) + stbi__f2f(1.501321110f)); - - // rounding biases in column/row passes, see stbi__idct_block for explanation. - __m128i bias_0 = _mm_set1_epi32(512); - __m128i bias_1 = _mm_set1_epi32(65536 + (128 << 17)); - - // load - row0 = _mm_load_si128((const __m128i*)(data + 0 * 8)); - row1 = _mm_load_si128((const __m128i*)(data + 1 * 8)); - row2 = _mm_load_si128((const __m128i*)(data + 2 * 8)); - row3 = _mm_load_si128((const __m128i*)(data + 3 * 8)); - row4 = _mm_load_si128((const __m128i*)(data + 4 * 8)); - row5 = _mm_load_si128((const __m128i*)(data + 5 * 8)); - row6 = _mm_load_si128((const __m128i*)(data + 6 * 8)); - row7 = _mm_load_si128((const __m128i*)(data + 7 * 8)); - - // column pass - dct_pass(bias_0, 10); - - { - // 16bit 8x8 transpose pass 1 - dct_interleave16(row0, row4); - dct_interleave16(row1, row5); - dct_interleave16(row2, row6); - dct_interleave16(row3, row7); - - // transpose pass 2 - dct_interleave16(row0, row2); - dct_interleave16(row1, row3); - dct_interleave16(row4, row6); - dct_interleave16(row5, row7); - - // transpose pass 3 - dct_interleave16(row0, row1); - dct_interleave16(row2, row3); - dct_interleave16(row4, row5); - dct_interleave16(row6, row7); - } - - // row pass - dct_pass(bias_1, 17); - - { - // pack - __m128i p0 = _mm_packus_epi16(row0, row1); // a0a1a2a3...a7b0b1b2b3...b7 - __m128i p1 = _mm_packus_epi16(row2, row3); - __m128i p2 = _mm_packus_epi16(row4, row5); - __m128i p3 = _mm_packus_epi16(row6, row7); - - // 8bit 8x8 transpose pass 1 - dct_interleave8(p0, p2); // a0e0a1e1... - dct_interleave8(p1, p3); // c0g0c1g1... - - // transpose pass 2 - dct_interleave8(p0, p1); // a0c0e0g0... - dct_interleave8(p2, p3); // b0d0f0h0... - - // transpose pass 3 - dct_interleave8(p0, p2); // a0b0c0d0... - dct_interleave8(p1, p3); // a4b4c4d4... - - // store - _mm_storel_epi64((__m128i*)out, p0); - out += out_stride; - _mm_storel_epi64((__m128i*)out, _mm_shuffle_epi32(p0, 0x4e)); - out += out_stride; - _mm_storel_epi64((__m128i*)out, p2); - out += out_stride; - _mm_storel_epi64((__m128i*)out, _mm_shuffle_epi32(p2, 0x4e)); - out += out_stride; - _mm_storel_epi64((__m128i*)out, p1); - out += out_stride; - _mm_storel_epi64((__m128i*)out, _mm_shuffle_epi32(p1, 0x4e)); - out += out_stride; - _mm_storel_epi64((__m128i*)out, p3); - out += out_stride; - _mm_storel_epi64((__m128i*)out, _mm_shuffle_epi32(p3, 0x4e)); - } - -#undef dct_const -#undef dct_rot -#undef dct_widen -#undef dct_wadd -#undef dct_wsub -#undef dct_bfly32o -#undef dct_interleave8 -#undef dct_interleave16 -#undef dct_pass -} - -#endif // STBI_SSE2 - -#ifdef STBI_NEON - -// NEON integer IDCT. should produce bit-identical -// results to the generic C version. -static void stbi__idct_simd(stbi_uc* out, int out_stride, short data[64]) { - int16x8_t row0, row1, row2, row3, row4, row5, row6, row7; - - int16x4_t rot0_0 = vdup_n_s16(stbi__f2f(0.5411961f)); - int16x4_t rot0_1 = vdup_n_s16(stbi__f2f(-1.847759065f)); - int16x4_t rot0_2 = vdup_n_s16(stbi__f2f(0.765366865f)); - int16x4_t rot1_0 = vdup_n_s16(stbi__f2f(1.175875602f)); - int16x4_t rot1_1 = vdup_n_s16(stbi__f2f(-0.899976223f)); - int16x4_t rot1_2 = vdup_n_s16(stbi__f2f(-2.562915447f)); - int16x4_t rot2_0 = vdup_n_s16(stbi__f2f(-1.961570560f)); - int16x4_t rot2_1 = vdup_n_s16(stbi__f2f(-0.390180644f)); - int16x4_t rot3_0 = vdup_n_s16(stbi__f2f(0.298631336f)); - int16x4_t rot3_1 = vdup_n_s16(stbi__f2f(2.053119869f)); - int16x4_t rot3_2 = vdup_n_s16(stbi__f2f(3.072711026f)); - int16x4_t rot3_3 = vdup_n_s16(stbi__f2f(1.501321110f)); - -#define dct_long_mul(out, inq, coeff) \ - int32x4_t out##_l = vmull_s16(vget_low_s16(inq), coeff); \ - int32x4_t out##_h = vmull_s16(vget_high_s16(inq), coeff) - -#define dct_long_mac(out, acc, inq, coeff) \ - int32x4_t out##_l = vmlal_s16(acc##_l, vget_low_s16(inq), coeff); \ - int32x4_t out##_h = vmlal_s16(acc##_h, vget_high_s16(inq), coeff) - -#define dct_widen(out, inq) \ - int32x4_t out##_l = vshll_n_s16(vget_low_s16(inq), 12); \ - int32x4_t out##_h = vshll_n_s16(vget_high_s16(inq), 12) - -// wide add -#define dct_wadd(out, a, b) \ - int32x4_t out##_l = vaddq_s32(a##_l, b##_l); \ - int32x4_t out##_h = vaddq_s32(a##_h, b##_h) - -// wide sub -#define dct_wsub(out, a, b) \ - int32x4_t out##_l = vsubq_s32(a##_l, b##_l); \ - int32x4_t out##_h = vsubq_s32(a##_h, b##_h) - -// butterfly a/b, then shift using "shiftop" by "s" and pack -#define dct_bfly32o(out0, out1, a, b, shiftop, s) \ - { \ - dct_wadd(sum, a, b); \ - dct_wsub(dif, a, b); \ - out0 = vcombine_s16(shiftop(sum_l, s), shiftop(sum_h, s)); \ - out1 = vcombine_s16(shiftop(dif_l, s), shiftop(dif_h, s)); \ - } - -#define dct_pass(shiftop, shift) \ - { \ - /* even part */ \ - int16x8_t sum26 = vaddq_s16(row2, row6); \ - dct_long_mul(p1e, sum26, rot0_0); \ - dct_long_mac(t2e, p1e, row6, rot0_1); \ - dct_long_mac(t3e, p1e, row2, rot0_2); \ - int16x8_t sum04 = vaddq_s16(row0, row4); \ - int16x8_t dif04 = vsubq_s16(row0, row4); \ - dct_widen(t0e, sum04); \ - dct_widen(t1e, dif04); \ - dct_wadd(x0, t0e, t3e); \ - dct_wsub(x3, t0e, t3e); \ - dct_wadd(x1, t1e, t2e); \ - dct_wsub(x2, t1e, t2e); \ - /* odd part */ \ - int16x8_t sum15 = vaddq_s16(row1, row5); \ - int16x8_t sum17 = vaddq_s16(row1, row7); \ - int16x8_t sum35 = vaddq_s16(row3, row5); \ - int16x8_t sum37 = vaddq_s16(row3, row7); \ - int16x8_t sumodd = vaddq_s16(sum17, sum35); \ - dct_long_mul(p5o, sumodd, rot1_0); \ - dct_long_mac(p1o, p5o, sum17, rot1_1); \ - dct_long_mac(p2o, p5o, sum35, rot1_2); \ - dct_long_mul(p3o, sum37, rot2_0); \ - dct_long_mul(p4o, sum15, rot2_1); \ - dct_wadd(sump13o, p1o, p3o); \ - dct_wadd(sump24o, p2o, p4o); \ - dct_wadd(sump23o, p2o, p3o); \ - dct_wadd(sump14o, p1o, p4o); \ - dct_long_mac(x4, sump13o, row7, rot3_0); \ - dct_long_mac(x5, sump24o, row5, rot3_1); \ - dct_long_mac(x6, sump23o, row3, rot3_2); \ - dct_long_mac(x7, sump14o, row1, rot3_3); \ - dct_bfly32o(row0, row7, x0, x7, shiftop, shift); \ - dct_bfly32o(row1, row6, x1, x6, shiftop, shift); \ - dct_bfly32o(row2, row5, x2, x5, shiftop, shift); \ - dct_bfly32o(row3, row4, x3, x4, shiftop, shift); \ - } - - // load - row0 = vld1q_s16(data + 0 * 8); - row1 = vld1q_s16(data + 1 * 8); - row2 = vld1q_s16(data + 2 * 8); - row3 = vld1q_s16(data + 3 * 8); - row4 = vld1q_s16(data + 4 * 8); - row5 = vld1q_s16(data + 5 * 8); - row6 = vld1q_s16(data + 6 * 8); - row7 = vld1q_s16(data + 7 * 8); - - // add DC bias - row0 = vaddq_s16(row0, vsetq_lane_s16(1024, vdupq_n_s16(0), 0)); - - // column pass - dct_pass(vrshrn_n_s32, 10); - - // 16bit 8x8 transpose - { -// these three map to a single VTRN.16, VTRN.32, and VSWP, respectively. -// whether compilers actually get this is another story, sadly. -#define dct_trn16(x, y) \ - { \ - int16x8x2_t t = vtrnq_s16(x, y); \ - x = t.val[0]; \ - y = t.val[1]; \ - } -#define dct_trn32(x, y) \ - { \ - int32x4x2_t t = vtrnq_s32(vreinterpretq_s32_s16(x), vreinterpretq_s32_s16(y)); \ - x = vreinterpretq_s16_s32(t.val[0]); \ - y = vreinterpretq_s16_s32(t.val[1]); \ - } -#define dct_trn64(x, y) \ - { \ - int16x8_t x0 = x; \ - int16x8_t y0 = y; \ - x = vcombine_s16(vget_low_s16(x0), vget_low_s16(y0)); \ - y = vcombine_s16(vget_high_s16(x0), vget_high_s16(y0)); \ - } - - // pass 1 - dct_trn16(row0, row1); // a0b0a2b2a4b4a6b6 - dct_trn16(row2, row3); - dct_trn16(row4, row5); - dct_trn16(row6, row7); - - // pass 2 - dct_trn32(row0, row2); // a0b0c0d0a4b4c4d4 - dct_trn32(row1, row3); - dct_trn32(row4, row6); - dct_trn32(row5, row7); - - // pass 3 - dct_trn64(row0, row4); // a0b0c0d0e0f0g0h0 - dct_trn64(row1, row5); - dct_trn64(row2, row6); - dct_trn64(row3, row7); - -#undef dct_trn16 -#undef dct_trn32 -#undef dct_trn64 - } - - // row pass - // vrshrn_n_s32 only supports shifts up to 16, we need - // 17. so do a non-rounding shift of 16 first then follow - // up with a rounding shift by 1. - dct_pass(vshrn_n_s32, 16); - - { - // pack and round - uint8x8_t p0 = vqrshrun_n_s16(row0, 1); - uint8x8_t p1 = vqrshrun_n_s16(row1, 1); - uint8x8_t p2 = vqrshrun_n_s16(row2, 1); - uint8x8_t p3 = vqrshrun_n_s16(row3, 1); - uint8x8_t p4 = vqrshrun_n_s16(row4, 1); - uint8x8_t p5 = vqrshrun_n_s16(row5, 1); - uint8x8_t p6 = vqrshrun_n_s16(row6, 1); - uint8x8_t p7 = vqrshrun_n_s16(row7, 1); - - // again, these can translate into one instruction, but often don't. -#define dct_trn8_8(x, y) \ - { \ - uint8x8x2_t t = vtrn_u8(x, y); \ - x = t.val[0]; \ - y = t.val[1]; \ - } -#define dct_trn8_16(x, y) \ - { \ - uint16x4x2_t t = vtrn_u16(vreinterpret_u16_u8(x), vreinterpret_u16_u8(y)); \ - x = vreinterpret_u8_u16(t.val[0]); \ - y = vreinterpret_u8_u16(t.val[1]); \ - } -#define dct_trn8_32(x, y) \ - { \ - uint32x2x2_t t = vtrn_u32(vreinterpret_u32_u8(x), vreinterpret_u32_u8(y)); \ - x = vreinterpret_u8_u32(t.val[0]); \ - y = vreinterpret_u8_u32(t.val[1]); \ - } - - // sadly can't use interleaved stores here since we only write - // 8 bytes to each scan line! - - // 8x8 8-bit transpose pass 1 - dct_trn8_8(p0, p1); - dct_trn8_8(p2, p3); - dct_trn8_8(p4, p5); - dct_trn8_8(p6, p7); - - // pass 2 - dct_trn8_16(p0, p2); - dct_trn8_16(p1, p3); - dct_trn8_16(p4, p6); - dct_trn8_16(p5, p7); - - // pass 3 - dct_trn8_32(p0, p4); - dct_trn8_32(p1, p5); - dct_trn8_32(p2, p6); - dct_trn8_32(p3, p7); - - // store - vst1_u8(out, p0); - out += out_stride; - vst1_u8(out, p1); - out += out_stride; - vst1_u8(out, p2); - out += out_stride; - vst1_u8(out, p3); - out += out_stride; - vst1_u8(out, p4); - out += out_stride; - vst1_u8(out, p5); - out += out_stride; - vst1_u8(out, p6); - out += out_stride; - vst1_u8(out, p7); - -#undef dct_trn8_8 -#undef dct_trn8_16 -#undef dct_trn8_32 - } - -#undef dct_long_mul -#undef dct_long_mac -#undef dct_widen -#undef dct_wadd -#undef dct_wsub -#undef dct_bfly32o -#undef dct_pass -} - -#endif // STBI_NEON - -#define STBI__MARKER_none 0xff -// if there's a pending marker from the entropy stream, return that -// otherwise, fetch from the stream and get a marker. if there's no -// marker, return 0xff, which is never a valid marker value -static stbi_uc stbi__get_marker(stbi__jpeg* j) { - stbi_uc x; - if (j->marker != STBI__MARKER_none) { - x = j->marker; - j->marker = STBI__MARKER_none; - return x; - } - x = stbi__get8(j->s); - if (x != 0xff) - return STBI__MARKER_none; - while (x == 0xff) - x = stbi__get8(j->s); // consume repeated 0xff fill bytes - return x; -} - -// in each scan, we'll have scan_n components, and the order -// of the components is specified by order[] -#define STBI__RESTART(x) ((x) >= 0xd0 && (x) <= 0xd7) - -// after a restart interval, stbi__jpeg_reset the entropy decoder and -// the dc prediction -static void stbi__jpeg_reset(stbi__jpeg* j) { - j->code_bits = 0; - j->code_buffer = 0; - j->nomore = 0; - j->img_comp[0].dc_pred = j->img_comp[1].dc_pred = j->img_comp[2].dc_pred = - j->img_comp[3].dc_pred = 0; - j->marker = STBI__MARKER_none; - j->todo = j->restart_interval ? j->restart_interval : 0x7fffffff; - j->eob_run = 0; - // no more than 1<<31 MCUs if no restart_interal? that's plenty safe, - // since we don't even allow 1<<30 pixels -} - -static int stbi__parse_entropy_coded_data(stbi__jpeg* z) { - stbi__jpeg_reset(z); - if (!z->progressive) { - if (z->scan_n == 1) { - int i, j; - STBI_SIMD_ALIGN(short, data[64]); - int n = z->order[0]; - // non-interleaved data, we just need to process one block at a time, - // in trivial scanline order - // number of blocks to do just depends on how many actual "pixels" this - // component has, independent of interleaved MCU blocking and such - int w = (z->img_comp[n].x + 7) >> 3; - int h = (z->img_comp[n].y + 7) >> 3; - for (j = 0; j < h; ++j) { - for (i = 0; i < w; ++i) { - int ha = z->img_comp[n].ha; - if (!stbi__jpeg_decode_block(z, data, z->huff_dc + z->img_comp[n].hd, z->huff_ac + ha, - z->fast_ac[ha], n, z->dequant[z->img_comp[n].tq])) - return 0; - z->idct_block_kernel(z->img_comp[n].data + z->img_comp[n].w2 * j * 8 + i * 8, - z->img_comp[n].w2, data); - // every data block is an MCU, so countdown the restart interval - if (--z->todo <= 0) { - if (z->code_bits < 24) - stbi__grow_buffer_unsafe(z); - // if it's NOT a restart, then just bail, so we get corrupt data - // rather than no data - if (!STBI__RESTART(z->marker)) - return 1; - stbi__jpeg_reset(z); - } - } - } - return 1; - } else { // interleaved - int i, j, k, x, y; - STBI_SIMD_ALIGN(short, data[64]); - for (j = 0; j < z->img_mcu_y; ++j) { - for (i = 0; i < z->img_mcu_x; ++i) { - // scan an interleaved mcu... process scan_n components in order - for (k = 0; k < z->scan_n; ++k) { - int n = z->order[k]; - // scan out an mcu's worth of this component; that's just determined - // by the basic H and V specified for the component - for (y = 0; y < z->img_comp[n].v; ++y) { - for (x = 0; x < z->img_comp[n].h; ++x) { - int x2 = (i * z->img_comp[n].h + x) * 8; - int y2 = (j * z->img_comp[n].v + y) * 8; - int ha = z->img_comp[n].ha; - if (!stbi__jpeg_decode_block(z, data, z->huff_dc + z->img_comp[n].hd, - z->huff_ac + ha, z->fast_ac[ha], n, - z->dequant[z->img_comp[n].tq])) - return 0; - z->idct_block_kernel(z->img_comp[n].data + z->img_comp[n].w2 * y2 + x2, - z->img_comp[n].w2, data); - } - } - } - // after all interleaved components, that's an interleaved MCU, - // so now count down the restart interval - if (--z->todo <= 0) { - if (z->code_bits < 24) - stbi__grow_buffer_unsafe(z); - if (!STBI__RESTART(z->marker)) - return 1; - stbi__jpeg_reset(z); - } - } - } - return 1; - } - } else { - if (z->scan_n == 1) { - int i, j; - int n = z->order[0]; - // non-interleaved data, we just need to process one block at a time, - // in trivial scanline order - // number of blocks to do just depends on how many actual "pixels" this - // component has, independent of interleaved MCU blocking and such - int w = (z->img_comp[n].x + 7) >> 3; - int h = (z->img_comp[n].y + 7) >> 3; - for (j = 0; j < h; ++j) { - for (i = 0; i < w; ++i) { - short* data = z->img_comp[n].coeff + 64 * (i + j * z->img_comp[n].coeff_w); - if (z->spec_start == 0) { - if (!stbi__jpeg_decode_block_prog_dc(z, data, &z->huff_dc[z->img_comp[n].hd], n)) - return 0; - } else { - int ha = z->img_comp[n].ha; - if (!stbi__jpeg_decode_block_prog_ac(z, data, &z->huff_ac[ha], z->fast_ac[ha])) - return 0; - } - // every data block is an MCU, so countdown the restart interval - if (--z->todo <= 0) { - if (z->code_bits < 24) - stbi__grow_buffer_unsafe(z); - if (!STBI__RESTART(z->marker)) - return 1; - stbi__jpeg_reset(z); - } - } - } - return 1; - } else { // interleaved - int i, j, k, x, y; - for (j = 0; j < z->img_mcu_y; ++j) { - for (i = 0; i < z->img_mcu_x; ++i) { - // scan an interleaved mcu... process scan_n components in order - for (k = 0; k < z->scan_n; ++k) { - int n = z->order[k]; - // scan out an mcu's worth of this component; that's just determined - // by the basic H and V specified for the component - for (y = 0; y < z->img_comp[n].v; ++y) { - for (x = 0; x < z->img_comp[n].h; ++x) { - int x2 = (i * z->img_comp[n].h + x); - int y2 = (j * z->img_comp[n].v + y); - short* data = z->img_comp[n].coeff + 64 * (x2 + y2 * z->img_comp[n].coeff_w); - if (!stbi__jpeg_decode_block_prog_dc(z, data, &z->huff_dc[z->img_comp[n].hd], n)) - return 0; - } - } - } - // after all interleaved components, that's an interleaved MCU, - // so now count down the restart interval - if (--z->todo <= 0) { - if (z->code_bits < 24) - stbi__grow_buffer_unsafe(z); - if (!STBI__RESTART(z->marker)) - return 1; - stbi__jpeg_reset(z); - } - } - } - return 1; - } - } -} - -static void stbi__jpeg_dequantize(short* data, stbi__uint16* dequant) { - int i; - for (i = 0; i < 64; ++i) - data[i] *= dequant[i]; -} - -static void stbi__jpeg_finish(stbi__jpeg* z) { - if (z->progressive) { - // dequantize and idct the data - int i, j, n; - for (n = 0; n < z->s->img_n; ++n) { - int w = (z->img_comp[n].x + 7) >> 3; - int h = (z->img_comp[n].y + 7) >> 3; - for (j = 0; j < h; ++j) { - for (i = 0; i < w; ++i) { - short* data = z->img_comp[n].coeff + 64 * (i + j * z->img_comp[n].coeff_w); - stbi__jpeg_dequantize(data, z->dequant[z->img_comp[n].tq]); - z->idct_block_kernel(z->img_comp[n].data + z->img_comp[n].w2 * j * 8 + i * 8, - z->img_comp[n].w2, data); - } - } - } - } -} - -static int stbi__process_marker(stbi__jpeg* z, int m) { - int L; - switch (m) { - case STBI__MARKER_none: // no marker found - return stbi__err("expected marker", "Corrupt JPEG"); - - case 0xDD: // DRI - specify restart interval - if (stbi__get16be(z->s) != 4) - return stbi__err("bad DRI len", "Corrupt JPEG"); - z->restart_interval = stbi__get16be(z->s); - return 1; - - case 0xDB: // DQT - define quantization table - L = stbi__get16be(z->s) - 2; - while (L > 0) { - int q = stbi__get8(z->s); - int p = q >> 4, sixteen = (p != 0); - int t = q & 15, i; - if (p != 0 && p != 1) - return stbi__err("bad DQT type", "Corrupt JPEG"); - if (t > 3) - return stbi__err("bad DQT table", "Corrupt JPEG"); - - for (i = 0; i < 64; ++i) - z->dequant[t][stbi__jpeg_dezigzag[i]] = - (stbi__uint16)(sixteen ? stbi__get16be(z->s) : stbi__get8(z->s)); - L -= (sixteen ? 129 : 65); - } - return L == 0; - - case 0xC4: // DHT - define huffman table - L = stbi__get16be(z->s) - 2; - while (L > 0) { - stbi_uc* v; - int sizes[16], i, n = 0; - int q = stbi__get8(z->s); - int tc = q >> 4; - int th = q & 15; - if (tc > 1 || th > 3) - return stbi__err("bad DHT header", "Corrupt JPEG"); - for (i = 0; i < 16; ++i) { - sizes[i] = stbi__get8(z->s); - n += sizes[i]; - } - L -= 17; - if (tc == 0) { - if (!stbi__build_huffman(z->huff_dc + th, sizes)) - return 0; - v = z->huff_dc[th].values; - } else { - if (!stbi__build_huffman(z->huff_ac + th, sizes)) - return 0; - v = z->huff_ac[th].values; - } - for (i = 0; i < n; ++i) - v[i] = stbi__get8(z->s); - if (tc != 0) - stbi__build_fast_ac(z->fast_ac[th], z->huff_ac + th); - L -= n; - } - return L == 0; - } - - // check for comment block or APP blocks - if ((m >= 0xE0 && m <= 0xEF) || m == 0xFE) { - L = stbi__get16be(z->s); - if (L < 2) { - if (m == 0xFE) - return stbi__err("bad COM len", "Corrupt JPEG"); - else - return stbi__err("bad APP len", "Corrupt JPEG"); - } - L -= 2; - - if (m == 0xE0 && L >= 5) { // JFIF APP0 segment - static const unsigned char tag[5] = {'J', 'F', 'I', 'F', '\0'}; - int ok = 1; - int i; - for (i = 0; i < 5; ++i) - if (stbi__get8(z->s) != tag[i]) - ok = 0; - L -= 5; - if (ok) - z->jfif = 1; - } else if (m == 0xEE && L >= 12) { // Adobe APP14 segment - static const unsigned char tag[6] = {'A', 'd', 'o', 'b', 'e', '\0'}; - int ok = 1; - int i; - for (i = 0; i < 6; ++i) - if (stbi__get8(z->s) != tag[i]) - ok = 0; - L -= 6; - if (ok) { - stbi__get8(z->s); // version - stbi__get16be(z->s); // flags0 - stbi__get16be(z->s); // flags1 - z->app14_color_transform = stbi__get8(z->s); // color transform - L -= 6; - } - } - - stbi__skip(z->s, L); - return 1; - } - - return stbi__err("unknown marker", "Corrupt JPEG"); -} - -// after we see SOS -static int stbi__process_scan_header(stbi__jpeg* z) { - int i; - int Ls = stbi__get16be(z->s); - z->scan_n = stbi__get8(z->s); - if (z->scan_n < 1 || z->scan_n > 4 || z->scan_n > (int)z->s->img_n) - return stbi__err("bad SOS component count", "Corrupt JPEG"); - if (Ls != 6 + 2 * z->scan_n) - return stbi__err("bad SOS len", "Corrupt JPEG"); - for (i = 0; i < z->scan_n; ++i) { - int id = stbi__get8(z->s), which; - int q = stbi__get8(z->s); - for (which = 0; which < z->s->img_n; ++which) - if (z->img_comp[which].id == id) - break; - if (which == z->s->img_n) - return 0; // no match - z->img_comp[which].hd = q >> 4; - if (z->img_comp[which].hd > 3) - return stbi__err("bad DC huff", "Corrupt JPEG"); - z->img_comp[which].ha = q & 15; - if (z->img_comp[which].ha > 3) - return stbi__err("bad AC huff", "Corrupt JPEG"); - z->order[i] = which; - } - - { - int aa; - z->spec_start = stbi__get8(z->s); - z->spec_end = stbi__get8(z->s); // should be 63, but might be 0 - aa = stbi__get8(z->s); - z->succ_high = (aa >> 4); - z->succ_low = (aa & 15); - if (z->progressive) { - if (z->spec_start > 63 || z->spec_end > 63 || z->spec_start > z->spec_end || - z->succ_high > 13 || z->succ_low > 13) - return stbi__err("bad SOS", "Corrupt JPEG"); - } else { - if (z->spec_start != 0) - return stbi__err("bad SOS", "Corrupt JPEG"); - if (z->succ_high != 0 || z->succ_low != 0) - return stbi__err("bad SOS", "Corrupt JPEG"); - z->spec_end = 63; - } - } - - return 1; -} - -static int stbi__free_jpeg_components(stbi__jpeg* z, int ncomp, int why) { - int i; - for (i = 0; i < ncomp; ++i) { - if (z->img_comp[i].raw_data) { - STBI_FREE(z->img_comp[i].raw_data); - z->img_comp[i].raw_data = NULL; - z->img_comp[i].data = NULL; - } - if (z->img_comp[i].raw_coeff) { - STBI_FREE(z->img_comp[i].raw_coeff); - z->img_comp[i].raw_coeff = 0; - z->img_comp[i].coeff = 0; - } - if (z->img_comp[i].linebuf) { - STBI_FREE(z->img_comp[i].linebuf); - z->img_comp[i].linebuf = NULL; - } - } - return why; -} - -static int stbi__process_frame_header(stbi__jpeg* z, int scan) { - stbi__context* s = z->s; - int Lf, p, i, q, h_max = 1, v_max = 1, c; - Lf = stbi__get16be(s); - if (Lf < 11) - return stbi__err("bad SOF len", "Corrupt JPEG"); // JPEG - p = stbi__get8(s); - if (p != 8) - return stbi__err("only 8-bit", "JPEG format not supported: 8-bit only"); // JPEG baseline - s->img_y = stbi__get16be(s); - if (s->img_y == 0) - return stbi__err("no header height", - "JPEG format not supported: delayed height"); // Legal, but we don't handle - // it--but neither does IJG - s->img_x = stbi__get16be(s); - if (s->img_x == 0) - return stbi__err("0 width", "Corrupt JPEG"); // JPEG requires - if (s->img_y > STBI_MAX_DIMENSIONS) - return stbi__err("too large", "Very large image (corrupt?)"); - if (s->img_x > STBI_MAX_DIMENSIONS) - return stbi__err("too large", "Very large image (corrupt?)"); - c = stbi__get8(s); - if (c != 3 && c != 1 && c != 4) - return stbi__err("bad component count", "Corrupt JPEG"); - s->img_n = c; - for (i = 0; i < c; ++i) { - z->img_comp[i].data = NULL; - z->img_comp[i].linebuf = NULL; - } - - if (Lf != 8 + 3 * s->img_n) - return stbi__err("bad SOF len", "Corrupt JPEG"); - - z->rgb = 0; - for (i = 0; i < s->img_n; ++i) { - static const unsigned char rgb[3] = {'R', 'G', 'B'}; - z->img_comp[i].id = stbi__get8(s); - if (s->img_n == 3 && z->img_comp[i].id == rgb[i]) - ++z->rgb; - q = stbi__get8(s); - z->img_comp[i].h = (q >> 4); - if (!z->img_comp[i].h || z->img_comp[i].h > 4) - return stbi__err("bad H", "Corrupt JPEG"); - z->img_comp[i].v = q & 15; - if (!z->img_comp[i].v || z->img_comp[i].v > 4) - return stbi__err("bad V", "Corrupt JPEG"); - z->img_comp[i].tq = stbi__get8(s); - if (z->img_comp[i].tq > 3) - return stbi__err("bad TQ", "Corrupt JPEG"); - } - - if (scan != STBI__SCAN_load) - return 1; - - if (!stbi__mad3sizes_valid(s->img_x, s->img_y, s->img_n, 0)) - return stbi__err("too large", "Image too large to decode"); - - for (i = 0; i < s->img_n; ++i) { - if (z->img_comp[i].h > h_max) - h_max = z->img_comp[i].h; - if (z->img_comp[i].v > v_max) - v_max = z->img_comp[i].v; - } - - // check that plane subsampling factors are integer ratios; our resamplers can't deal with - // fractional ratios and I've never seen a non-corrupted JPEG file actually use them - for (i = 0; i < s->img_n; ++i) { - if (h_max % z->img_comp[i].h != 0) - return stbi__err("bad H", "Corrupt JPEG"); - if (v_max % z->img_comp[i].v != 0) - return stbi__err("bad V", "Corrupt JPEG"); - } - - // compute interleaved mcu info - z->img_h_max = h_max; - z->img_v_max = v_max; - z->img_mcu_w = h_max * 8; - z->img_mcu_h = v_max * 8; - // these sizes can't be more than 17 bits - z->img_mcu_x = (s->img_x + z->img_mcu_w - 1) / z->img_mcu_w; - z->img_mcu_y = (s->img_y + z->img_mcu_h - 1) / z->img_mcu_h; - - for (i = 0; i < s->img_n; ++i) { - // number of effective pixels (e.g. for non-interleaved MCU) - z->img_comp[i].x = (s->img_x * z->img_comp[i].h + h_max - 1) / h_max; - z->img_comp[i].y = (s->img_y * z->img_comp[i].v + v_max - 1) / v_max; - // to simplify generation, we'll allocate enough memory to decode - // the bogus oversized data from using interleaved MCUs and their - // big blocks (e.g. a 16x16 iMCU on an image of width 33); we won't - // discard the extra data until colorspace conversion - // - // img_mcu_x, img_mcu_y: <=17 bits; comp[i].h and .v are <=4 (checked earlier) - // so these muls can't overflow with 32-bit ints (which we require) - z->img_comp[i].w2 = z->img_mcu_x * z->img_comp[i].h * 8; - z->img_comp[i].h2 = z->img_mcu_y * z->img_comp[i].v * 8; - z->img_comp[i].coeff = 0; - z->img_comp[i].raw_coeff = 0; - z->img_comp[i].linebuf = NULL; - z->img_comp[i].raw_data = stbi__malloc_mad2(z->img_comp[i].w2, z->img_comp[i].h2, 15); - if (z->img_comp[i].raw_data == NULL) - return stbi__free_jpeg_components(z, i + 1, stbi__err("outofmem", "Out of memory")); - // align blocks for idct using mmx/sse - z->img_comp[i].data = (stbi_uc*)(((size_t)z->img_comp[i].raw_data + 15) & ~15); - if (z->progressive) { - // w2, h2 are multiples of 8 (see above) - z->img_comp[i].coeff_w = z->img_comp[i].w2 / 8; - z->img_comp[i].coeff_h = z->img_comp[i].h2 / 8; - z->img_comp[i].raw_coeff = - stbi__malloc_mad3(z->img_comp[i].w2, z->img_comp[i].h2, sizeof(short), 15); - if (z->img_comp[i].raw_coeff == NULL) - return stbi__free_jpeg_components(z, i + 1, stbi__err("outofmem", "Out of memory")); - z->img_comp[i].coeff = (short*)(((size_t)z->img_comp[i].raw_coeff + 15) & ~15); - } - } - - return 1; -} - -// use comparisons since in some cases we handle more than one case (e.g. SOF) -#define stbi__DNL(x) ((x) == 0xdc) -#define stbi__SOI(x) ((x) == 0xd8) -#define stbi__EOI(x) ((x) == 0xd9) -#define stbi__SOF(x) ((x) == 0xc0 || (x) == 0xc1 || (x) == 0xc2) -#define stbi__SOS(x) ((x) == 0xda) - -#define stbi__SOF_progressive(x) ((x) == 0xc2) - -static int stbi__decode_jpeg_header(stbi__jpeg* z, int scan) { - int m; - z->jfif = 0; - z->app14_color_transform = -1; // valid values are 0,1,2 - z->marker = STBI__MARKER_none; // initialize cached marker to empty - m = stbi__get_marker(z); - if (!stbi__SOI(m)) - return stbi__err("no SOI", "Corrupt JPEG"); - if (scan == STBI__SCAN_type) - return 1; - m = stbi__get_marker(z); - while (!stbi__SOF(m)) { - if (!stbi__process_marker(z, m)) - return 0; - m = stbi__get_marker(z); - while (m == STBI__MARKER_none) { - // some files have extra padding after their blocks, so ok, we'll scan - if (stbi__at_eof(z->s)) - return stbi__err("no SOF", "Corrupt JPEG"); - m = stbi__get_marker(z); - } - } - z->progressive = stbi__SOF_progressive(m); - if (!stbi__process_frame_header(z, scan)) - return 0; - return 1; -} - -// decode image to YCbCr format -static int stbi__decode_jpeg_image(stbi__jpeg* j) { - int m; - for (m = 0; m < 4; m++) { - j->img_comp[m].raw_data = NULL; - j->img_comp[m].raw_coeff = NULL; - } - j->restart_interval = 0; - if (!stbi__decode_jpeg_header(j, STBI__SCAN_load)) - return 0; - m = stbi__get_marker(j); - while (!stbi__EOI(m)) { - if (stbi__SOS(m)) { - if (!stbi__process_scan_header(j)) - return 0; - if (!stbi__parse_entropy_coded_data(j)) - return 0; - if (j->marker == STBI__MARKER_none) { - // handle 0s at the end of image data from IP Kamera 9060 - while (!stbi__at_eof(j->s)) { - int x = stbi__get8(j->s); - if (x == 255) { - j->marker = stbi__get8(j->s); - break; - } - } - // if we reach eof without hitting a marker, stbi__get_marker() below will fail and we'll - // eventually return 0 - } - } else if (stbi__DNL(m)) { - int Ld = stbi__get16be(j->s); - stbi__uint32 NL = stbi__get16be(j->s); - if (Ld != 4) - return stbi__err("bad DNL len", "Corrupt JPEG"); - if (NL != j->s->img_y) - return stbi__err("bad DNL height", "Corrupt JPEG"); - } else { - if (!stbi__process_marker(j, m)) - return 0; - } - m = stbi__get_marker(j); - } - if (j->progressive) - stbi__jpeg_finish(j); - return 1; -} - -// static jfif-centered resampling (across block boundaries) - -typedef stbi_uc* (*resample_row_func)(stbi_uc* out, stbi_uc* in0, stbi_uc* in1, int w, int hs); - -#define stbi__div4(x) ((stbi_uc)((x) >> 2)) - -static stbi_uc* resample_row_1(stbi_uc* out, stbi_uc* in_near, stbi_uc* in_far, int w, int hs) { - STBI_NOTUSED(out); - STBI_NOTUSED(in_far); - STBI_NOTUSED(w); - STBI_NOTUSED(hs); - return in_near; -} - -static stbi_uc* stbi__resample_row_v_2(stbi_uc* out, - stbi_uc* in_near, - stbi_uc* in_far, - int w, - int hs) { - // need to generate two samples vertically for every one in input - int i; - STBI_NOTUSED(hs); - for (i = 0; i < w; ++i) - out[i] = stbi__div4(3 * in_near[i] + in_far[i] + 2); - return out; -} - -static stbi_uc* stbi__resample_row_h_2(stbi_uc* out, - stbi_uc* in_near, - stbi_uc* in_far, - int w, - int hs) { - // need to generate two samples horizontally for every one in input - int i; - stbi_uc* input = in_near; - - if (w == 1) { - // if only one sample, can't do any interpolation - out[0] = out[1] = input[0]; - return out; - } - - out[0] = input[0]; - out[1] = stbi__div4(input[0] * 3 + input[1] + 2); - for (i = 1; i < w - 1; ++i) { - int n = 3 * input[i] + 2; - out[i * 2 + 0] = stbi__div4(n + input[i - 1]); - out[i * 2 + 1] = stbi__div4(n + input[i + 1]); - } - out[i * 2 + 0] = stbi__div4(input[w - 2] * 3 + input[w - 1] + 2); - out[i * 2 + 1] = input[w - 1]; - - STBI_NOTUSED(in_far); - STBI_NOTUSED(hs); - - return out; -} - -#define stbi__div16(x) ((stbi_uc)((x) >> 4)) - -static stbi_uc* stbi__resample_row_hv_2(stbi_uc* out, - stbi_uc* in_near, - stbi_uc* in_far, - int w, - int hs) { - // need to generate 2x2 samples for every one in input - int i, t0, t1; - if (w == 1) { - out[0] = out[1] = stbi__div4(3 * in_near[0] + in_far[0] + 2); - return out; - } - - t1 = 3 * in_near[0] + in_far[0]; - out[0] = stbi__div4(t1 + 2); - for (i = 1; i < w; ++i) { - t0 = t1; - t1 = 3 * in_near[i] + in_far[i]; - out[i * 2 - 1] = stbi__div16(3 * t0 + t1 + 8); - out[i * 2] = stbi__div16(3 * t1 + t0 + 8); - } - out[w * 2 - 1] = stbi__div4(t1 + 2); - - STBI_NOTUSED(hs); - - return out; -} - -#if defined(STBI_SSE2) || defined(STBI_NEON) -static stbi_uc* stbi__resample_row_hv_2_simd(stbi_uc* out, - stbi_uc* in_near, - stbi_uc* in_far, - int w, - int hs) { - // need to generate 2x2 samples for every one in input - int i = 0, t0, t1; - - if (w == 1) { - out[0] = out[1] = stbi__div4(3 * in_near[0] + in_far[0] + 2); - return out; - } - - t1 = 3 * in_near[0] + in_far[0]; - // process groups of 8 pixels for as long as we can. - // note we can't handle the last pixel in a row in this loop - // because we need to handle the filter boundary conditions. - for (; i < ((w - 1) & ~7); i += 8) { -#if defined(STBI_SSE2) - // load and perform the vertical filtering pass - // this uses 3*x + y = 4*x + (y - x) - __m128i zero = _mm_setzero_si128(); - __m128i farb = _mm_loadl_epi64((__m128i*)(in_far + i)); - __m128i nearb = _mm_loadl_epi64((__m128i*)(in_near + i)); - __m128i farw = _mm_unpacklo_epi8(farb, zero); - __m128i nearw = _mm_unpacklo_epi8(nearb, zero); - __m128i diff = _mm_sub_epi16(farw, nearw); - __m128i nears = _mm_slli_epi16(nearw, 2); - __m128i curr = _mm_add_epi16(nears, diff); // current row - - // horizontal filter works the same based on shifted vers of current - // row. "prev" is current row shifted right by 1 pixel; we need to - // insert the previous pixel value (from t1). - // "next" is current row shifted left by 1 pixel, with first pixel - // of next block of 8 pixels added in. - __m128i prv0 = _mm_slli_si128(curr, 2); - __m128i nxt0 = _mm_srli_si128(curr, 2); - __m128i prev = _mm_insert_epi16(prv0, t1, 0); - __m128i next = _mm_insert_epi16(nxt0, 3 * in_near[i + 8] + in_far[i + 8], 7); - - // horizontal filter, polyphase implementation since it's convenient: - // even pixels = 3*cur + prev = cur*4 + (prev - cur) - // odd pixels = 3*cur + next = cur*4 + (next - cur) - // note the shared term. - __m128i bias = _mm_set1_epi16(8); - __m128i curs = _mm_slli_epi16(curr, 2); - __m128i prvd = _mm_sub_epi16(prev, curr); - __m128i nxtd = _mm_sub_epi16(next, curr); - __m128i curb = _mm_add_epi16(curs, bias); - __m128i even = _mm_add_epi16(prvd, curb); - __m128i odd = _mm_add_epi16(nxtd, curb); - - // interleave even and odd pixels, then undo scaling. - __m128i int0 = _mm_unpacklo_epi16(even, odd); - __m128i int1 = _mm_unpackhi_epi16(even, odd); - __m128i de0 = _mm_srli_epi16(int0, 4); - __m128i de1 = _mm_srli_epi16(int1, 4); - - // pack and write output - __m128i outv = _mm_packus_epi16(de0, de1); - _mm_storeu_si128((__m128i*)(out + i * 2), outv); -#elif defined(STBI_NEON) - // load and perform the vertical filtering pass - // this uses 3*x + y = 4*x + (y - x) - uint8x8_t farb = vld1_u8(in_far + i); - uint8x8_t nearb = vld1_u8(in_near + i); - int16x8_t diff = vreinterpretq_s16_u16(vsubl_u8(farb, nearb)); - int16x8_t nears = vreinterpretq_s16_u16(vshll_n_u8(nearb, 2)); - int16x8_t curr = vaddq_s16(nears, diff); // current row - - // horizontal filter works the same based on shifted vers of current - // row. "prev" is current row shifted right by 1 pixel; we need to - // insert the previous pixel value (from t1). - // "next" is current row shifted left by 1 pixel, with first pixel - // of next block of 8 pixels added in. - int16x8_t prv0 = vextq_s16(curr, curr, 7); - int16x8_t nxt0 = vextq_s16(curr, curr, 1); - int16x8_t prev = vsetq_lane_s16(t1, prv0, 0); - int16x8_t next = vsetq_lane_s16(3 * in_near[i + 8] + in_far[i + 8], nxt0, 7); - - // horizontal filter, polyphase implementation since it's convenient: - // even pixels = 3*cur + prev = cur*4 + (prev - cur) - // odd pixels = 3*cur + next = cur*4 + (next - cur) - // note the shared term. - int16x8_t curs = vshlq_n_s16(curr, 2); - int16x8_t prvd = vsubq_s16(prev, curr); - int16x8_t nxtd = vsubq_s16(next, curr); - int16x8_t even = vaddq_s16(curs, prvd); - int16x8_t odd = vaddq_s16(curs, nxtd); - - // undo scaling and round, then store with even/odd phases interleaved - uint8x8x2_t o; - o.val[0] = vqrshrun_n_s16(even, 4); - o.val[1] = vqrshrun_n_s16(odd, 4); - vst2_u8(out + i * 2, o); -#endif - - // "previous" value for next iter - t1 = 3 * in_near[i + 7] + in_far[i + 7]; - } - - t0 = t1; - t1 = 3 * in_near[i] + in_far[i]; - out[i * 2] = stbi__div16(3 * t1 + t0 + 8); - - for (++i; i < w; ++i) { - t0 = t1; - t1 = 3 * in_near[i] + in_far[i]; - out[i * 2 - 1] = stbi__div16(3 * t0 + t1 + 8); - out[i * 2] = stbi__div16(3 * t1 + t0 + 8); - } - out[w * 2 - 1] = stbi__div4(t1 + 2); - - STBI_NOTUSED(hs); - - return out; -} -#endif - -static stbi_uc* stbi__resample_row_generic(stbi_uc* out, - stbi_uc* in_near, - stbi_uc* in_far, - int w, - int hs) { - // resample with nearest-neighbor - int i, j; - STBI_NOTUSED(in_far); - for (i = 0; i < w; ++i) - for (j = 0; j < hs; ++j) - out[i * hs + j] = in_near[i]; - return out; -} - -// this is a reduced-precision calculation of YCbCr-to-RGB introduced -// to make sure the code produces the same results in both SIMD and scalar -#define stbi__float2fixed(x) (((int)((x)*4096.0f + 0.5f)) << 8) -static void stbi__YCbCr_to_RGB_row(stbi_uc* out, - const stbi_uc* y, - const stbi_uc* pcb, - const stbi_uc* pcr, - int count, - int step) { - int i; - for (i = 0; i < count; ++i) { - int y_fixed = (y[i] << 20) + (1 << 19); // rounding - int r, g, b; - int cr = pcr[i] - 128; - int cb = pcb[i] - 128; - r = y_fixed + cr * stbi__float2fixed(1.40200f); - g = y_fixed + (cr * -stbi__float2fixed(0.71414f)) + - ((cb * -stbi__float2fixed(0.34414f)) & 0xffff0000); - b = y_fixed + cb * stbi__float2fixed(1.77200f); - r >>= 20; - g >>= 20; - b >>= 20; - if ((unsigned)r > 255) { - if (r < 0) - r = 0; - else - r = 255; - } - if ((unsigned)g > 255) { - if (g < 0) - g = 0; - else - g = 255; - } - if ((unsigned)b > 255) { - if (b < 0) - b = 0; - else - b = 255; - } - out[0] = (stbi_uc)r; - out[1] = (stbi_uc)g; - out[2] = (stbi_uc)b; - out[3] = 255; - out += step; - } -} - -#if defined(STBI_SSE2) || defined(STBI_NEON) -static void stbi__YCbCr_to_RGB_simd(stbi_uc* out, - stbi_uc const* y, - stbi_uc const* pcb, - stbi_uc const* pcr, - int count, - int step) { - int i = 0; - -#ifdef STBI_SSE2 - // step == 3 is pretty ugly on the final interleave, and i'm not convinced - // it's useful in practice (you wouldn't use it for textures, for example). - // so just accelerate step == 4 case. - if (step == 4) { - // this is a fairly straightforward implementation and not super-optimized. - __m128i signflip = _mm_set1_epi8(-0x80); - __m128i cr_const0 = _mm_set1_epi16((short)(1.40200f * 4096.0f + 0.5f)); - __m128i cr_const1 = _mm_set1_epi16(-(short)(0.71414f * 4096.0f + 0.5f)); - __m128i cb_const0 = _mm_set1_epi16(-(short)(0.34414f * 4096.0f + 0.5f)); - __m128i cb_const1 = _mm_set1_epi16((short)(1.77200f * 4096.0f + 0.5f)); - __m128i y_bias = _mm_set1_epi8((char)(unsigned char)128); - __m128i xw = _mm_set1_epi16(255); // alpha channel - - for (; i + 7 < count; i += 8) { - // load - __m128i y_bytes = _mm_loadl_epi64((__m128i*)(y + i)); - __m128i cr_bytes = _mm_loadl_epi64((__m128i*)(pcr + i)); - __m128i cb_bytes = _mm_loadl_epi64((__m128i*)(pcb + i)); - __m128i cr_biased = _mm_xor_si128(cr_bytes, signflip); // -128 - __m128i cb_biased = _mm_xor_si128(cb_bytes, signflip); // -128 - - // unpack to short (and left-shift cr, cb by 8) - __m128i yw = _mm_unpacklo_epi8(y_bias, y_bytes); - __m128i crw = _mm_unpacklo_epi8(_mm_setzero_si128(), cr_biased); - __m128i cbw = _mm_unpacklo_epi8(_mm_setzero_si128(), cb_biased); - - // color transform - __m128i yws = _mm_srli_epi16(yw, 4); - __m128i cr0 = _mm_mulhi_epi16(cr_const0, crw); - __m128i cb0 = _mm_mulhi_epi16(cb_const0, cbw); - __m128i cb1 = _mm_mulhi_epi16(cbw, cb_const1); - __m128i cr1 = _mm_mulhi_epi16(crw, cr_const1); - __m128i rws = _mm_add_epi16(cr0, yws); - __m128i gwt = _mm_add_epi16(cb0, yws); - __m128i bws = _mm_add_epi16(yws, cb1); - __m128i gws = _mm_add_epi16(gwt, cr1); - - // descale - __m128i rw = _mm_srai_epi16(rws, 4); - __m128i bw = _mm_srai_epi16(bws, 4); - __m128i gw = _mm_srai_epi16(gws, 4); - - // back to byte, set up for transpose - __m128i brb = _mm_packus_epi16(rw, bw); - __m128i gxb = _mm_packus_epi16(gw, xw); - - // transpose to interleave channels - __m128i t0 = _mm_unpacklo_epi8(brb, gxb); - __m128i t1 = _mm_unpackhi_epi8(brb, gxb); - __m128i o0 = _mm_unpacklo_epi16(t0, t1); - __m128i o1 = _mm_unpackhi_epi16(t0, t1); - - // store - _mm_storeu_si128((__m128i*)(out + 0), o0); - _mm_storeu_si128((__m128i*)(out + 16), o1); - out += 32; - } - } -#endif - -#ifdef STBI_NEON - // in this version, step=3 support would be easy to add. but is there demand? - if (step == 4) { - // this is a fairly straightforward implementation and not super-optimized. - uint8x8_t signflip = vdup_n_u8(0x80); - int16x8_t cr_const0 = vdupq_n_s16((short)(1.40200f * 4096.0f + 0.5f)); - int16x8_t cr_const1 = vdupq_n_s16(-(short)(0.71414f * 4096.0f + 0.5f)); - int16x8_t cb_const0 = vdupq_n_s16(-(short)(0.34414f * 4096.0f + 0.5f)); - int16x8_t cb_const1 = vdupq_n_s16((short)(1.77200f * 4096.0f + 0.5f)); - - for (; i + 7 < count; i += 8) { - // load - uint8x8_t y_bytes = vld1_u8(y + i); - uint8x8_t cr_bytes = vld1_u8(pcr + i); - uint8x8_t cb_bytes = vld1_u8(pcb + i); - int8x8_t cr_biased = vreinterpret_s8_u8(vsub_u8(cr_bytes, signflip)); - int8x8_t cb_biased = vreinterpret_s8_u8(vsub_u8(cb_bytes, signflip)); - - // expand to s16 - int16x8_t yws = vreinterpretq_s16_u16(vshll_n_u8(y_bytes, 4)); - int16x8_t crw = vshll_n_s8(cr_biased, 7); - int16x8_t cbw = vshll_n_s8(cb_biased, 7); - - // color transform - int16x8_t cr0 = vqdmulhq_s16(crw, cr_const0); - int16x8_t cb0 = vqdmulhq_s16(cbw, cb_const0); - int16x8_t cr1 = vqdmulhq_s16(crw, cr_const1); - int16x8_t cb1 = vqdmulhq_s16(cbw, cb_const1); - int16x8_t rws = vaddq_s16(yws, cr0); - int16x8_t gws = vaddq_s16(vaddq_s16(yws, cb0), cr1); - int16x8_t bws = vaddq_s16(yws, cb1); - - // undo scaling, round, convert to byte - uint8x8x4_t o; - o.val[0] = vqrshrun_n_s16(rws, 4); - o.val[1] = vqrshrun_n_s16(gws, 4); - o.val[2] = vqrshrun_n_s16(bws, 4); - o.val[3] = vdup_n_u8(255); - - // store, interleaving r/g/b/a - vst4_u8(out, o); - out += 8 * 4; - } - } -#endif - - for (; i < count; ++i) { - int y_fixed = (y[i] << 20) + (1 << 19); // rounding - int r, g, b; - int cr = pcr[i] - 128; - int cb = pcb[i] - 128; - r = y_fixed + cr * stbi__float2fixed(1.40200f); - g = y_fixed + cr * -stbi__float2fixed(0.71414f) + - ((cb * -stbi__float2fixed(0.34414f)) & 0xffff0000); - b = y_fixed + cb * stbi__float2fixed(1.77200f); - r >>= 20; - g >>= 20; - b >>= 20; - if ((unsigned)r > 255) { - if (r < 0) - r = 0; - else - r = 255; - } - if ((unsigned)g > 255) { - if (g < 0) - g = 0; - else - g = 255; - } - if ((unsigned)b > 255) { - if (b < 0) - b = 0; - else - b = 255; - } - out[0] = (stbi_uc)r; - out[1] = (stbi_uc)g; - out[2] = (stbi_uc)b; - out[3] = 255; - out += step; - } -} -#endif - -// set up the kernels -static void stbi__setup_jpeg(stbi__jpeg* j) { - j->idct_block_kernel = stbi__idct_block; - j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_row; - j->resample_row_hv_2_kernel = stbi__resample_row_hv_2; - -#ifdef STBI_SSE2 - if (stbi__sse2_available()) { - j->idct_block_kernel = stbi__idct_simd; - j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_simd; - j->resample_row_hv_2_kernel = stbi__resample_row_hv_2_simd; - } -#endif - -#ifdef STBI_NEON - j->idct_block_kernel = stbi__idct_simd; - j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_simd; - j->resample_row_hv_2_kernel = stbi__resample_row_hv_2_simd; -#endif -} - -// clean up the temporary component buffers -static void stbi__cleanup_jpeg(stbi__jpeg* j) { - stbi__free_jpeg_components(j, j->s->img_n, 0); -} - -typedef struct { - resample_row_func resample; - stbi_uc *line0, *line1; - int hs, vs; // expansion factor in each axis - int w_lores; // horizontal pixels pre-expansion - int ystep; // how far through vertical expansion we are - int ypos; // which pre-expansion row we're on -} stbi__resample; - -// fast 0..255 * 0..255 => 0..255 rounded multiplication -static stbi_uc stbi__blinn_8x8(stbi_uc x, stbi_uc y) { - unsigned int t = x * y + 128; - return (stbi_uc)((t + (t >> 8)) >> 8); -} - -static stbi_uc* load_jpeg_image(stbi__jpeg* z, int* out_x, int* out_y, int* comp, int req_comp) { - int n, decode_n, is_rgb; - z->s->img_n = 0; // make stbi__cleanup_jpeg safe - - // validate req_comp - if (req_comp < 0 || req_comp > 4) - return stbi__errpuc("bad req_comp", "Internal error"); - - // load a jpeg image from whichever source, but leave in YCbCr format - if (!stbi__decode_jpeg_image(z)) { - stbi__cleanup_jpeg(z); - return NULL; - } - - // determine actual number of components to generate - n = req_comp ? req_comp : z->s->img_n >= 3 ? 3 : 1; - - is_rgb = z->s->img_n == 3 && (z->rgb == 3 || (z->app14_color_transform == 0 && !z->jfif)); - - if (z->s->img_n == 3 && n < 3 && !is_rgb) - decode_n = 1; - else - decode_n = z->s->img_n; - - // nothing to do if no components requested; check this now to avoid - // accessing uninitialized coutput[0] later - if (decode_n <= 0) { - stbi__cleanup_jpeg(z); - return NULL; - } - - // resample and color-convert - { - int k; - unsigned int i, j; - stbi_uc* output; - stbi_uc* coutput[4] = {NULL, NULL, NULL, NULL}; - - stbi__resample res_comp[4]; - - for (k = 0; k < decode_n; ++k) { - stbi__resample* r = &res_comp[k]; - - // allocate line buffer big enough for upsampling off the edges - // with upsample factor of 4 - z->img_comp[k].linebuf = (stbi_uc*)stbi__malloc(z->s->img_x + 3); - if (!z->img_comp[k].linebuf) { - stbi__cleanup_jpeg(z); - return stbi__errpuc("outofmem", "Out of memory"); - } - - r->hs = z->img_h_max / z->img_comp[k].h; - r->vs = z->img_v_max / z->img_comp[k].v; - r->ystep = r->vs >> 1; - r->w_lores = (z->s->img_x + r->hs - 1) / r->hs; - r->ypos = 0; - r->line0 = r->line1 = z->img_comp[k].data; - - if (r->hs == 1 && r->vs == 1) - r->resample = resample_row_1; - else if (r->hs == 1 && r->vs == 2) - r->resample = stbi__resample_row_v_2; - else if (r->hs == 2 && r->vs == 1) - r->resample = stbi__resample_row_h_2; - else if (r->hs == 2 && r->vs == 2) - r->resample = z->resample_row_hv_2_kernel; - else - r->resample = stbi__resample_row_generic; - } - - // can't error after this so, this is safe - output = (stbi_uc*)stbi__malloc_mad3(n, z->s->img_x, z->s->img_y, 1); - if (!output) { - stbi__cleanup_jpeg(z); - return stbi__errpuc("outofmem", "Out of memory"); - } - - // now go ahead and resample - for (j = 0; j < z->s->img_y; ++j) { - stbi_uc* out = output + n * z->s->img_x * j; - for (k = 0; k < decode_n; ++k) { - stbi__resample* r = &res_comp[k]; - int y_bot = r->ystep >= (r->vs >> 1); - coutput[k] = r->resample(z->img_comp[k].linebuf, y_bot ? r->line1 : r->line0, - y_bot ? r->line0 : r->line1, r->w_lores, r->hs); - if (++r->ystep >= r->vs) { - r->ystep = 0; - r->line0 = r->line1; - if (++r->ypos < z->img_comp[k].y) - r->line1 += z->img_comp[k].w2; - } - } - if (n >= 3) { - stbi_uc* y = coutput[0]; - if (z->s->img_n == 3) { - if (is_rgb) { - for (i = 0; i < z->s->img_x; ++i) { - out[0] = y[i]; - out[1] = coutput[1][i]; - out[2] = coutput[2][i]; - out[3] = 255; - out += n; - } - } else { - z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); - } - } else if (z->s->img_n == 4) { - if (z->app14_color_transform == 0) { // CMYK - for (i = 0; i < z->s->img_x; ++i) { - stbi_uc m = coutput[3][i]; - out[0] = stbi__blinn_8x8(coutput[0][i], m); - out[1] = stbi__blinn_8x8(coutput[1][i], m); - out[2] = stbi__blinn_8x8(coutput[2][i], m); - out[3] = 255; - out += n; - } - } else if (z->app14_color_transform == 2) { // YCCK - z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); - for (i = 0; i < z->s->img_x; ++i) { - stbi_uc m = coutput[3][i]; - out[0] = stbi__blinn_8x8(255 - out[0], m); - out[1] = stbi__blinn_8x8(255 - out[1], m); - out[2] = stbi__blinn_8x8(255 - out[2], m); - out += n; - } - } else { // YCbCr + alpha? Ignore the fourth channel for now - z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); - } - } else - for (i = 0; i < z->s->img_x; ++i) { - out[0] = out[1] = out[2] = y[i]; - out[3] = 255; // not used if n==3 - out += n; - } - } else { - if (is_rgb) { - if (n == 1) - for (i = 0; i < z->s->img_x; ++i) - *out++ = stbi__compute_y(coutput[0][i], coutput[1][i], coutput[2][i]); - else { - for (i = 0; i < z->s->img_x; ++i, out += 2) { - out[0] = stbi__compute_y(coutput[0][i], coutput[1][i], coutput[2][i]); - out[1] = 255; - } - } - } else if (z->s->img_n == 4 && z->app14_color_transform == 0) { - for (i = 0; i < z->s->img_x; ++i) { - stbi_uc m = coutput[3][i]; - stbi_uc r = stbi__blinn_8x8(coutput[0][i], m); - stbi_uc g = stbi__blinn_8x8(coutput[1][i], m); - stbi_uc b = stbi__blinn_8x8(coutput[2][i], m); - out[0] = stbi__compute_y(r, g, b); - out[1] = 255; - out += n; - } - } else if (z->s->img_n == 4 && z->app14_color_transform == 2) { - for (i = 0; i < z->s->img_x; ++i) { - out[0] = stbi__blinn_8x8(255 - coutput[0][i], coutput[3][i]); - out[1] = 255; - out += n; - } - } else { - stbi_uc* y = coutput[0]; - if (n == 1) - for (i = 0; i < z->s->img_x; ++i) - out[i] = y[i]; - else - for (i = 0; i < z->s->img_x; ++i) { - *out++ = y[i]; - *out++ = 255; - } - } - } - } - stbi__cleanup_jpeg(z); - *out_x = z->s->img_x; - *out_y = z->s->img_y; - if (comp) - *comp = z->s->img_n >= 3 ? 3 : 1; // report original components, not output - return output; - } -} - -static void* stbi__jpeg_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri) { - unsigned char* result; - stbi__jpeg* j = (stbi__jpeg*)stbi__malloc(sizeof(stbi__jpeg)); - if (!j) - return stbi__errpuc("outofmem", "Out of memory"); - STBI_NOTUSED(ri); - j->s = s; - stbi__setup_jpeg(j); - result = load_jpeg_image(j, x, y, comp, req_comp); - STBI_FREE(j); - return result; -} - -static int stbi__jpeg_test(stbi__context* s) { - int r; - stbi__jpeg* j = (stbi__jpeg*)stbi__malloc(sizeof(stbi__jpeg)); - if (!j) - return stbi__err("outofmem", "Out of memory"); - j->s = s; - stbi__setup_jpeg(j); - r = stbi__decode_jpeg_header(j, STBI__SCAN_type); - stbi__rewind(s); - STBI_FREE(j); - return r; -} - -static int stbi__jpeg_info_raw(stbi__jpeg* j, int* x, int* y, int* comp) { - if (!stbi__decode_jpeg_header(j, STBI__SCAN_header)) { - stbi__rewind(j->s); - return 0; - } - if (x) - *x = j->s->img_x; - if (y) - *y = j->s->img_y; - if (comp) - *comp = j->s->img_n >= 3 ? 3 : 1; - return 1; -} - -static int stbi__jpeg_info(stbi__context* s, int* x, int* y, int* comp) { - int result; - stbi__jpeg* j = (stbi__jpeg*)(stbi__malloc(sizeof(stbi__jpeg))); - if (!j) - return stbi__err("outofmem", "Out of memory"); - j->s = s; - result = stbi__jpeg_info_raw(j, x, y, comp); - STBI_FREE(j); - return result; -} -#endif - -// public domain zlib decode v0.2 Sean Barrett 2006-11-18 -// simple implementation -// - all input must be provided in an upfront buffer -// - all output is written to a single output buffer (can malloc/realloc) -// performance -// - fast huffman - -#ifndef STBI_NO_ZLIB - -// fast-way is faster to check than jpeg huffman, but slow way is slower -#define STBI__ZFAST_BITS 9 // accelerate all cases in default tables -#define STBI__ZFAST_MASK ((1 << STBI__ZFAST_BITS) - 1) -#define STBI__ZNSYMS 288 // number of symbols in literal/length alphabet - -// zlib-style huffman encoding -// (jpegs packs from left, zlib from right, so can't share code) -typedef struct { - stbi__uint16 fast[1 << STBI__ZFAST_BITS]; - stbi__uint16 firstcode[16]; - int maxcode[17]; - stbi__uint16 firstsymbol[16]; - stbi_uc size[STBI__ZNSYMS]; - stbi__uint16 value[STBI__ZNSYMS]; -} stbi__zhuffman; - -stbi_inline static int stbi__bitreverse16(int n) { - n = ((n & 0xAAAA) >> 1) | ((n & 0x5555) << 1); - n = ((n & 0xCCCC) >> 2) | ((n & 0x3333) << 2); - n = ((n & 0xF0F0) >> 4) | ((n & 0x0F0F) << 4); - n = ((n & 0xFF00) >> 8) | ((n & 0x00FF) << 8); - return n; -} - -stbi_inline static int stbi__bit_reverse(int v, int bits) { - STBI_ASSERT(bits <= 16); - // to bit reverse n bits, reverse 16 and shift - // e.g. 11 bits, bit reverse and shift away 5 - return stbi__bitreverse16(v) >> (16 - bits); -} - -static int stbi__zbuild_huffman(stbi__zhuffman* z, const stbi_uc* sizelist, int num) { - int i, k = 0; - int code, next_code[16], sizes[17]; - - // DEFLATE spec for generating codes - memset(sizes, 0, sizeof(sizes)); - memset(z->fast, 0, sizeof(z->fast)); - for (i = 0; i < num; ++i) - ++sizes[sizelist[i]]; - sizes[0] = 0; - for (i = 1; i < 16; ++i) - if (sizes[i] > (1 << i)) - return stbi__err("bad sizes", "Corrupt PNG"); - code = 0; - for (i = 1; i < 16; ++i) { - next_code[i] = code; - z->firstcode[i] = (stbi__uint16)code; - z->firstsymbol[i] = (stbi__uint16)k; - code = (code + sizes[i]); - if (sizes[i]) - if (code - 1 >= (1 << i)) - return stbi__err("bad codelengths", "Corrupt PNG"); - z->maxcode[i] = code << (16 - i); // preshift for inner loop - code <<= 1; - k += sizes[i]; - } - z->maxcode[16] = 0x10000; // sentinel - for (i = 0; i < num; ++i) { - int s = sizelist[i]; - if (s) { - int c = next_code[s] - z->firstcode[s] + z->firstsymbol[s]; - stbi__uint16 fastv = (stbi__uint16)((s << 9) | i); - z->size[c] = (stbi_uc)s; - z->value[c] = (stbi__uint16)i; - if (s <= STBI__ZFAST_BITS) { - int j = stbi__bit_reverse(next_code[s], s); - while (j < (1 << STBI__ZFAST_BITS)) { - z->fast[j] = fastv; - j += (1 << s); - } - } - ++next_code[s]; - } - } - return 1; -} - -// zlib-from-memory implementation for PNG reading -// because PNG allows splitting the zlib stream arbitrarily, -// and it's annoying structurally to have PNG call ZLIB call PNG, -// we require PNG read all the IDATs and combine them into a single -// memory buffer - -typedef struct { - stbi_uc *zbuffer, *zbuffer_end; - int num_bits; - stbi__uint32 code_buffer; - - char* zout; - char* zout_start; - char* zout_end; - int z_expandable; - - stbi__zhuffman z_length, z_distance; -} stbi__zbuf; - -stbi_inline static int stbi__zeof(stbi__zbuf* z) { - return (z->zbuffer >= z->zbuffer_end); -} - -stbi_inline static stbi_uc stbi__zget8(stbi__zbuf* z) { - return stbi__zeof(z) ? 0 : *z->zbuffer++; -} - -static void stbi__fill_bits(stbi__zbuf* z) { - do { - if (z->code_buffer >= (1U << z->num_bits)) { - z->zbuffer = z->zbuffer_end; /* treat this as EOF so we fail. */ - return; - } - z->code_buffer |= (unsigned int)stbi__zget8(z) << z->num_bits; - z->num_bits += 8; - } while (z->num_bits <= 24); -} - -stbi_inline static unsigned int stbi__zreceive(stbi__zbuf* z, int n) { - unsigned int k; - if (z->num_bits < n) - stbi__fill_bits(z); - k = z->code_buffer & ((1 << n) - 1); - z->code_buffer >>= n; - z->num_bits -= n; - return k; -} - -static int stbi__zhuffman_decode_slowpath(stbi__zbuf* a, stbi__zhuffman* z) { - int b, s, k; - // not resolved by fast table, so compute it the slow way - // use jpeg approach, which requires MSbits at top - k = stbi__bit_reverse(a->code_buffer, 16); - for (s = STBI__ZFAST_BITS + 1;; ++s) - if (k < z->maxcode[s]) - break; - if (s >= 16) - return -1; // invalid code! - // code size is s, so: - b = (k >> (16 - s)) - z->firstcode[s] + z->firstsymbol[s]; - if (b >= STBI__ZNSYMS) - return -1; // some data was corrupt somewhere! - if (z->size[b] != s) - return -1; // was originally an assert, but report failure instead. - a->code_buffer >>= s; - a->num_bits -= s; - return z->value[b]; -} - -stbi_inline static int stbi__zhuffman_decode(stbi__zbuf* a, stbi__zhuffman* z) { - int b, s; - if (a->num_bits < 16) { - if (stbi__zeof(a)) { - return -1; /* report error for unexpected end of data. */ - } - stbi__fill_bits(a); - } - b = z->fast[a->code_buffer & STBI__ZFAST_MASK]; - if (b) { - s = b >> 9; - a->code_buffer >>= s; - a->num_bits -= s; - return b & 511; - } - return stbi__zhuffman_decode_slowpath(a, z); -} - -static int stbi__zexpand(stbi__zbuf* z, char* zout, int n) // need to make room for n bytes -{ - char* q; - unsigned int cur, limit, old_limit; - z->zout = zout; - if (!z->z_expandable) - return stbi__err("output buffer limit", "Corrupt PNG"); - cur = (unsigned int)(z->zout - z->zout_start); - limit = old_limit = (unsigned)(z->zout_end - z->zout_start); - if (UINT_MAX - cur < (unsigned)n) - return stbi__err("outofmem", "Out of memory"); - while (cur + n > limit) { - if (limit > UINT_MAX / 2) - return stbi__err("outofmem", "Out of memory"); - limit *= 2; - } - q = (char*)STBI_REALLOC_SIZED(z->zout_start, old_limit, limit); - STBI_NOTUSED(old_limit); - if (q == NULL) - return stbi__err("outofmem", "Out of memory"); - z->zout_start = q; - z->zout = q + cur; - z->zout_end = q + limit; - return 1; -} - -static const int stbi__zlength_base[31] = {3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, - 17, 19, 23, 27, 31, 35, 43, 51, 59, 67, 83, - 99, 115, 131, 163, 195, 227, 258, 0, 0}; - -static const int stbi__zlength_extra[31] = {0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, - 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0, 0, 0}; - -static const int stbi__zdist_base[32] = { - 1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193, - 257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145, 8193, 12289, 16385, 24577, 0, 0}; - -static const int stbi__zdist_extra[32] = {0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, - 6, 7, 7, 8, 8, 9, 9, 10, 10, 11, 11, 12, 12, 13, 13}; - -static int stbi__parse_huffman_block(stbi__zbuf* a) { - char* zout = a->zout; - for (;;) { - int z = stbi__zhuffman_decode(a, &a->z_length); - if (z < 256) { - if (z < 0) - return stbi__err("bad huffman code", "Corrupt PNG"); // error in huffman codes - if (zout >= a->zout_end) { - if (!stbi__zexpand(a, zout, 1)) - return 0; - zout = a->zout; - } - *zout++ = (char)z; - } else { - stbi_uc* p; - int len, dist; - if (z == 256) { - a->zout = zout; - return 1; - } - z -= 257; - len = stbi__zlength_base[z]; - if (stbi__zlength_extra[z]) - len += stbi__zreceive(a, stbi__zlength_extra[z]); - z = stbi__zhuffman_decode(a, &a->z_distance); - if (z < 0) - return stbi__err("bad huffman code", "Corrupt PNG"); - dist = stbi__zdist_base[z]; - if (stbi__zdist_extra[z]) - dist += stbi__zreceive(a, stbi__zdist_extra[z]); - if (zout - a->zout_start < dist) - return stbi__err("bad dist", "Corrupt PNG"); - if (zout + len > a->zout_end) { - if (!stbi__zexpand(a, zout, len)) - return 0; - zout = a->zout; - } - p = (stbi_uc*)(zout - dist); - if (dist == 1) { // run of one byte; common in images. - stbi_uc v = *p; - if (len) { - do - *zout++ = v; - while (--len); - } - } else { - if (len) { - do - *zout++ = *p++; - while (--len); - } - } - } - } -} - -static int stbi__compute_huffman_codes(stbi__zbuf* a) { - static const stbi_uc length_dezigzag[19] = {16, 17, 18, 0, 8, 7, 9, 6, 10, 5, - 11, 4, 12, 3, 13, 2, 14, 1, 15}; - stbi__zhuffman z_codelength; - stbi_uc lencodes[286 + 32 + 137]; // padding for maximum single op - stbi_uc codelength_sizes[19]; - int i, n; - - int hlit = stbi__zreceive(a, 5) + 257; - int hdist = stbi__zreceive(a, 5) + 1; - int hclen = stbi__zreceive(a, 4) + 4; - int ntot = hlit + hdist; - - memset(codelength_sizes, 0, sizeof(codelength_sizes)); - for (i = 0; i < hclen; ++i) { - int s = stbi__zreceive(a, 3); - codelength_sizes[length_dezigzag[i]] = (stbi_uc)s; - } - if (!stbi__zbuild_huffman(&z_codelength, codelength_sizes, 19)) - return 0; - - n = 0; - while (n < ntot) { - int c = stbi__zhuffman_decode(a, &z_codelength); - if (c < 0 || c >= 19) - return stbi__err("bad codelengths", "Corrupt PNG"); - if (c < 16) - lencodes[n++] = (stbi_uc)c; - else { - stbi_uc fill = 0; - if (c == 16) { - c = stbi__zreceive(a, 2) + 3; - if (n == 0) - return stbi__err("bad codelengths", "Corrupt PNG"); - fill = lencodes[n - 1]; - } else if (c == 17) { - c = stbi__zreceive(a, 3) + 3; - } else if (c == 18) { - c = stbi__zreceive(a, 7) + 11; - } else { - return stbi__err("bad codelengths", "Corrupt PNG"); - } - if (ntot - n < c) - return stbi__err("bad codelengths", "Corrupt PNG"); - memset(lencodes + n, fill, c); - n += c; - } - } - if (n != ntot) - return stbi__err("bad codelengths", "Corrupt PNG"); - if (!stbi__zbuild_huffman(&a->z_length, lencodes, hlit)) - return 0; - if (!stbi__zbuild_huffman(&a->z_distance, lencodes + hlit, hdist)) - return 0; - return 1; -} - -static int stbi__parse_uncompressed_block(stbi__zbuf* a) { - stbi_uc header[4]; - int len, nlen, k; - if (a->num_bits & 7) - stbi__zreceive(a, a->num_bits & 7); // discard - // drain the bit-packed data into header - k = 0; - while (a->num_bits > 0) { - header[k++] = (stbi_uc)(a->code_buffer & 255); // suppress MSVC run-time check - a->code_buffer >>= 8; - a->num_bits -= 8; - } - if (a->num_bits < 0) - return stbi__err("zlib corrupt", "Corrupt PNG"); - // now fill header the normal way - while (k < 4) - header[k++] = stbi__zget8(a); - len = header[1] * 256 + header[0]; - nlen = header[3] * 256 + header[2]; - if (nlen != (len ^ 0xffff)) - return stbi__err("zlib corrupt", "Corrupt PNG"); - if (a->zbuffer + len > a->zbuffer_end) - return stbi__err("read past buffer", "Corrupt PNG"); - if (a->zout + len > a->zout_end) - if (!stbi__zexpand(a, a->zout, len)) - return 0; - memcpy(a->zout, a->zbuffer, len); - a->zbuffer += len; - a->zout += len; - return 1; -} - -static int stbi__parse_zlib_header(stbi__zbuf* a) { - int cmf = stbi__zget8(a); - int cm = cmf & 15; - /* int cinfo = cmf >> 4; */ - int flg = stbi__zget8(a); - if (stbi__zeof(a)) - return stbi__err("bad zlib header", "Corrupt PNG"); // zlib spec - if ((cmf * 256 + flg) % 31 != 0) - return stbi__err("bad zlib header", "Corrupt PNG"); // zlib spec - if (flg & 32) - return stbi__err("no preset dict", "Corrupt PNG"); // preset dictionary not allowed in png - if (cm != 8) - return stbi__err("bad compression", "Corrupt PNG"); // DEFLATE required for png - // window = 1 << (8 + cinfo)... but who cares, we fully buffer output - return 1; -} - -static const stbi_uc stbi__zdefault_length[STBI__ZNSYMS] = { - 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, - 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, - 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, - 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, - 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, - 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, - 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, - 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, - 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 8, 8, 8, 8, 8, 8, 8, 8}; -static const stbi_uc stbi__zdefault_distance[32] = {5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, - 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5}; -/* -Init algorithm: -{ - int i; // use <= to match clearly with spec - for (i=0; i <= 143; ++i) stbi__zdefault_length[i] = 8; - for ( ; i <= 255; ++i) stbi__zdefault_length[i] = 9; - for ( ; i <= 279; ++i) stbi__zdefault_length[i] = 7; - for ( ; i <= 287; ++i) stbi__zdefault_length[i] = 8; - - for (i=0; i <= 31; ++i) stbi__zdefault_distance[i] = 5; -} -*/ - -static int stbi__parse_zlib(stbi__zbuf* a, int parse_header) { - int final, type; - if (parse_header) - if (!stbi__parse_zlib_header(a)) - return 0; - a->num_bits = 0; - a->code_buffer = 0; - do { - final = stbi__zreceive(a, 1); - type = stbi__zreceive(a, 2); - if (type == 0) { - if (!stbi__parse_uncompressed_block(a)) - return 0; - } else if (type == 3) { - return 0; - } else { - if (type == 1) { - // use fixed code lengths - if (!stbi__zbuild_huffman(&a->z_length, stbi__zdefault_length, STBI__ZNSYMS)) - return 0; - if (!stbi__zbuild_huffman(&a->z_distance, stbi__zdefault_distance, 32)) - return 0; - } else { - if (!stbi__compute_huffman_codes(a)) - return 0; - } - if (!stbi__parse_huffman_block(a)) - return 0; - } - } while (!final); - return 1; -} - -static int stbi__do_zlib(stbi__zbuf* a, char* obuf, int olen, int exp, int parse_header) { - a->zout_start = obuf; - a->zout = obuf; - a->zout_end = obuf + olen; - a->z_expandable = exp; - - return stbi__parse_zlib(a, parse_header); -} - -STBIDEF char* stbi_zlib_decode_malloc_guesssize(const char* buffer, - int len, - int initial_size, - int* outlen) { - stbi__zbuf a; - char* p = (char*)stbi__malloc(initial_size); - if (p == NULL) - return NULL; - a.zbuffer = (stbi_uc*)buffer; - a.zbuffer_end = (stbi_uc*)buffer + len; - if (stbi__do_zlib(&a, p, initial_size, 1, 1)) { - if (outlen) - *outlen = (int)(a.zout - a.zout_start); - return a.zout_start; - } else { - STBI_FREE(a.zout_start); - return NULL; - } -} - -STBIDEF char* stbi_zlib_decode_malloc(char const* buffer, int len, int* outlen) { - return stbi_zlib_decode_malloc_guesssize(buffer, len, 16384, outlen); -} - -STBIDEF char* stbi_zlib_decode_malloc_guesssize_headerflag(const char* buffer, - int len, - int initial_size, - int* outlen, - int parse_header) { - stbi__zbuf a; - char* p = (char*)stbi__malloc(initial_size); - if (p == NULL) - return NULL; - a.zbuffer = (stbi_uc*)buffer; - a.zbuffer_end = (stbi_uc*)buffer + len; - if (stbi__do_zlib(&a, p, initial_size, 1, parse_header)) { - if (outlen) - *outlen = (int)(a.zout - a.zout_start); - return a.zout_start; - } else { - STBI_FREE(a.zout_start); - return NULL; - } -} - -STBIDEF int stbi_zlib_decode_buffer(char* obuffer, int olen, char const* ibuffer, int ilen) { - stbi__zbuf a; - a.zbuffer = (stbi_uc*)ibuffer; - a.zbuffer_end = (stbi_uc*)ibuffer + ilen; - if (stbi__do_zlib(&a, obuffer, olen, 0, 1)) - return (int)(a.zout - a.zout_start); - else - return -1; -} - -STBIDEF char* stbi_zlib_decode_noheader_malloc(char const* buffer, int len, int* outlen) { - stbi__zbuf a; - char* p = (char*)stbi__malloc(16384); - if (p == NULL) - return NULL; - a.zbuffer = (stbi_uc*)buffer; - a.zbuffer_end = (stbi_uc*)buffer + len; - if (stbi__do_zlib(&a, p, 16384, 1, 0)) { - if (outlen) - *outlen = (int)(a.zout - a.zout_start); - return a.zout_start; - } else { - STBI_FREE(a.zout_start); - return NULL; - } -} - -STBIDEF int stbi_zlib_decode_noheader_buffer(char* obuffer, - int olen, - const char* ibuffer, - int ilen) { - stbi__zbuf a; - a.zbuffer = (stbi_uc*)ibuffer; - a.zbuffer_end = (stbi_uc*)ibuffer + ilen; - if (stbi__do_zlib(&a, obuffer, olen, 0, 0)) - return (int)(a.zout - a.zout_start); - else - return -1; -} -#endif - -// public domain "baseline" PNG decoder v0.10 Sean Barrett 2006-11-18 -// simple implementation -// - only 8-bit samples -// - no CRC checking -// - allocates lots of intermediate memory -// - avoids problem of streaming data between subsystems -// - avoids explicit window management -// performance -// - uses stb_zlib, a PD zlib implementation with fast huffman decoding - -#ifndef STBI_NO_PNG -typedef struct { - stbi__uint32 length; - stbi__uint32 type; -} stbi__pngchunk; - -static stbi__pngchunk stbi__get_chunk_header(stbi__context* s) { - stbi__pngchunk c; - c.length = stbi__get32be(s); - c.type = stbi__get32be(s); - return c; -} - -static int stbi__check_png_header(stbi__context* s) { - static const stbi_uc png_sig[8] = {137, 80, 78, 71, 13, 10, 26, 10}; - int i; - for (i = 0; i < 8; ++i) - if (stbi__get8(s) != png_sig[i]) - return stbi__err("bad png sig", "Not a PNG"); - return 1; -} - -typedef struct { - stbi__context* s; - stbi_uc *idata, *expanded, *out; - int depth; -} stbi__png; - -enum { - STBI__F_none = 0, - STBI__F_sub = 1, - STBI__F_up = 2, - STBI__F_avg = 3, - STBI__F_paeth = 4, - // synthetic filters used for first scanline to avoid needing a dummy row of 0s - STBI__F_avg_first, - STBI__F_paeth_first -}; - -static stbi_uc first_row_filter[5] = {STBI__F_none, STBI__F_sub, STBI__F_none, STBI__F_avg_first, - STBI__F_paeth_first}; - -static int stbi__paeth(int a, int b, int c) { - int p = a + b - c; - int pa = abs(p - a); - int pb = abs(p - b); - int pc = abs(p - c); - if (pa <= pb && pa <= pc) - return a; - if (pb <= pc) - return b; - return c; -} - -static const stbi_uc stbi__depth_scale_table[9] = {0, 0xff, 0x55, 0, 0x11, 0, 0, 0, 0x01}; - -// create the png data from post-deflated data -static int stbi__create_png_image_raw(stbi__png* a, - stbi_uc* raw, - stbi__uint32 raw_len, - int out_n, - stbi__uint32 x, - stbi__uint32 y, - int depth, - int color) { - int bytes = (depth == 16 ? 2 : 1); - stbi__context* s = a->s; - stbi__uint32 i, j, stride = x * out_n * bytes; - stbi__uint32 img_len, img_width_bytes; - int k; - int img_n = s->img_n; // copy it into a local for later - - int output_bytes = out_n * bytes; - int filter_bytes = img_n * bytes; - int width = x; - - STBI_ASSERT(out_n == s->img_n || out_n == s->img_n + 1); - a->out = - (stbi_uc*)stbi__malloc_mad3(x, y, output_bytes, 0); // extra bytes to write off the end into - if (!a->out) - return stbi__err("outofmem", "Out of memory"); - - if (!stbi__mad3sizes_valid(img_n, x, depth, 7)) - return stbi__err("too large", "Corrupt PNG"); - img_width_bytes = (((img_n * x * depth) + 7) >> 3); - img_len = (img_width_bytes + 1) * y; - - // we used to check for exact match between raw_len and img_len on non-interlaced PNGs, - // but issue #276 reported a PNG in the wild that had extra data at the end (all zeros), - // so just check for raw_len < img_len always. - if (raw_len < img_len) - return stbi__err("not enough pixels", "Corrupt PNG"); - - for (j = 0; j < y; ++j) { - stbi_uc* cur = a->out + stride * j; - stbi_uc* prior; - int filter = *raw++; - - if (filter > 4) - return stbi__err("invalid filter", "Corrupt PNG"); - - if (depth < 8) { - if (img_width_bytes > x) - return stbi__err("invalid width", "Corrupt PNG"); - cur += x * out_n - img_width_bytes; // store output to the rightmost img_len bytes, so we can - // decode in place - filter_bytes = 1; - width = img_width_bytes; - } - prior = cur - stride; // bugfix: need to compute this after 'cur +=' computation above - - // if first row, use special filter that doesn't sample previous row - if (j == 0) - filter = first_row_filter[filter]; - - // handle first byte explicitly - for (k = 0; k < filter_bytes; ++k) { - switch (filter) { - case STBI__F_none: - cur[k] = raw[k]; - break; - case STBI__F_sub: - cur[k] = raw[k]; - break; - case STBI__F_up: - cur[k] = STBI__BYTECAST(raw[k] + prior[k]); - break; - case STBI__F_avg: - cur[k] = STBI__BYTECAST(raw[k] + (prior[k] >> 1)); - break; - case STBI__F_paeth: - cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(0, prior[k], 0)); - break; - case STBI__F_avg_first: - cur[k] = raw[k]; - break; - case STBI__F_paeth_first: - cur[k] = raw[k]; - break; - } - } - - if (depth == 8) { - if (img_n != out_n) - cur[img_n] = 255; // first pixel - raw += img_n; - cur += out_n; - prior += out_n; - } else if (depth == 16) { - if (img_n != out_n) { - cur[filter_bytes] = 255; // first pixel top byte - cur[filter_bytes + 1] = 255; // first pixel bottom byte - } - raw += filter_bytes; - cur += output_bytes; - prior += output_bytes; - } else { - raw += 1; - cur += 1; - prior += 1; - } - - // this is a little gross, so that we don't switch per-pixel or per-component - if (depth < 8 || img_n == out_n) { - int nk = (width - 1) * filter_bytes; -#define STBI__CASE(f) \ - case f: \ - for (k = 0; k < nk; ++k) - switch (filter) { - // "none" filter turns into a memcpy here; make that explicit. - case STBI__F_none: - memcpy(cur, raw, nk); - break; - STBI__CASE(STBI__F_sub) { cur[k] = STBI__BYTECAST(raw[k] + cur[k - filter_bytes]); } - break; - STBI__CASE(STBI__F_up) { cur[k] = STBI__BYTECAST(raw[k] + prior[k]); } - break; - STBI__CASE(STBI__F_avg) { - cur[k] = STBI__BYTECAST(raw[k] + ((prior[k] + cur[k - filter_bytes]) >> 1)); - } - break; - STBI__CASE(STBI__F_paeth) { - cur[k] = STBI__BYTECAST( - raw[k] + stbi__paeth(cur[k - filter_bytes], prior[k], prior[k - filter_bytes])); - } - break; - STBI__CASE(STBI__F_avg_first) { - cur[k] = STBI__BYTECAST(raw[k] + (cur[k - filter_bytes] >> 1)); - } - break; - STBI__CASE(STBI__F_paeth_first) { - cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k - filter_bytes], 0, 0)); - } - break; - } -#undef STBI__CASE - raw += nk; - } else { - STBI_ASSERT(img_n + 1 == out_n); -#define STBI__CASE(f) \ - case f: \ - for (i = x - 1; i >= 1; --i, cur[filter_bytes] = 255, raw += filter_bytes, \ - cur += output_bytes, prior += output_bytes) \ - for (k = 0; k < filter_bytes; ++k) - switch (filter) { - STBI__CASE(STBI__F_none) { cur[k] = raw[k]; } - break; - STBI__CASE(STBI__F_sub) { cur[k] = STBI__BYTECAST(raw[k] + cur[k - output_bytes]); } - break; - STBI__CASE(STBI__F_up) { cur[k] = STBI__BYTECAST(raw[k] + prior[k]); } - break; - STBI__CASE(STBI__F_avg) { - cur[k] = STBI__BYTECAST(raw[k] + ((prior[k] + cur[k - output_bytes]) >> 1)); - } - break; - STBI__CASE(STBI__F_paeth) { - cur[k] = STBI__BYTECAST( - raw[k] + stbi__paeth(cur[k - output_bytes], prior[k], prior[k - output_bytes])); - } - break; - STBI__CASE(STBI__F_avg_first) { - cur[k] = STBI__BYTECAST(raw[k] + (cur[k - output_bytes] >> 1)); - } - break; - STBI__CASE(STBI__F_paeth_first) { - cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k - output_bytes], 0, 0)); - } - break; - } -#undef STBI__CASE - - // the loop above sets the high byte of the pixels' alpha, but for - // 16 bit png files we also need the low byte set. we'll do that here. - if (depth == 16) { - cur = a->out + stride * j; // start at the beginning of the row again - for (i = 0; i < x; ++i, cur += output_bytes) { - cur[filter_bytes + 1] = 255; - } - } - } - } - - // we make a separate pass to expand bits to pixels; for performance, - // this could run two scanlines behind the above code, so it won't - // intefere with filtering but will still be in the cache. - if (depth < 8) { - for (j = 0; j < y; ++j) { - stbi_uc* cur = a->out + stride * j; - stbi_uc* in = a->out + stride * j + x * out_n - img_width_bytes; - // unpack 1/2/4-bit into a 8-bit buffer. allows us to keep the common 8-bit path optimal at - // minimal cost for 1/2/4-bit png guarante byte alignment, if width is not multiple of 8/4/2 - // we'll decode dummy trailing data that will be skipped in the later loop - stbi_uc scale = (color == 0) ? stbi__depth_scale_table[depth] - : 1; // scale grayscale values to 0..255 range - - // note that the final byte might overshoot and write more data than desired. - // we can allocate enough data that this never writes out of memory, but it - // could also overwrite the next scanline. can it overwrite non-empty data - // on the next scanline? yes, consider 1-pixel-wide scanlines with 1-bit-per-pixel. - // so we need to explicitly clamp the final ones - - if (depth == 4) { - for (k = x * img_n; k >= 2; k -= 2, ++in) { - *cur++ = scale * ((*in >> 4)); - *cur++ = scale * ((*in) & 0x0f); - } - if (k > 0) - *cur++ = scale * ((*in >> 4)); - } else if (depth == 2) { - for (k = x * img_n; k >= 4; k -= 4, ++in) { - *cur++ = scale * ((*in >> 6)); - *cur++ = scale * ((*in >> 4) & 0x03); - *cur++ = scale * ((*in >> 2) & 0x03); - *cur++ = scale * ((*in) & 0x03); - } - if (k > 0) - *cur++ = scale * ((*in >> 6)); - if (k > 1) - *cur++ = scale * ((*in >> 4) & 0x03); - if (k > 2) - *cur++ = scale * ((*in >> 2) & 0x03); - } else if (depth == 1) { - for (k = x * img_n; k >= 8; k -= 8, ++in) { - *cur++ = scale * ((*in >> 7)); - *cur++ = scale * ((*in >> 6) & 0x01); - *cur++ = scale * ((*in >> 5) & 0x01); - *cur++ = scale * ((*in >> 4) & 0x01); - *cur++ = scale * ((*in >> 3) & 0x01); - *cur++ = scale * ((*in >> 2) & 0x01); - *cur++ = scale * ((*in >> 1) & 0x01); - *cur++ = scale * ((*in) & 0x01); - } - if (k > 0) - *cur++ = scale * ((*in >> 7)); - if (k > 1) - *cur++ = scale * ((*in >> 6) & 0x01); - if (k > 2) - *cur++ = scale * ((*in >> 5) & 0x01); - if (k > 3) - *cur++ = scale * ((*in >> 4) & 0x01); - if (k > 4) - *cur++ = scale * ((*in >> 3) & 0x01); - if (k > 5) - *cur++ = scale * ((*in >> 2) & 0x01); - if (k > 6) - *cur++ = scale * ((*in >> 1) & 0x01); - } - if (img_n != out_n) { - int q; - // insert alpha = 255 - cur = a->out + stride * j; - if (img_n == 1) { - for (q = x - 1; q >= 0; --q) { - cur[q * 2 + 1] = 255; - cur[q * 2 + 0] = cur[q]; - } - } else { - STBI_ASSERT(img_n == 3); - for (q = x - 1; q >= 0; --q) { - cur[q * 4 + 3] = 255; - cur[q * 4 + 2] = cur[q * 3 + 2]; - cur[q * 4 + 1] = cur[q * 3 + 1]; - cur[q * 4 + 0] = cur[q * 3 + 0]; - } - } - } - } - } else if (depth == 16) { - // force the image data from big-endian to platform-native. - // this is done in a separate pass due to the decoding relying - // on the data being untouched, but could probably be done - // per-line during decode if care is taken. - stbi_uc* cur = a->out; - stbi__uint16* cur16 = (stbi__uint16*)cur; - - for (i = 0; i < x * y * out_n; ++i, cur16++, cur += 2) { - *cur16 = (cur[0] << 8) | cur[1]; - } - } - - return 1; -} - -static int stbi__create_png_image(stbi__png* a, - stbi_uc* image_data, - stbi__uint32 image_data_len, - int out_n, - int depth, - int color, - int interlaced) { - int bytes = (depth == 16 ? 2 : 1); - int out_bytes = out_n * bytes; - stbi_uc* final; - int p; - if (!interlaced) - return stbi__create_png_image_raw(a, image_data, image_data_len, out_n, a->s->img_x, - a->s->img_y, depth, color); - - // de-interlacing - final = (stbi_uc*)stbi__malloc_mad3(a->s->img_x, a->s->img_y, out_bytes, 0); - if (!final) - return stbi__err("outofmem", "Out of memory"); - for (p = 0; p < 7; ++p) { - int xorig[] = {0, 4, 0, 2, 0, 1, 0}; - int yorig[] = {0, 0, 4, 0, 2, 0, 1}; - int xspc[] = {8, 8, 4, 4, 2, 2, 1}; - int yspc[] = {8, 8, 8, 4, 4, 2, 2}; - int i, j, x, y; - // pass1_x[4] = 0, pass1_x[5] = 1, pass1_x[12] = 1 - x = (a->s->img_x - xorig[p] + xspc[p] - 1) / xspc[p]; - y = (a->s->img_y - yorig[p] + yspc[p] - 1) / yspc[p]; - if (x && y) { - stbi__uint32 img_len = ((((a->s->img_n * x * depth) + 7) >> 3) + 1) * y; - if (!stbi__create_png_image_raw(a, image_data, image_data_len, out_n, x, y, depth, color)) { - STBI_FREE(final); - return 0; - } - for (j = 0; j < y; ++j) { - for (i = 0; i < x; ++i) { - int out_y = j * yspc[p] + yorig[p]; - int out_x = i * xspc[p] + xorig[p]; - memcpy(final + out_y * a->s->img_x * out_bytes + out_x * out_bytes, - a->out + (j * x + i) * out_bytes, out_bytes); - } - } - STBI_FREE(a->out); - image_data += img_len; - image_data_len -= img_len; - } - } - a->out = final; - - return 1; -} - -static int stbi__compute_transparency(stbi__png* z, stbi_uc tc[3], int out_n) { - stbi__context* s = z->s; - stbi__uint32 i, pixel_count = s->img_x * s->img_y; - stbi_uc* p = z->out; - - // compute color-based transparency, assuming we've - // already got 255 as the alpha value in the output - STBI_ASSERT(out_n == 2 || out_n == 4); - - if (out_n == 2) { - for (i = 0; i < pixel_count; ++i) { - p[1] = (p[0] == tc[0] ? 0 : 255); - p += 2; - } - } else { - for (i = 0; i < pixel_count; ++i) { - if (p[0] == tc[0] && p[1] == tc[1] && p[2] == tc[2]) - p[3] = 0; - p += 4; - } - } - return 1; -} - -static int stbi__compute_transparency16(stbi__png* z, stbi__uint16 tc[3], int out_n) { - stbi__context* s = z->s; - stbi__uint32 i, pixel_count = s->img_x * s->img_y; - stbi__uint16* p = (stbi__uint16*)z->out; - - // compute color-based transparency, assuming we've - // already got 65535 as the alpha value in the output - STBI_ASSERT(out_n == 2 || out_n == 4); - - if (out_n == 2) { - for (i = 0; i < pixel_count; ++i) { - p[1] = (p[0] == tc[0] ? 0 : 65535); - p += 2; - } - } else { - for (i = 0; i < pixel_count; ++i) { - if (p[0] == tc[0] && p[1] == tc[1] && p[2] == tc[2]) - p[3] = 0; - p += 4; - } - } - return 1; -} - -static int stbi__expand_png_palette(stbi__png* a, stbi_uc* palette, int len, int pal_img_n) { - stbi__uint32 i, pixel_count = a->s->img_x * a->s->img_y; - stbi_uc *p, *temp_out, *orig = a->out; - - p = (stbi_uc*)stbi__malloc_mad2(pixel_count, pal_img_n, 0); - if (p == NULL) - return stbi__err("outofmem", "Out of memory"); - - // between here and free(out) below, exitting would leak - temp_out = p; - - if (pal_img_n == 3) { - for (i = 0; i < pixel_count; ++i) { - int n = orig[i] * 4; - p[0] = palette[n]; - p[1] = palette[n + 1]; - p[2] = palette[n + 2]; - p += 3; - } - } else { - for (i = 0; i < pixel_count; ++i) { - int n = orig[i] * 4; - p[0] = palette[n]; - p[1] = palette[n + 1]; - p[2] = palette[n + 2]; - p[3] = palette[n + 3]; - p += 4; - } - } - STBI_FREE(a->out); - a->out = temp_out; - - STBI_NOTUSED(len); - - return 1; -} - -static int stbi__unpremultiply_on_load_global = 0; -static int stbi__de_iphone_flag_global = 0; - -STBIDEF void stbi_set_unpremultiply_on_load(int flag_true_if_should_unpremultiply) { - stbi__unpremultiply_on_load_global = flag_true_if_should_unpremultiply; -} - -STBIDEF void stbi_convert_iphone_png_to_rgb(int flag_true_if_should_convert) { - stbi__de_iphone_flag_global = flag_true_if_should_convert; -} - -#ifndef STBI_THREAD_LOCAL -#define stbi__unpremultiply_on_load stbi__unpremultiply_on_load_global -#define stbi__de_iphone_flag stbi__de_iphone_flag_global -#else -static STBI_THREAD_LOCAL int stbi__unpremultiply_on_load_local, stbi__unpremultiply_on_load_set; -static STBI_THREAD_LOCAL int stbi__de_iphone_flag_local, stbi__de_iphone_flag_set; - -STBIDEF void stbi__unpremultiply_on_load_thread(int flag_true_if_should_unpremultiply) { - stbi__unpremultiply_on_load_local = flag_true_if_should_unpremultiply; - stbi__unpremultiply_on_load_set = 1; -} - -STBIDEF void stbi_convert_iphone_png_to_rgb_thread(int flag_true_if_should_convert) { - stbi__de_iphone_flag_local = flag_true_if_should_convert; - stbi__de_iphone_flag_set = 1; -} - -#define stbi__unpremultiply_on_load \ - (stbi__unpremultiply_on_load_set ? stbi__unpremultiply_on_load_local \ - : stbi__unpremultiply_on_load_global) -#define stbi__de_iphone_flag \ - (stbi__de_iphone_flag_set ? stbi__de_iphone_flag_local : stbi__de_iphone_flag_global) -#endif // STBI_THREAD_LOCAL - -static void stbi__de_iphone(stbi__png* z) { - stbi__context* s = z->s; - stbi__uint32 i, pixel_count = s->img_x * s->img_y; - stbi_uc* p = z->out; - - if (s->img_out_n == 3) { // convert bgr to rgb - for (i = 0; i < pixel_count; ++i) { - stbi_uc t = p[0]; - p[0] = p[2]; - p[2] = t; - p += 3; - } - } else { - STBI_ASSERT(s->img_out_n == 4); - if (stbi__unpremultiply_on_load) { - // convert bgr to rgb and unpremultiply - for (i = 0; i < pixel_count; ++i) { - stbi_uc a = p[3]; - stbi_uc t = p[0]; - if (a) { - stbi_uc half = a / 2; - p[0] = (p[2] * 255 + half) / a; - p[1] = (p[1] * 255 + half) / a; - p[2] = (t * 255 + half) / a; - } else { - p[0] = p[2]; - p[2] = t; - } - p += 4; - } - } else { - // convert bgr to rgb - for (i = 0; i < pixel_count; ++i) { - stbi_uc t = p[0]; - p[0] = p[2]; - p[2] = t; - p += 4; - } - } - } -} - -#define STBI__PNG_TYPE(a, b, c, d) \ - (((unsigned)(a) << 24) + ((unsigned)(b) << 16) + ((unsigned)(c) << 8) + (unsigned)(d)) - -static int stbi__parse_png_file(stbi__png* z, int scan, int req_comp) { - stbi_uc palette[1024], pal_img_n = 0; - stbi_uc has_trans = 0, tc[3] = {0}; - stbi__uint16 tc16[3]; - stbi__uint32 ioff = 0, idata_limit = 0, i, pal_len = 0; - int first = 1, k, interlace = 0, color = 0, is_iphone = 0; - stbi__context* s = z->s; - - z->expanded = NULL; - z->idata = NULL; - z->out = NULL; - - if (!stbi__check_png_header(s)) - return 0; - - if (scan == STBI__SCAN_type) - return 1; - - for (;;) { - stbi__pngchunk c = stbi__get_chunk_header(s); - switch (c.type) { - case STBI__PNG_TYPE('C', 'g', 'B', 'I'): - is_iphone = 1; - stbi__skip(s, c.length); - break; - case STBI__PNG_TYPE('I', 'H', 'D', 'R'): { - int comp, filter; - if (!first) - return stbi__err("multiple IHDR", "Corrupt PNG"); - first = 0; - if (c.length != 13) - return stbi__err("bad IHDR len", "Corrupt PNG"); - s->img_x = stbi__get32be(s); - s->img_y = stbi__get32be(s); - if (s->img_y > STBI_MAX_DIMENSIONS) - return stbi__err("too large", "Very large image (corrupt?)"); - if (s->img_x > STBI_MAX_DIMENSIONS) - return stbi__err("too large", "Very large image (corrupt?)"); - z->depth = stbi__get8(s); - if (z->depth != 1 && z->depth != 2 && z->depth != 4 && z->depth != 8 && z->depth != 16) - return stbi__err("1/2/4/8/16-bit only", "PNG not supported: 1/2/4/8/16-bit only"); - color = stbi__get8(s); - if (color > 6) - return stbi__err("bad ctype", "Corrupt PNG"); - if (color == 3 && z->depth == 16) - return stbi__err("bad ctype", "Corrupt PNG"); - if (color == 3) - pal_img_n = 3; - else if (color & 1) - return stbi__err("bad ctype", "Corrupt PNG"); - comp = stbi__get8(s); - if (comp) - return stbi__err("bad comp method", "Corrupt PNG"); - filter = stbi__get8(s); - if (filter) - return stbi__err("bad filter method", "Corrupt PNG"); - interlace = stbi__get8(s); - if (interlace > 1) - return stbi__err("bad interlace method", "Corrupt PNG"); - if (!s->img_x || !s->img_y) - return stbi__err("0-pixel image", "Corrupt PNG"); - if (!pal_img_n) { - s->img_n = (color & 2 ? 3 : 1) + (color & 4 ? 1 : 0); - if ((1 << 30) / s->img_x / s->img_n < s->img_y) - return stbi__err("too large", "Image too large to decode"); - if (scan == STBI__SCAN_header) - return 1; - } else { - // if paletted, then pal_n is our final components, and - // img_n is # components to decompress/filter. - s->img_n = 1; - if ((1 << 30) / s->img_x / 4 < s->img_y) - return stbi__err("too large", "Corrupt PNG"); - // if SCAN_header, have to scan to see if we have a tRNS - } - break; - } - - case STBI__PNG_TYPE('P', 'L', 'T', 'E'): { - if (first) - return stbi__err("first not IHDR", "Corrupt PNG"); - if (c.length > 256 * 3) - return stbi__err("invalid PLTE", "Corrupt PNG"); - pal_len = c.length / 3; - if (pal_len * 3 != c.length) - return stbi__err("invalid PLTE", "Corrupt PNG"); - for (i = 0; i < pal_len; ++i) { - palette[i * 4 + 0] = stbi__get8(s); - palette[i * 4 + 1] = stbi__get8(s); - palette[i * 4 + 2] = stbi__get8(s); - palette[i * 4 + 3] = 255; - } - break; - } - - case STBI__PNG_TYPE('t', 'R', 'N', 'S'): { - if (first) - return stbi__err("first not IHDR", "Corrupt PNG"); - if (z->idata) - return stbi__err("tRNS after IDAT", "Corrupt PNG"); - if (pal_img_n) { - if (scan == STBI__SCAN_header) { - s->img_n = 4; - return 1; - } - if (pal_len == 0) - return stbi__err("tRNS before PLTE", "Corrupt PNG"); - if (c.length > pal_len) - return stbi__err("bad tRNS len", "Corrupt PNG"); - pal_img_n = 4; - for (i = 0; i < c.length; ++i) - palette[i * 4 + 3] = stbi__get8(s); - } else { - if (!(s->img_n & 1)) - return stbi__err("tRNS with alpha", "Corrupt PNG"); - if (c.length != (stbi__uint32)s->img_n * 2) - return stbi__err("bad tRNS len", "Corrupt PNG"); - has_trans = 1; - if (z->depth == 16) { - for (k = 0; k < s->img_n; ++k) - tc16[k] = (stbi__uint16)stbi__get16be(s); // copy the values as-is - } else { - for (k = 0; k < s->img_n; ++k) - tc[k] = (stbi_uc)(stbi__get16be(s) & 255) * - stbi__depth_scale_table[z->depth]; // non 8-bit images will be larger - } - } - break; - } - - case STBI__PNG_TYPE('I', 'D', 'A', 'T'): { - if (first) - return stbi__err("first not IHDR", "Corrupt PNG"); - if (pal_img_n && !pal_len) - return stbi__err("no PLTE", "Corrupt PNG"); - if (scan == STBI__SCAN_header) { - s->img_n = pal_img_n; - return 1; - } - if ((int)(ioff + c.length) < (int)ioff) - return 0; - if (ioff + c.length > idata_limit) { - stbi__uint32 idata_limit_old = idata_limit; - stbi_uc* p; - if (idata_limit == 0) - idata_limit = c.length > 4096 ? c.length : 4096; - while (ioff + c.length > idata_limit) - idata_limit *= 2; - STBI_NOTUSED(idata_limit_old); - p = (stbi_uc*)STBI_REALLOC_SIZED(z->idata, idata_limit_old, idata_limit); - if (p == NULL) - return stbi__err("outofmem", "Out of memory"); - z->idata = p; - } - if (!stbi__getn(s, z->idata + ioff, c.length)) - return stbi__err("outofdata", "Corrupt PNG"); - ioff += c.length; - break; - } - - case STBI__PNG_TYPE('I', 'E', 'N', 'D'): { - stbi__uint32 raw_len, bpl; - if (first) - return stbi__err("first not IHDR", "Corrupt PNG"); - if (scan != STBI__SCAN_load) - return 1; - if (z->idata == NULL) - return stbi__err("no IDAT", "Corrupt PNG"); - // initial guess for decoded data size to avoid unnecessary reallocs - bpl = (s->img_x * z->depth + 7) / 8; // bytes per line, per component - raw_len = bpl * s->img_y * s->img_n /* pixels */ + s->img_y /* filter mode per row */; - z->expanded = (stbi_uc*)stbi_zlib_decode_malloc_guesssize_headerflag( - (char*)z->idata, ioff, raw_len, (int*)&raw_len, !is_iphone); - if (z->expanded == NULL) - return 0; // zlib should set error - STBI_FREE(z->idata); - z->idata = NULL; - if ((req_comp == s->img_n + 1 && req_comp != 3 && !pal_img_n) || has_trans) - s->img_out_n = s->img_n + 1; - else - s->img_out_n = s->img_n; - if (!stbi__create_png_image(z, z->expanded, raw_len, s->img_out_n, z->depth, color, - interlace)) - return 0; - if (has_trans) { - if (z->depth == 16) { - if (!stbi__compute_transparency16(z, tc16, s->img_out_n)) - return 0; - } else { - if (!stbi__compute_transparency(z, tc, s->img_out_n)) - return 0; - } - } - if (is_iphone && stbi__de_iphone_flag && s->img_out_n > 2) - stbi__de_iphone(z); - if (pal_img_n) { - // pal_img_n == 3 or 4 - s->img_n = pal_img_n; // record the actual colors we had - s->img_out_n = pal_img_n; - if (req_comp >= 3) - s->img_out_n = req_comp; - if (!stbi__expand_png_palette(z, palette, pal_len, s->img_out_n)) - return 0; - } else if (has_trans) { - // non-paletted image with tRNS -> source image has (constant) alpha - ++s->img_n; - } - STBI_FREE(z->expanded); - z->expanded = NULL; - // end of PNG chunk, read and skip CRC - stbi__get32be(s); - return 1; - } - - default: - // if critical, fail - if (first) - return stbi__err("first not IHDR", "Corrupt PNG"); - if ((c.type & (1 << 29)) == 0) { -#ifndef STBI_NO_FAILURE_STRINGS - // not threadsafe - static char invalid_chunk[] = "XXXX PNG chunk not known"; - invalid_chunk[0] = STBI__BYTECAST(c.type >> 24); - invalid_chunk[1] = STBI__BYTECAST(c.type >> 16); - invalid_chunk[2] = STBI__BYTECAST(c.type >> 8); - invalid_chunk[3] = STBI__BYTECAST(c.type >> 0); -#endif - return stbi__err(invalid_chunk, "PNG not supported: unknown PNG chunk type"); - } - stbi__skip(s, c.length); - break; - } - // end of PNG chunk, read and skip CRC - stbi__get32be(s); - } -} - -static void* stbi__do_png(stbi__png* p, - int* x, - int* y, - int* n, - int req_comp, - stbi__result_info* ri) { - void* result = NULL; - if (req_comp < 0 || req_comp > 4) - return stbi__errpuc("bad req_comp", "Internal error"); - if (stbi__parse_png_file(p, STBI__SCAN_load, req_comp)) { - if (p->depth <= 8) - ri->bits_per_channel = 8; - else if (p->depth == 16) - ri->bits_per_channel = 16; - else - return stbi__errpuc("bad bits_per_channel", "PNG not supported: unsupported color depth"); - result = p->out; - p->out = NULL; - if (req_comp && req_comp != p->s->img_out_n) { - if (ri->bits_per_channel == 8) - result = stbi__convert_format((unsigned char*)result, p->s->img_out_n, req_comp, - p->s->img_x, p->s->img_y); - else - result = stbi__convert_format16((stbi__uint16*)result, p->s->img_out_n, req_comp, - p->s->img_x, p->s->img_y); - p->s->img_out_n = req_comp; - if (result == NULL) - return result; - } - *x = p->s->img_x; - *y = p->s->img_y; - if (n) - *n = p->s->img_n; - } - STBI_FREE(p->out); - p->out = NULL; - STBI_FREE(p->expanded); - p->expanded = NULL; - STBI_FREE(p->idata); - p->idata = NULL; - - return result; -} - -static void* stbi__png_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri) { - stbi__png p; - p.s = s; - return stbi__do_png(&p, x, y, comp, req_comp, ri); -} - -static int stbi__png_test(stbi__context* s) { - int r; - r = stbi__check_png_header(s); - stbi__rewind(s); - return r; -} - -static int stbi__png_info_raw(stbi__png* p, int* x, int* y, int* comp) { - if (!stbi__parse_png_file(p, STBI__SCAN_header, 0)) { - stbi__rewind(p->s); - return 0; - } - if (x) - *x = p->s->img_x; - if (y) - *y = p->s->img_y; - if (comp) - *comp = p->s->img_n; - return 1; -} - -static int stbi__png_info(stbi__context* s, int* x, int* y, int* comp) { - stbi__png p; - p.s = s; - return stbi__png_info_raw(&p, x, y, comp); -} - -static int stbi__png_is16(stbi__context* s) { - stbi__png p; - p.s = s; - if (!stbi__png_info_raw(&p, NULL, NULL, NULL)) - return 0; - if (p.depth != 16) { - stbi__rewind(p.s); - return 0; - } - return 1; -} -#endif - -// Microsoft/Windows BMP image - -#ifndef STBI_NO_BMP -static int stbi__bmp_test_raw(stbi__context* s) { - int r; - int sz; - if (stbi__get8(s) != 'B') - return 0; - if (stbi__get8(s) != 'M') - return 0; - stbi__get32le(s); // discard filesize - stbi__get16le(s); // discard reserved - stbi__get16le(s); // discard reserved - stbi__get32le(s); // discard data offset - sz = stbi__get32le(s); - r = (sz == 12 || sz == 40 || sz == 56 || sz == 108 || sz == 124); - return r; -} - -static int stbi__bmp_test(stbi__context* s) { - int r = stbi__bmp_test_raw(s); - stbi__rewind(s); - return r; -} - -// returns 0..31 for the highest set bit -static int stbi__high_bit(unsigned int z) { - int n = 0; - if (z == 0) - return -1; - if (z >= 0x10000) { - n += 16; - z >>= 16; - } - if (z >= 0x00100) { - n += 8; - z >>= 8; - } - if (z >= 0x00010) { - n += 4; - z >>= 4; - } - if (z >= 0x00004) { - n += 2; - z >>= 2; - } - if (z >= 0x00002) { - n += 1; /* >>= 1;*/ - } - return n; -} - -static int stbi__bitcount(unsigned int a) { - a = (a & 0x55555555) + ((a >> 1) & 0x55555555); // max 2 - a = (a & 0x33333333) + ((a >> 2) & 0x33333333); // max 4 - a = (a + (a >> 4)) & 0x0f0f0f0f; // max 8 per 4, now 8 bits - a = (a + (a >> 8)); // max 16 per 8 bits - a = (a + (a >> 16)); // max 32 per 8 bits - return a & 0xff; -} - -// extract an arbitrarily-aligned N-bit value (N=bits) -// from v, and then make it 8-bits long and fractionally -// extend it to full full range. -static int stbi__shiftsigned(unsigned int v, int shift, int bits) { - static unsigned int mul_table[9] = { - 0, - 0xff /*0b11111111*/, - 0x55 /*0b01010101*/, - 0x49 /*0b01001001*/, - 0x11 /*0b00010001*/, - 0x21 /*0b00100001*/, - 0x41 /*0b01000001*/, - 0x81 /*0b10000001*/, - 0x01 /*0b00000001*/, - }; - static unsigned int shift_table[9] = { - 0, 0, 0, 1, 0, 2, 4, 6, 0, - }; - if (shift < 0) - v <<= -shift; - else - v >>= shift; - STBI_ASSERT(v < 256); - v >>= (8 - bits); - STBI_ASSERT(bits >= 0 && bits <= 8); - return (int)((unsigned)v * mul_table[bits]) >> shift_table[bits]; -} - -typedef struct { - int bpp, offset, hsz; - unsigned int mr, mg, mb, ma, all_a; - int extra_read; -} stbi__bmp_data; - -static int stbi__bmp_set_mask_defaults(stbi__bmp_data* info, int compress) { - // BI_BITFIELDS specifies masks explicitly, don't override - if (compress == 3) - return 1; - - if (compress == 0) { - if (info->bpp == 16) { - info->mr = 31u << 10; - info->mg = 31u << 5; - info->mb = 31u << 0; - } else if (info->bpp == 32) { - info->mr = 0xffu << 16; - info->mg = 0xffu << 8; - info->mb = 0xffu << 0; - info->ma = 0xffu << 24; - info->all_a = 0; // if all_a is 0 at end, then we loaded alpha channel but it was all 0 - } else { - // otherwise, use defaults, which is all-0 - info->mr = info->mg = info->mb = info->ma = 0; - } - return 1; - } - return 0; // error -} - -static void* stbi__bmp_parse_header(stbi__context* s, stbi__bmp_data* info) { - int hsz; - if (stbi__get8(s) != 'B' || stbi__get8(s) != 'M') - return stbi__errpuc("not BMP", "Corrupt BMP"); - stbi__get32le(s); // discard filesize - stbi__get16le(s); // discard reserved - stbi__get16le(s); // discard reserved - info->offset = stbi__get32le(s); - info->hsz = hsz = stbi__get32le(s); - info->mr = info->mg = info->mb = info->ma = 0; - info->extra_read = 14; - - if (info->offset < 0) - return stbi__errpuc("bad BMP", "bad BMP"); - - if (hsz != 12 && hsz != 40 && hsz != 56 && hsz != 108 && hsz != 124) - return stbi__errpuc("unknown BMP", "BMP type not supported: unknown"); - if (hsz == 12) { - s->img_x = stbi__get16le(s); - s->img_y = stbi__get16le(s); - } else { - s->img_x = stbi__get32le(s); - s->img_y = stbi__get32le(s); - } - if (stbi__get16le(s) != 1) - return stbi__errpuc("bad BMP", "bad BMP"); - info->bpp = stbi__get16le(s); - if (hsz != 12) { - int compress = stbi__get32le(s); - if (compress == 1 || compress == 2) - return stbi__errpuc("BMP RLE", "BMP type not supported: RLE"); - if (compress >= 4) - return stbi__errpuc( - "BMP JPEG/PNG", - "BMP type not supported: unsupported compression"); // this includes PNG/JPEG modes - if (compress == 3 && info->bpp != 16 && info->bpp != 32) - return stbi__errpuc("bad BMP", "bad BMP"); // bitfields requires 16 or 32 bits/pixel - stbi__get32le(s); // discard sizeof - stbi__get32le(s); // discard hres - stbi__get32le(s); // discard vres - stbi__get32le(s); // discard colorsused - stbi__get32le(s); // discard max important - if (hsz == 40 || hsz == 56) { - if (hsz == 56) { - stbi__get32le(s); - stbi__get32le(s); - stbi__get32le(s); - stbi__get32le(s); - } - if (info->bpp == 16 || info->bpp == 32) { - if (compress == 0) { - stbi__bmp_set_mask_defaults(info, compress); - } else if (compress == 3) { - info->mr = stbi__get32le(s); - info->mg = stbi__get32le(s); - info->mb = stbi__get32le(s); - info->extra_read += 12; - // not documented, but generated by photoshop and handled by mspaint - if (info->mr == info->mg && info->mg == info->mb) { - // ?!?!? - return stbi__errpuc("bad BMP", "bad BMP"); - } - } else - return stbi__errpuc("bad BMP", "bad BMP"); - } - } else { - // V4/V5 header - int i; - if (hsz != 108 && hsz != 124) - return stbi__errpuc("bad BMP", "bad BMP"); - info->mr = stbi__get32le(s); - info->mg = stbi__get32le(s); - info->mb = stbi__get32le(s); - info->ma = stbi__get32le(s); - if (compress != 3) // override mr/mg/mb unless in BI_BITFIELDS mode, as per docs - stbi__bmp_set_mask_defaults(info, compress); - stbi__get32le(s); // discard color space - for (i = 0; i < 12; ++i) - stbi__get32le(s); // discard color space parameters - if (hsz == 124) { - stbi__get32le(s); // discard rendering intent - stbi__get32le(s); // discard offset of profile data - stbi__get32le(s); // discard size of profile data - stbi__get32le(s); // discard reserved - } - } - } - return (void*)1; -} - -static void* stbi__bmp_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri) { - stbi_uc* out; - unsigned int mr = 0, mg = 0, mb = 0, ma = 0, all_a; - stbi_uc pal[256][4]; - int psize = 0, i, j, width; - int flip_vertically, pad, target; - stbi__bmp_data info; - STBI_NOTUSED(ri); - - info.all_a = 255; - if (stbi__bmp_parse_header(s, &info) == NULL) - return NULL; // error code already set - - flip_vertically = ((int)s->img_y) > 0; - s->img_y = abs((int)s->img_y); - - if (s->img_y > STBI_MAX_DIMENSIONS) - return stbi__errpuc("too large", "Very large image (corrupt?)"); - if (s->img_x > STBI_MAX_DIMENSIONS) - return stbi__errpuc("too large", "Very large image (corrupt?)"); - - mr = info.mr; - mg = info.mg; - mb = info.mb; - ma = info.ma; - all_a = info.all_a; - - if (info.hsz == 12) { - if (info.bpp < 24) - psize = (info.offset - info.extra_read - 24) / 3; - } else { - if (info.bpp < 16) - psize = (info.offset - info.extra_read - info.hsz) >> 2; - } - if (psize == 0) { - if (info.offset != s->callback_already_read + (s->img_buffer - s->img_buffer_original)) { - return stbi__errpuc("bad offset", "Corrupt BMP"); - } - } - - if (info.bpp == 24 && ma == 0xff000000) - s->img_n = 3; - else - s->img_n = ma ? 4 : 3; - if (req_comp && req_comp >= 3) // we can directly decode 3 or 4 - target = req_comp; - else - target = s->img_n; // if they want monochrome, we'll post-convert - - // sanity-check size - if (!stbi__mad3sizes_valid(target, s->img_x, s->img_y, 0)) - return stbi__errpuc("too large", "Corrupt BMP"); - - out = (stbi_uc*)stbi__malloc_mad3(target, s->img_x, s->img_y, 0); - if (!out) - return stbi__errpuc("outofmem", "Out of memory"); - if (info.bpp < 16) { - int z = 0; - if (psize == 0 || psize > 256) { - STBI_FREE(out); - return stbi__errpuc("invalid", "Corrupt BMP"); - } - for (i = 0; i < psize; ++i) { - pal[i][2] = stbi__get8(s); - pal[i][1] = stbi__get8(s); - pal[i][0] = stbi__get8(s); - if (info.hsz != 12) - stbi__get8(s); - pal[i][3] = 255; - } - stbi__skip(s, info.offset - info.extra_read - info.hsz - psize * (info.hsz == 12 ? 3 : 4)); - if (info.bpp == 1) - width = (s->img_x + 7) >> 3; - else if (info.bpp == 4) - width = (s->img_x + 1) >> 1; - else if (info.bpp == 8) - width = s->img_x; - else { - STBI_FREE(out); - return stbi__errpuc("bad bpp", "Corrupt BMP"); - } - pad = (-width) & 3; - if (info.bpp == 1) { - for (j = 0; j < (int)s->img_y; ++j) { - int bit_offset = 7, v = stbi__get8(s); - for (i = 0; i < (int)s->img_x; ++i) { - int color = (v >> bit_offset) & 0x1; - out[z++] = pal[color][0]; - out[z++] = pal[color][1]; - out[z++] = pal[color][2]; - if (target == 4) - out[z++] = 255; - if (i + 1 == (int)s->img_x) - break; - if ((--bit_offset) < 0) { - bit_offset = 7; - v = stbi__get8(s); - } - } - stbi__skip(s, pad); - } - } else { - for (j = 0; j < (int)s->img_y; ++j) { - for (i = 0; i < (int)s->img_x; i += 2) { - int v = stbi__get8(s), v2 = 0; - if (info.bpp == 4) { - v2 = v & 15; - v >>= 4; - } - out[z++] = pal[v][0]; - out[z++] = pal[v][1]; - out[z++] = pal[v][2]; - if (target == 4) - out[z++] = 255; - if (i + 1 == (int)s->img_x) - break; - v = (info.bpp == 8) ? stbi__get8(s) : v2; - out[z++] = pal[v][0]; - out[z++] = pal[v][1]; - out[z++] = pal[v][2]; - if (target == 4) - out[z++] = 255; - } - stbi__skip(s, pad); - } - } - } else { - int rshift = 0, gshift = 0, bshift = 0, ashift = 0, rcount = 0, gcount = 0, bcount = 0, - acount = 0; - int z = 0; - int easy = 0; - stbi__skip(s, info.offset - info.extra_read - info.hsz); - if (info.bpp == 24) - width = 3 * s->img_x; - else if (info.bpp == 16) - width = 2 * s->img_x; - else /* bpp = 32 and pad = 0 */ - width = 0; - pad = (-width) & 3; - if (info.bpp == 24) { - easy = 1; - } else if (info.bpp == 32) { - if (mb == 0xff && mg == 0xff00 && mr == 0x00ff0000 && ma == 0xff000000) - easy = 2; - } - if (!easy) { - if (!mr || !mg || !mb) { - STBI_FREE(out); - return stbi__errpuc("bad masks", "Corrupt BMP"); - } - // right shift amt to put high bit in position #7 - rshift = stbi__high_bit(mr) - 7; - rcount = stbi__bitcount(mr); - gshift = stbi__high_bit(mg) - 7; - gcount = stbi__bitcount(mg); - bshift = stbi__high_bit(mb) - 7; - bcount = stbi__bitcount(mb); - ashift = stbi__high_bit(ma) - 7; - acount = stbi__bitcount(ma); - if (rcount > 8 || gcount > 8 || bcount > 8 || acount > 8) { - STBI_FREE(out); - return stbi__errpuc("bad masks", "Corrupt BMP"); - } - } - for (j = 0; j < (int)s->img_y; ++j) { - if (easy) { - for (i = 0; i < (int)s->img_x; ++i) { - unsigned char a; - out[z + 2] = stbi__get8(s); - out[z + 1] = stbi__get8(s); - out[z + 0] = stbi__get8(s); - z += 3; - a = (easy == 2 ? stbi__get8(s) : 255); - all_a |= a; - if (target == 4) - out[z++] = a; - } - } else { - int bpp = info.bpp; - for (i = 0; i < (int)s->img_x; ++i) { - stbi__uint32 v = (bpp == 16 ? (stbi__uint32)stbi__get16le(s) : stbi__get32le(s)); - unsigned int a; - out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mr, rshift, rcount)); - out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mg, gshift, gcount)); - out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mb, bshift, bcount)); - a = (ma ? stbi__shiftsigned(v & ma, ashift, acount) : 255); - all_a |= a; - if (target == 4) - out[z++] = STBI__BYTECAST(a); - } - } - stbi__skip(s, pad); - } - } - - // if alpha channel is all 0s, replace with all 255s - if (target == 4 && all_a == 0) - for (i = 4 * s->img_x * s->img_y - 1; i >= 0; i -= 4) - out[i] = 255; - - if (flip_vertically) { - stbi_uc t; - for (j = 0; j < (int)s->img_y >> 1; ++j) { - stbi_uc* p1 = out + j * s->img_x * target; - stbi_uc* p2 = out + (s->img_y - 1 - j) * s->img_x * target; - for (i = 0; i < (int)s->img_x * target; ++i) { - t = p1[i]; - p1[i] = p2[i]; - p2[i] = t; - } - } - } - - if (req_comp && req_comp != target) { - out = stbi__convert_format(out, target, req_comp, s->img_x, s->img_y); - if (out == NULL) - return out; // stbi__convert_format frees input on failure - } - - *x = s->img_x; - *y = s->img_y; - if (comp) - *comp = s->img_n; - return out; -} -#endif - -// Targa Truevision - TGA -// by Jonathan Dummer -#ifndef STBI_NO_TGA -// returns STBI_rgb or whatever, 0 on error -static int stbi__tga_get_comp(int bits_per_pixel, int is_grey, int* is_rgb16) { - // only RGB or RGBA (incl. 16bit) or grey allowed - if (is_rgb16) - *is_rgb16 = 0; - switch (bits_per_pixel) { - case 8: - return STBI_grey; - case 16: - if (is_grey) - return STBI_grey_alpha; - // fallthrough - case 15: - if (is_rgb16) - *is_rgb16 = 1; - return STBI_rgb; - case 24: // fallthrough - case 32: - return bits_per_pixel / 8; - default: - return 0; - } -} - -static int stbi__tga_info(stbi__context* s, int* x, int* y, int* comp) { - int tga_w, tga_h, tga_comp, tga_image_type, tga_bits_per_pixel, tga_colormap_bpp; - int sz, tga_colormap_type; - stbi__get8(s); // discard Offset - tga_colormap_type = stbi__get8(s); // colormap type - if (tga_colormap_type > 1) { - stbi__rewind(s); - return 0; // only RGB or indexed allowed - } - tga_image_type = stbi__get8(s); // image type - if (tga_colormap_type == 1) { // colormapped (paletted) image - if (tga_image_type != 1 && tga_image_type != 9) { - stbi__rewind(s); - return 0; - } - stbi__skip(s, 4); // skip index of first colormap entry and number of entries - sz = stbi__get8(s); // check bits per palette color entry - if ((sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32)) { - stbi__rewind(s); - return 0; - } - stbi__skip(s, 4); // skip image x and y origin - tga_colormap_bpp = sz; - } else { // "normal" image w/o colormap - only RGB or grey allowed, +/- RLE - if ((tga_image_type != 2) && (tga_image_type != 3) && (tga_image_type != 10) && - (tga_image_type != 11)) { - stbi__rewind(s); - return 0; // only RGB or grey allowed, +/- RLE - } - stbi__skip(s, 9); // skip colormap specification and image x/y origin - tga_colormap_bpp = 0; - } - tga_w = stbi__get16le(s); - if (tga_w < 1) { - stbi__rewind(s); - return 0; // test width - } - tga_h = stbi__get16le(s); - if (tga_h < 1) { - stbi__rewind(s); - return 0; // test height - } - tga_bits_per_pixel = stbi__get8(s); // bits per pixel - stbi__get8(s); // ignore alpha bits - if (tga_colormap_bpp != 0) { - if ((tga_bits_per_pixel != 8) && (tga_bits_per_pixel != 16)) { - // when using a colormap, tga_bits_per_pixel is the size of the indexes - // I don't think anything but 8 or 16bit indexes makes sense - stbi__rewind(s); - return 0; - } - tga_comp = stbi__tga_get_comp(tga_colormap_bpp, 0, NULL); - } else { - tga_comp = stbi__tga_get_comp(tga_bits_per_pixel, - (tga_image_type == 3) || (tga_image_type == 11), NULL); - } - if (!tga_comp) { - stbi__rewind(s); - return 0; - } - if (x) - *x = tga_w; - if (y) - *y = tga_h; - if (comp) - *comp = tga_comp; - return 1; // seems to have passed everything -} - -static int stbi__tga_test(stbi__context* s) { - int res = 0; - int sz, tga_color_type; - stbi__get8(s); // discard Offset - tga_color_type = stbi__get8(s); // color type - if (tga_color_type > 1) - goto errorEnd; // only RGB or indexed allowed - sz = stbi__get8(s); // image type - if (tga_color_type == 1) { // colormapped (paletted) image - if (sz != 1 && sz != 9) - goto errorEnd; // colortype 1 demands image type 1 or 9 - stbi__skip(s, 4); // skip index of first colormap entry and number of entries - sz = stbi__get8(s); // check bits per palette color entry - if ((sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32)) - goto errorEnd; - stbi__skip(s, 4); // skip image x and y origin - } else { // "normal" image w/o colormap - if ((sz != 2) && (sz != 3) && (sz != 10) && (sz != 11)) - goto errorEnd; // only RGB or grey allowed, +/- RLE - stbi__skip(s, 9); // skip colormap specification and image x/y origin - } - if (stbi__get16le(s) < 1) - goto errorEnd; // test width - if (stbi__get16le(s) < 1) - goto errorEnd; // test height - sz = stbi__get8(s); // bits per pixel - if ((tga_color_type == 1) && (sz != 8) && (sz != 16)) - goto errorEnd; // for colormapped images, bpp is size of an index - if ((sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32)) - goto errorEnd; - - res = 1; // if we got this far, everything's good and we can return 1 instead of 0 - -errorEnd: - stbi__rewind(s); - return res; -} - -// read 16bit value and convert to 24bit RGB -static void stbi__tga_read_rgb16(stbi__context* s, stbi_uc* out) { - stbi__uint16 px = (stbi__uint16)stbi__get16le(s); - stbi__uint16 fiveBitMask = 31; - // we have 3 channels with 5bits each - int r = (px >> 10) & fiveBitMask; - int g = (px >> 5) & fiveBitMask; - int b = px & fiveBitMask; - // Note that this saves the data in RGB(A) order, so it doesn't need to be swapped later - out[0] = (stbi_uc)((r * 255) / 31); - out[1] = (stbi_uc)((g * 255) / 31); - out[2] = (stbi_uc)((b * 255) / 31); - - // some people claim that the most significant bit might be used for alpha - // (possibly if an alpha-bit is set in the "image descriptor byte") - // but that only made 16bit test images completely translucent.. - // so let's treat all 15 and 16bit TGAs as RGB with no alpha. -} - -static void* stbi__tga_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri) { - // read in the TGA header stuff - int tga_offset = stbi__get8(s); - int tga_indexed = stbi__get8(s); - int tga_image_type = stbi__get8(s); - int tga_is_RLE = 0; - int tga_palette_start = stbi__get16le(s); - int tga_palette_len = stbi__get16le(s); - int tga_palette_bits = stbi__get8(s); - int tga_x_origin = stbi__get16le(s); - int tga_y_origin = stbi__get16le(s); - int tga_width = stbi__get16le(s); - int tga_height = stbi__get16le(s); - int tga_bits_per_pixel = stbi__get8(s); - int tga_comp, tga_rgb16 = 0; - int tga_inverted = stbi__get8(s); - // int tga_alpha_bits = tga_inverted & 15; // the 4 lowest bits - unused (useless?) - // image data - unsigned char* tga_data; - unsigned char* tga_palette = NULL; - int i, j; - unsigned char raw_data[4] = {0}; - int RLE_count = 0; - int RLE_repeating = 0; - int read_next_pixel = 1; - STBI_NOTUSED(ri); - STBI_NOTUSED(tga_x_origin); // @TODO - STBI_NOTUSED(tga_y_origin); // @TODO - - if (tga_height > STBI_MAX_DIMENSIONS) - return stbi__errpuc("too large", "Very large image (corrupt?)"); - if (tga_width > STBI_MAX_DIMENSIONS) - return stbi__errpuc("too large", "Very large image (corrupt?)"); - - // do a tiny bit of precessing - if (tga_image_type >= 8) { - tga_image_type -= 8; - tga_is_RLE = 1; - } - tga_inverted = 1 - ((tga_inverted >> 5) & 1); - - // If I'm paletted, then I'll use the number of bits from the palette - if (tga_indexed) - tga_comp = stbi__tga_get_comp(tga_palette_bits, 0, &tga_rgb16); - else - tga_comp = stbi__tga_get_comp(tga_bits_per_pixel, (tga_image_type == 3), &tga_rgb16); - - if (!tga_comp) // shouldn't really happen, stbi__tga_test() should have ensured basic consistency - return stbi__errpuc("bad format", "Can't find out TGA pixelformat"); - - // tga info - *x = tga_width; - *y = tga_height; - if (comp) - *comp = tga_comp; - - if (!stbi__mad3sizes_valid(tga_width, tga_height, tga_comp, 0)) - return stbi__errpuc("too large", "Corrupt TGA"); - - tga_data = (unsigned char*)stbi__malloc_mad3(tga_width, tga_height, tga_comp, 0); - if (!tga_data) - return stbi__errpuc("outofmem", "Out of memory"); - - // skip to the data's starting position (offset usually = 0) - stbi__skip(s, tga_offset); - - if (!tga_indexed && !tga_is_RLE && !tga_rgb16) { - for (i = 0; i < tga_height; ++i) { - int row = tga_inverted ? tga_height - i - 1 : i; - stbi_uc* tga_row = tga_data + row * tga_width * tga_comp; - stbi__getn(s, tga_row, tga_width * tga_comp); - } - } else { - // do I need to load a palette? - if (tga_indexed) { - if (tga_palette_len == 0) { /* you have to have at least one entry! */ - STBI_FREE(tga_data); - return stbi__errpuc("bad palette", "Corrupt TGA"); - } - - // any data to skip? (offset usually = 0) - stbi__skip(s, tga_palette_start); - // load the palette - tga_palette = (unsigned char*)stbi__malloc_mad2(tga_palette_len, tga_comp, 0); - if (!tga_palette) { - STBI_FREE(tga_data); - return stbi__errpuc("outofmem", "Out of memory"); - } - if (tga_rgb16) { - stbi_uc* pal_entry = tga_palette; - STBI_ASSERT(tga_comp == STBI_rgb); - for (i = 0; i < tga_palette_len; ++i) { - stbi__tga_read_rgb16(s, pal_entry); - pal_entry += tga_comp; - } - } else if (!stbi__getn(s, tga_palette, tga_palette_len * tga_comp)) { - STBI_FREE(tga_data); - STBI_FREE(tga_palette); - return stbi__errpuc("bad palette", "Corrupt TGA"); - } - } - // load the data - for (i = 0; i < tga_width * tga_height; ++i) { - // if I'm in RLE mode, do I need to get a RLE stbi__pngchunk? - if (tga_is_RLE) { - if (RLE_count == 0) { - // yep, get the next byte as a RLE command - int RLE_cmd = stbi__get8(s); - RLE_count = 1 + (RLE_cmd & 127); - RLE_repeating = RLE_cmd >> 7; - read_next_pixel = 1; - } else if (!RLE_repeating) { - read_next_pixel = 1; - } - } else { - read_next_pixel = 1; - } - // OK, if I need to read a pixel, do it now - if (read_next_pixel) { - // load however much data we did have - if (tga_indexed) { - // read in index, then perform the lookup - int pal_idx = (tga_bits_per_pixel == 8) ? stbi__get8(s) : stbi__get16le(s); - if (pal_idx >= tga_palette_len) { - // invalid index - pal_idx = 0; - } - pal_idx *= tga_comp; - for (j = 0; j < tga_comp; ++j) { - raw_data[j] = tga_palette[pal_idx + j]; - } - } else if (tga_rgb16) { - STBI_ASSERT(tga_comp == STBI_rgb); - stbi__tga_read_rgb16(s, raw_data); - } else { - // read in the data raw - for (j = 0; j < tga_comp; ++j) { - raw_data[j] = stbi__get8(s); - } - } - // clear the reading flag for the next pixel - read_next_pixel = 0; - } // end of reading a pixel - - // copy data - for (j = 0; j < tga_comp; ++j) - tga_data[i * tga_comp + j] = raw_data[j]; - - // in case we're in RLE mode, keep counting down - --RLE_count; - } - // do I need to invert the image? - if (tga_inverted) { - for (j = 0; j * 2 < tga_height; ++j) { - int index1 = j * tga_width * tga_comp; - int index2 = (tga_height - 1 - j) * tga_width * tga_comp; - for (i = tga_width * tga_comp; i > 0; --i) { - unsigned char temp = tga_data[index1]; - tga_data[index1] = tga_data[index2]; - tga_data[index2] = temp; - ++index1; - ++index2; - } - } - } - // clear my palette, if I had one - if (tga_palette != NULL) { - STBI_FREE(tga_palette); - } - } - - // swap RGB - if the source data was RGB16, it already is in the right order - if (tga_comp >= 3 && !tga_rgb16) { - unsigned char* tga_pixel = tga_data; - for (i = 0; i < tga_width * tga_height; ++i) { - unsigned char temp = tga_pixel[0]; - tga_pixel[0] = tga_pixel[2]; - tga_pixel[2] = temp; - tga_pixel += tga_comp; - } - } - - // convert to target component count - if (req_comp && req_comp != tga_comp) - tga_data = stbi__convert_format(tga_data, tga_comp, req_comp, tga_width, tga_height); - - // the things I do to get rid of an error message, and yet keep - // Microsoft's C compilers happy... [8^( - tga_palette_start = tga_palette_len = tga_palette_bits = tga_x_origin = tga_y_origin = 0; - STBI_NOTUSED(tga_palette_start); - // OK, done - return tga_data; -} -#endif - -// ************************************************************************************************* -// Photoshop PSD loader -- PD by Thatcher Ulrich, integration by Nicolas Schulz, tweaked by STB - -#ifndef STBI_NO_PSD -static int stbi__psd_test(stbi__context* s) { - int r = (stbi__get32be(s) == 0x38425053); - stbi__rewind(s); - return r; -} - -static int stbi__psd_decode_rle(stbi__context* s, stbi_uc* p, int pixelCount) { - int count, nleft, len; - - count = 0; - while ((nleft = pixelCount - count) > 0) { - len = stbi__get8(s); - if (len == 128) { - // No-op. - } else if (len < 128) { - // Copy next len+1 bytes literally. - len++; - if (len > nleft) - return 0; // corrupt data - count += len; - while (len) { - *p = stbi__get8(s); - p += 4; - len--; - } - } else if (len > 128) { - stbi_uc val; - // Next -len+1 bytes in the dest are replicated from next source byte. - // (Interpret len as a negative 8-bit int.) - len = 257 - len; - if (len > nleft) - return 0; // corrupt data - val = stbi__get8(s); - count += len; - while (len) { - *p = val; - p += 4; - len--; - } - } - } - - return 1; -} - -static void* stbi__psd_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri, - int bpc) { - int pixelCount; - int channelCount, compression; - int channel, i; - int bitdepth; - int w, h; - stbi_uc* out; - STBI_NOTUSED(ri); - - // Check identifier - if (stbi__get32be(s) != 0x38425053) // "8BPS" - return stbi__errpuc("not PSD", "Corrupt PSD image"); - - // Check file type version. - if (stbi__get16be(s) != 1) - return stbi__errpuc("wrong version", "Unsupported version of PSD image"); - - // Skip 6 reserved bytes. - stbi__skip(s, 6); - - // Read the number of channels (R, G, B, A, etc). - channelCount = stbi__get16be(s); - if (channelCount < 0 || channelCount > 16) - return stbi__errpuc("wrong channel count", "Unsupported number of channels in PSD image"); - - // Read the rows and columns of the image. - h = stbi__get32be(s); - w = stbi__get32be(s); - - if (h > STBI_MAX_DIMENSIONS) - return stbi__errpuc("too large", "Very large image (corrupt?)"); - if (w > STBI_MAX_DIMENSIONS) - return stbi__errpuc("too large", "Very large image (corrupt?)"); - - // Make sure the depth is 8 bits. - bitdepth = stbi__get16be(s); - if (bitdepth != 8 && bitdepth != 16) - return stbi__errpuc("unsupported bit depth", "PSD bit depth is not 8 or 16 bit"); - - // Make sure the color mode is RGB. - // Valid options are: - // 0: Bitmap - // 1: Grayscale - // 2: Indexed color - // 3: RGB color - // 4: CMYK color - // 7: Multichannel - // 8: Duotone - // 9: Lab color - if (stbi__get16be(s) != 3) - return stbi__errpuc("wrong color format", "PSD is not in RGB color format"); - - // Skip the Mode Data. (It's the palette for indexed color; other info for other modes.) - stbi__skip(s, stbi__get32be(s)); - - // Skip the image resources. (resolution, pen tool paths, etc) - stbi__skip(s, stbi__get32be(s)); - - // Skip the reserved data. - stbi__skip(s, stbi__get32be(s)); - - // Find out if the data is compressed. - // Known values: - // 0: no compression - // 1: RLE compressed - compression = stbi__get16be(s); - if (compression > 1) - return stbi__errpuc("bad compression", "PSD has an unknown compression format"); - - // Check size - if (!stbi__mad3sizes_valid(4, w, h, 0)) - return stbi__errpuc("too large", "Corrupt PSD"); - - // Create the destination image. - - if (!compression && bitdepth == 16 && bpc == 16) { - out = (stbi_uc*)stbi__malloc_mad3(8, w, h, 0); - ri->bits_per_channel = 16; - } else - out = (stbi_uc*)stbi__malloc(4 * w * h); - - if (!out) - return stbi__errpuc("outofmem", "Out of memory"); - pixelCount = w * h; - - // Initialize the data to zero. - // memset( out, 0, pixelCount * 4 ); - - // Finally, the image data. - if (compression) { - // RLE as used by .PSD and .TIFF - // Loop until you get the number of unpacked bytes you are expecting: - // Read the next source byte into n. - // If n is between 0 and 127 inclusive, copy the next n+1 bytes literally. - // Else if n is between -127 and -1 inclusive, copy the next byte -n+1 times. - // Else if n is 128, noop. - // Endloop - - // The RLE-compressed data is preceded by a 2-byte data count for each row in the data, - // which we're going to just skip. - stbi__skip(s, h * channelCount * 2); - - // Read the RLE data by channel. - for (channel = 0; channel < 4; channel++) { - stbi_uc* p; - - p = out + channel; - if (channel >= channelCount) { - // Fill this channel with default data. - for (i = 0; i < pixelCount; i++, p += 4) - *p = (channel == 3 ? 255 : 0); - } else { - // Read the RLE data. - if (!stbi__psd_decode_rle(s, p, pixelCount)) { - STBI_FREE(out); - return stbi__errpuc("corrupt", "bad RLE data"); - } - } - } - - } else { - // We're at the raw image data. It's each channel in order (Red, Green, Blue, Alpha, ...) - // where each channel consists of an 8-bit (or 16-bit) value for each pixel in the image. - - // Read the data by channel. - for (channel = 0; channel < 4; channel++) { - if (channel >= channelCount) { - // Fill this channel with default data. - if (bitdepth == 16 && bpc == 16) { - stbi__uint16* q = ((stbi__uint16*)out) + channel; - stbi__uint16 val = channel == 3 ? 65535 : 0; - for (i = 0; i < pixelCount; i++, q += 4) - *q = val; - } else { - stbi_uc* p = out + channel; - stbi_uc val = channel == 3 ? 255 : 0; - for (i = 0; i < pixelCount; i++, p += 4) - *p = val; - } - } else { - if (ri->bits_per_channel == 16) { // output bpc - stbi__uint16* q = ((stbi__uint16*)out) + channel; - for (i = 0; i < pixelCount; i++, q += 4) - *q = (stbi__uint16)stbi__get16be(s); - } else { - stbi_uc* p = out + channel; - if (bitdepth == 16) { // input bpc - for (i = 0; i < pixelCount; i++, p += 4) - *p = (stbi_uc)(stbi__get16be(s) >> 8); - } else { - for (i = 0; i < pixelCount; i++, p += 4) - *p = stbi__get8(s); - } - } - } - } - } - - // remove weird white matte from PSD - if (channelCount >= 4) { - if (ri->bits_per_channel == 16) { - for (i = 0; i < w * h; ++i) { - stbi__uint16* pixel = (stbi__uint16*)out + 4 * i; - if (pixel[3] != 0 && pixel[3] != 65535) { - float a = pixel[3] / 65535.0f; - float ra = 1.0f / a; - float inv_a = 65535.0f * (1 - ra); - pixel[0] = (stbi__uint16)(pixel[0] * ra + inv_a); - pixel[1] = (stbi__uint16)(pixel[1] * ra + inv_a); - pixel[2] = (stbi__uint16)(pixel[2] * ra + inv_a); - } - } - } else { - for (i = 0; i < w * h; ++i) { - unsigned char* pixel = out + 4 * i; - if (pixel[3] != 0 && pixel[3] != 255) { - float a = pixel[3] / 255.0f; - float ra = 1.0f / a; - float inv_a = 255.0f * (1 - ra); - pixel[0] = (unsigned char)(pixel[0] * ra + inv_a); - pixel[1] = (unsigned char)(pixel[1] * ra + inv_a); - pixel[2] = (unsigned char)(pixel[2] * ra + inv_a); - } - } - } - } - - // convert to desired output format - if (req_comp && req_comp != 4) { - if (ri->bits_per_channel == 16) - out = (stbi_uc*)stbi__convert_format16((stbi__uint16*)out, 4, req_comp, w, h); - else - out = stbi__convert_format(out, 4, req_comp, w, h); - if (out == NULL) - return out; // stbi__convert_format frees input on failure - } - - if (comp) - *comp = 4; - *y = h; - *x = w; - - return out; -} -#endif - -// ************************************************************************************************* -// Softimage PIC loader -// by Tom Seddon -// -// See http://softimage.wiki.softimage.com/index.php/INFO:_PIC_file_format -// See http://ozviz.wasp.uwa.edu.au/~pbourke/dataformats/softimagepic/ - -#ifndef STBI_NO_PIC -static int stbi__pic_is4(stbi__context* s, const char* str) { - int i; - for (i = 0; i < 4; ++i) - if (stbi__get8(s) != (stbi_uc)str[i]) - return 0; - - return 1; -} - -static int stbi__pic_test_core(stbi__context* s) { - int i; - - if (!stbi__pic_is4(s, "\x53\x80\xF6\x34")) - return 0; - - for (i = 0; i < 84; ++i) - stbi__get8(s); - - if (!stbi__pic_is4(s, "PICT")) - return 0; - - return 1; -} - -typedef struct { - stbi_uc size, type, channel; -} stbi__pic_packet; - -static stbi_uc* stbi__readval(stbi__context* s, int channel, stbi_uc* dest) { - int mask = 0x80, i; - - for (i = 0; i < 4; ++i, mask >>= 1) { - if (channel & mask) { - if (stbi__at_eof(s)) - return stbi__errpuc("bad file", "PIC file too short"); - dest[i] = stbi__get8(s); - } - } - - return dest; -} - -static void stbi__copyval(int channel, stbi_uc* dest, const stbi_uc* src) { - int mask = 0x80, i; - - for (i = 0; i < 4; ++i, mask >>= 1) - if (channel & mask) - dest[i] = src[i]; -} - -static stbi_uc* stbi__pic_load_core(stbi__context* s, - int width, - int height, - int* comp, - stbi_uc* result) { - int act_comp = 0, num_packets = 0, y, chained; - stbi__pic_packet packets[10]; - - // this will (should...) cater for even some bizarre stuff like having data - // for the same channel in multiple packets. - do { - stbi__pic_packet* packet; - - if (num_packets == sizeof(packets) / sizeof(packets[0])) - return stbi__errpuc("bad format", "too many packets"); - - packet = &packets[num_packets++]; - - chained = stbi__get8(s); - packet->size = stbi__get8(s); - packet->type = stbi__get8(s); - packet->channel = stbi__get8(s); - - act_comp |= packet->channel; - - if (stbi__at_eof(s)) - return stbi__errpuc("bad file", "file too short (reading packets)"); - if (packet->size != 8) - return stbi__errpuc("bad format", "packet isn't 8bpp"); - } while (chained); - - *comp = (act_comp & 0x10 ? 4 : 3); // has alpha channel? - - for (y = 0; y < height; ++y) { - int packet_idx; - - for (packet_idx = 0; packet_idx < num_packets; ++packet_idx) { - stbi__pic_packet* packet = &packets[packet_idx]; - stbi_uc* dest = result + y * width * 4; - - switch (packet->type) { - default: - return stbi__errpuc("bad format", "packet has bad compression type"); - - case 0: { // uncompressed - int x; - - for (x = 0; x < width; ++x, dest += 4) - if (!stbi__readval(s, packet->channel, dest)) - return 0; - break; - } - - case 1: // Pure RLE - { - int left = width, i; - - while (left > 0) { - stbi_uc count, value[4]; - - count = stbi__get8(s); - if (stbi__at_eof(s)) - return stbi__errpuc("bad file", "file too short (pure read count)"); - - if (count > left) - count = (stbi_uc)left; - - if (!stbi__readval(s, packet->channel, value)) - return 0; - - for (i = 0; i < count; ++i, dest += 4) - stbi__copyval(packet->channel, dest, value); - left -= count; - } - } break; - - case 2: { // Mixed RLE - int left = width; - while (left > 0) { - int count = stbi__get8(s), i; - if (stbi__at_eof(s)) - return stbi__errpuc("bad file", "file too short (mixed read count)"); - - if (count >= 128) { // Repeated - stbi_uc value[4]; - - if (count == 128) - count = stbi__get16be(s); - else - count -= 127; - if (count > left) - return stbi__errpuc("bad file", "scanline overrun"); - - if (!stbi__readval(s, packet->channel, value)) - return 0; - - for (i = 0; i < count; ++i, dest += 4) - stbi__copyval(packet->channel, dest, value); - } else { // Raw - ++count; - if (count > left) - return stbi__errpuc("bad file", "scanline overrun"); - - for (i = 0; i < count; ++i, dest += 4) - if (!stbi__readval(s, packet->channel, dest)) - return 0; - } - left -= count; - } - break; - } - } - } - } - - return result; -} - -static void* stbi__pic_load(stbi__context* s, - int* px, - int* py, - int* comp, - int req_comp, - stbi__result_info* ri) { - stbi_uc* result; - int i, x, y, internal_comp; - STBI_NOTUSED(ri); - - if (!comp) - comp = &internal_comp; - - for (i = 0; i < 92; ++i) - stbi__get8(s); - - x = stbi__get16be(s); - y = stbi__get16be(s); - - if (y > STBI_MAX_DIMENSIONS) - return stbi__errpuc("too large", "Very large image (corrupt?)"); - if (x > STBI_MAX_DIMENSIONS) - return stbi__errpuc("too large", "Very large image (corrupt?)"); - - if (stbi__at_eof(s)) - return stbi__errpuc("bad file", "file too short (pic header)"); - if (!stbi__mad3sizes_valid(x, y, 4, 0)) - return stbi__errpuc("too large", "PIC image too large to decode"); - - stbi__get32be(s); // skip `ratio' - stbi__get16be(s); // skip `fields' - stbi__get16be(s); // skip `pad' - - // intermediate buffer is RGBA - result = (stbi_uc*)stbi__malloc_mad3(x, y, 4, 0); - if (!result) - return stbi__errpuc("outofmem", "Out of memory"); - memset(result, 0xff, x * y * 4); - - if (!stbi__pic_load_core(s, x, y, comp, result)) { - STBI_FREE(result); - result = 0; - } - *px = x; - *py = y; - if (req_comp == 0) - req_comp = *comp; - result = stbi__convert_format(result, 4, req_comp, x, y); - - return result; -} - -static int stbi__pic_test(stbi__context* s) { - int r = stbi__pic_test_core(s); - stbi__rewind(s); - return r; -} -#endif - -// ************************************************************************************************* -// GIF loader -- public domain by Jean-Marc Lienher -- simplified/shrunk by stb - -#ifndef STBI_NO_GIF -typedef struct { - stbi__int16 prefix; - stbi_uc first; - stbi_uc suffix; -} stbi__gif_lzw; - -typedef struct { - int w, h; - stbi_uc* out; // output buffer (always 4 components) - stbi_uc* background; // The current "background" as far as a gif is concerned - stbi_uc* history; - int flags, bgindex, ratio, transparent, eflags; - stbi_uc pal[256][4]; - stbi_uc lpal[256][4]; - stbi__gif_lzw codes[8192]; - stbi_uc* color_table; - int parse, step; - int lflags; - int start_x, start_y; - int max_x, max_y; - int cur_x, cur_y; - int line_size; - int delay; -} stbi__gif; - -static int stbi__gif_test_raw(stbi__context* s) { - int sz; - if (stbi__get8(s) != 'G' || stbi__get8(s) != 'I' || stbi__get8(s) != 'F' || stbi__get8(s) != '8') - return 0; - sz = stbi__get8(s); - if (sz != '9' && sz != '7') - return 0; - if (stbi__get8(s) != 'a') - return 0; - return 1; -} - -static int stbi__gif_test(stbi__context* s) { - int r = stbi__gif_test_raw(s); - stbi__rewind(s); - return r; -} - -static void stbi__gif_parse_colortable(stbi__context* s, - stbi_uc pal[256][4], - int num_entries, - int transp) { - int i; - for (i = 0; i < num_entries; ++i) { - pal[i][2] = stbi__get8(s); - pal[i][1] = stbi__get8(s); - pal[i][0] = stbi__get8(s); - pal[i][3] = transp == i ? 0 : 255; - } -} - -static int stbi__gif_header(stbi__context* s, stbi__gif* g, int* comp, int is_info) { - stbi_uc version; - if (stbi__get8(s) != 'G' || stbi__get8(s) != 'I' || stbi__get8(s) != 'F' || stbi__get8(s) != '8') - return stbi__err("not GIF", "Corrupt GIF"); - - version = stbi__get8(s); - if (version != '7' && version != '9') - return stbi__err("not GIF", "Corrupt GIF"); - if (stbi__get8(s) != 'a') - return stbi__err("not GIF", "Corrupt GIF"); - - stbi__g_failure_reason = ""; - g->w = stbi__get16le(s); - g->h = stbi__get16le(s); - g->flags = stbi__get8(s); - g->bgindex = stbi__get8(s); - g->ratio = stbi__get8(s); - g->transparent = -1; - - if (g->w > STBI_MAX_DIMENSIONS) - return stbi__err("too large", "Very large image (corrupt?)"); - if (g->h > STBI_MAX_DIMENSIONS) - return stbi__err("too large", "Very large image (corrupt?)"); - - if (comp != 0) - *comp = 4; // can't actually tell whether it's 3 or 4 until we parse the comments - - if (is_info) - return 1; - - if (g->flags & 0x80) - stbi__gif_parse_colortable(s, g->pal, 2 << (g->flags & 7), -1); - - return 1; -} - -static int stbi__gif_info_raw(stbi__context* s, int* x, int* y, int* comp) { - stbi__gif* g = (stbi__gif*)stbi__malloc(sizeof(stbi__gif)); - if (!g) - return stbi__err("outofmem", "Out of memory"); - if (!stbi__gif_header(s, g, comp, 1)) { - STBI_FREE(g); - stbi__rewind(s); - return 0; - } - if (x) - *x = g->w; - if (y) - *y = g->h; - STBI_FREE(g); - return 1; -} - -static void stbi__out_gif_code(stbi__gif* g, stbi__uint16 code) { - stbi_uc *p, *c; - int idx; - - // recurse to decode the prefixes, since the linked-list is backwards, - // and working backwards through an interleaved image would be nasty - if (g->codes[code].prefix >= 0) - stbi__out_gif_code(g, g->codes[code].prefix); - - if (g->cur_y >= g->max_y) - return; - - idx = g->cur_x + g->cur_y; - p = &g->out[idx]; - g->history[idx / 4] = 1; - - c = &g->color_table[g->codes[code].suffix * 4]; - if (c[3] > 128) { // don't render transparent pixels; - p[0] = c[2]; - p[1] = c[1]; - p[2] = c[0]; - p[3] = c[3]; - } - g->cur_x += 4; - - if (g->cur_x >= g->max_x) { - g->cur_x = g->start_x; - g->cur_y += g->step; - - while (g->cur_y >= g->max_y && g->parse > 0) { - g->step = (1 << g->parse) * g->line_size; - g->cur_y = g->start_y + (g->step >> 1); - --g->parse; - } - } -} - -static stbi_uc* stbi__process_gif_raster(stbi__context* s, stbi__gif* g) { - stbi_uc lzw_cs; - stbi__int32 len, init_code; - stbi__uint32 first; - stbi__int32 codesize, codemask, avail, oldcode, bits, valid_bits, clear; - stbi__gif_lzw* p; - - lzw_cs = stbi__get8(s); - if (lzw_cs > 12) - return NULL; - clear = 1 << lzw_cs; - first = 1; - codesize = lzw_cs + 1; - codemask = (1 << codesize) - 1; - bits = 0; - valid_bits = 0; - for (init_code = 0; init_code < clear; init_code++) { - g->codes[init_code].prefix = -1; - g->codes[init_code].first = (stbi_uc)init_code; - g->codes[init_code].suffix = (stbi_uc)init_code; - } - - // support no starting clear code - avail = clear + 2; - oldcode = -1; - - len = 0; - for (;;) { - if (valid_bits < codesize) { - if (len == 0) { - len = stbi__get8(s); // start new block - if (len == 0) - return g->out; - } - --len; - bits |= (stbi__int32)stbi__get8(s) << valid_bits; - valid_bits += 8; - } else { - stbi__int32 code = bits & codemask; - bits >>= codesize; - valid_bits -= codesize; - // @OPTIMIZE: is there some way we can accelerate the non-clear path? - if (code == clear) { // clear code - codesize = lzw_cs + 1; - codemask = (1 << codesize) - 1; - avail = clear + 2; - oldcode = -1; - first = 0; - } else if (code == clear + 1) { // end of stream code - stbi__skip(s, len); - while ((len = stbi__get8(s)) > 0) - stbi__skip(s, len); - return g->out; - } else if (code <= avail) { - if (first) { - return stbi__errpuc("no clear code", "Corrupt GIF"); - } - - if (oldcode >= 0) { - p = &g->codes[avail++]; - if (avail > 8192) { - return stbi__errpuc("too many codes", "Corrupt GIF"); - } - - p->prefix = (stbi__int16)oldcode; - p->first = g->codes[oldcode].first; - p->suffix = (code == avail) ? p->first : g->codes[code].first; - } else if (code == avail) - return stbi__errpuc("illegal code in raster", "Corrupt GIF"); - - stbi__out_gif_code(g, (stbi__uint16)code); - - if ((avail & codemask) == 0 && avail <= 0x0FFF) { - codesize++; - codemask = (1 << codesize) - 1; - } - - oldcode = code; - } else { - return stbi__errpuc("illegal code in raster", "Corrupt GIF"); - } - } - } -} - -// this function is designed to support animated gifs, although stb_image doesn't support it -// two back is the image from two frames ago, used for a very specific disposal format -static stbi_uc* stbi__gif_load_next(stbi__context* s, - stbi__gif* g, - int* comp, - int req_comp, - stbi_uc* two_back) { - int dispose; - int first_frame; - int pi; - int pcount; - STBI_NOTUSED(req_comp); - - // on first frame, any non-written pixels get the background colour (non-transparent) - first_frame = 0; - if (g->out == 0) { - if (!stbi__gif_header(s, g, comp, 0)) - return 0; // stbi__g_failure_reason set by stbi__gif_header - if (!stbi__mad3sizes_valid(4, g->w, g->h, 0)) - return stbi__errpuc("too large", "GIF image is too large"); - pcount = g->w * g->h; - g->out = (stbi_uc*)stbi__malloc(4 * pcount); - g->background = (stbi_uc*)stbi__malloc(4 * pcount); - g->history = (stbi_uc*)stbi__malloc(pcount); - if (!g->out || !g->background || !g->history) - return stbi__errpuc("outofmem", "Out of memory"); - - // image is treated as "transparent" at the start - ie, nothing overwrites the current - // background; background colour is only used for pixels that are not rendered first frame, - // after that "background" color refers to the color that was there the previous frame. - memset(g->out, 0x00, 4 * pcount); - memset(g->background, 0x00, 4 * pcount); // state of the background (starts transparent) - memset(g->history, 0x00, pcount); // pixels that were affected previous frame - first_frame = 1; - } else { - // second frame - how do we dispose of the previous one? - dispose = (g->eflags & 0x1C) >> 2; - pcount = g->w * g->h; - - if ((dispose == 3) && (two_back == 0)) { - dispose = 2; // if I don't have an image to revert back to, default to the old background - } - - if (dispose == 3) { // use previous graphic - for (pi = 0; pi < pcount; ++pi) { - if (g->history[pi]) { - memcpy(&g->out[pi * 4], &two_back[pi * 4], 4); - } - } - } else if (dispose == 2) { - // restore what was changed last frame to background before that frame; - for (pi = 0; pi < pcount; ++pi) { - if (g->history[pi]) { - memcpy(&g->out[pi * 4], &g->background[pi * 4], 4); - } - } - } else { - // This is a non-disposal case eithe way, so just - // leave the pixels as is, and they will become the new background - // 1: do not dispose - // 0: not specified. - } - - // background is what out is after the undoing of the previou frame; - memcpy(g->background, g->out, 4 * g->w * g->h); - } - - // clear my history; - memset(g->history, 0x00, g->w * g->h); // pixels that were affected previous frame - - for (;;) { - int tag = stbi__get8(s); - switch (tag) { - case 0x2C: /* Image Descriptor */ - { - stbi__int32 x, y, w, h; - stbi_uc* o; - - x = stbi__get16le(s); - y = stbi__get16le(s); - w = stbi__get16le(s); - h = stbi__get16le(s); - if (((x + w) > (g->w)) || ((y + h) > (g->h))) - return stbi__errpuc("bad Image Descriptor", "Corrupt GIF"); - - g->line_size = g->w * 4; - g->start_x = x * 4; - g->start_y = y * g->line_size; - g->max_x = g->start_x + w * 4; - g->max_y = g->start_y + h * g->line_size; - g->cur_x = g->start_x; - g->cur_y = g->start_y; - - // if the width of the specified rectangle is 0, that means - // we may not see *any* pixels or the image is malformed; - // to make sure this is caught, move the current y down to - // max_y (which is what out_gif_code checks). - if (w == 0) - g->cur_y = g->max_y; - - g->lflags = stbi__get8(s); - - if (g->lflags & 0x40) { - g->step = 8 * g->line_size; // first interlaced spacing - g->parse = 3; - } else { - g->step = g->line_size; - g->parse = 0; - } - - if (g->lflags & 0x80) { - stbi__gif_parse_colortable(s, g->lpal, 2 << (g->lflags & 7), - g->eflags & 0x01 ? g->transparent : -1); - g->color_table = (stbi_uc*)g->lpal; - } else if (g->flags & 0x80) { - g->color_table = (stbi_uc*)g->pal; - } else - return stbi__errpuc("missing color table", "Corrupt GIF"); - - o = stbi__process_gif_raster(s, g); - if (!o) - return NULL; - - // if this was the first frame, - pcount = g->w * g->h; - if (first_frame && (g->bgindex > 0)) { - // if first frame, any pixel not drawn to gets the background color - for (pi = 0; pi < pcount; ++pi) { - if (g->history[pi] == 0) { - g->pal[g->bgindex][3] = 255; // just in case it was made transparent, undo that; It - // will be reset next frame if need be; - memcpy(&g->out[pi * 4], &g->pal[g->bgindex], 4); - } - } - } - - return o; - } - - case 0x21: // Comment Extension. - { - int len; - int ext = stbi__get8(s); - if (ext == 0xF9) { // Graphic Control Extension. - len = stbi__get8(s); - if (len == 4) { - g->eflags = stbi__get8(s); - g->delay = 10 * stbi__get16le(s); // delay - 1/100th of a second, saving as 1/1000ths. - - // unset old transparent - if (g->transparent >= 0) { - g->pal[g->transparent][3] = 255; - } - if (g->eflags & 0x01) { - g->transparent = stbi__get8(s); - if (g->transparent >= 0) { - g->pal[g->transparent][3] = 0; - } - } else { - // don't need transparent - stbi__skip(s, 1); - g->transparent = -1; - } - } else { - stbi__skip(s, len); - break; - } - } - while ((len = stbi__get8(s)) != 0) { - stbi__skip(s, len); - } - break; - } - - case 0x3B: // gif stream termination code - return (stbi_uc*)s; // using '1' causes warning on some compilers - - default: - return stbi__errpuc("unknown code", "Corrupt GIF"); - } - } -} - -static void* stbi__load_gif_main_outofmem(stbi__gif* g, stbi_uc* out, int** delays) { - STBI_FREE(g->out); - STBI_FREE(g->history); - STBI_FREE(g->background); - - if (out) - STBI_FREE(out); - if (delays && *delays) - STBI_FREE(*delays); - return stbi__errpuc("outofmem", "Out of memory"); -} - -static void* stbi__load_gif_main(stbi__context* s, - int** delays, - int* x, - int* y, - int* z, - int* comp, - int req_comp) { - if (stbi__gif_test(s)) { - int layers = 0; - stbi_uc* u = 0; - stbi_uc* out = 0; - stbi_uc* two_back = 0; - stbi__gif g; - int stride; - int out_size = 0; - int delays_size = 0; - - STBI_NOTUSED(out_size); - STBI_NOTUSED(delays_size); - - memset(&g, 0, sizeof(g)); - if (delays) { - *delays = 0; - } - - do { - u = stbi__gif_load_next(s, &g, comp, req_comp, two_back); - if (u == (stbi_uc*)s) - u = 0; // end of animated gif marker - - if (u) { - *x = g.w; - *y = g.h; - ++layers; - stride = g.w * g.h * 4; - - if (out) { - void* tmp = (stbi_uc*)STBI_REALLOC_SIZED(out, out_size, layers * stride); - if (!tmp) - return stbi__load_gif_main_outofmem(&g, out, delays); - else { - out = (stbi_uc*)tmp; - out_size = layers * stride; - } - - if (delays) { - int* new_delays = (int*)STBI_REALLOC_SIZED(*delays, delays_size, sizeof(int) * layers); - if (!new_delays) - return stbi__load_gif_main_outofmem(&g, out, delays); - *delays = new_delays; - delays_size = layers * sizeof(int); - } - } else { - out = (stbi_uc*)stbi__malloc(layers * stride); - if (!out) - return stbi__load_gif_main_outofmem(&g, out, delays); - out_size = layers * stride; - if (delays) { - *delays = (int*)stbi__malloc(layers * sizeof(int)); - if (!*delays) - return stbi__load_gif_main_outofmem(&g, out, delays); - delays_size = layers * sizeof(int); - } - } - memcpy(out + ((layers - 1) * stride), u, stride); - if (layers >= 2) { - two_back = out - 2 * stride; - } - - if (delays) { - (*delays)[layers - 1U] = g.delay; - } - } - } while (u != 0); - - // free temp buffer; - STBI_FREE(g.out); - STBI_FREE(g.history); - STBI_FREE(g.background); - - // do the final conversion after loading everything; - if (req_comp && req_comp != 4) - out = stbi__convert_format(out, 4, req_comp, layers * g.w, g.h); - - *z = layers; - return out; - } else { - return stbi__errpuc("not GIF", "Image was not as a gif type."); - } -} - -static void* stbi__gif_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri) { - stbi_uc* u = 0; - stbi__gif g; - memset(&g, 0, sizeof(g)); - STBI_NOTUSED(ri); - - u = stbi__gif_load_next(s, &g, comp, req_comp, 0); - if (u == (stbi_uc*)s) - u = 0; // end of animated gif marker - if (u) { - *x = g.w; - *y = g.h; - - // moved conversion to after successful load so that the same - // can be done for multiple frames. - if (req_comp && req_comp != 4) - u = stbi__convert_format(u, 4, req_comp, g.w, g.h); - } else if (g.out) { - // if there was an error and we allocated an image buffer, free it! - STBI_FREE(g.out); - } - - // free buffers needed for multiple frame loading; - STBI_FREE(g.history); - STBI_FREE(g.background); - - return u; -} - -static int stbi__gif_info(stbi__context* s, int* x, int* y, int* comp) { - return stbi__gif_info_raw(s, x, y, comp); -} -#endif - -// ************************************************************************************************* -// Radiance RGBE HDR loader -// originally by Nicolas Schulz -#ifndef STBI_NO_HDR -static int stbi__hdr_test_core(stbi__context* s, const char* signature) { - int i; - for (i = 0; signature[i]; ++i) - if (stbi__get8(s) != signature[i]) - return 0; - stbi__rewind(s); - return 1; -} - -static int stbi__hdr_test(stbi__context* s) { - int r = stbi__hdr_test_core(s, "#?RADIANCE\n"); - stbi__rewind(s); - if (!r) { - r = stbi__hdr_test_core(s, "#?RGBE\n"); - stbi__rewind(s); - } - return r; -} - -#define STBI__HDR_BUFLEN 1024 -static char* stbi__hdr_gettoken(stbi__context* z, char* buffer) { - int len = 0; - char c = '\0'; - - c = (char)stbi__get8(z); - - while (!stbi__at_eof(z) && c != '\n') { - buffer[len++] = c; - if (len == STBI__HDR_BUFLEN - 1) { - // flush to end of line - while (!stbi__at_eof(z) && stbi__get8(z) != '\n') - ; - break; - } - c = (char)stbi__get8(z); - } - - buffer[len] = 0; - return buffer; -} - -static void stbi__hdr_convert(float* output, stbi_uc* input, int req_comp) { - if (input[3] != 0) { - float f1; - // Exponent - f1 = (float)ldexp(1.0f, input[3] - (int)(128 + 8)); - if (req_comp <= 2) - output[0] = (input[0] + input[1] + input[2]) * f1 / 3; - else { - output[0] = input[0] * f1; - output[1] = input[1] * f1; - output[2] = input[2] * f1; - } - if (req_comp == 2) - output[1] = 1; - if (req_comp == 4) - output[3] = 1; - } else { - switch (req_comp) { - case 4: - output[3] = 1; /* fallthrough */ - case 3: - output[0] = output[1] = output[2] = 0; - break; - case 2: - output[1] = 1; /* fallthrough */ - case 1: - output[0] = 0; - break; - } - } -} - -static float* stbi__hdr_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri) { - char buffer[STBI__HDR_BUFLEN]; - char* token; - int valid = 0; - int width, height; - stbi_uc* scanline; - float* hdr_data; - int len; - unsigned char count, value; - int i, j, k, c1, c2, z; - const char* headerToken; - STBI_NOTUSED(ri); - - // Check identifier - headerToken = stbi__hdr_gettoken(s, buffer); - if (strcmp(headerToken, "#?RADIANCE") != 0 && strcmp(headerToken, "#?RGBE") != 0) - return stbi__errpf("not HDR", "Corrupt HDR image"); - - // Parse header - for (;;) { - token = stbi__hdr_gettoken(s, buffer); - if (token[0] == 0) - break; - if (strcmp(token, "FORMAT=32-bit_rle_rgbe") == 0) - valid = 1; - } - - if (!valid) - return stbi__errpf("unsupported format", "Unsupported HDR format"); - - // Parse width and height - // can't use sscanf() if we're not using stdio! - token = stbi__hdr_gettoken(s, buffer); - if (strncmp(token, "-Y ", 3)) - return stbi__errpf("unsupported data layout", "Unsupported HDR format"); - token += 3; - height = (int)strtol(token, &token, 10); - while (*token == ' ') - ++token; - if (strncmp(token, "+X ", 3)) - return stbi__errpf("unsupported data layout", "Unsupported HDR format"); - token += 3; - width = (int)strtol(token, NULL, 10); - - if (height > STBI_MAX_DIMENSIONS) - return stbi__errpf("too large", "Very large image (corrupt?)"); - if (width > STBI_MAX_DIMENSIONS) - return stbi__errpf("too large", "Very large image (corrupt?)"); - - *x = width; - *y = height; - - if (comp) - *comp = 3; - if (req_comp == 0) - req_comp = 3; - - if (!stbi__mad4sizes_valid(width, height, req_comp, sizeof(float), 0)) - return stbi__errpf("too large", "HDR image is too large"); - - // Read data - hdr_data = (float*)stbi__malloc_mad4(width, height, req_comp, sizeof(float), 0); - if (!hdr_data) - return stbi__errpf("outofmem", "Out of memory"); - - // Load image data - // image data is stored as some number of sca - if (width < 8 || width >= 32768) { - // Read flat data - for (j = 0; j < height; ++j) { - for (i = 0; i < width; ++i) { - stbi_uc rgbe[4]; - main_decode_loop: - stbi__getn(s, rgbe, 4); - stbi__hdr_convert(hdr_data + j * width * req_comp + i * req_comp, rgbe, req_comp); - } - } - } else { - // Read RLE-encoded data - scanline = NULL; - - for (j = 0; j < height; ++j) { - c1 = stbi__get8(s); - c2 = stbi__get8(s); - len = stbi__get8(s); - if (c1 != 2 || c2 != 2 || (len & 0x80)) { - // not run-length encoded, so we have to actually use THIS data as a decoded - // pixel (note this can't be a valid pixel--one of RGB must be >= 128) - stbi_uc rgbe[4]; - rgbe[0] = (stbi_uc)c1; - rgbe[1] = (stbi_uc)c2; - rgbe[2] = (stbi_uc)len; - rgbe[3] = (stbi_uc)stbi__get8(s); - stbi__hdr_convert(hdr_data, rgbe, req_comp); - i = 1; - j = 0; - STBI_FREE(scanline); - goto main_decode_loop; // yes, this makes no sense - } - len <<= 8; - len |= stbi__get8(s); - if (len != width) { - STBI_FREE(hdr_data); - STBI_FREE(scanline); - return stbi__errpf("invalid decoded scanline length", "corrupt HDR"); - } - if (scanline == NULL) { - scanline = (stbi_uc*)stbi__malloc_mad2(width, 4, 0); - if (!scanline) { - STBI_FREE(hdr_data); - return stbi__errpf("outofmem", "Out of memory"); - } - } - - for (k = 0; k < 4; ++k) { - int nleft; - i = 0; - while ((nleft = width - i) > 0) { - count = stbi__get8(s); - if (count > 128) { - // Run - value = stbi__get8(s); - count -= 128; - if (count > nleft) { - STBI_FREE(hdr_data); - STBI_FREE(scanline); - return stbi__errpf("corrupt", "bad RLE data in HDR"); - } - for (z = 0; z < count; ++z) - scanline[i++ * 4 + k] = value; - } else { - // Dump - if (count > nleft) { - STBI_FREE(hdr_data); - STBI_FREE(scanline); - return stbi__errpf("corrupt", "bad RLE data in HDR"); - } - for (z = 0; z < count; ++z) - scanline[i++ * 4 + k] = stbi__get8(s); - } - } - } - for (i = 0; i < width; ++i) - stbi__hdr_convert(hdr_data + (j * width + i) * req_comp, scanline + i * 4, req_comp); - } - if (scanline) - STBI_FREE(scanline); - } - - return hdr_data; -} - -static int stbi__hdr_info(stbi__context* s, int* x, int* y, int* comp) { - char buffer[STBI__HDR_BUFLEN]; - char* token; - int valid = 0; - int dummy; - - if (!x) - x = &dummy; - if (!y) - y = &dummy; - if (!comp) - comp = &dummy; - - if (stbi__hdr_test(s) == 0) { - stbi__rewind(s); - return 0; - } - - for (;;) { - token = stbi__hdr_gettoken(s, buffer); - if (token[0] == 0) - break; - if (strcmp(token, "FORMAT=32-bit_rle_rgbe") == 0) - valid = 1; - } - - if (!valid) { - stbi__rewind(s); - return 0; - } - token = stbi__hdr_gettoken(s, buffer); - if (strncmp(token, "-Y ", 3)) { - stbi__rewind(s); - return 0; - } - token += 3; - *y = (int)strtol(token, &token, 10); - while (*token == ' ') - ++token; - if (strncmp(token, "+X ", 3)) { - stbi__rewind(s); - return 0; - } - token += 3; - *x = (int)strtol(token, NULL, 10); - *comp = 3; - return 1; -} -#endif // STBI_NO_HDR - -#ifndef STBI_NO_BMP -static int stbi__bmp_info(stbi__context* s, int* x, int* y, int* comp) { - void* p; - stbi__bmp_data info; - - info.all_a = 255; - p = stbi__bmp_parse_header(s, &info); - if (p == NULL) { - stbi__rewind(s); - return 0; - } - if (x) - *x = s->img_x; - if (y) - *y = s->img_y; - if (comp) { - if (info.bpp == 24 && info.ma == 0xff000000) - *comp = 3; - else - *comp = info.ma ? 4 : 3; - } - return 1; -} -#endif - -#ifndef STBI_NO_PSD -static int stbi__psd_info(stbi__context* s, int* x, int* y, int* comp) { - int channelCount, dummy, depth; - if (!x) - x = &dummy; - if (!y) - y = &dummy; - if (!comp) - comp = &dummy; - if (stbi__get32be(s) != 0x38425053) { - stbi__rewind(s); - return 0; - } - if (stbi__get16be(s) != 1) { - stbi__rewind(s); - return 0; - } - stbi__skip(s, 6); - channelCount = stbi__get16be(s); - if (channelCount < 0 || channelCount > 16) { - stbi__rewind(s); - return 0; - } - *y = stbi__get32be(s); - *x = stbi__get32be(s); - depth = stbi__get16be(s); - if (depth != 8 && depth != 16) { - stbi__rewind(s); - return 0; - } - if (stbi__get16be(s) != 3) { - stbi__rewind(s); - return 0; - } - *comp = 4; - return 1; -} - -static int stbi__psd_is16(stbi__context* s) { - int channelCount, depth; - if (stbi__get32be(s) != 0x38425053) { - stbi__rewind(s); - return 0; - } - if (stbi__get16be(s) != 1) { - stbi__rewind(s); - return 0; - } - stbi__skip(s, 6); - channelCount = stbi__get16be(s); - if (channelCount < 0 || channelCount > 16) { - stbi__rewind(s); - return 0; - } - STBI_NOTUSED(stbi__get32be(s)); - STBI_NOTUSED(stbi__get32be(s)); - depth = stbi__get16be(s); - if (depth != 16) { - stbi__rewind(s); - return 0; - } - return 1; -} -#endif - -#ifndef STBI_NO_PIC -static int stbi__pic_info(stbi__context* s, int* x, int* y, int* comp) { - int act_comp = 0, num_packets = 0, chained, dummy; - stbi__pic_packet packets[10]; - - if (!x) - x = &dummy; - if (!y) - y = &dummy; - if (!comp) - comp = &dummy; - - if (!stbi__pic_is4(s, "\x53\x80\xF6\x34")) { - stbi__rewind(s); - return 0; - } - - stbi__skip(s, 88); - - *x = stbi__get16be(s); - *y = stbi__get16be(s); - if (stbi__at_eof(s)) { - stbi__rewind(s); - return 0; - } - if ((*x) != 0 && (1 << 28) / (*x) < (*y)) { - stbi__rewind(s); - return 0; - } - - stbi__skip(s, 8); - - do { - stbi__pic_packet* packet; - - if (num_packets == sizeof(packets) / sizeof(packets[0])) - return 0; - - packet = &packets[num_packets++]; - chained = stbi__get8(s); - packet->size = stbi__get8(s); - packet->type = stbi__get8(s); - packet->channel = stbi__get8(s); - act_comp |= packet->channel; - - if (stbi__at_eof(s)) { - stbi__rewind(s); - return 0; - } - if (packet->size != 8) { - stbi__rewind(s); - return 0; - } - } while (chained); - - *comp = (act_comp & 0x10 ? 4 : 3); - - return 1; -} -#endif - -// ************************************************************************************************* -// Portable Gray Map and Portable Pixel Map loader -// by Ken Miller -// -// PGM: http://netpbm.sourceforge.net/doc/pgm.html -// PPM: http://netpbm.sourceforge.net/doc/ppm.html -// -// Known limitations: -// Does not support comments in the header section -// Does not support ASCII image data (formats P2 and P3) - -#ifndef STBI_NO_PNM - -static int stbi__pnm_test(stbi__context* s) { - char p, t; - p = (char)stbi__get8(s); - t = (char)stbi__get8(s); - if (p != 'P' || (t != '5' && t != '6')) { - stbi__rewind(s); - return 0; - } - return 1; -} - -static void* stbi__pnm_load(stbi__context* s, - int* x, - int* y, - int* comp, - int req_comp, - stbi__result_info* ri) { - stbi_uc* out; - STBI_NOTUSED(ri); - - ri->bits_per_channel = stbi__pnm_info(s, (int*)&s->img_x, (int*)&s->img_y, (int*)&s->img_n); - if (ri->bits_per_channel == 0) - return 0; - - if (s->img_y > STBI_MAX_DIMENSIONS) - return stbi__errpuc("too large", "Very large image (corrupt?)"); - if (s->img_x > STBI_MAX_DIMENSIONS) - return stbi__errpuc("too large", "Very large image (corrupt?)"); - - *x = s->img_x; - *y = s->img_y; - if (comp) - *comp = s->img_n; - - if (!stbi__mad4sizes_valid(s->img_n, s->img_x, s->img_y, ri->bits_per_channel / 8, 0)) - return stbi__errpuc("too large", "PNM too large"); - - out = (stbi_uc*)stbi__malloc_mad4(s->img_n, s->img_x, s->img_y, ri->bits_per_channel / 8, 0); - if (!out) - return stbi__errpuc("outofmem", "Out of memory"); - stbi__getn(s, out, s->img_n * s->img_x * s->img_y * (ri->bits_per_channel / 8)); - - if (req_comp && req_comp != s->img_n) { - out = stbi__convert_format(out, s->img_n, req_comp, s->img_x, s->img_y); - if (out == NULL) - return out; // stbi__convert_format frees input on failure - } - return out; -} - -static int stbi__pnm_isspace(char c) { - return c == ' ' || c == '\t' || c == '\n' || c == '\v' || c == '\f' || c == '\r'; -} - -static void stbi__pnm_skip_whitespace(stbi__context* s, char* c) { - for (;;) { - while (!stbi__at_eof(s) && stbi__pnm_isspace(*c)) - *c = (char)stbi__get8(s); - - if (stbi__at_eof(s) || *c != '#') - break; - - while (!stbi__at_eof(s) && *c != '\n' && *c != '\r') - *c = (char)stbi__get8(s); - } -} - -static int stbi__pnm_isdigit(char c) { - return c >= '0' && c <= '9'; -} - -static int stbi__pnm_getinteger(stbi__context* s, char* c) { - int value = 0; - - while (!stbi__at_eof(s) && stbi__pnm_isdigit(*c)) { - value = value * 10 + (*c - '0'); - *c = (char)stbi__get8(s); - } - - return value; -} - -static int stbi__pnm_info(stbi__context* s, int* x, int* y, int* comp) { - int maxv, dummy; - char c, p, t; - - if (!x) - x = &dummy; - if (!y) - y = &dummy; - if (!comp) - comp = &dummy; - - stbi__rewind(s); - - // Get identifier - p = (char)stbi__get8(s); - t = (char)stbi__get8(s); - if (p != 'P' || (t != '5' && t != '6')) { - stbi__rewind(s); - return 0; - } - - *comp = (t == '6') ? 3 : 1; // '5' is 1-component .pgm; '6' is 3-component .ppm - - c = (char)stbi__get8(s); - stbi__pnm_skip_whitespace(s, &c); - - *x = stbi__pnm_getinteger(s, &c); // read width - stbi__pnm_skip_whitespace(s, &c); - - *y = stbi__pnm_getinteger(s, &c); // read height - stbi__pnm_skip_whitespace(s, &c); - - maxv = stbi__pnm_getinteger(s, &c); // read max value - if (maxv > 65535) - return stbi__err("max value > 65535", "PPM image supports only 8-bit and 16-bit images"); - else if (maxv > 255) - return 16; - else - return 8; -} - -static int stbi__pnm_is16(stbi__context* s) { - if (stbi__pnm_info(s, NULL, NULL, NULL) == 16) - return 1; - return 0; -} -#endif - -static int stbi__info_main(stbi__context* s, int* x, int* y, int* comp) { -#ifndef STBI_NO_JPEG - if (stbi__jpeg_info(s, x, y, comp)) - return 1; -#endif - -#ifndef STBI_NO_PNG - if (stbi__png_info(s, x, y, comp)) - return 1; -#endif - -#ifndef STBI_NO_GIF - if (stbi__gif_info(s, x, y, comp)) - return 1; -#endif - -#ifndef STBI_NO_BMP - if (stbi__bmp_info(s, x, y, comp)) - return 1; -#endif - -#ifndef STBI_NO_PSD - if (stbi__psd_info(s, x, y, comp)) - return 1; -#endif - -#ifndef STBI_NO_PIC - if (stbi__pic_info(s, x, y, comp)) - return 1; -#endif - -#ifndef STBI_NO_PNM - if (stbi__pnm_info(s, x, y, comp)) - return 1; -#endif - -#ifndef STBI_NO_HDR - if (stbi__hdr_info(s, x, y, comp)) - return 1; -#endif - -// test tga last because it's a crappy test! -#ifndef STBI_NO_TGA - if (stbi__tga_info(s, x, y, comp)) - return 1; -#endif - return stbi__err("unknown image type", "Image not of any known type, or corrupt"); -} - -static int stbi__is_16_main(stbi__context* s) { -#ifndef STBI_NO_PNG - if (stbi__png_is16(s)) - return 1; -#endif - -#ifndef STBI_NO_PSD - if (stbi__psd_is16(s)) - return 1; -#endif - -#ifndef STBI_NO_PNM - if (stbi__pnm_is16(s)) - return 1; -#endif - return 0; -} - -#ifndef STBI_NO_STDIO -STBIDEF int stbi_info(char const* filename, int* x, int* y, int* comp) { - FILE* f = stbi__fopen(filename, "rb"); - int result; - if (!f) - return stbi__err("can't fopen", "Unable to open file"); - result = stbi_info_from_file(f, x, y, comp); - fclose(f); - return result; -} - -STBIDEF int stbi_info_from_file(FILE* f, int* x, int* y, int* comp) { - int r; - stbi__context s; - long pos = ftell(f); - stbi__start_file(&s, f); - r = stbi__info_main(&s, x, y, comp); - fseek(f, pos, SEEK_SET); - return r; -} - -STBIDEF int stbi_is_16_bit(char const* filename) { - FILE* f = stbi__fopen(filename, "rb"); - int result; - if (!f) - return stbi__err("can't fopen", "Unable to open file"); - result = stbi_is_16_bit_from_file(f); - fclose(f); - return result; -} - -STBIDEF int stbi_is_16_bit_from_file(FILE* f) { - int r; - stbi__context s; - long pos = ftell(f); - stbi__start_file(&s, f); - r = stbi__is_16_main(&s); - fseek(f, pos, SEEK_SET); - return r; -} -#endif // !STBI_NO_STDIO - -STBIDEF int stbi_info_from_memory(stbi_uc const* buffer, int len, int* x, int* y, int* comp) { - stbi__context s; - stbi__start_mem(&s, buffer, len); - return stbi__info_main(&s, x, y, comp); -} - -STBIDEF int stbi_info_from_callbacks(stbi_io_callbacks const* c, - void* user, - int* x, - int* y, - int* comp) { - stbi__context s; - stbi__start_callbacks(&s, (stbi_io_callbacks*)c, user); - return stbi__info_main(&s, x, y, comp); -} - -STBIDEF int stbi_is_16_bit_from_memory(stbi_uc const* buffer, int len) { - stbi__context s; - stbi__start_mem(&s, buffer, len); - return stbi__is_16_main(&s); -} - -STBIDEF int stbi_is_16_bit_from_callbacks(stbi_io_callbacks const* c, void* user) { - stbi__context s; - stbi__start_callbacks(&s, (stbi_io_callbacks*)c, user); - return stbi__is_16_main(&s); -} - -#endif // STB_IMAGE_IMPLEMENTATION - -/* - revision history: - 2.20 (2019-02-07) support utf8 filenames in Windows; fix warnings and platform ifdefs - 2.19 (2018-02-11) fix warning - 2.18 (2018-01-30) fix warnings - 2.17 (2018-01-29) change sbti__shiftsigned to avoid clang -O2 bug - 1-bit BMP - *_is_16_bit api - avoid warnings - 2.16 (2017-07-23) all functions have 16-bit variants; - STBI_NO_STDIO works again; - compilation fixes; - fix rounding in unpremultiply; - optimize vertical flip; - disable raw_len validation; - documentation fixes - 2.15 (2017-03-18) fix png-1,2,4 bug; now all Imagenet JPGs decode; - warning fixes; disable run-time SSE detection on gcc; - uniform handling of optional "return" values; - thread-safe initialization of zlib tables - 2.14 (2017-03-03) remove deprecated STBI_JPEG_OLD; fixes for Imagenet JPGs - 2.13 (2016-11-29) add 16-bit API, only supported for PNG right now - 2.12 (2016-04-02) fix typo in 2.11 PSD fix that caused crashes - 2.11 (2016-04-02) allocate large structures on the stack - remove white matting for transparent PSD - fix reported channel count for PNG & BMP - re-enable SSE2 in non-gcc 64-bit - support RGB-formatted JPEG - read 16-bit PNGs (only as 8-bit) - 2.10 (2016-01-22) avoid warning introduced in 2.09 by STBI_REALLOC_SIZED - 2.09 (2016-01-16) allow comments in PNM files - 16-bit-per-pixel TGA (not bit-per-component) - info() for TGA could break due to .hdr handling - info() for BMP to shares code instead of sloppy parse - can use STBI_REALLOC_SIZED if allocator doesn't support realloc - code cleanup - 2.08 (2015-09-13) fix to 2.07 cleanup, reading RGB PSD as RGBA - 2.07 (2015-09-13) fix compiler warnings - partial animated GIF support - limited 16-bpc PSD support - #ifdef unused functions - bug with < 92 byte PIC,PNM,HDR,TGA - 2.06 (2015-04-19) fix bug where PSD returns wrong '*comp' value - 2.05 (2015-04-19) fix bug in progressive JPEG handling, fix warning - 2.04 (2015-04-15) try to re-enable SIMD on MinGW 64-bit - 2.03 (2015-04-12) extra corruption checking (mmozeiko) - stbi_set_flip_vertically_on_load (nguillemot) - fix NEON support; fix mingw support - 2.02 (2015-01-19) fix incorrect assert, fix warning - 2.01 (2015-01-17) fix various warnings; suppress SIMD on gcc 32-bit without -msse2 - 2.00b (2014-12-25) fix STBI_MALLOC in progressive JPEG - 2.00 (2014-12-25) optimize JPG, including x86 SSE2 & NEON SIMD (ryg) - progressive JPEG (stb) - PGM/PPM support (Ken Miller) - STBI_MALLOC,STBI_REALLOC,STBI_FREE - GIF bugfix -- seemingly never worked - STBI_NO_*, STBI_ONLY_* - 1.48 (2014-12-14) fix incorrectly-named assert() - 1.47 (2014-12-14) 1/2/4-bit PNG support, both direct and paletted (Omar Cornut & stb) - optimize PNG (ryg) - fix bug in interlaced PNG with user-specified channel count (stb) - 1.46 (2014-08-26) - fix broken tRNS chunk (colorkey-style transparency) in non-paletted PNG - 1.45 (2014-08-16) - fix MSVC-ARM internal compiler error by wrapping malloc - 1.44 (2014-08-07) - various warning fixes from Ronny Chevalier - 1.43 (2014-07-15) - fix MSVC-only compiler problem in code changed in 1.42 - 1.42 (2014-07-09) - don't define _CRT_SECURE_NO_WARNINGS (affects user code) - fixes to stbi__cleanup_jpeg path - added STBI_ASSERT to avoid requiring assert.h - 1.41 (2014-06-25) - fix search&replace from 1.36 that messed up comments/error messages - 1.40 (2014-06-22) - fix gcc struct-initialization warning - 1.39 (2014-06-15) - fix to TGA optimization when req_comp != number of components in TGA; - fix to GIF loading because BMP wasn't rewinding (whoops, no GIFs in my test suite) - add support for BMP version 5 (more ignored fields) - 1.38 (2014-06-06) - suppress MSVC warnings on integer casts truncating values - fix accidental rename of 'skip' field of I/O - 1.37 (2014-06-04) - remove duplicate typedef - 1.36 (2014-06-03) - convert to header file single-file library - if de-iphone isn't set, load iphone images color-swapped instead of returning NULL - 1.35 (2014-05-27) - various warnings - fix broken STBI_SIMD path - fix bug where stbi_load_from_file no longer left file pointer in correct place - fix broken non-easy path for 32-bit BMP (possibly never used) - TGA optimization by Arseny Kapoulkine - 1.34 (unknown) - use STBI_NOTUSED in stbi__resample_row_generic(), fix one more leak in tga failure - case 1.33 (2011-07-14) make stbi_is_hdr work in STBI_NO_HDR (as specified), minor - compiler-friendly improvements 1.32 (2011-07-13) support for "info" function for all supported - filetypes (SpartanJ) 1.31 (2011-06-20) a few more leak fixes, bug in PNG handling (SpartanJ) - 1.30 (2011-06-11) - added ability to load files via callbacks to accomidate custom input streams (Ben - Wenger) removed deprecated format-specific test/load functions removed support for installable - file formats (stbi_loader) -- would have been broken for IO callbacks anyway error cases in bmp - and tga give messages and don't leak (Raymond Barbiero, grisha) fix inefficiency in decoding - 32-bit BMP (David Woo) 1.29 (2010-08-16) various warning fixes from Aurelien Pocheville 1.28 - (2010-08-01) fix bug in GIF palette transparency (SpartanJ) 1.27 (2010-08-01) cast-to-stbi_uc to - fix warnings 1.26 (2010-07-24) fix bug in file buffering for PNG reported by SpartanJ 1.25 - (2010-07-17) refix trans_data warning (Won Chun) 1.24 (2010-07-12) perf improvements reading - from files on platforms with lock-heavy fgetc() minor perf improvements for jpeg deprecated - type-specific functions so we'll get feedback if they're needed attempt to fix trans_data warning - (Won Chun) 1.23 fixed bug in iPhone support 1.22 (2010-07-10) removed image *writing* support - stbi_info support from Jetro Lauha - GIF support from Jean-Marc Lienher - iPhone PNG-extensions from James Brown - warning-fixes from Nicolas Schulz and Janez Zemva (i.stbi__err. Janez (U+017D)emva) - 1.21 fix use of 'stbi_uc' in header (reported by jon blow) - 1.20 added support for Softimage PIC, by Tom Seddon - 1.19 bug in interlaced PNG corruption check (found by ryg) - 1.18 (2008-08-02) - fix a threading bug (local mutable static) - 1.17 support interlaced PNG - 1.16 major bugfix - stbi__convert_format converted one too many pixels - 1.15 initialize some fields for thread safety - 1.14 fix threadsafe conversion bug - header-file-only version (#define STBI_HEADER_FILE_ONLY before including) - 1.13 threadsafe - 1.12 const qualifiers in the API - 1.11 Support installable IDCT, colorspace conversion routines - 1.10 Fixes for 64-bit (don't use "unsigned long") - optimized upsampling by Fabian "ryg" Giesen - 1.09 Fix format-conversion for PSD code (bad global variables!) - 1.08 Thatcher Ulrich's PSD code integrated by Nicolas Schulz - 1.07 attempt to fix C++ warning/errors again - 1.06 attempt to fix C++ warning/errors again - 1.05 fix TGA loading to return correct *comp and use good luminance calc - 1.04 default float alpha is 1, not 255; use 'void *' for stbi_image_free - 1.03 bugfixes to STBI_NO_STDIO, STBI_NO_HDR - 1.02 support for (subset of) HDR files, float interface for preferred access to them - 1.01 fix bug: possible bug in handling right-side up bmps... not sure - fix bug: the stbi__bmp_load() and stbi__tga_load() functions didn't work at all - 1.00 interface to zlib that skips zlib header - 0.99 correct handling of alpha in palette - 0.98 TGA loader by lonesock; dynamically add loaders (untested) - 0.97 jpeg errors on too large a file; also catch another malloc failure - 0.96 fix detection of invalid v value - particleman@mollyrocket forum - 0.95 during header scan, seek to markers in case of padding - 0.94 STBI_NO_STDIO to disable stdio usage; rename all #defines the same - 0.93 handle jpegtran output; verbose errors - 0.92 read 4,8,16,24,32-bit BMP files of several formats - 0.91 output 24-bit Windows 3.0 BMP files - 0.90 fix a few more warnings; bump version number to approach 1.0 - 0.61 bugfixes due to Marc LeBlanc, Christopher Lloyd - 0.60 fix compiling as c++ - 0.59 fix warnings: merge Dave Moore's -Wall fixes - 0.58 fix bug: zlib uncompressed mode len/nlen was wrong endian - 0.57 fix bug: jpg last huffman symbol before marker was >9 bits but less than 16 available - 0.56 fix bug: zlib uncompressed mode len vs. nlen - 0.55 fix bug: restart_interval not initialized to 0 - 0.54 allow NULL for 'int *comp' - 0.53 fix bug in png 3->4; speedup png decoding - 0.52 png handles req_comp=3,4 directly; minor cleanup; jpeg comments - 0.51 obey req_comp requests, 1-component jpegs return as 1-component, - on 'test' only check type, not whether we support this variant - 0.50 (2006-11-19) - first released version -*/ - -/* ------------------------------------------------------------------------------- -This software is available under 2 licenses -- choose whichever you prefer. ------------------------------------------------------------------------------- -ALTERNATIVE A - MIT License -Copyright (c) 2017 Sean Barrett -Permission is hereby granted, free of charge, to any person obtaining a copy of -this software and associated documentation files (the "Software"), to deal in -the Software without restriction, including without limitation the rights to -use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies -of the Software, and to permit persons to whom the Software is furnished to do -so, subject to the following conditions: -The above copyright notice and this permission notice shall be included in all -copies or substantial portions of the Software. -THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER -LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE -SOFTWARE. ------------------------------------------------------------------------------- -ALTERNATIVE B - Public Domain (www.unlicense.org) -This is free and unencumbered software released into the public domain. -Anyone is free to copy, modify, publish, use, compile, sell, or distribute this -software, either in source code form or as a compiled binary, for any purpose, -commercial or non-commercial, and by any means. -In jurisdictions that recognize copyright laws, the author or authors of this -software dedicate any and all copyright interest in the software to the public -domain. We make this dedication for the benefit of the public at large and to -the detriment of our heirs and successors. We intend this dedication to be an -overt act of relinquishment in perpetuity of all present and future rights to -this software under copyright law. -THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE -AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN -ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION -WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. ------------------------------------------------------------------------------- -*/ diff --git a/third-party/stb_image/CMakeLists.txt b/third-party/stb_image/CMakeLists.txt new file mode 100644 index 0000000000..f390f23512 --- /dev/null +++ b/third-party/stb_image/CMakeLists.txt @@ -0,0 +1,9 @@ +if (UNIX) + set(CMAKE_CXX_FLAGS "-O3") +else () + set(CMAKE_CXX_FLAGS "/EHsc") +endif (UNIX) + + +add_library(stb_image stb_image.cpp) + diff --git a/third-party/stb_image/stb_image.cpp b/third-party/stb_image/stb_image.cpp new file mode 100644 index 0000000000..80d8f9af64 --- /dev/null +++ b/third-party/stb_image/stb_image.cpp @@ -0,0 +1,4 @@ +#define STB_IMAGE_IMPLEMENTATION +#define STB_IMAGE_WRITE_IMPLEMENTATION +#include "stb_image_write.h" +#include "stb_image.h" \ No newline at end of file diff --git a/third-party/stb_image/stb_image.h b/third-party/stb_image/stb_image.h new file mode 100644 index 0000000000..db2b1bcdf4 --- /dev/null +++ b/third-party/stb_image/stb_image.h @@ -0,0 +1,7530 @@ +/* stb_image - v2.21 - public domain image loader - http://nothings.org/stb + no warranty implied; use at your own risk + + Do this: + #define STB_IMAGE_IMPLEMENTATION + before you include this file in *one* C or C++ file to create the implementation. + + // i.e. it should look like this: + #include ... + #include ... + #include ... + #define STB_IMAGE_IMPLEMENTATION + #include "stb_image.h" + + You can #define STBI_ASSERT(x) before the #include to avoid using assert.h. + And #define STBI_MALLOC, STBI_REALLOC, and STBI_FREE to avoid using malloc,realloc,free + + + QUICK NOTES: + Primarily of interest to game developers and other people who can + avoid problematic images and only need the trivial interface + + JPEG baseline & progressive (12 bpc/arithmetic not supported, same as stock IJG lib) + PNG 1/2/4/8/16-bit-per-channel + + TGA (not sure what subset, if a subset) + BMP non-1bpp, non-RLE + PSD (composited view only, no extra channels, 8/16 bit-per-channel) + + GIF (*comp always reports as 4-channel) + HDR (radiance rgbE format) + PIC (Softimage PIC) + PNM (PPM and PGM binary only) + + Animated GIF still needs a proper API, but here's one way to do it: + http://gist.github.com/urraka/685d9a6340b26b830d49 + + - decode from memory or through FILE (define STBI_NO_STDIO to remove code) + - decode from arbitrary I/O callbacks + - SIMD acceleration on x86/x64 (SSE2) and ARM (NEON) + + Full documentation under "DOCUMENTATION" below. + + +LICENSE + + See end of file for license information. + +RECENT REVISION HISTORY: + + 2.21 (2019-02-25) fix typo in comment + 2.20 (2019-02-07) support utf8 filenames in Windows; fix warnings and platform ifdefs + 2.19 (2018-02-11) fix warning + 2.18 (2018-01-30) fix warnings + 2.17 (2018-01-29) bugfix, 1-bit BMP, 16-bitness query, fix warnings + 2.16 (2017-07-23) all functions have 16-bit variants; optimizations; bugfixes + 2.15 (2017-03-18) fix png-1,2,4; all Imagenet JPGs; no runtime SSE detection on GCC + 2.14 (2017-03-03) remove deprecated STBI_JPEG_OLD; fixes for Imagenet JPGs + 2.13 (2016-12-04) experimental 16-bit API, only for PNG so far; fixes + 2.12 (2016-04-02) fix typo in 2.11 PSD fix that caused crashes + 2.11 (2016-04-02) 16-bit PNGS; enable SSE2 in non-gcc x64 + RGB-format JPEG; remove white matting in PSD; + allocate large structures on the stack; + correct channel count for PNG & BMP + 2.10 (2016-01-22) avoid warning introduced in 2.09 + 2.09 (2016-01-16) 16-bit TGA; comments in PNM files; STBI_REALLOC_SIZED + + See end of file for full revision history. + + + ============================ Contributors ========================= + + Image formats Extensions, features + Sean Barrett (jpeg, png, bmp) Jetro Lauha (stbi_info) + Nicolas Schulz (hdr, psd) Martin "SpartanJ" Golini (stbi_info) + Jonathan Dummer (tga) James "moose2000" Brown (iPhone PNG) + Jean-Marc Lienher (gif) Ben "Disch" Wenger (io callbacks) + Tom Seddon (pic) Omar Cornut (1/2/4-bit PNG) + Thatcher Ulrich (psd) Nicolas Guillemot (vertical flip) + Ken Miller (pgm, ppm) Richard Mitton (16-bit PSD) + github:urraka (animated gif) Junggon Kim (PNM comments) + Christopher Forseth (animated gif) Daniel Gibson (16-bit TGA) + socks-the-fox (16-bit PNG) + Jeremy Sawicki (handle all ImageNet JPGs) + Optimizations & bugfixes Mikhail Morozov (1-bit BMP) + Fabian "ryg" Giesen Anael Seghezzi (is-16-bit query) + Arseny Kapoulkine + John-Mark Allen + Carmelo J Fdez-Aguera + + Bug & warning fixes + Marc LeBlanc David Woo Guillaume George Martins Mozeiko + Christpher Lloyd Jerry Jansson Joseph Thomson Phil Jordan + Dave Moore Roy Eltham Hayaki Saito Nathan Reed + Won Chun Luke Graham Johan Duparc Nick Verigakis + the Horde3D community Thomas Ruf Ronny Chevalier github:rlyeh + Janez Zemva John Bartholomew Michal Cichon github:romigrou + Jonathan Blow Ken Hamada Tero Hanninen github:svdijk + Laurent Gomila Cort Stratton Sergio Gonzalez github:snagar + Aruelien Pocheville Thibault Reuille Cass Everitt github:Zelex + Ryamond Barbiero Paul Du Bois Engin Manap github:grim210 + Aldo Culquicondor Philipp Wiesemann Dale Weiler github:sammyhw + Oriol Ferrer Mesia Josh Tobin Matthew Gregan github:phprus + Julian Raschke Gregory Mullen Baldur Karlsson github:poppolopoppo + Christian Floisand Kevin Schmidt JR Smith github:darealshinji + Blazej Dariusz Roszkowski github:Michaelangel007 +*/ + +#ifndef STBI_INCLUDE_STB_IMAGE_H +#define STBI_INCLUDE_STB_IMAGE_H + +// DOCUMENTATION +// +// Limitations: +// - no 12-bit-per-channel JPEG +// - no JPEGs with arithmetic coding +// - GIF always returns *comp=4 +// +// Basic usage (see HDR discussion below for HDR usage): +// int x,y,n; +// unsigned char *data = stbi_load(filename, &x, &y, &n, 0); +// // ... process data if not NULL ... +// // ... x = width, y = height, n = # 8-bit components per pixel ... +// // ... replace '0' with '1'..'4' to force that many components per pixel +// // ... but 'n' will always be the number that it would have been if you said 0 +// stbi_image_free(data) +// +// Standard parameters: +// int *x -- outputs image width in pixels +// int *y -- outputs image height in pixels +// int *channels_in_file -- outputs # of image components in image file +// int desired_channels -- if non-zero, # of image components requested in result +// +// The return value from an image loader is an 'unsigned char *' which points +// to the pixel data, or NULL on an allocation failure or if the image is +// corrupt or invalid. The pixel data consists of *y scanlines of *x pixels, +// with each pixel consisting of N interleaved 8-bit components; the first +// pixel pointed to is top-left-most in the image. There is no padding between +// image scanlines or between pixels, regardless of format. The number of +// components N is 'desired_channels' if desired_channels is non-zero, or +// *channels_in_file otherwise. If desired_channels is non-zero, +// *channels_in_file has the number of components that _would_ have been +// output otherwise. E.g. if you set desired_channels to 4, you will always +// get RGBA output, but you can check *channels_in_file to see if it's trivially +// opaque because e.g. there were only 3 channels in the source image. +// +// An output image with N components has the following components interleaved +// in this order in each pixel: +// +// N=#comp components +// 1 grey +// 2 grey, alpha +// 3 red, green, blue +// 4 red, green, blue, alpha +// +// If image loading fails for any reason, the return value will be NULL, +// and *x, *y, *channels_in_file will be unchanged. The function +// stbi_failure_reason() can be queried for an extremely brief, end-user +// unfriendly explanation of why the load failed. Define STBI_NO_FAILURE_STRINGS +// to avoid compiling these strings at all, and STBI_FAILURE_USERMSG to get slightly +// more user-friendly ones. +// +// Paletted PNG, BMP, GIF, and PIC images are automatically depalettized. +// +// =========================================================================== +// +// UNICODE: +// +// If compiling for Windows and you wish to use Unicode filenames, compile +// with +// #define STBI_WINDOWS_UTF8 +// and pass utf8-encoded filenames. Call stbi_convert_wchar_to_utf8 to convert +// Windows wchar_t filenames to utf8. +// +// =========================================================================== +// +// Philosophy +// +// stb libraries are designed with the following priorities: +// +// 1. easy to use +// 2. easy to maintain +// 3. good performance +// +// Sometimes I let "good performance" creep up in priority over "easy to maintain", +// and for best performance I may provide less-easy-to-use APIs that give higher +// performance, in addition to the easy-to-use ones. Nevertheless, it's important +// to keep in mind that from the standpoint of you, a client of this library, +// all you care about is #1 and #3, and stb libraries DO NOT emphasize #3 above all. +// +// Some secondary priorities arise directly from the first two, some of which +// provide more explicit reasons why performance can't be emphasized. +// +// - Portable ("ease of use") +// - Small source code footprint ("easy to maintain") +// - No dependencies ("ease of use") +// +// =========================================================================== +// +// I/O callbacks +// +// I/O callbacks allow you to read from arbitrary sources, like packaged +// files or some other source. Data read from callbacks are processed +// through a small internal buffer (currently 128 bytes) to try to reduce +// overhead. +// +// The three functions you must define are "read" (reads some bytes of data), +// "skip" (skips some bytes of data), "eof" (reports if the stream is at the end). +// +// =========================================================================== +// +// SIMD support +// +// The JPEG decoder will try to automatically use SIMD kernels on x86 when +// supported by the compiler. For ARM Neon support, you must explicitly +// request it. +// +// (The old do-it-yourself SIMD API is no longer supported in the current +// code.) +// +// On x86, SSE2 will automatically be used when available based on a run-time +// test; if not, the generic C versions are used as a fall-back. On ARM targets, +// the typical path is to have separate builds for NEON and non-NEON devices +// (at least this is true for iOS and Android). Therefore, the NEON support is +// toggled by a build flag: define STBI_NEON to get NEON loops. +// +// If for some reason you do not want to use any of SIMD code, or if +// you have issues compiling it, you can disable it entirely by +// defining STBI_NO_SIMD. +// +// =========================================================================== +// +// HDR image support (disable by defining STBI_NO_HDR) +// +// stb_image supports loading HDR images in general, and currently the Radiance +// .HDR file format specifically. You can still load any file through the existing +// interface; if you attempt to load an HDR file, it will be automatically remapped +// to LDR, assuming gamma 2.2 and an arbitrary scale factor defaulting to 1; +// both of these constants can be reconfigured through this interface: +// +// stbi_hdr_to_ldr_gamma(2.2f); +// stbi_hdr_to_ldr_scale(1.0f); +// +// (note, do not use _inverse_ constants; stbi_image will invert them +// appropriately). +// +// Additionally, there is a new, parallel interface for loading files as +// (linear) floats to preserve the full dynamic range: +// +// float *data = stbi_loadf(filename, &x, &y, &n, 0); +// +// If you load LDR images through this interface, those images will +// be promoted to floating point values, run through the inverse of +// constants corresponding to the above: +// +// stbi_ldr_to_hdr_scale(1.0f); +// stbi_ldr_to_hdr_gamma(2.2f); +// +// Finally, given a filename (or an open file or memory block--see header +// file for details) containing image data, you can query for the "most +// appropriate" interface to use (that is, whether the image is HDR or +// not), using: +// +// stbi_is_hdr(char *filename); +// +// =========================================================================== +// +// iPhone PNG support: +// +// By default we convert iphone-formatted PNGs back to RGB, even though +// they are internally encoded differently. You can disable this conversion +// by calling stbi_convert_iphone_png_to_rgb(0), in which case +// you will always just get the native iphone "format" through (which +// is BGR stored in RGB). +// +// Call stbi_set_unpremultiply_on_load(1) as well to force a divide per +// pixel to remove any premultiplied alpha *only* if the image file explicitly +// says there's premultiplied data (currently only happens in iPhone images, +// and only if iPhone convert-to-rgb processing is on). +// +// =========================================================================== +// +// ADDITIONAL CONFIGURATION +// +// - You can suppress implementation of any of the decoders to reduce +// your code footprint by #defining one or more of the following +// symbols before creating the implementation. +// +// STBI_NO_JPEG +// STBI_NO_PNG +// STBI_NO_BMP +// STBI_NO_PSD +// STBI_NO_TGA +// STBI_NO_GIF +// STBI_NO_HDR +// STBI_NO_PIC +// STBI_NO_PNM (.ppm and .pgm) +// +// - You can request *only* certain decoders and suppress all other ones +// (this will be more forward-compatible, as addition of new decoders +// doesn't require you to disable them explicitly): +// +// STBI_ONLY_JPEG +// STBI_ONLY_PNG +// STBI_ONLY_BMP +// STBI_ONLY_PSD +// STBI_ONLY_TGA +// STBI_ONLY_GIF +// STBI_ONLY_HDR +// STBI_ONLY_PIC +// STBI_ONLY_PNM (.ppm and .pgm) +// +// - If you use STBI_NO_PNG (or _ONLY_ without PNG), and you still +// want the zlib decoder to be available, #define STBI_SUPPORT_ZLIB +// + + +#ifndef STBI_NO_STDIO +#include +#endif // STBI_NO_STDIO + +#define STBI_VERSION 1 + +enum +{ + STBI_default = 0, // only used for desired_channels + + STBI_grey = 1, + STBI_grey_alpha = 2, + STBI_rgb = 3, + STBI_rgb_alpha = 4 +}; + +#include +typedef unsigned char stbi_uc; +typedef unsigned short stbi_us; + +#ifdef __cplusplus +extern "C" { +#endif + +#ifdef STB_IMAGE_STATIC +#define STBIDEF static +#else +#define STBIDEF extern +#endif + +////////////////////////////////////////////////////////////////////////////// +// +// PRIMARY API - works on images of any type +// + +// +// load image by filename, open file, or memory buffer +// + +typedef struct +{ + int (*read) (void *user,char *data,int size); // fill 'data' with 'size' bytes. return number of bytes actually read + void (*skip) (void *user,int n); // skip the next 'n' bytes, or 'unget' the last -n bytes if negative + int (*eof) (void *user); // returns nonzero if we are at end of file/data +} stbi_io_callbacks; + +//////////////////////////////////// +// +// 8-bits-per-channel interface +// + +STBIDEF stbi_uc *stbi_load_from_memory (stbi_uc const *buffer, int len , int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_uc *stbi_load_from_callbacks(stbi_io_callbacks const *clbk , void *user, int *x, int *y, int *channels_in_file, int desired_channels); + +#ifndef STBI_NO_STDIO +STBIDEF stbi_uc *stbi_load (char const *filename, int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_uc *stbi_load_from_file (FILE *f, int *x, int *y, int *channels_in_file, int desired_channels); +// for stbi_load_from_file, file pointer is left pointing immediately after image +#endif + +#ifndef STBI_NO_GIF +STBIDEF stbi_uc *stbi_load_gif_from_memory(stbi_uc const *buffer, int len, int **delays, int *x, int *y, int *z, int *comp, int req_comp); +#endif + +#ifdef STBI_WINDOWS_UTF8 +STBIDEF int stbi_convert_wchar_to_utf8(char *buffer, size_t bufferlen, const wchar_t* input); +#endif + +//////////////////////////////////// +// +// 16-bits-per-channel interface +// + +STBIDEF stbi_us *stbi_load_16_from_memory (stbi_uc const *buffer, int len, int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_us *stbi_load_16_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *channels_in_file, int desired_channels); + +#ifndef STBI_NO_STDIO +STBIDEF stbi_us *stbi_load_16 (char const *filename, int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_us *stbi_load_from_file_16(FILE *f, int *x, int *y, int *channels_in_file, int desired_channels); +#endif + +//////////////////////////////////// +// +// float-per-channel interface +// +#ifndef STBI_NO_LINEAR +STBIDEF float *stbi_loadf_from_memory (stbi_uc const *buffer, int len, int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF float *stbi_loadf_from_callbacks (stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *channels_in_file, int desired_channels); + +#ifndef STBI_NO_STDIO +STBIDEF float *stbi_loadf (char const *filename, int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF float *stbi_loadf_from_file (FILE *f, int *x, int *y, int *channels_in_file, int desired_channels); +#endif +#endif + +#ifndef STBI_NO_HDR +STBIDEF void stbi_hdr_to_ldr_gamma(float gamma); +STBIDEF void stbi_hdr_to_ldr_scale(float scale); +#endif // STBI_NO_HDR + +#ifndef STBI_NO_LINEAR +STBIDEF void stbi_ldr_to_hdr_gamma(float gamma); +STBIDEF void stbi_ldr_to_hdr_scale(float scale); +#endif // STBI_NO_LINEAR + +// stbi_is_hdr is always defined, but always returns false if STBI_NO_HDR +STBIDEF int stbi_is_hdr_from_callbacks(stbi_io_callbacks const *clbk, void *user); +STBIDEF int stbi_is_hdr_from_memory(stbi_uc const *buffer, int len); +#ifndef STBI_NO_STDIO +STBIDEF int stbi_is_hdr (char const *filename); +STBIDEF int stbi_is_hdr_from_file(FILE *f); +#endif // STBI_NO_STDIO + + +// get a VERY brief reason for failure +// NOT THREADSAFE +STBIDEF const char *stbi_failure_reason (void); + +// free the loaded image -- this is just free() +STBIDEF void stbi_image_free (void *retval_from_stbi_load); + +// get image dimensions & components without fully decoding +STBIDEF int stbi_info_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp); +STBIDEF int stbi_info_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *comp); +STBIDEF int stbi_is_16_bit_from_memory(stbi_uc const *buffer, int len); +STBIDEF int stbi_is_16_bit_from_callbacks(stbi_io_callbacks const *clbk, void *user); + +#ifndef STBI_NO_STDIO +STBIDEF int stbi_info (char const *filename, int *x, int *y, int *comp); +STBIDEF int stbi_info_from_file (FILE *f, int *x, int *y, int *comp); +STBIDEF int stbi_is_16_bit (char const *filename); +STBIDEF int stbi_is_16_bit_from_file(FILE *f); +#endif + + + +// for image formats that explicitly notate that they have premultiplied alpha, +// we just return the colors as stored in the file. set this flag to force +// unpremultiplication. results are undefined if the unpremultiply overflow. +STBIDEF void stbi_set_unpremultiply_on_load(int flag_true_if_should_unpremultiply); + +// indicate whether we should process iphone images back to canonical format, +// or just pass them through "as-is" +STBIDEF void stbi_convert_iphone_png_to_rgb(int flag_true_if_should_convert); + +// flip the image vertically, so the first pixel in the output array is the bottom left +STBIDEF void stbi_set_flip_vertically_on_load(int flag_true_if_should_flip); + +// ZLIB client - used by PNG, available for other purposes + +STBIDEF char *stbi_zlib_decode_malloc_guesssize(const char *buffer, int len, int initial_size, int *outlen); +STBIDEF char *stbi_zlib_decode_malloc_guesssize_headerflag(const char *buffer, int len, int initial_size, int *outlen, int parse_header); +STBIDEF char *stbi_zlib_decode_malloc(const char *buffer, int len, int *outlen); +STBIDEF int stbi_zlib_decode_buffer(char *obuffer, int olen, const char *ibuffer, int ilen); + +STBIDEF char *stbi_zlib_decode_noheader_malloc(const char *buffer, int len, int *outlen); +STBIDEF int stbi_zlib_decode_noheader_buffer(char *obuffer, int olen, const char *ibuffer, int ilen); + + +#ifdef __cplusplus +} +#endif + +// +// +//// end header file ///////////////////////////////////////////////////// +#endif // STBI_INCLUDE_STB_IMAGE_H + +#ifdef STB_IMAGE_IMPLEMENTATION + +#if defined(STBI_ONLY_JPEG) || defined(STBI_ONLY_PNG) || defined(STBI_ONLY_BMP) \ + || defined(STBI_ONLY_TGA) || defined(STBI_ONLY_GIF) || defined(STBI_ONLY_PSD) \ + || defined(STBI_ONLY_HDR) || defined(STBI_ONLY_PIC) || defined(STBI_ONLY_PNM) \ + || defined(STBI_ONLY_ZLIB) + #ifndef STBI_ONLY_JPEG + #define STBI_NO_JPEG + #endif + #ifndef STBI_ONLY_PNG + #define STBI_NO_PNG + #endif + #ifndef STBI_ONLY_BMP + #define STBI_NO_BMP + #endif + #ifndef STBI_ONLY_PSD + #define STBI_NO_PSD + #endif + #ifndef STBI_ONLY_TGA + #define STBI_NO_TGA + #endif + #ifndef STBI_ONLY_GIF + #define STBI_NO_GIF + #endif + #ifndef STBI_ONLY_HDR + #define STBI_NO_HDR + #endif + #ifndef STBI_ONLY_PIC + #define STBI_NO_PIC + #endif + #ifndef STBI_ONLY_PNM + #define STBI_NO_PNM + #endif +#endif + +#if defined(STBI_NO_PNG) && !defined(STBI_SUPPORT_ZLIB) && !defined(STBI_NO_ZLIB) +#define STBI_NO_ZLIB +#endif + + +#include +#include // ptrdiff_t on osx +#include +#include +#include + +#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) +#include // ldexp, pow +#endif + +#ifndef STBI_NO_STDIO +#include +#endif + +#ifndef STBI_ASSERT +#include +#define STBI_ASSERT(x) assert(x) +#endif + +#ifdef __cplusplus +#define STBI_EXTERN extern "C" +#else +#define STBI_EXTERN extern +#endif + + +#ifndef _MSC_VER + #ifdef __cplusplus + #define stbi_inline inline + #else + #define stbi_inline + #endif +#else + #define stbi_inline __forceinline +#endif + + +#ifdef _MSC_VER +typedef unsigned short stbi__uint16; +typedef signed short stbi__int16; +typedef unsigned int stbi__uint32; +typedef signed int stbi__int32; +#else +#include +typedef uint16_t stbi__uint16; +typedef int16_t stbi__int16; +typedef uint32_t stbi__uint32; +typedef int32_t stbi__int32; +#endif + +// should produce compiler error if size is wrong +typedef unsigned char validate_uint32[sizeof(stbi__uint32)==4 ? 1 : -1]; + +#ifdef _MSC_VER +#define STBI_NOTUSED(v) (void)(v) +#else +#define STBI_NOTUSED(v) (void)sizeof(v) +#endif + +#ifdef _MSC_VER +#define STBI_HAS_LROTL +#endif + +#ifdef STBI_HAS_LROTL + #define stbi_lrot(x,y) _lrotl(x,y) +#else + #define stbi_lrot(x,y) (((x) << (y)) | ((x) >> (32 - (y)))) +#endif + +#if defined(STBI_MALLOC) && defined(STBI_FREE) && (defined(STBI_REALLOC) || defined(STBI_REALLOC_SIZED)) +// ok +#elif !defined(STBI_MALLOC) && !defined(STBI_FREE) && !defined(STBI_REALLOC) && !defined(STBI_REALLOC_SIZED) +// ok +#else +#error "Must define all or none of STBI_MALLOC, STBI_FREE, and STBI_REALLOC (or STBI_REALLOC_SIZED)." +#endif + +#ifndef STBI_MALLOC +#define STBI_MALLOC(sz) malloc(sz) +#define STBI_REALLOC(p,newsz) realloc(p,newsz) +#define STBI_FREE(p) free(p) +#endif + +#ifndef STBI_REALLOC_SIZED +#define STBI_REALLOC_SIZED(p,oldsz,newsz) STBI_REALLOC(p,newsz) +#endif + +// x86/x64 detection +#if defined(__x86_64__) || defined(_M_X64) +#define STBI__X64_TARGET +#elif defined(__i386) || defined(_M_IX86) +#define STBI__X86_TARGET +#endif + +#if defined(__GNUC__) && defined(STBI__X86_TARGET) && !defined(__SSE2__) && !defined(STBI_NO_SIMD) +// gcc doesn't support sse2 intrinsics unless you compile with -msse2, +// which in turn means it gets to use SSE2 everywhere. This is unfortunate, +// but previous attempts to provide the SSE2 functions with runtime +// detection caused numerous issues. The way architecture extensions are +// exposed in GCC/Clang is, sadly, not really suited for one-file libs. +// New behavior: if compiled with -msse2, we use SSE2 without any +// detection; if not, we don't use it at all. +#define STBI_NO_SIMD +#endif + +#if defined(__MINGW32__) && defined(STBI__X86_TARGET) && !defined(STBI_MINGW_ENABLE_SSE2) && !defined(STBI_NO_SIMD) +// Note that __MINGW32__ doesn't actually mean 32-bit, so we have to avoid STBI__X64_TARGET +// +// 32-bit MinGW wants ESP to be 16-byte aligned, but this is not in the +// Windows ABI and VC++ as well as Windows DLLs don't maintain that invariant. +// As a result, enabling SSE2 on 32-bit MinGW is dangerous when not +// simultaneously enabling "-mstackrealign". +// +// See https://github.com/nothings/stb/issues/81 for more information. +// +// So default to no SSE2 on 32-bit MinGW. If you've read this far and added +// -mstackrealign to your build settings, feel free to #define STBI_MINGW_ENABLE_SSE2. +#define STBI_NO_SIMD +#endif + +#if !defined(STBI_NO_SIMD) && (defined(STBI__X86_TARGET) || defined(STBI__X64_TARGET)) +#define STBI_SSE2 +#include + +#ifdef _MSC_VER + +#if _MSC_VER >= 1400 // not VC6 +#include // __cpuid +static int stbi__cpuid3(void) +{ + int info[4]; + __cpuid(info,1); + return info[3]; +} +#else +static int stbi__cpuid3(void) +{ + int res; + __asm { + mov eax,1 + cpuid + mov res,edx + } + return res; +} +#endif + +#define STBI_SIMD_ALIGN(type, name) __declspec(align(16)) type name + +#if !defined(STBI_NO_JPEG) && defined(STBI_SSE2) +static int stbi__sse2_available(void) +{ + int info3 = stbi__cpuid3(); + return ((info3 >> 26) & 1) != 0; +} +#endif + +#else // assume GCC-style if not VC++ +#define STBI_SIMD_ALIGN(type, name) type name __attribute__((aligned(16))) + +#if !defined(STBI_NO_JPEG) && defined(STBI_SSE2) +static int stbi__sse2_available(void) +{ + // If we're even attempting to compile this on GCC/Clang, that means + // -msse2 is on, which means the compiler is allowed to use SSE2 + // instructions at will, and so are we. + return 1; +} +#endif + +#endif +#endif + +// ARM NEON +#if defined(STBI_NO_SIMD) && defined(STBI_NEON) +#undef STBI_NEON +#endif + +#ifdef STBI_NEON +#include +// assume GCC or Clang on ARM targets +#define STBI_SIMD_ALIGN(type, name) type name __attribute__((aligned(16))) +#endif + +#ifndef STBI_SIMD_ALIGN +#define STBI_SIMD_ALIGN(type, name) type name +#endif + +/////////////////////////////////////////////// +// +// stbi__context struct and start_xxx functions + +// stbi__context structure is our basic context used by all images, so it +// contains all the IO context, plus some basic image information +typedef struct +{ + stbi__uint32 img_x, img_y; + int img_n, img_out_n; + + stbi_io_callbacks io; + void *io_user_data; + + int read_from_callbacks; + int buflen; + stbi_uc buffer_start[128]; + + stbi_uc *img_buffer, *img_buffer_end; + stbi_uc *img_buffer_original, *img_buffer_original_end; +} stbi__context; + + +static void stbi__refill_buffer(stbi__context *s); + +// initialize a memory-decode context +static void stbi__start_mem(stbi__context *s, stbi_uc const *buffer, int len) +{ + s->io.read = NULL; + s->read_from_callbacks = 0; + s->img_buffer = s->img_buffer_original = (stbi_uc *) buffer; + s->img_buffer_end = s->img_buffer_original_end = (stbi_uc *) buffer+len; +} + +// initialize a callback-based context +static void stbi__start_callbacks(stbi__context *s, stbi_io_callbacks *c, void *user) +{ + s->io = *c; + s->io_user_data = user; + s->buflen = sizeof(s->buffer_start); + s->read_from_callbacks = 1; + s->img_buffer_original = s->buffer_start; + stbi__refill_buffer(s); + s->img_buffer_original_end = s->img_buffer_end; +} + +#ifndef STBI_NO_STDIO + +static int stbi__stdio_read(void *user, char *data, int size) +{ + return (int) fread(data,1,size,(FILE*) user); +} + +static void stbi__stdio_skip(void *user, int n) +{ + fseek((FILE*) user, n, SEEK_CUR); +} + +static int stbi__stdio_eof(void *user) +{ + return feof((FILE*) user); +} + +static stbi_io_callbacks stbi__stdio_callbacks = +{ + stbi__stdio_read, + stbi__stdio_skip, + stbi__stdio_eof, +}; + +static void stbi__start_file(stbi__context *s, FILE *f) +{ + stbi__start_callbacks(s, &stbi__stdio_callbacks, (void *) f); +} + +//static void stop_file(stbi__context *s) { } + +#endif // !STBI_NO_STDIO + +static void stbi__rewind(stbi__context *s) +{ + // conceptually rewind SHOULD rewind to the beginning of the stream, + // but we just rewind to the beginning of the initial buffer, because + // we only use it after doing 'test', which only ever looks at at most 92 bytes + s->img_buffer = s->img_buffer_original; + s->img_buffer_end = s->img_buffer_original_end; +} + +enum +{ + STBI_ORDER_RGB, + STBI_ORDER_BGR +}; + +typedef struct +{ + int bits_per_channel; + int num_channels; + int channel_order; +} stbi__result_info; + +#ifndef STBI_NO_JPEG +static int stbi__jpeg_test(stbi__context *s); +static void *stbi__jpeg_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__jpeg_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_PNG +static int stbi__png_test(stbi__context *s); +static void *stbi__png_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__png_info(stbi__context *s, int *x, int *y, int *comp); +static int stbi__png_is16(stbi__context *s); +#endif + +#ifndef STBI_NO_BMP +static int stbi__bmp_test(stbi__context *s); +static void *stbi__bmp_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__bmp_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_TGA +static int stbi__tga_test(stbi__context *s); +static void *stbi__tga_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__tga_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_PSD +static int stbi__psd_test(stbi__context *s); +static void *stbi__psd_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri, int bpc); +static int stbi__psd_info(stbi__context *s, int *x, int *y, int *comp); +static int stbi__psd_is16(stbi__context *s); +#endif + +#ifndef STBI_NO_HDR +static int stbi__hdr_test(stbi__context *s); +static float *stbi__hdr_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__hdr_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_PIC +static int stbi__pic_test(stbi__context *s); +static void *stbi__pic_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__pic_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_GIF +static int stbi__gif_test(stbi__context *s); +static void *stbi__gif_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static void *stbi__load_gif_main(stbi__context *s, int **delays, int *x, int *y, int *z, int *comp, int req_comp); +static int stbi__gif_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +#ifndef STBI_NO_PNM +static int stbi__pnm_test(stbi__context *s); +static void *stbi__pnm_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__pnm_info(stbi__context *s, int *x, int *y, int *comp); +#endif + +// this is not threadsafe +static const char *stbi__g_failure_reason; + +STBIDEF const char *stbi_failure_reason(void) +{ + return stbi__g_failure_reason; +} + +static int stbi__err(const char *str) +{ + stbi__g_failure_reason = str; + return 0; +} + +static void *stbi__malloc(size_t size) +{ + return STBI_MALLOC(size); +} + +// stb_image uses ints pervasively, including for offset calculations. +// therefore the largest decoded image size we can support with the +// current code, even on 64-bit targets, is INT_MAX. this is not a +// significant limitation for the intended use case. +// +// we do, however, need to make sure our size calculations don't +// overflow. hence a few helper functions for size calculations that +// multiply integers together, making sure that they're non-negative +// and no overflow occurs. + +// return 1 if the sum is valid, 0 on overflow. +// negative terms are considered invalid. +static int stbi__addsizes_valid(int a, int b) +{ + if (b < 0) return 0; + // now 0 <= b <= INT_MAX, hence also + // 0 <= INT_MAX - b <= INTMAX. + // And "a + b <= INT_MAX" (which might overflow) is the + // same as a <= INT_MAX - b (no overflow) + return a <= INT_MAX - b; +} + +// returns 1 if the product is valid, 0 on overflow. +// negative factors are considered invalid. +static int stbi__mul2sizes_valid(int a, int b) +{ + if (a < 0 || b < 0) return 0; + if (b == 0) return 1; // mul-by-0 is always safe + // portable way to check for no overflows in a*b + return a <= INT_MAX/b; +} + +// returns 1 if "a*b + add" has no negative terms/factors and doesn't overflow +static int stbi__mad2sizes_valid(int a, int b, int add) +{ + return stbi__mul2sizes_valid(a, b) && stbi__addsizes_valid(a*b, add); +} + +// returns 1 if "a*b*c + add" has no negative terms/factors and doesn't overflow +static int stbi__mad3sizes_valid(int a, int b, int c, int add) +{ + return stbi__mul2sizes_valid(a, b) && stbi__mul2sizes_valid(a*b, c) && + stbi__addsizes_valid(a*b*c, add); +} + +// returns 1 if "a*b*c*d + add" has no negative terms/factors and doesn't overflow +#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) +static int stbi__mad4sizes_valid(int a, int b, int c, int d, int add) +{ + return stbi__mul2sizes_valid(a, b) && stbi__mul2sizes_valid(a*b, c) && + stbi__mul2sizes_valid(a*b*c, d) && stbi__addsizes_valid(a*b*c*d, add); +} +#endif + +// mallocs with size overflow checking +static void *stbi__malloc_mad2(int a, int b, int add) +{ + if (!stbi__mad2sizes_valid(a, b, add)) return NULL; + return stbi__malloc(a*b + add); +} + +static void *stbi__malloc_mad3(int a, int b, int c, int add) +{ + if (!stbi__mad3sizes_valid(a, b, c, add)) return NULL; + return stbi__malloc(a*b*c + add); +} + +#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) +static void *stbi__malloc_mad4(int a, int b, int c, int d, int add) +{ + if (!stbi__mad4sizes_valid(a, b, c, d, add)) return NULL; + return stbi__malloc(a*b*c*d + add); +} +#endif + +// stbi__err - error +// stbi__errpf - error returning pointer to float +// stbi__errpuc - error returning pointer to unsigned char + +#ifdef STBI_NO_FAILURE_STRINGS + #define stbi__err(x,y) 0 +#elif defined(STBI_FAILURE_USERMSG) + #define stbi__err(x,y) stbi__err(y) +#else + #define stbi__err(x,y) stbi__err(x) +#endif + +#define stbi__errpf(x,y) ((float *)(size_t) (stbi__err(x,y)?NULL:NULL)) +#define stbi__errpuc(x,y) ((unsigned char *)(size_t) (stbi__err(x,y)?NULL:NULL)) + +STBIDEF void stbi_image_free(void *retval_from_stbi_load) +{ + STBI_FREE(retval_from_stbi_load); +} + +#ifndef STBI_NO_LINEAR +static float *stbi__ldr_to_hdr(stbi_uc *data, int x, int y, int comp); +#endif + +#ifndef STBI_NO_HDR +static stbi_uc *stbi__hdr_to_ldr(float *data, int x, int y, int comp); +#endif + +static int stbi__vertically_flip_on_load = 0; + +STBIDEF void stbi_set_flip_vertically_on_load(int flag_true_if_should_flip) +{ + stbi__vertically_flip_on_load = flag_true_if_should_flip; +} + +static void *stbi__load_main(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri, int bpc) +{ + memset(ri, 0, sizeof(*ri)); // make sure it's initialized if we add new fields + ri->bits_per_channel = 8; // default is 8 so most paths don't have to be changed + ri->channel_order = STBI_ORDER_RGB; // all current input & output are this, but this is here so we can add BGR order + ri->num_channels = 0; + + #ifndef STBI_NO_JPEG + if (stbi__jpeg_test(s)) return stbi__jpeg_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_PNG + if (stbi__png_test(s)) return stbi__png_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_BMP + if (stbi__bmp_test(s)) return stbi__bmp_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_GIF + if (stbi__gif_test(s)) return stbi__gif_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_PSD + if (stbi__psd_test(s)) return stbi__psd_load(s,x,y,comp,req_comp, ri, bpc); + #endif + #ifndef STBI_NO_PIC + if (stbi__pic_test(s)) return stbi__pic_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_PNM + if (stbi__pnm_test(s)) return stbi__pnm_load(s,x,y,comp,req_comp, ri); + #endif + + #ifndef STBI_NO_HDR + if (stbi__hdr_test(s)) { + float *hdr = stbi__hdr_load(s, x,y,comp,req_comp, ri); + return stbi__hdr_to_ldr(hdr, *x, *y, req_comp ? req_comp : *comp); + } + #endif + + #ifndef STBI_NO_TGA + // test tga last because it's a crappy test! + if (stbi__tga_test(s)) + return stbi__tga_load(s,x,y,comp,req_comp, ri); + #endif + + return stbi__errpuc("unknown image type", "Image not of any known type, or corrupt"); +} + +static stbi_uc *stbi__convert_16_to_8(stbi__uint16 *orig, int w, int h, int channels) +{ + int i; + int img_len = w * h * channels; + stbi_uc *reduced; + + reduced = (stbi_uc *) stbi__malloc(img_len); + if (reduced == NULL) return stbi__errpuc("outofmem", "Out of memory"); + + for (i = 0; i < img_len; ++i) + reduced[i] = (stbi_uc)((orig[i] >> 8) & 0xFF); // top half of each byte is sufficient approx of 16->8 bit scaling + + STBI_FREE(orig); + return reduced; +} + +static stbi__uint16 *stbi__convert_8_to_16(stbi_uc *orig, int w, int h, int channels) +{ + int i; + int img_len = w * h * channels; + stbi__uint16 *enlarged; + + enlarged = (stbi__uint16 *) stbi__malloc(img_len*2); + if (enlarged == NULL) return (stbi__uint16 *) stbi__errpuc("outofmem", "Out of memory"); + + for (i = 0; i < img_len; ++i) + enlarged[i] = (stbi__uint16)((orig[i] << 8) + orig[i]); // replicate to high and low byte, maps 0->0, 255->0xffff + + STBI_FREE(orig); + return enlarged; +} + +static void stbi__vertical_flip(void *image, int w, int h, int bytes_per_pixel) +{ + int row; + size_t bytes_per_row = (size_t)w * bytes_per_pixel; + stbi_uc temp[2048]; + stbi_uc *bytes = (stbi_uc *)image; + + for (row = 0; row < (h>>1); row++) { + stbi_uc *row0 = bytes + row*bytes_per_row; + stbi_uc *row1 = bytes + (h - row - 1)*bytes_per_row; + // swap row0 with row1 + size_t bytes_left = bytes_per_row; + while (bytes_left) { + size_t bytes_copy = (bytes_left < sizeof(temp)) ? bytes_left : sizeof(temp); + memcpy(temp, row0, bytes_copy); + memcpy(row0, row1, bytes_copy); + memcpy(row1, temp, bytes_copy); + row0 += bytes_copy; + row1 += bytes_copy; + bytes_left -= bytes_copy; + } + } +} + +#ifndef STBI_NO_GIF +static void stbi__vertical_flip_slices(void *image, int w, int h, int z, int bytes_per_pixel) +{ + int slice; + int slice_size = w * h * bytes_per_pixel; + + stbi_uc *bytes = (stbi_uc *)image; + for (slice = 0; slice < z; ++slice) { + stbi__vertical_flip(bytes, w, h, bytes_per_pixel); + bytes += slice_size; + } +} +#endif + +static unsigned char *stbi__load_and_postprocess_8bit(stbi__context *s, int *x, int *y, int *comp, int req_comp) +{ + stbi__result_info ri; + void *result = stbi__load_main(s, x, y, comp, req_comp, &ri, 8); + + if (result == NULL) + return NULL; + + if (ri.bits_per_channel != 8) { + STBI_ASSERT(ri.bits_per_channel == 16); + result = stbi__convert_16_to_8((stbi__uint16 *) result, *x, *y, req_comp == 0 ? *comp : req_comp); + ri.bits_per_channel = 8; + } + + // @TODO: move stbi__convert_format to here + + if (stbi__vertically_flip_on_load) { + int channels = req_comp ? req_comp : *comp; + stbi__vertical_flip(result, *x, *y, channels * sizeof(stbi_uc)); + } + + return (unsigned char *) result; +} + +static stbi__uint16 *stbi__load_and_postprocess_16bit(stbi__context *s, int *x, int *y, int *comp, int req_comp) +{ + stbi__result_info ri; + void *result = stbi__load_main(s, x, y, comp, req_comp, &ri, 16); + + if (result == NULL) + return NULL; + + if (ri.bits_per_channel != 16) { + STBI_ASSERT(ri.bits_per_channel == 8); + result = stbi__convert_8_to_16((stbi_uc *) result, *x, *y, req_comp == 0 ? *comp : req_comp); + ri.bits_per_channel = 16; + } + + // @TODO: move stbi__convert_format16 to here + // @TODO: special case RGB-to-Y (and RGBA-to-YA) for 8-bit-to-16-bit case to keep more precision + + if (stbi__vertically_flip_on_load) { + int channels = req_comp ? req_comp : *comp; + stbi__vertical_flip(result, *x, *y, channels * sizeof(stbi__uint16)); + } + + return (stbi__uint16 *) result; +} + +#if !defined(STBI_NO_HDR) && !defined(STBI_NO_LINEAR) +static void stbi__float_postprocess(float *result, int *x, int *y, int *comp, int req_comp) +{ + if (stbi__vertically_flip_on_load && result != NULL) { + int channels = req_comp ? req_comp : *comp; + stbi__vertical_flip(result, *x, *y, channels * sizeof(float)); + } +} +#endif + +#ifndef STBI_NO_STDIO + +#if defined(_MSC_VER) && defined(STBI_WINDOWS_UTF8) +STBI_EXTERN __declspec(dllimport) int __stdcall MultiByteToWideChar(unsigned int cp, unsigned long flags, const char *str, int cbmb, wchar_t *widestr, int cchwide); +STBI_EXTERN __declspec(dllimport) int __stdcall WideCharToMultiByte(unsigned int cp, unsigned long flags, const wchar_t *widestr, int cchwide, char *str, int cbmb, const char *defchar, int *used_default); +#endif + +#if defined(_MSC_VER) && defined(STBI_WINDOWS_UTF8) +STBIDEF int stbi_convert_wchar_to_utf8(char *buffer, size_t bufferlen, const wchar_t* input) +{ + return WideCharToMultiByte(65001 /* UTF8 */, 0, input, -1, buffer, bufferlen, NULL, NULL); +} +#endif + +static FILE *stbi__fopen(char const *filename, char const *mode) +{ + FILE *f; +#if defined(_MSC_VER) && defined(STBI_WINDOWS_UTF8) + wchar_t wMode[64]; + wchar_t wFilename[1024]; + if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, filename, -1, wFilename, sizeof(wFilename))) + return 0; + + if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, mode, -1, wMode, sizeof(wMode))) + return 0; + +#if _MSC_VER >= 1400 + if (0 != _wfopen_s(&f, wFilename, wMode)) + f = 0; +#else + f = _wfopen(wFilename, wMode); +#endif + +#elif defined(_MSC_VER) && _MSC_VER >= 1400 + if (0 != fopen_s(&f, filename, mode)) + f=0; +#else + f = fopen(filename, mode); +#endif + return f; +} + + +STBIDEF stbi_uc *stbi_load(char const *filename, int *x, int *y, int *comp, int req_comp) +{ + FILE *f = stbi__fopen(filename, "rb"); + unsigned char *result; + if (!f) return stbi__errpuc("can't fopen", "Unable to open file"); + result = stbi_load_from_file(f,x,y,comp,req_comp); + fclose(f); + return result; +} + +STBIDEF stbi_uc *stbi_load_from_file(FILE *f, int *x, int *y, int *comp, int req_comp) +{ + unsigned char *result; + stbi__context s; + stbi__start_file(&s,f); + result = stbi__load_and_postprocess_8bit(&s,x,y,comp,req_comp); + if (result) { + // need to 'unget' all the characters in the IO buffer + fseek(f, - (int) (s.img_buffer_end - s.img_buffer), SEEK_CUR); + } + return result; +} + +STBIDEF stbi__uint16 *stbi_load_from_file_16(FILE *f, int *x, int *y, int *comp, int req_comp) +{ + stbi__uint16 *result; + stbi__context s; + stbi__start_file(&s,f); + result = stbi__load_and_postprocess_16bit(&s,x,y,comp,req_comp); + if (result) { + // need to 'unget' all the characters in the IO buffer + fseek(f, - (int) (s.img_buffer_end - s.img_buffer), SEEK_CUR); + } + return result; +} + +STBIDEF stbi_us *stbi_load_16(char const *filename, int *x, int *y, int *comp, int req_comp) +{ + FILE *f = stbi__fopen(filename, "rb"); + stbi__uint16 *result; + if (!f) return (stbi_us *) stbi__errpuc("can't fopen", "Unable to open file"); + result = stbi_load_from_file_16(f,x,y,comp,req_comp); + fclose(f); + return result; +} + + +#endif //!STBI_NO_STDIO + +STBIDEF stbi_us *stbi_load_16_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *channels_in_file, int desired_channels) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__load_and_postprocess_16bit(&s,x,y,channels_in_file,desired_channels); +} + +STBIDEF stbi_us *stbi_load_16_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *channels_in_file, int desired_channels) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *)clbk, user); + return stbi__load_and_postprocess_16bit(&s,x,y,channels_in_file,desired_channels); +} + +STBIDEF stbi_uc *stbi_load_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__load_and_postprocess_8bit(&s,x,y,comp,req_comp); +} + +STBIDEF stbi_uc *stbi_load_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) clbk, user); + return stbi__load_and_postprocess_8bit(&s,x,y,comp,req_comp); +} + +#ifndef STBI_NO_GIF +STBIDEF stbi_uc *stbi_load_gif_from_memory(stbi_uc const *buffer, int len, int **delays, int *x, int *y, int *z, int *comp, int req_comp) +{ + unsigned char *result; + stbi__context s; + stbi__start_mem(&s,buffer,len); + + result = (unsigned char*) stbi__load_gif_main(&s, delays, x, y, z, comp, req_comp); + if (stbi__vertically_flip_on_load) { + stbi__vertical_flip_slices( result, *x, *y, *z, *comp ); + } + + return result; +} +#endif + +#ifndef STBI_NO_LINEAR +static float *stbi__loadf_main(stbi__context *s, int *x, int *y, int *comp, int req_comp) +{ + unsigned char *data; + #ifndef STBI_NO_HDR + if (stbi__hdr_test(s)) { + stbi__result_info ri; + float *hdr_data = stbi__hdr_load(s,x,y,comp,req_comp, &ri); + if (hdr_data) + stbi__float_postprocess(hdr_data,x,y,comp,req_comp); + return hdr_data; + } + #endif + data = stbi__load_and_postprocess_8bit(s, x, y, comp, req_comp); + if (data) + return stbi__ldr_to_hdr(data, *x, *y, req_comp ? req_comp : *comp); + return stbi__errpf("unknown image type", "Image not of any known type, or corrupt"); +} + +STBIDEF float *stbi_loadf_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__loadf_main(&s,x,y,comp,req_comp); +} + +STBIDEF float *stbi_loadf_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) clbk, user); + return stbi__loadf_main(&s,x,y,comp,req_comp); +} + +#ifndef STBI_NO_STDIO +STBIDEF float *stbi_loadf(char const *filename, int *x, int *y, int *comp, int req_comp) +{ + float *result; + FILE *f = stbi__fopen(filename, "rb"); + if (!f) return stbi__errpf("can't fopen", "Unable to open file"); + result = stbi_loadf_from_file(f,x,y,comp,req_comp); + fclose(f); + return result; +} + +STBIDEF float *stbi_loadf_from_file(FILE *f, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_file(&s,f); + return stbi__loadf_main(&s,x,y,comp,req_comp); +} +#endif // !STBI_NO_STDIO + +#endif // !STBI_NO_LINEAR + +// these is-hdr-or-not is defined independent of whether STBI_NO_LINEAR is +// defined, for API simplicity; if STBI_NO_LINEAR is defined, it always +// reports false! + +STBIDEF int stbi_is_hdr_from_memory(stbi_uc const *buffer, int len) +{ + #ifndef STBI_NO_HDR + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__hdr_test(&s); + #else + STBI_NOTUSED(buffer); + STBI_NOTUSED(len); + return 0; + #endif +} + +#ifndef STBI_NO_STDIO +STBIDEF int stbi_is_hdr (char const *filename) +{ + FILE *f = stbi__fopen(filename, "rb"); + int result=0; + if (f) { + result = stbi_is_hdr_from_file(f); + fclose(f); + } + return result; +} + +STBIDEF int stbi_is_hdr_from_file(FILE *f) +{ + #ifndef STBI_NO_HDR + long pos = ftell(f); + int res; + stbi__context s; + stbi__start_file(&s,f); + res = stbi__hdr_test(&s); + fseek(f, pos, SEEK_SET); + return res; + #else + STBI_NOTUSED(f); + return 0; + #endif +} +#endif // !STBI_NO_STDIO + +STBIDEF int stbi_is_hdr_from_callbacks(stbi_io_callbacks const *clbk, void *user) +{ + #ifndef STBI_NO_HDR + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) clbk, user); + return stbi__hdr_test(&s); + #else + STBI_NOTUSED(clbk); + STBI_NOTUSED(user); + return 0; + #endif +} + +#ifndef STBI_NO_LINEAR +static float stbi__l2h_gamma=2.2f, stbi__l2h_scale=1.0f; + +STBIDEF void stbi_ldr_to_hdr_gamma(float gamma) { stbi__l2h_gamma = gamma; } +STBIDEF void stbi_ldr_to_hdr_scale(float scale) { stbi__l2h_scale = scale; } +#endif + +static float stbi__h2l_gamma_i=1.0f/2.2f, stbi__h2l_scale_i=1.0f; + +STBIDEF void stbi_hdr_to_ldr_gamma(float gamma) { stbi__h2l_gamma_i = 1/gamma; } +STBIDEF void stbi_hdr_to_ldr_scale(float scale) { stbi__h2l_scale_i = 1/scale; } + + +////////////////////////////////////////////////////////////////////////////// +// +// Common code used by all image loaders +// + +enum +{ + STBI__SCAN_load=0, + STBI__SCAN_type, + STBI__SCAN_header +}; + +static void stbi__refill_buffer(stbi__context *s) +{ + int n = (s->io.read)(s->io_user_data,(char*)s->buffer_start,s->buflen); + if (n == 0) { + // at end of file, treat same as if from memory, but need to handle case + // where s->img_buffer isn't pointing to safe memory, e.g. 0-byte file + s->read_from_callbacks = 0; + s->img_buffer = s->buffer_start; + s->img_buffer_end = s->buffer_start+1; + *s->img_buffer = 0; + } else { + s->img_buffer = s->buffer_start; + s->img_buffer_end = s->buffer_start + n; + } +} + +stbi_inline static stbi_uc stbi__get8(stbi__context *s) +{ + if (s->img_buffer < s->img_buffer_end) + return *s->img_buffer++; + if (s->read_from_callbacks) { + stbi__refill_buffer(s); + return *s->img_buffer++; + } + return 0; +} + +stbi_inline static int stbi__at_eof(stbi__context *s) +{ + if (s->io.read) { + if (!(s->io.eof)(s->io_user_data)) return 0; + // if feof() is true, check if buffer = end + // special case: we've only got the special 0 character at the end + if (s->read_from_callbacks == 0) return 1; + } + + return s->img_buffer >= s->img_buffer_end; +} + +static void stbi__skip(stbi__context *s, int n) +{ + if (n < 0) { + s->img_buffer = s->img_buffer_end; + return; + } + if (s->io.read) { + int blen = (int) (s->img_buffer_end - s->img_buffer); + if (blen < n) { + s->img_buffer = s->img_buffer_end; + (s->io.skip)(s->io_user_data, n - blen); + return; + } + } + s->img_buffer += n; +} + +static int stbi__getn(stbi__context *s, stbi_uc *buffer, int n) +{ + if (s->io.read) { + int blen = (int) (s->img_buffer_end - s->img_buffer); + if (blen < n) { + int res, count; + + memcpy(buffer, s->img_buffer, blen); + + count = (s->io.read)(s->io_user_data, (char*) buffer + blen, n - blen); + res = (count == (n-blen)); + s->img_buffer = s->img_buffer_end; + return res; + } + } + + if (s->img_buffer+n <= s->img_buffer_end) { + memcpy(buffer, s->img_buffer, n); + s->img_buffer += n; + return 1; + } else + return 0; +} + +static int stbi__get16be(stbi__context *s) +{ + int z = stbi__get8(s); + return (z << 8) + stbi__get8(s); +} + +static stbi__uint32 stbi__get32be(stbi__context *s) +{ + stbi__uint32 z = stbi__get16be(s); + return (z << 16) + stbi__get16be(s); +} + +#if defined(STBI_NO_BMP) && defined(STBI_NO_TGA) && defined(STBI_NO_GIF) +// nothing +#else +static int stbi__get16le(stbi__context *s) +{ + int z = stbi__get8(s); + return z + (stbi__get8(s) << 8); +} +#endif + +#ifndef STBI_NO_BMP +static stbi__uint32 stbi__get32le(stbi__context *s) +{ + stbi__uint32 z = stbi__get16le(s); + return z + (stbi__get16le(s) << 16); +} +#endif + +#define STBI__BYTECAST(x) ((stbi_uc) ((x) & 255)) // truncate int to byte without warnings + + +////////////////////////////////////////////////////////////////////////////// +// +// generic converter from built-in img_n to req_comp +// individual types do this automatically as much as possible (e.g. jpeg +// does all cases internally since it needs to colorspace convert anyway, +// and it never has alpha, so very few cases ). png can automatically +// interleave an alpha=255 channel, but falls back to this for other cases +// +// assume data buffer is malloced, so malloc a new one and free that one +// only failure mode is malloc failing + +static stbi_uc stbi__compute_y(int r, int g, int b) +{ + return (stbi_uc) (((r*77) + (g*150) + (29*b)) >> 8); +} + +static unsigned char *stbi__convert_format(unsigned char *data, int img_n, int req_comp, unsigned int x, unsigned int y) +{ + int i,j; + unsigned char *good; + + if (req_comp == img_n) return data; + STBI_ASSERT(req_comp >= 1 && req_comp <= 4); + + good = (unsigned char *) stbi__malloc_mad3(req_comp, x, y, 0); + if (good == NULL) { + STBI_FREE(data); + return stbi__errpuc("outofmem", "Out of memory"); + } + + for (j=0; j < (int) y; ++j) { + unsigned char *src = data + j * x * img_n ; + unsigned char *dest = good + j * x * req_comp; + + #define STBI__COMBO(a,b) ((a)*8+(b)) + #define STBI__CASE(a,b) case STBI__COMBO(a,b): for(i=x-1; i >= 0; --i, src += a, dest += b) + // convert source image with img_n components to one with req_comp components; + // avoid switch per pixel, so use switch per scanline and massive macros + switch (STBI__COMBO(img_n, req_comp)) { + STBI__CASE(1,2) { dest[0]=src[0]; dest[1]=255; } break; + STBI__CASE(1,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(1,4) { dest[0]=dest[1]=dest[2]=src[0]; dest[3]=255; } break; + STBI__CASE(2,1) { dest[0]=src[0]; } break; + STBI__CASE(2,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(2,4) { dest[0]=dest[1]=dest[2]=src[0]; dest[3]=src[1]; } break; + STBI__CASE(3,4) { dest[0]=src[0];dest[1]=src[1];dest[2]=src[2];dest[3]=255; } break; + STBI__CASE(3,1) { dest[0]=stbi__compute_y(src[0],src[1],src[2]); } break; + STBI__CASE(3,2) { dest[0]=stbi__compute_y(src[0],src[1],src[2]); dest[1] = 255; } break; + STBI__CASE(4,1) { dest[0]=stbi__compute_y(src[0],src[1],src[2]); } break; + STBI__CASE(4,2) { dest[0]=stbi__compute_y(src[0],src[1],src[2]); dest[1] = src[3]; } break; + STBI__CASE(4,3) { dest[0]=src[0];dest[1]=src[1];dest[2]=src[2]; } break; + default: STBI_ASSERT(0); + } + #undef STBI__CASE + } + + STBI_FREE(data); + return good; +} + +static stbi__uint16 stbi__compute_y_16(int r, int g, int b) +{ + return (stbi__uint16) (((r*77) + (g*150) + (29*b)) >> 8); +} + +static stbi__uint16 *stbi__convert_format16(stbi__uint16 *data, int img_n, int req_comp, unsigned int x, unsigned int y) +{ + int i,j; + stbi__uint16 *good; + + if (req_comp == img_n) return data; + STBI_ASSERT(req_comp >= 1 && req_comp <= 4); + + good = (stbi__uint16 *) stbi__malloc(req_comp * x * y * 2); + if (good == NULL) { + STBI_FREE(data); + return (stbi__uint16 *) stbi__errpuc("outofmem", "Out of memory"); + } + + for (j=0; j < (int) y; ++j) { + stbi__uint16 *src = data + j * x * img_n ; + stbi__uint16 *dest = good + j * x * req_comp; + + #define STBI__COMBO(a,b) ((a)*8+(b)) + #define STBI__CASE(a,b) case STBI__COMBO(a,b): for(i=x-1; i >= 0; --i, src += a, dest += b) + // convert source image with img_n components to one with req_comp components; + // avoid switch per pixel, so use switch per scanline and massive macros + switch (STBI__COMBO(img_n, req_comp)) { + STBI__CASE(1,2) { dest[0]=src[0]; dest[1]=0xffff; } break; + STBI__CASE(1,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(1,4) { dest[0]=dest[1]=dest[2]=src[0]; dest[3]=0xffff; } break; + STBI__CASE(2,1) { dest[0]=src[0]; } break; + STBI__CASE(2,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(2,4) { dest[0]=dest[1]=dest[2]=src[0]; dest[3]=src[1]; } break; + STBI__CASE(3,4) { dest[0]=src[0];dest[1]=src[1];dest[2]=src[2];dest[3]=0xffff; } break; + STBI__CASE(3,1) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]); } break; + STBI__CASE(3,2) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]); dest[1] = 0xffff; } break; + STBI__CASE(4,1) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]); } break; + STBI__CASE(4,2) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]); dest[1] = src[3]; } break; + STBI__CASE(4,3) { dest[0]=src[0];dest[1]=src[1];dest[2]=src[2]; } break; + default: STBI_ASSERT(0); + } + #undef STBI__CASE + } + + STBI_FREE(data); + return good; +} + +#ifndef STBI_NO_LINEAR +static float *stbi__ldr_to_hdr(stbi_uc *data, int x, int y, int comp) +{ + int i,k,n; + float *output; + if (!data) return NULL; + output = (float *) stbi__malloc_mad4(x, y, comp, sizeof(float), 0); + if (output == NULL) { STBI_FREE(data); return stbi__errpf("outofmem", "Out of memory"); } + // compute number of non-alpha components + if (comp & 1) n = comp; else n = comp-1; + for (i=0; i < x*y; ++i) { + for (k=0; k < n; ++k) { + output[i*comp + k] = (float) (pow(data[i*comp+k]/255.0f, stbi__l2h_gamma) * stbi__l2h_scale); + } + } + if (n < comp) { + for (i=0; i < x*y; ++i) { + output[i*comp + n] = data[i*comp + n]/255.0f; + } + } + STBI_FREE(data); + return output; +} +#endif + +#ifndef STBI_NO_HDR +#define stbi__float2int(x) ((int) (x)) +static stbi_uc *stbi__hdr_to_ldr(float *data, int x, int y, int comp) +{ + int i,k,n; + stbi_uc *output; + if (!data) return NULL; + output = (stbi_uc *) stbi__malloc_mad3(x, y, comp, 0); + if (output == NULL) { STBI_FREE(data); return stbi__errpuc("outofmem", "Out of memory"); } + // compute number of non-alpha components + if (comp & 1) n = comp; else n = comp-1; + for (i=0; i < x*y; ++i) { + for (k=0; k < n; ++k) { + float z = (float) pow(data[i*comp+k]*stbi__h2l_scale_i, stbi__h2l_gamma_i) * 255 + 0.5f; + if (z < 0) z = 0; + if (z > 255) z = 255; + output[i*comp + k] = (stbi_uc) stbi__float2int(z); + } + if (k < comp) { + float z = data[i*comp+k] * 255 + 0.5f; + if (z < 0) z = 0; + if (z > 255) z = 255; + output[i*comp + k] = (stbi_uc) stbi__float2int(z); + } + } + STBI_FREE(data); + return output; +} +#endif + +////////////////////////////////////////////////////////////////////////////// +// +// "baseline" JPEG/JFIF decoder +// +// simple implementation +// - doesn't support delayed output of y-dimension +// - simple interface (only one output format: 8-bit interleaved RGB) +// - doesn't try to recover corrupt jpegs +// - doesn't allow partial loading, loading multiple at once +// - still fast on x86 (copying globals into locals doesn't help x86) +// - allocates lots of intermediate memory (full size of all components) +// - non-interleaved case requires this anyway +// - allows good upsampling (see next) +// high-quality +// - upsampled channels are bilinearly interpolated, even across blocks +// - quality integer IDCT derived from IJG's 'slow' +// performance +// - fast huffman; reasonable integer IDCT +// - some SIMD kernels for common paths on targets with SSE2/NEON +// - uses a lot of intermediate memory, could cache poorly + +#ifndef STBI_NO_JPEG + +// huffman decoding acceleration +#define FAST_BITS 9 // larger handles more cases; smaller stomps less cache + +typedef struct +{ + stbi_uc fast[1 << FAST_BITS]; + // weirdly, repacking this into AoS is a 10% speed loss, instead of a win + stbi__uint16 code[256]; + stbi_uc values[256]; + stbi_uc size[257]; + unsigned int maxcode[18]; + int delta[17]; // old 'firstsymbol' - old 'firstcode' +} stbi__huffman; + +typedef struct +{ + stbi__context *s; + stbi__huffman huff_dc[4]; + stbi__huffman huff_ac[4]; + stbi__uint16 dequant[4][64]; + stbi__int16 fast_ac[4][1 << FAST_BITS]; + +// sizes for components, interleaved MCUs + int img_h_max, img_v_max; + int img_mcu_x, img_mcu_y; + int img_mcu_w, img_mcu_h; + +// definition of jpeg image component + struct + { + int id; + int h,v; + int tq; + int hd,ha; + int dc_pred; + + int x,y,w2,h2; + stbi_uc *data; + void *raw_data, *raw_coeff; + stbi_uc *linebuf; + short *coeff; // progressive only + int coeff_w, coeff_h; // number of 8x8 coefficient blocks + } img_comp[4]; + + stbi__uint32 code_buffer; // jpeg entropy-coded buffer + int code_bits; // number of valid bits + unsigned char marker; // marker seen while filling entropy buffer + int nomore; // flag if we saw a marker so must stop + + int progressive; + int spec_start; + int spec_end; + int succ_high; + int succ_low; + int eob_run; + int jfif; + int app14_color_transform; // Adobe APP14 tag + int rgb; + + int scan_n, order[4]; + int restart_interval, todo; + +// kernels + void (*idct_block_kernel)(stbi_uc *out, int out_stride, short data[64]); + void (*YCbCr_to_RGB_kernel)(stbi_uc *out, const stbi_uc *y, const stbi_uc *pcb, const stbi_uc *pcr, int count, int step); + stbi_uc *(*resample_row_hv_2_kernel)(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs); +} stbi__jpeg; + +static int stbi__build_huffman(stbi__huffman *h, int *count) +{ + int i,j,k=0; + unsigned int code; + // build size list for each symbol (from JPEG spec) + for (i=0; i < 16; ++i) + for (j=0; j < count[i]; ++j) + h->size[k++] = (stbi_uc) (i+1); + h->size[k] = 0; + + // compute actual symbols (from jpeg spec) + code = 0; + k = 0; + for(j=1; j <= 16; ++j) { + // compute delta to add to code to compute symbol id + h->delta[j] = k - code; + if (h->size[k] == j) { + while (h->size[k] == j) + h->code[k++] = (stbi__uint16) (code++); + if (code-1 >= (1u << j)) return stbi__err("bad code lengths","Corrupt JPEG"); + } + // compute largest code + 1 for this size, preshifted as needed later + h->maxcode[j] = code << (16-j); + code <<= 1; + } + h->maxcode[j] = 0xffffffff; + + // build non-spec acceleration table; 255 is flag for not-accelerated + memset(h->fast, 255, 1 << FAST_BITS); + for (i=0; i < k; ++i) { + int s = h->size[i]; + if (s <= FAST_BITS) { + int c = h->code[i] << (FAST_BITS-s); + int m = 1 << (FAST_BITS-s); + for (j=0; j < m; ++j) { + h->fast[c+j] = (stbi_uc) i; + } + } + } + return 1; +} + +// build a table that decodes both magnitude and value of small ACs in +// one go. +static void stbi__build_fast_ac(stbi__int16 *fast_ac, stbi__huffman *h) +{ + int i; + for (i=0; i < (1 << FAST_BITS); ++i) { + stbi_uc fast = h->fast[i]; + fast_ac[i] = 0; + if (fast < 255) { + int rs = h->values[fast]; + int run = (rs >> 4) & 15; + int magbits = rs & 15; + int len = h->size[fast]; + + if (magbits && len + magbits <= FAST_BITS) { + // magnitude code followed by receive_extend code + int k = ((i << len) & ((1 << FAST_BITS) - 1)) >> (FAST_BITS - magbits); + int m = 1 << (magbits - 1); + if (k < m) k += (~0U << magbits) + 1; + // if the result is small enough, we can fit it in fast_ac table + if (k >= -128 && k <= 127) + fast_ac[i] = (stbi__int16) ((k * 256) + (run * 16) + (len + magbits)); + } + } + } +} + +static void stbi__grow_buffer_unsafe(stbi__jpeg *j) +{ + do { + unsigned int b = j->nomore ? 0 : stbi__get8(j->s); + if (b == 0xff) { + int c = stbi__get8(j->s); + while (c == 0xff) c = stbi__get8(j->s); // consume fill bytes + if (c != 0) { + j->marker = (unsigned char) c; + j->nomore = 1; + return; + } + } + j->code_buffer |= b << (24 - j->code_bits); + j->code_bits += 8; + } while (j->code_bits <= 24); +} + +// (1 << n) - 1 +static const stbi__uint32 stbi__bmask[17]={0,1,3,7,15,31,63,127,255,511,1023,2047,4095,8191,16383,32767,65535}; + +// decode a jpeg huffman value from the bitstream +stbi_inline static int stbi__jpeg_huff_decode(stbi__jpeg *j, stbi__huffman *h) +{ + unsigned int temp; + int c,k; + + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + + // look at the top FAST_BITS and determine what symbol ID it is, + // if the code is <= FAST_BITS + c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS)-1); + k = h->fast[c]; + if (k < 255) { + int s = h->size[k]; + if (s > j->code_bits) + return -1; + j->code_buffer <<= s; + j->code_bits -= s; + return h->values[k]; + } + + // naive test is to shift the code_buffer down so k bits are + // valid, then test against maxcode. To speed this up, we've + // preshifted maxcode left so that it has (16-k) 0s at the + // end; in other words, regardless of the number of bits, it + // wants to be compared against something shifted to have 16; + // that way we don't need to shift inside the loop. + temp = j->code_buffer >> 16; + for (k=FAST_BITS+1 ; ; ++k) + if (temp < h->maxcode[k]) + break; + if (k == 17) { + // error! code not found + j->code_bits -= 16; + return -1; + } + + if (k > j->code_bits) + return -1; + + // convert the huffman code to the symbol id + c = ((j->code_buffer >> (32 - k)) & stbi__bmask[k]) + h->delta[k]; + STBI_ASSERT((((j->code_buffer) >> (32 - h->size[c])) & stbi__bmask[h->size[c]]) == h->code[c]); + + // convert the id to a symbol + j->code_bits -= k; + j->code_buffer <<= k; + return h->values[c]; +} + +// bias[n] = (-1<code_bits < n) stbi__grow_buffer_unsafe(j); + + sgn = (stbi__int32)j->code_buffer >> 31; // sign bit is always in MSB + k = stbi_lrot(j->code_buffer, n); + STBI_ASSERT(n >= 0 && n < (int) (sizeof(stbi__bmask)/sizeof(*stbi__bmask))); + j->code_buffer = k & ~stbi__bmask[n]; + k &= stbi__bmask[n]; + j->code_bits -= n; + return k + (stbi__jbias[n] & ~sgn); +} + +// get some unsigned bits +stbi_inline static int stbi__jpeg_get_bits(stbi__jpeg *j, int n) +{ + unsigned int k; + if (j->code_bits < n) stbi__grow_buffer_unsafe(j); + k = stbi_lrot(j->code_buffer, n); + j->code_buffer = k & ~stbi__bmask[n]; + k &= stbi__bmask[n]; + j->code_bits -= n; + return k; +} + +stbi_inline static int stbi__jpeg_get_bit(stbi__jpeg *j) +{ + unsigned int k; + if (j->code_bits < 1) stbi__grow_buffer_unsafe(j); + k = j->code_buffer; + j->code_buffer <<= 1; + --j->code_bits; + return k & 0x80000000; +} + +// given a value that's at position X in the zigzag stream, +// where does it appear in the 8x8 matrix coded as row-major? +static const stbi_uc stbi__jpeg_dezigzag[64+15] = +{ + 0, 1, 8, 16, 9, 2, 3, 10, + 17, 24, 32, 25, 18, 11, 4, 5, + 12, 19, 26, 33, 40, 48, 41, 34, + 27, 20, 13, 6, 7, 14, 21, 28, + 35, 42, 49, 56, 57, 50, 43, 36, + 29, 22, 15, 23, 30, 37, 44, 51, + 58, 59, 52, 45, 38, 31, 39, 46, + 53, 60, 61, 54, 47, 55, 62, 63, + // let corrupt input sample past end + 63, 63, 63, 63, 63, 63, 63, 63, + 63, 63, 63, 63, 63, 63, 63 +}; + +// decode one 64-entry block-- +static int stbi__jpeg_decode_block(stbi__jpeg *j, short data[64], stbi__huffman *hdc, stbi__huffman *hac, stbi__int16 *fac, int b, stbi__uint16 *dequant) +{ + int diff,dc,k; + int t; + + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + t = stbi__jpeg_huff_decode(j, hdc); + if (t < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + + // 0 all the ac values now so we can do it 32-bits at a time + memset(data,0,64*sizeof(data[0])); + + diff = t ? stbi__extend_receive(j, t) : 0; + dc = j->img_comp[b].dc_pred + diff; + j->img_comp[b].dc_pred = dc; + data[0] = (short) (dc * dequant[0]); + + // decode AC components, see JPEG spec + k = 1; + do { + unsigned int zig; + int c,r,s; + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS)-1); + r = fac[c]; + if (r) { // fast-AC path + k += (r >> 4) & 15; // run + s = r & 15; // combined length + j->code_buffer <<= s; + j->code_bits -= s; + // decode into unzigzag'd location + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) ((r >> 8) * dequant[zig]); + } else { + int rs = stbi__jpeg_huff_decode(j, hac); + if (rs < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + s = rs & 15; + r = rs >> 4; + if (s == 0) { + if (rs != 0xf0) break; // end block + k += 16; + } else { + k += r; + // decode into unzigzag'd location + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) (stbi__extend_receive(j,s) * dequant[zig]); + } + } + } while (k < 64); + return 1; +} + +static int stbi__jpeg_decode_block_prog_dc(stbi__jpeg *j, short data[64], stbi__huffman *hdc, int b) +{ + int diff,dc; + int t; + if (j->spec_end != 0) return stbi__err("can't merge dc and ac", "Corrupt JPEG"); + + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + + if (j->succ_high == 0) { + // first scan for DC coefficient, must be first + memset(data,0,64*sizeof(data[0])); // 0 all the ac values now + t = stbi__jpeg_huff_decode(j, hdc); + diff = t ? stbi__extend_receive(j, t) : 0; + + dc = j->img_comp[b].dc_pred + diff; + j->img_comp[b].dc_pred = dc; + data[0] = (short) (dc << j->succ_low); + } else { + // refinement scan for DC coefficient + if (stbi__jpeg_get_bit(j)) + data[0] += (short) (1 << j->succ_low); + } + return 1; +} + +// @OPTIMIZE: store non-zigzagged during the decode passes, +// and only de-zigzag when dequantizing +static int stbi__jpeg_decode_block_prog_ac(stbi__jpeg *j, short data[64], stbi__huffman *hac, stbi__int16 *fac) +{ + int k; + if (j->spec_start == 0) return stbi__err("can't merge dc and ac", "Corrupt JPEG"); + + if (j->succ_high == 0) { + int shift = j->succ_low; + + if (j->eob_run) { + --j->eob_run; + return 1; + } + + k = j->spec_start; + do { + unsigned int zig; + int c,r,s; + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS)-1); + r = fac[c]; + if (r) { // fast-AC path + k += (r >> 4) & 15; // run + s = r & 15; // combined length + j->code_buffer <<= s; + j->code_bits -= s; + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) ((r >> 8) << shift); + } else { + int rs = stbi__jpeg_huff_decode(j, hac); + if (rs < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + s = rs & 15; + r = rs >> 4; + if (s == 0) { + if (r < 15) { + j->eob_run = (1 << r); + if (r) + j->eob_run += stbi__jpeg_get_bits(j, r); + --j->eob_run; + break; + } + k += 16; + } else { + k += r; + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) (stbi__extend_receive(j,s) << shift); + } + } + } while (k <= j->spec_end); + } else { + // refinement scan for these AC coefficients + + short bit = (short) (1 << j->succ_low); + + if (j->eob_run) { + --j->eob_run; + for (k = j->spec_start; k <= j->spec_end; ++k) { + short *p = &data[stbi__jpeg_dezigzag[k]]; + if (*p != 0) + if (stbi__jpeg_get_bit(j)) + if ((*p & bit)==0) { + if (*p > 0) + *p += bit; + else + *p -= bit; + } + } + } else { + k = j->spec_start; + do { + int r,s; + int rs = stbi__jpeg_huff_decode(j, hac); // @OPTIMIZE see if we can use the fast path here, advance-by-r is so slow, eh + if (rs < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + s = rs & 15; + r = rs >> 4; + if (s == 0) { + if (r < 15) { + j->eob_run = (1 << r) - 1; + if (r) + j->eob_run += stbi__jpeg_get_bits(j, r); + r = 64; // force end of block + } else { + // r=15 s=0 should write 16 0s, so we just do + // a run of 15 0s and then write s (which is 0), + // so we don't have to do anything special here + } + } else { + if (s != 1) return stbi__err("bad huffman code", "Corrupt JPEG"); + // sign bit + if (stbi__jpeg_get_bit(j)) + s = bit; + else + s = -bit; + } + + // advance by r + while (k <= j->spec_end) { + short *p = &data[stbi__jpeg_dezigzag[k++]]; + if (*p != 0) { + if (stbi__jpeg_get_bit(j)) + if ((*p & bit)==0) { + if (*p > 0) + *p += bit; + else + *p -= bit; + } + } else { + if (r == 0) { + *p = (short) s; + break; + } + --r; + } + } + } while (k <= j->spec_end); + } + } + return 1; +} + +// take a -128..127 value and stbi__clamp it and convert to 0..255 +stbi_inline static stbi_uc stbi__clamp(int x) +{ + // trick to use a single test to catch both cases + if ((unsigned int) x > 255) { + if (x < 0) return 0; + if (x > 255) return 255; + } + return (stbi_uc) x; +} + +#define stbi__f2f(x) ((int) (((x) * 4096 + 0.5))) +#define stbi__fsh(x) ((x) * 4096) + +// derived from jidctint -- DCT_ISLOW +#define STBI__IDCT_1D(s0,s1,s2,s3,s4,s5,s6,s7) \ + int t0,t1,t2,t3,p1,p2,p3,p4,p5,x0,x1,x2,x3; \ + p2 = s2; \ + p3 = s6; \ + p1 = (p2+p3) * stbi__f2f(0.5411961f); \ + t2 = p1 + p3*stbi__f2f(-1.847759065f); \ + t3 = p1 + p2*stbi__f2f( 0.765366865f); \ + p2 = s0; \ + p3 = s4; \ + t0 = stbi__fsh(p2+p3); \ + t1 = stbi__fsh(p2-p3); \ + x0 = t0+t3; \ + x3 = t0-t3; \ + x1 = t1+t2; \ + x2 = t1-t2; \ + t0 = s7; \ + t1 = s5; \ + t2 = s3; \ + t3 = s1; \ + p3 = t0+t2; \ + p4 = t1+t3; \ + p1 = t0+t3; \ + p2 = t1+t2; \ + p5 = (p3+p4)*stbi__f2f( 1.175875602f); \ + t0 = t0*stbi__f2f( 0.298631336f); \ + t1 = t1*stbi__f2f( 2.053119869f); \ + t2 = t2*stbi__f2f( 3.072711026f); \ + t3 = t3*stbi__f2f( 1.501321110f); \ + p1 = p5 + p1*stbi__f2f(-0.899976223f); \ + p2 = p5 + p2*stbi__f2f(-2.562915447f); \ + p3 = p3*stbi__f2f(-1.961570560f); \ + p4 = p4*stbi__f2f(-0.390180644f); \ + t3 += p1+p4; \ + t2 += p2+p3; \ + t1 += p2+p4; \ + t0 += p1+p3; + +static void stbi__idct_block(stbi_uc *out, int out_stride, short data[64]) +{ + int i,val[64],*v=val; + stbi_uc *o; + short *d = data; + + // columns + for (i=0; i < 8; ++i,++d, ++v) { + // if all zeroes, shortcut -- this avoids dequantizing 0s and IDCTing + if (d[ 8]==0 && d[16]==0 && d[24]==0 && d[32]==0 + && d[40]==0 && d[48]==0 && d[56]==0) { + // no shortcut 0 seconds + // (1|2|3|4|5|6|7)==0 0 seconds + // all separate -0.047 seconds + // 1 && 2|3 && 4|5 && 6|7: -0.047 seconds + int dcterm = d[0]*4; + v[0] = v[8] = v[16] = v[24] = v[32] = v[40] = v[48] = v[56] = dcterm; + } else { + STBI__IDCT_1D(d[ 0],d[ 8],d[16],d[24],d[32],d[40],d[48],d[56]) + // constants scaled things up by 1<<12; let's bring them back + // down, but keep 2 extra bits of precision + x0 += 512; x1 += 512; x2 += 512; x3 += 512; + v[ 0] = (x0+t3) >> 10; + v[56] = (x0-t3) >> 10; + v[ 8] = (x1+t2) >> 10; + v[48] = (x1-t2) >> 10; + v[16] = (x2+t1) >> 10; + v[40] = (x2-t1) >> 10; + v[24] = (x3+t0) >> 10; + v[32] = (x3-t0) >> 10; + } + } + + for (i=0, v=val, o=out; i < 8; ++i,v+=8,o+=out_stride) { + // no fast case since the first 1D IDCT spread components out + STBI__IDCT_1D(v[0],v[1],v[2],v[3],v[4],v[5],v[6],v[7]) + // constants scaled things up by 1<<12, plus we had 1<<2 from first + // loop, plus horizontal and vertical each scale by sqrt(8) so together + // we've got an extra 1<<3, so 1<<17 total we need to remove. + // so we want to round that, which means adding 0.5 * 1<<17, + // aka 65536. Also, we'll end up with -128 to 127 that we want + // to encode as 0..255 by adding 128, so we'll add that before the shift + x0 += 65536 + (128<<17); + x1 += 65536 + (128<<17); + x2 += 65536 + (128<<17); + x3 += 65536 + (128<<17); + // tried computing the shifts into temps, or'ing the temps to see + // if any were out of range, but that was slower + o[0] = stbi__clamp((x0+t3) >> 17); + o[7] = stbi__clamp((x0-t3) >> 17); + o[1] = stbi__clamp((x1+t2) >> 17); + o[6] = stbi__clamp((x1-t2) >> 17); + o[2] = stbi__clamp((x2+t1) >> 17); + o[5] = stbi__clamp((x2-t1) >> 17); + o[3] = stbi__clamp((x3+t0) >> 17); + o[4] = stbi__clamp((x3-t0) >> 17); + } +} + +#ifdef STBI_SSE2 +// sse2 integer IDCT. not the fastest possible implementation but it +// produces bit-identical results to the generic C version so it's +// fully "transparent". +static void stbi__idct_simd(stbi_uc *out, int out_stride, short data[64]) +{ + // This is constructed to match our regular (generic) integer IDCT exactly. + __m128i row0, row1, row2, row3, row4, row5, row6, row7; + __m128i tmp; + + // dot product constant: even elems=x, odd elems=y + #define dct_const(x,y) _mm_setr_epi16((x),(y),(x),(y),(x),(y),(x),(y)) + + // out(0) = c0[even]*x + c0[odd]*y (c0, x, y 16-bit, out 32-bit) + // out(1) = c1[even]*x + c1[odd]*y + #define dct_rot(out0,out1, x,y,c0,c1) \ + __m128i c0##lo = _mm_unpacklo_epi16((x),(y)); \ + __m128i c0##hi = _mm_unpackhi_epi16((x),(y)); \ + __m128i out0##_l = _mm_madd_epi16(c0##lo, c0); \ + __m128i out0##_h = _mm_madd_epi16(c0##hi, c0); \ + __m128i out1##_l = _mm_madd_epi16(c0##lo, c1); \ + __m128i out1##_h = _mm_madd_epi16(c0##hi, c1) + + // out = in << 12 (in 16-bit, out 32-bit) + #define dct_widen(out, in) \ + __m128i out##_l = _mm_srai_epi32(_mm_unpacklo_epi16(_mm_setzero_si128(), (in)), 4); \ + __m128i out##_h = _mm_srai_epi32(_mm_unpackhi_epi16(_mm_setzero_si128(), (in)), 4) + + // wide add + #define dct_wadd(out, a, b) \ + __m128i out##_l = _mm_add_epi32(a##_l, b##_l); \ + __m128i out##_h = _mm_add_epi32(a##_h, b##_h) + + // wide sub + #define dct_wsub(out, a, b) \ + __m128i out##_l = _mm_sub_epi32(a##_l, b##_l); \ + __m128i out##_h = _mm_sub_epi32(a##_h, b##_h) + + // butterfly a/b, add bias, then shift by "s" and pack + #define dct_bfly32o(out0, out1, a,b,bias,s) \ + { \ + __m128i abiased_l = _mm_add_epi32(a##_l, bias); \ + __m128i abiased_h = _mm_add_epi32(a##_h, bias); \ + dct_wadd(sum, abiased, b); \ + dct_wsub(dif, abiased, b); \ + out0 = _mm_packs_epi32(_mm_srai_epi32(sum_l, s), _mm_srai_epi32(sum_h, s)); \ + out1 = _mm_packs_epi32(_mm_srai_epi32(dif_l, s), _mm_srai_epi32(dif_h, s)); \ + } + + // 8-bit interleave step (for transposes) + #define dct_interleave8(a, b) \ + tmp = a; \ + a = _mm_unpacklo_epi8(a, b); \ + b = _mm_unpackhi_epi8(tmp, b) + + // 16-bit interleave step (for transposes) + #define dct_interleave16(a, b) \ + tmp = a; \ + a = _mm_unpacklo_epi16(a, b); \ + b = _mm_unpackhi_epi16(tmp, b) + + #define dct_pass(bias,shift) \ + { \ + /* even part */ \ + dct_rot(t2e,t3e, row2,row6, rot0_0,rot0_1); \ + __m128i sum04 = _mm_add_epi16(row0, row4); \ + __m128i dif04 = _mm_sub_epi16(row0, row4); \ + dct_widen(t0e, sum04); \ + dct_widen(t1e, dif04); \ + dct_wadd(x0, t0e, t3e); \ + dct_wsub(x3, t0e, t3e); \ + dct_wadd(x1, t1e, t2e); \ + dct_wsub(x2, t1e, t2e); \ + /* odd part */ \ + dct_rot(y0o,y2o, row7,row3, rot2_0,rot2_1); \ + dct_rot(y1o,y3o, row5,row1, rot3_0,rot3_1); \ + __m128i sum17 = _mm_add_epi16(row1, row7); \ + __m128i sum35 = _mm_add_epi16(row3, row5); \ + dct_rot(y4o,y5o, sum17,sum35, rot1_0,rot1_1); \ + dct_wadd(x4, y0o, y4o); \ + dct_wadd(x5, y1o, y5o); \ + dct_wadd(x6, y2o, y5o); \ + dct_wadd(x7, y3o, y4o); \ + dct_bfly32o(row0,row7, x0,x7,bias,shift); \ + dct_bfly32o(row1,row6, x1,x6,bias,shift); \ + dct_bfly32o(row2,row5, x2,x5,bias,shift); \ + dct_bfly32o(row3,row4, x3,x4,bias,shift); \ + } + + __m128i rot0_0 = dct_const(stbi__f2f(0.5411961f), stbi__f2f(0.5411961f) + stbi__f2f(-1.847759065f)); + __m128i rot0_1 = dct_const(stbi__f2f(0.5411961f) + stbi__f2f( 0.765366865f), stbi__f2f(0.5411961f)); + __m128i rot1_0 = dct_const(stbi__f2f(1.175875602f) + stbi__f2f(-0.899976223f), stbi__f2f(1.175875602f)); + __m128i rot1_1 = dct_const(stbi__f2f(1.175875602f), stbi__f2f(1.175875602f) + stbi__f2f(-2.562915447f)); + __m128i rot2_0 = dct_const(stbi__f2f(-1.961570560f) + stbi__f2f( 0.298631336f), stbi__f2f(-1.961570560f)); + __m128i rot2_1 = dct_const(stbi__f2f(-1.961570560f), stbi__f2f(-1.961570560f) + stbi__f2f( 3.072711026f)); + __m128i rot3_0 = dct_const(stbi__f2f(-0.390180644f) + stbi__f2f( 2.053119869f), stbi__f2f(-0.390180644f)); + __m128i rot3_1 = dct_const(stbi__f2f(-0.390180644f), stbi__f2f(-0.390180644f) + stbi__f2f( 1.501321110f)); + + // rounding biases in column/row passes, see stbi__idct_block for explanation. + __m128i bias_0 = _mm_set1_epi32(512); + __m128i bias_1 = _mm_set1_epi32(65536 + (128<<17)); + + // load + row0 = _mm_load_si128((const __m128i *) (data + 0*8)); + row1 = _mm_load_si128((const __m128i *) (data + 1*8)); + row2 = _mm_load_si128((const __m128i *) (data + 2*8)); + row3 = _mm_load_si128((const __m128i *) (data + 3*8)); + row4 = _mm_load_si128((const __m128i *) (data + 4*8)); + row5 = _mm_load_si128((const __m128i *) (data + 5*8)); + row6 = _mm_load_si128((const __m128i *) (data + 6*8)); + row7 = _mm_load_si128((const __m128i *) (data + 7*8)); + + // column pass + dct_pass(bias_0, 10); + + { + // 16bit 8x8 transpose pass 1 + dct_interleave16(row0, row4); + dct_interleave16(row1, row5); + dct_interleave16(row2, row6); + dct_interleave16(row3, row7); + + // transpose pass 2 + dct_interleave16(row0, row2); + dct_interleave16(row1, row3); + dct_interleave16(row4, row6); + dct_interleave16(row5, row7); + + // transpose pass 3 + dct_interleave16(row0, row1); + dct_interleave16(row2, row3); + dct_interleave16(row4, row5); + dct_interleave16(row6, row7); + } + + // row pass + dct_pass(bias_1, 17); + + { + // pack + __m128i p0 = _mm_packus_epi16(row0, row1); // a0a1a2a3...a7b0b1b2b3...b7 + __m128i p1 = _mm_packus_epi16(row2, row3); + __m128i p2 = _mm_packus_epi16(row4, row5); + __m128i p3 = _mm_packus_epi16(row6, row7); + + // 8bit 8x8 transpose pass 1 + dct_interleave8(p0, p2); // a0e0a1e1... + dct_interleave8(p1, p3); // c0g0c1g1... + + // transpose pass 2 + dct_interleave8(p0, p1); // a0c0e0g0... + dct_interleave8(p2, p3); // b0d0f0h0... + + // transpose pass 3 + dct_interleave8(p0, p2); // a0b0c0d0... + dct_interleave8(p1, p3); // a4b4c4d4... + + // store + _mm_storel_epi64((__m128i *) out, p0); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p0, 0x4e)); out += out_stride; + _mm_storel_epi64((__m128i *) out, p2); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p2, 0x4e)); out += out_stride; + _mm_storel_epi64((__m128i *) out, p1); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p1, 0x4e)); out += out_stride; + _mm_storel_epi64((__m128i *) out, p3); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p3, 0x4e)); + } + +#undef dct_const +#undef dct_rot +#undef dct_widen +#undef dct_wadd +#undef dct_wsub +#undef dct_bfly32o +#undef dct_interleave8 +#undef dct_interleave16 +#undef dct_pass +} + +#endif // STBI_SSE2 + +#ifdef STBI_NEON + +// NEON integer IDCT. should produce bit-identical +// results to the generic C version. +static void stbi__idct_simd(stbi_uc *out, int out_stride, short data[64]) +{ + int16x8_t row0, row1, row2, row3, row4, row5, row6, row7; + + int16x4_t rot0_0 = vdup_n_s16(stbi__f2f(0.5411961f)); + int16x4_t rot0_1 = vdup_n_s16(stbi__f2f(-1.847759065f)); + int16x4_t rot0_2 = vdup_n_s16(stbi__f2f( 0.765366865f)); + int16x4_t rot1_0 = vdup_n_s16(stbi__f2f( 1.175875602f)); + int16x4_t rot1_1 = vdup_n_s16(stbi__f2f(-0.899976223f)); + int16x4_t rot1_2 = vdup_n_s16(stbi__f2f(-2.562915447f)); + int16x4_t rot2_0 = vdup_n_s16(stbi__f2f(-1.961570560f)); + int16x4_t rot2_1 = vdup_n_s16(stbi__f2f(-0.390180644f)); + int16x4_t rot3_0 = vdup_n_s16(stbi__f2f( 0.298631336f)); + int16x4_t rot3_1 = vdup_n_s16(stbi__f2f( 2.053119869f)); + int16x4_t rot3_2 = vdup_n_s16(stbi__f2f( 3.072711026f)); + int16x4_t rot3_3 = vdup_n_s16(stbi__f2f( 1.501321110f)); + +#define dct_long_mul(out, inq, coeff) \ + int32x4_t out##_l = vmull_s16(vget_low_s16(inq), coeff); \ + int32x4_t out##_h = vmull_s16(vget_high_s16(inq), coeff) + +#define dct_long_mac(out, acc, inq, coeff) \ + int32x4_t out##_l = vmlal_s16(acc##_l, vget_low_s16(inq), coeff); \ + int32x4_t out##_h = vmlal_s16(acc##_h, vget_high_s16(inq), coeff) + +#define dct_widen(out, inq) \ + int32x4_t out##_l = vshll_n_s16(vget_low_s16(inq), 12); \ + int32x4_t out##_h = vshll_n_s16(vget_high_s16(inq), 12) + +// wide add +#define dct_wadd(out, a, b) \ + int32x4_t out##_l = vaddq_s32(a##_l, b##_l); \ + int32x4_t out##_h = vaddq_s32(a##_h, b##_h) + +// wide sub +#define dct_wsub(out, a, b) \ + int32x4_t out##_l = vsubq_s32(a##_l, b##_l); \ + int32x4_t out##_h = vsubq_s32(a##_h, b##_h) + +// butterfly a/b, then shift using "shiftop" by "s" and pack +#define dct_bfly32o(out0,out1, a,b,shiftop,s) \ + { \ + dct_wadd(sum, a, b); \ + dct_wsub(dif, a, b); \ + out0 = vcombine_s16(shiftop(sum_l, s), shiftop(sum_h, s)); \ + out1 = vcombine_s16(shiftop(dif_l, s), shiftop(dif_h, s)); \ + } + +#define dct_pass(shiftop, shift) \ + { \ + /* even part */ \ + int16x8_t sum26 = vaddq_s16(row2, row6); \ + dct_long_mul(p1e, sum26, rot0_0); \ + dct_long_mac(t2e, p1e, row6, rot0_1); \ + dct_long_mac(t3e, p1e, row2, rot0_2); \ + int16x8_t sum04 = vaddq_s16(row0, row4); \ + int16x8_t dif04 = vsubq_s16(row0, row4); \ + dct_widen(t0e, sum04); \ + dct_widen(t1e, dif04); \ + dct_wadd(x0, t0e, t3e); \ + dct_wsub(x3, t0e, t3e); \ + dct_wadd(x1, t1e, t2e); \ + dct_wsub(x2, t1e, t2e); \ + /* odd part */ \ + int16x8_t sum15 = vaddq_s16(row1, row5); \ + int16x8_t sum17 = vaddq_s16(row1, row7); \ + int16x8_t sum35 = vaddq_s16(row3, row5); \ + int16x8_t sum37 = vaddq_s16(row3, row7); \ + int16x8_t sumodd = vaddq_s16(sum17, sum35); \ + dct_long_mul(p5o, sumodd, rot1_0); \ + dct_long_mac(p1o, p5o, sum17, rot1_1); \ + dct_long_mac(p2o, p5o, sum35, rot1_2); \ + dct_long_mul(p3o, sum37, rot2_0); \ + dct_long_mul(p4o, sum15, rot2_1); \ + dct_wadd(sump13o, p1o, p3o); \ + dct_wadd(sump24o, p2o, p4o); \ + dct_wadd(sump23o, p2o, p3o); \ + dct_wadd(sump14o, p1o, p4o); \ + dct_long_mac(x4, sump13o, row7, rot3_0); \ + dct_long_mac(x5, sump24o, row5, rot3_1); \ + dct_long_mac(x6, sump23o, row3, rot3_2); \ + dct_long_mac(x7, sump14o, row1, rot3_3); \ + dct_bfly32o(row0,row7, x0,x7,shiftop,shift); \ + dct_bfly32o(row1,row6, x1,x6,shiftop,shift); \ + dct_bfly32o(row2,row5, x2,x5,shiftop,shift); \ + dct_bfly32o(row3,row4, x3,x4,shiftop,shift); \ + } + + // load + row0 = vld1q_s16(data + 0*8); + row1 = vld1q_s16(data + 1*8); + row2 = vld1q_s16(data + 2*8); + row3 = vld1q_s16(data + 3*8); + row4 = vld1q_s16(data + 4*8); + row5 = vld1q_s16(data + 5*8); + row6 = vld1q_s16(data + 6*8); + row7 = vld1q_s16(data + 7*8); + + // add DC bias + row0 = vaddq_s16(row0, vsetq_lane_s16(1024, vdupq_n_s16(0), 0)); + + // column pass + dct_pass(vrshrn_n_s32, 10); + + // 16bit 8x8 transpose + { +// these three map to a single VTRN.16, VTRN.32, and VSWP, respectively. +// whether compilers actually get this is another story, sadly. +#define dct_trn16(x, y) { int16x8x2_t t = vtrnq_s16(x, y); x = t.val[0]; y = t.val[1]; } +#define dct_trn32(x, y) { int32x4x2_t t = vtrnq_s32(vreinterpretq_s32_s16(x), vreinterpretq_s32_s16(y)); x = vreinterpretq_s16_s32(t.val[0]); y = vreinterpretq_s16_s32(t.val[1]); } +#define dct_trn64(x, y) { int16x8_t x0 = x; int16x8_t y0 = y; x = vcombine_s16(vget_low_s16(x0), vget_low_s16(y0)); y = vcombine_s16(vget_high_s16(x0), vget_high_s16(y0)); } + + // pass 1 + dct_trn16(row0, row1); // a0b0a2b2a4b4a6b6 + dct_trn16(row2, row3); + dct_trn16(row4, row5); + dct_trn16(row6, row7); + + // pass 2 + dct_trn32(row0, row2); // a0b0c0d0a4b4c4d4 + dct_trn32(row1, row3); + dct_trn32(row4, row6); + dct_trn32(row5, row7); + + // pass 3 + dct_trn64(row0, row4); // a0b0c0d0e0f0g0h0 + dct_trn64(row1, row5); + dct_trn64(row2, row6); + dct_trn64(row3, row7); + +#undef dct_trn16 +#undef dct_trn32 +#undef dct_trn64 + } + + // row pass + // vrshrn_n_s32 only supports shifts up to 16, we need + // 17. so do a non-rounding shift of 16 first then follow + // up with a rounding shift by 1. + dct_pass(vshrn_n_s32, 16); + + { + // pack and round + uint8x8_t p0 = vqrshrun_n_s16(row0, 1); + uint8x8_t p1 = vqrshrun_n_s16(row1, 1); + uint8x8_t p2 = vqrshrun_n_s16(row2, 1); + uint8x8_t p3 = vqrshrun_n_s16(row3, 1); + uint8x8_t p4 = vqrshrun_n_s16(row4, 1); + uint8x8_t p5 = vqrshrun_n_s16(row5, 1); + uint8x8_t p6 = vqrshrun_n_s16(row6, 1); + uint8x8_t p7 = vqrshrun_n_s16(row7, 1); + + // again, these can translate into one instruction, but often don't. +#define dct_trn8_8(x, y) { uint8x8x2_t t = vtrn_u8(x, y); x = t.val[0]; y = t.val[1]; } +#define dct_trn8_16(x, y) { uint16x4x2_t t = vtrn_u16(vreinterpret_u16_u8(x), vreinterpret_u16_u8(y)); x = vreinterpret_u8_u16(t.val[0]); y = vreinterpret_u8_u16(t.val[1]); } +#define dct_trn8_32(x, y) { uint32x2x2_t t = vtrn_u32(vreinterpret_u32_u8(x), vreinterpret_u32_u8(y)); x = vreinterpret_u8_u32(t.val[0]); y = vreinterpret_u8_u32(t.val[1]); } + + // sadly can't use interleaved stores here since we only write + // 8 bytes to each scan line! + + // 8x8 8-bit transpose pass 1 + dct_trn8_8(p0, p1); + dct_trn8_8(p2, p3); + dct_trn8_8(p4, p5); + dct_trn8_8(p6, p7); + + // pass 2 + dct_trn8_16(p0, p2); + dct_trn8_16(p1, p3); + dct_trn8_16(p4, p6); + dct_trn8_16(p5, p7); + + // pass 3 + dct_trn8_32(p0, p4); + dct_trn8_32(p1, p5); + dct_trn8_32(p2, p6); + dct_trn8_32(p3, p7); + + // store + vst1_u8(out, p0); out += out_stride; + vst1_u8(out, p1); out += out_stride; + vst1_u8(out, p2); out += out_stride; + vst1_u8(out, p3); out += out_stride; + vst1_u8(out, p4); out += out_stride; + vst1_u8(out, p5); out += out_stride; + vst1_u8(out, p6); out += out_stride; + vst1_u8(out, p7); + +#undef dct_trn8_8 +#undef dct_trn8_16 +#undef dct_trn8_32 + } + +#undef dct_long_mul +#undef dct_long_mac +#undef dct_widen +#undef dct_wadd +#undef dct_wsub +#undef dct_bfly32o +#undef dct_pass +} + +#endif // STBI_NEON + +#define STBI__MARKER_none 0xff +// if there's a pending marker from the entropy stream, return that +// otherwise, fetch from the stream and get a marker. if there's no +// marker, return 0xff, which is never a valid marker value +static stbi_uc stbi__get_marker(stbi__jpeg *j) +{ + stbi_uc x; + if (j->marker != STBI__MARKER_none) { x = j->marker; j->marker = STBI__MARKER_none; return x; } + x = stbi__get8(j->s); + if (x != 0xff) return STBI__MARKER_none; + while (x == 0xff) + x = stbi__get8(j->s); // consume repeated 0xff fill bytes + return x; +} + +// in each scan, we'll have scan_n components, and the order +// of the components is specified by order[] +#define STBI__RESTART(x) ((x) >= 0xd0 && (x) <= 0xd7) + +// after a restart interval, stbi__jpeg_reset the entropy decoder and +// the dc prediction +static void stbi__jpeg_reset(stbi__jpeg *j) +{ + j->code_bits = 0; + j->code_buffer = 0; + j->nomore = 0; + j->img_comp[0].dc_pred = j->img_comp[1].dc_pred = j->img_comp[2].dc_pred = j->img_comp[3].dc_pred = 0; + j->marker = STBI__MARKER_none; + j->todo = j->restart_interval ? j->restart_interval : 0x7fffffff; + j->eob_run = 0; + // no more than 1<<31 MCUs if no restart_interal? that's plenty safe, + // since we don't even allow 1<<30 pixels +} + +static int stbi__parse_entropy_coded_data(stbi__jpeg *z) +{ + stbi__jpeg_reset(z); + if (!z->progressive) { + if (z->scan_n == 1) { + int i,j; + STBI_SIMD_ALIGN(short, data[64]); + int n = z->order[0]; + // non-interleaved data, we just need to process one block at a time, + // in trivial scanline order + // number of blocks to do just depends on how many actual "pixels" this + // component has, independent of interleaved MCU blocking and such + int w = (z->img_comp[n].x+7) >> 3; + int h = (z->img_comp[n].y+7) >> 3; + for (j=0; j < h; ++j) { + for (i=0; i < w; ++i) { + int ha = z->img_comp[n].ha; + if (!stbi__jpeg_decode_block(z, data, z->huff_dc+z->img_comp[n].hd, z->huff_ac+ha, z->fast_ac[ha], n, z->dequant[z->img_comp[n].tq])) return 0; + z->idct_block_kernel(z->img_comp[n].data+z->img_comp[n].w2*j*8+i*8, z->img_comp[n].w2, data); + // every data block is an MCU, so countdown the restart interval + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + // if it's NOT a restart, then just bail, so we get corrupt data + // rather than no data + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } else { // interleaved + int i,j,k,x,y; + STBI_SIMD_ALIGN(short, data[64]); + for (j=0; j < z->img_mcu_y; ++j) { + for (i=0; i < z->img_mcu_x; ++i) { + // scan an interleaved mcu... process scan_n components in order + for (k=0; k < z->scan_n; ++k) { + int n = z->order[k]; + // scan out an mcu's worth of this component; that's just determined + // by the basic H and V specified for the component + for (y=0; y < z->img_comp[n].v; ++y) { + for (x=0; x < z->img_comp[n].h; ++x) { + int x2 = (i*z->img_comp[n].h + x)*8; + int y2 = (j*z->img_comp[n].v + y)*8; + int ha = z->img_comp[n].ha; + if (!stbi__jpeg_decode_block(z, data, z->huff_dc+z->img_comp[n].hd, z->huff_ac+ha, z->fast_ac[ha], n, z->dequant[z->img_comp[n].tq])) return 0; + z->idct_block_kernel(z->img_comp[n].data+z->img_comp[n].w2*y2+x2, z->img_comp[n].w2, data); + } + } + } + // after all interleaved components, that's an interleaved MCU, + // so now count down the restart interval + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } + } else { + if (z->scan_n == 1) { + int i,j; + int n = z->order[0]; + // non-interleaved data, we just need to process one block at a time, + // in trivial scanline order + // number of blocks to do just depends on how many actual "pixels" this + // component has, independent of interleaved MCU blocking and such + int w = (z->img_comp[n].x+7) >> 3; + int h = (z->img_comp[n].y+7) >> 3; + for (j=0; j < h; ++j) { + for (i=0; i < w; ++i) { + short *data = z->img_comp[n].coeff + 64 * (i + j * z->img_comp[n].coeff_w); + if (z->spec_start == 0) { + if (!stbi__jpeg_decode_block_prog_dc(z, data, &z->huff_dc[z->img_comp[n].hd], n)) + return 0; + } else { + int ha = z->img_comp[n].ha; + if (!stbi__jpeg_decode_block_prog_ac(z, data, &z->huff_ac[ha], z->fast_ac[ha])) + return 0; + } + // every data block is an MCU, so countdown the restart interval + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } else { // interleaved + int i,j,k,x,y; + for (j=0; j < z->img_mcu_y; ++j) { + for (i=0; i < z->img_mcu_x; ++i) { + // scan an interleaved mcu... process scan_n components in order + for (k=0; k < z->scan_n; ++k) { + int n = z->order[k]; + // scan out an mcu's worth of this component; that's just determined + // by the basic H and V specified for the component + for (y=0; y < z->img_comp[n].v; ++y) { + for (x=0; x < z->img_comp[n].h; ++x) { + int x2 = (i*z->img_comp[n].h + x); + int y2 = (j*z->img_comp[n].v + y); + short *data = z->img_comp[n].coeff + 64 * (x2 + y2 * z->img_comp[n].coeff_w); + if (!stbi__jpeg_decode_block_prog_dc(z, data, &z->huff_dc[z->img_comp[n].hd], n)) + return 0; + } + } + } + // after all interleaved components, that's an interleaved MCU, + // so now count down the restart interval + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } + } +} + +static void stbi__jpeg_dequantize(short *data, stbi__uint16 *dequant) +{ + int i; + for (i=0; i < 64; ++i) + data[i] *= dequant[i]; +} + +static void stbi__jpeg_finish(stbi__jpeg *z) +{ + if (z->progressive) { + // dequantize and idct the data + int i,j,n; + for (n=0; n < z->s->img_n; ++n) { + int w = (z->img_comp[n].x+7) >> 3; + int h = (z->img_comp[n].y+7) >> 3; + for (j=0; j < h; ++j) { + for (i=0; i < w; ++i) { + short *data = z->img_comp[n].coeff + 64 * (i + j * z->img_comp[n].coeff_w); + stbi__jpeg_dequantize(data, z->dequant[z->img_comp[n].tq]); + z->idct_block_kernel(z->img_comp[n].data+z->img_comp[n].w2*j*8+i*8, z->img_comp[n].w2, data); + } + } + } + } +} + +static int stbi__process_marker(stbi__jpeg *z, int m) +{ + int L; + switch (m) { + case STBI__MARKER_none: // no marker found + return stbi__err("expected marker","Corrupt JPEG"); + + case 0xDD: // DRI - specify restart interval + if (stbi__get16be(z->s) != 4) return stbi__err("bad DRI len","Corrupt JPEG"); + z->restart_interval = stbi__get16be(z->s); + return 1; + + case 0xDB: // DQT - define quantization table + L = stbi__get16be(z->s)-2; + while (L > 0) { + int q = stbi__get8(z->s); + int p = q >> 4, sixteen = (p != 0); + int t = q & 15,i; + if (p != 0 && p != 1) return stbi__err("bad DQT type","Corrupt JPEG"); + if (t > 3) return stbi__err("bad DQT table","Corrupt JPEG"); + + for (i=0; i < 64; ++i) + z->dequant[t][stbi__jpeg_dezigzag[i]] = (stbi__uint16)(sixteen ? stbi__get16be(z->s) : stbi__get8(z->s)); + L -= (sixteen ? 129 : 65); + } + return L==0; + + case 0xC4: // DHT - define huffman table + L = stbi__get16be(z->s)-2; + while (L > 0) { + stbi_uc *v; + int sizes[16],i,n=0; + int q = stbi__get8(z->s); + int tc = q >> 4; + int th = q & 15; + if (tc > 1 || th > 3) return stbi__err("bad DHT header","Corrupt JPEG"); + for (i=0; i < 16; ++i) { + sizes[i] = stbi__get8(z->s); + n += sizes[i]; + } + L -= 17; + if (tc == 0) { + if (!stbi__build_huffman(z->huff_dc+th, sizes)) return 0; + v = z->huff_dc[th].values; + } else { + if (!stbi__build_huffman(z->huff_ac+th, sizes)) return 0; + v = z->huff_ac[th].values; + } + for (i=0; i < n; ++i) + v[i] = stbi__get8(z->s); + if (tc != 0) + stbi__build_fast_ac(z->fast_ac[th], z->huff_ac + th); + L -= n; + } + return L==0; + } + + // check for comment block or APP blocks + if ((m >= 0xE0 && m <= 0xEF) || m == 0xFE) { + L = stbi__get16be(z->s); + if (L < 2) { + if (m == 0xFE) + return stbi__err("bad COM len","Corrupt JPEG"); + else + return stbi__err("bad APP len","Corrupt JPEG"); + } + L -= 2; + + if (m == 0xE0 && L >= 5) { // JFIF APP0 segment + static const unsigned char tag[5] = {'J','F','I','F','\0'}; + int ok = 1; + int i; + for (i=0; i < 5; ++i) + if (stbi__get8(z->s) != tag[i]) + ok = 0; + L -= 5; + if (ok) + z->jfif = 1; + } else if (m == 0xEE && L >= 12) { // Adobe APP14 segment + static const unsigned char tag[6] = {'A','d','o','b','e','\0'}; + int ok = 1; + int i; + for (i=0; i < 6; ++i) + if (stbi__get8(z->s) != tag[i]) + ok = 0; + L -= 6; + if (ok) { + stbi__get8(z->s); // version + stbi__get16be(z->s); // flags0 + stbi__get16be(z->s); // flags1 + z->app14_color_transform = stbi__get8(z->s); // color transform + L -= 6; + } + } + + stbi__skip(z->s, L); + return 1; + } + + return stbi__err("unknown marker","Corrupt JPEG"); +} + +// after we see SOS +static int stbi__process_scan_header(stbi__jpeg *z) +{ + int i; + int Ls = stbi__get16be(z->s); + z->scan_n = stbi__get8(z->s); + if (z->scan_n < 1 || z->scan_n > 4 || z->scan_n > (int) z->s->img_n) return stbi__err("bad SOS component count","Corrupt JPEG"); + if (Ls != 6+2*z->scan_n) return stbi__err("bad SOS len","Corrupt JPEG"); + for (i=0; i < z->scan_n; ++i) { + int id = stbi__get8(z->s), which; + int q = stbi__get8(z->s); + for (which = 0; which < z->s->img_n; ++which) + if (z->img_comp[which].id == id) + break; + if (which == z->s->img_n) return 0; // no match + z->img_comp[which].hd = q >> 4; if (z->img_comp[which].hd > 3) return stbi__err("bad DC huff","Corrupt JPEG"); + z->img_comp[which].ha = q & 15; if (z->img_comp[which].ha > 3) return stbi__err("bad AC huff","Corrupt JPEG"); + z->order[i] = which; + } + + { + int aa; + z->spec_start = stbi__get8(z->s); + z->spec_end = stbi__get8(z->s); // should be 63, but might be 0 + aa = stbi__get8(z->s); + z->succ_high = (aa >> 4); + z->succ_low = (aa & 15); + if (z->progressive) { + if (z->spec_start > 63 || z->spec_end > 63 || z->spec_start > z->spec_end || z->succ_high > 13 || z->succ_low > 13) + return stbi__err("bad SOS", "Corrupt JPEG"); + } else { + if (z->spec_start != 0) return stbi__err("bad SOS","Corrupt JPEG"); + if (z->succ_high != 0 || z->succ_low != 0) return stbi__err("bad SOS","Corrupt JPEG"); + z->spec_end = 63; + } + } + + return 1; +} + +static int stbi__free_jpeg_components(stbi__jpeg *z, int ncomp, int why) +{ + int i; + for (i=0; i < ncomp; ++i) { + if (z->img_comp[i].raw_data) { + STBI_FREE(z->img_comp[i].raw_data); + z->img_comp[i].raw_data = NULL; + z->img_comp[i].data = NULL; + } + if (z->img_comp[i].raw_coeff) { + STBI_FREE(z->img_comp[i].raw_coeff); + z->img_comp[i].raw_coeff = 0; + z->img_comp[i].coeff = 0; + } + if (z->img_comp[i].linebuf) { + STBI_FREE(z->img_comp[i].linebuf); + z->img_comp[i].linebuf = NULL; + } + } + return why; +} + +static int stbi__process_frame_header(stbi__jpeg *z, int scan) +{ + stbi__context *s = z->s; + int Lf,p,i,q, h_max=1,v_max=1,c; + Lf = stbi__get16be(s); if (Lf < 11) return stbi__err("bad SOF len","Corrupt JPEG"); // JPEG + p = stbi__get8(s); if (p != 8) return stbi__err("only 8-bit","JPEG format not supported: 8-bit only"); // JPEG baseline + s->img_y = stbi__get16be(s); if (s->img_y == 0) return stbi__err("no header height", "JPEG format not supported: delayed height"); // Legal, but we don't handle it--but neither does IJG + s->img_x = stbi__get16be(s); if (s->img_x == 0) return stbi__err("0 width","Corrupt JPEG"); // JPEG requires + c = stbi__get8(s); + if (c != 3 && c != 1 && c != 4) return stbi__err("bad component count","Corrupt JPEG"); + s->img_n = c; + for (i=0; i < c; ++i) { + z->img_comp[i].data = NULL; + z->img_comp[i].linebuf = NULL; + } + + if (Lf != 8+3*s->img_n) return stbi__err("bad SOF len","Corrupt JPEG"); + + z->rgb = 0; + for (i=0; i < s->img_n; ++i) { + static const unsigned char rgb[3] = { 'R', 'G', 'B' }; + z->img_comp[i].id = stbi__get8(s); + if (s->img_n == 3 && z->img_comp[i].id == rgb[i]) + ++z->rgb; + q = stbi__get8(s); + z->img_comp[i].h = (q >> 4); if (!z->img_comp[i].h || z->img_comp[i].h > 4) return stbi__err("bad H","Corrupt JPEG"); + z->img_comp[i].v = q & 15; if (!z->img_comp[i].v || z->img_comp[i].v > 4) return stbi__err("bad V","Corrupt JPEG"); + z->img_comp[i].tq = stbi__get8(s); if (z->img_comp[i].tq > 3) return stbi__err("bad TQ","Corrupt JPEG"); + } + + if (scan != STBI__SCAN_load) return 1; + + if (!stbi__mad3sizes_valid(s->img_x, s->img_y, s->img_n, 0)) return stbi__err("too large", "Image too large to decode"); + + for (i=0; i < s->img_n; ++i) { + if (z->img_comp[i].h > h_max) h_max = z->img_comp[i].h; + if (z->img_comp[i].v > v_max) v_max = z->img_comp[i].v; + } + + // compute interleaved mcu info + z->img_h_max = h_max; + z->img_v_max = v_max; + z->img_mcu_w = h_max * 8; + z->img_mcu_h = v_max * 8; + // these sizes can't be more than 17 bits + z->img_mcu_x = (s->img_x + z->img_mcu_w-1) / z->img_mcu_w; + z->img_mcu_y = (s->img_y + z->img_mcu_h-1) / z->img_mcu_h; + + for (i=0; i < s->img_n; ++i) { + // number of effective pixels (e.g. for non-interleaved MCU) + z->img_comp[i].x = (s->img_x * z->img_comp[i].h + h_max-1) / h_max; + z->img_comp[i].y = (s->img_y * z->img_comp[i].v + v_max-1) / v_max; + // to simplify generation, we'll allocate enough memory to decode + // the bogus oversized data from using interleaved MCUs and their + // big blocks (e.g. a 16x16 iMCU on an image of width 33); we won't + // discard the extra data until colorspace conversion + // + // img_mcu_x, img_mcu_y: <=17 bits; comp[i].h and .v are <=4 (checked earlier) + // so these muls can't overflow with 32-bit ints (which we require) + z->img_comp[i].w2 = z->img_mcu_x * z->img_comp[i].h * 8; + z->img_comp[i].h2 = z->img_mcu_y * z->img_comp[i].v * 8; + z->img_comp[i].coeff = 0; + z->img_comp[i].raw_coeff = 0; + z->img_comp[i].linebuf = NULL; + z->img_comp[i].raw_data = stbi__malloc_mad2(z->img_comp[i].w2, z->img_comp[i].h2, 15); + if (z->img_comp[i].raw_data == NULL) + return stbi__free_jpeg_components(z, i+1, stbi__err("outofmem", "Out of memory")); + // align blocks for idct using mmx/sse + z->img_comp[i].data = (stbi_uc*) (((size_t) z->img_comp[i].raw_data + 15) & ~15); + if (z->progressive) { + // w2, h2 are multiples of 8 (see above) + z->img_comp[i].coeff_w = z->img_comp[i].w2 / 8; + z->img_comp[i].coeff_h = z->img_comp[i].h2 / 8; + z->img_comp[i].raw_coeff = stbi__malloc_mad3(z->img_comp[i].w2, z->img_comp[i].h2, sizeof(short), 15); + if (z->img_comp[i].raw_coeff == NULL) + return stbi__free_jpeg_components(z, i+1, stbi__err("outofmem", "Out of memory")); + z->img_comp[i].coeff = (short*) (((size_t) z->img_comp[i].raw_coeff + 15) & ~15); + } + } + + return 1; +} + +// use comparisons since in some cases we handle more than one case (e.g. SOF) +#define stbi__DNL(x) ((x) == 0xdc) +#define stbi__SOI(x) ((x) == 0xd8) +#define stbi__EOI(x) ((x) == 0xd9) +#define stbi__SOF(x) ((x) == 0xc0 || (x) == 0xc1 || (x) == 0xc2) +#define stbi__SOS(x) ((x) == 0xda) + +#define stbi__SOF_progressive(x) ((x) == 0xc2) + +static int stbi__decode_jpeg_header(stbi__jpeg *z, int scan) +{ + int m; + z->jfif = 0; + z->app14_color_transform = -1; // valid values are 0,1,2 + z->marker = STBI__MARKER_none; // initialize cached marker to empty + m = stbi__get_marker(z); + if (!stbi__SOI(m)) return stbi__err("no SOI","Corrupt JPEG"); + if (scan == STBI__SCAN_type) return 1; + m = stbi__get_marker(z); + while (!stbi__SOF(m)) { + if (!stbi__process_marker(z,m)) return 0; + m = stbi__get_marker(z); + while (m == STBI__MARKER_none) { + // some files have extra padding after their blocks, so ok, we'll scan + if (stbi__at_eof(z->s)) return stbi__err("no SOF", "Corrupt JPEG"); + m = stbi__get_marker(z); + } + } + z->progressive = stbi__SOF_progressive(m); + if (!stbi__process_frame_header(z, scan)) return 0; + return 1; +} + +// decode image to YCbCr format +static int stbi__decode_jpeg_image(stbi__jpeg *j) +{ + int m; + for (m = 0; m < 4; m++) { + j->img_comp[m].raw_data = NULL; + j->img_comp[m].raw_coeff = NULL; + } + j->restart_interval = 0; + if (!stbi__decode_jpeg_header(j, STBI__SCAN_load)) return 0; + m = stbi__get_marker(j); + while (!stbi__EOI(m)) { + if (stbi__SOS(m)) { + if (!stbi__process_scan_header(j)) return 0; + if (!stbi__parse_entropy_coded_data(j)) return 0; + if (j->marker == STBI__MARKER_none ) { + // handle 0s at the end of image data from IP Kamera 9060 + while (!stbi__at_eof(j->s)) { + int x = stbi__get8(j->s); + if (x == 255) { + j->marker = stbi__get8(j->s); + break; + } + } + // if we reach eof without hitting a marker, stbi__get_marker() below will fail and we'll eventually return 0 + } + } else if (stbi__DNL(m)) { + int Ld = stbi__get16be(j->s); + stbi__uint32 NL = stbi__get16be(j->s); + if (Ld != 4) return stbi__err("bad DNL len", "Corrupt JPEG"); + if (NL != j->s->img_y) return stbi__err("bad DNL height", "Corrupt JPEG"); + } else { + if (!stbi__process_marker(j, m)) return 0; + } + m = stbi__get_marker(j); + } + if (j->progressive) + stbi__jpeg_finish(j); + return 1; +} + +// static jfif-centered resampling (across block boundaries) + +typedef stbi_uc *(*resample_row_func)(stbi_uc *out, stbi_uc *in0, stbi_uc *in1, + int w, int hs); + +#define stbi__div4(x) ((stbi_uc) ((x) >> 2)) + +static stbi_uc *resample_row_1(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + STBI_NOTUSED(out); + STBI_NOTUSED(in_far); + STBI_NOTUSED(w); + STBI_NOTUSED(hs); + return in_near; +} + +static stbi_uc* stbi__resample_row_v_2(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // need to generate two samples vertically for every one in input + int i; + STBI_NOTUSED(hs); + for (i=0; i < w; ++i) + out[i] = stbi__div4(3*in_near[i] + in_far[i] + 2); + return out; +} + +static stbi_uc* stbi__resample_row_h_2(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // need to generate two samples horizontally for every one in input + int i; + stbi_uc *input = in_near; + + if (w == 1) { + // if only one sample, can't do any interpolation + out[0] = out[1] = input[0]; + return out; + } + + out[0] = input[0]; + out[1] = stbi__div4(input[0]*3 + input[1] + 2); + for (i=1; i < w-1; ++i) { + int n = 3*input[i]+2; + out[i*2+0] = stbi__div4(n+input[i-1]); + out[i*2+1] = stbi__div4(n+input[i+1]); + } + out[i*2+0] = stbi__div4(input[w-2]*3 + input[w-1] + 2); + out[i*2+1] = input[w-1]; + + STBI_NOTUSED(in_far); + STBI_NOTUSED(hs); + + return out; +} + +#define stbi__div16(x) ((stbi_uc) ((x) >> 4)) + +static stbi_uc *stbi__resample_row_hv_2(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // need to generate 2x2 samples for every one in input + int i,t0,t1; + if (w == 1) { + out[0] = out[1] = stbi__div4(3*in_near[0] + in_far[0] + 2); + return out; + } + + t1 = 3*in_near[0] + in_far[0]; + out[0] = stbi__div4(t1+2); + for (i=1; i < w; ++i) { + t0 = t1; + t1 = 3*in_near[i]+in_far[i]; + out[i*2-1] = stbi__div16(3*t0 + t1 + 8); + out[i*2 ] = stbi__div16(3*t1 + t0 + 8); + } + out[w*2-1] = stbi__div4(t1+2); + + STBI_NOTUSED(hs); + + return out; +} + +#if defined(STBI_SSE2) || defined(STBI_NEON) +static stbi_uc *stbi__resample_row_hv_2_simd(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // need to generate 2x2 samples for every one in input + int i=0,t0,t1; + + if (w == 1) { + out[0] = out[1] = stbi__div4(3*in_near[0] + in_far[0] + 2); + return out; + } + + t1 = 3*in_near[0] + in_far[0]; + // process groups of 8 pixels for as long as we can. + // note we can't handle the last pixel in a row in this loop + // because we need to handle the filter boundary conditions. + for (; i < ((w-1) & ~7); i += 8) { +#if defined(STBI_SSE2) + // load and perform the vertical filtering pass + // this uses 3*x + y = 4*x + (y - x) + __m128i zero = _mm_setzero_si128(); + __m128i farb = _mm_loadl_epi64((__m128i *) (in_far + i)); + __m128i nearb = _mm_loadl_epi64((__m128i *) (in_near + i)); + __m128i farw = _mm_unpacklo_epi8(farb, zero); + __m128i nearw = _mm_unpacklo_epi8(nearb, zero); + __m128i diff = _mm_sub_epi16(farw, nearw); + __m128i nears = _mm_slli_epi16(nearw, 2); + __m128i curr = _mm_add_epi16(nears, diff); // current row + + // horizontal filter works the same based on shifted vers of current + // row. "prev" is current row shifted right by 1 pixel; we need to + // insert the previous pixel value (from t1). + // "next" is current row shifted left by 1 pixel, with first pixel + // of next block of 8 pixels added in. + __m128i prv0 = _mm_slli_si128(curr, 2); + __m128i nxt0 = _mm_srli_si128(curr, 2); + __m128i prev = _mm_insert_epi16(prv0, t1, 0); + __m128i next = _mm_insert_epi16(nxt0, 3*in_near[i+8] + in_far[i+8], 7); + + // horizontal filter, polyphase implementation since it's convenient: + // even pixels = 3*cur + prev = cur*4 + (prev - cur) + // odd pixels = 3*cur + next = cur*4 + (next - cur) + // note the shared term. + __m128i bias = _mm_set1_epi16(8); + __m128i curs = _mm_slli_epi16(curr, 2); + __m128i prvd = _mm_sub_epi16(prev, curr); + __m128i nxtd = _mm_sub_epi16(next, curr); + __m128i curb = _mm_add_epi16(curs, bias); + __m128i even = _mm_add_epi16(prvd, curb); + __m128i odd = _mm_add_epi16(nxtd, curb); + + // interleave even and odd pixels, then undo scaling. + __m128i int0 = _mm_unpacklo_epi16(even, odd); + __m128i int1 = _mm_unpackhi_epi16(even, odd); + __m128i de0 = _mm_srli_epi16(int0, 4); + __m128i de1 = _mm_srli_epi16(int1, 4); + + // pack and write output + __m128i outv = _mm_packus_epi16(de0, de1); + _mm_storeu_si128((__m128i *) (out + i*2), outv); +#elif defined(STBI_NEON) + // load and perform the vertical filtering pass + // this uses 3*x + y = 4*x + (y - x) + uint8x8_t farb = vld1_u8(in_far + i); + uint8x8_t nearb = vld1_u8(in_near + i); + int16x8_t diff = vreinterpretq_s16_u16(vsubl_u8(farb, nearb)); + int16x8_t nears = vreinterpretq_s16_u16(vshll_n_u8(nearb, 2)); + int16x8_t curr = vaddq_s16(nears, diff); // current row + + // horizontal filter works the same based on shifted vers of current + // row. "prev" is current row shifted right by 1 pixel; we need to + // insert the previous pixel value (from t1). + // "next" is current row shifted left by 1 pixel, with first pixel + // of next block of 8 pixels added in. + int16x8_t prv0 = vextq_s16(curr, curr, 7); + int16x8_t nxt0 = vextq_s16(curr, curr, 1); + int16x8_t prev = vsetq_lane_s16(t1, prv0, 0); + int16x8_t next = vsetq_lane_s16(3*in_near[i+8] + in_far[i+8], nxt0, 7); + + // horizontal filter, polyphase implementation since it's convenient: + // even pixels = 3*cur + prev = cur*4 + (prev - cur) + // odd pixels = 3*cur + next = cur*4 + (next - cur) + // note the shared term. + int16x8_t curs = vshlq_n_s16(curr, 2); + int16x8_t prvd = vsubq_s16(prev, curr); + int16x8_t nxtd = vsubq_s16(next, curr); + int16x8_t even = vaddq_s16(curs, prvd); + int16x8_t odd = vaddq_s16(curs, nxtd); + + // undo scaling and round, then store with even/odd phases interleaved + uint8x8x2_t o; + o.val[0] = vqrshrun_n_s16(even, 4); + o.val[1] = vqrshrun_n_s16(odd, 4); + vst2_u8(out + i*2, o); +#endif + + // "previous" value for next iter + t1 = 3*in_near[i+7] + in_far[i+7]; + } + + t0 = t1; + t1 = 3*in_near[i] + in_far[i]; + out[i*2] = stbi__div16(3*t1 + t0 + 8); + + for (++i; i < w; ++i) { + t0 = t1; + t1 = 3*in_near[i]+in_far[i]; + out[i*2-1] = stbi__div16(3*t0 + t1 + 8); + out[i*2 ] = stbi__div16(3*t1 + t0 + 8); + } + out[w*2-1] = stbi__div4(t1+2); + + STBI_NOTUSED(hs); + + return out; +} +#endif + +static stbi_uc *stbi__resample_row_generic(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + // resample with nearest-neighbor + int i,j; + STBI_NOTUSED(in_far); + for (i=0; i < w; ++i) + for (j=0; j < hs; ++j) + out[i*hs+j] = in_near[i]; + return out; +} + +// this is a reduced-precision calculation of YCbCr-to-RGB introduced +// to make sure the code produces the same results in both SIMD and scalar +#define stbi__float2fixed(x) (((int) ((x) * 4096.0f + 0.5f)) << 8) +static void stbi__YCbCr_to_RGB_row(stbi_uc *out, const stbi_uc *y, const stbi_uc *pcb, const stbi_uc *pcr, int count, int step) +{ + int i; + for (i=0; i < count; ++i) { + int y_fixed = (y[i] << 20) + (1<<19); // rounding + int r,g,b; + int cr = pcr[i] - 128; + int cb = pcb[i] - 128; + r = y_fixed + cr* stbi__float2fixed(1.40200f); + g = y_fixed + (cr*-stbi__float2fixed(0.71414f)) + ((cb*-stbi__float2fixed(0.34414f)) & 0xffff0000); + b = y_fixed + cb* stbi__float2fixed(1.77200f); + r >>= 20; + g >>= 20; + b >>= 20; + if ((unsigned) r > 255) { if (r < 0) r = 0; else r = 255; } + if ((unsigned) g > 255) { if (g < 0) g = 0; else g = 255; } + if ((unsigned) b > 255) { if (b < 0) b = 0; else b = 255; } + out[0] = (stbi_uc)r; + out[1] = (stbi_uc)g; + out[2] = (stbi_uc)b; + out[3] = 255; + out += step; + } +} + +#if defined(STBI_SSE2) || defined(STBI_NEON) +static void stbi__YCbCr_to_RGB_simd(stbi_uc *out, stbi_uc const *y, stbi_uc const *pcb, stbi_uc const *pcr, int count, int step) +{ + int i = 0; + +#ifdef STBI_SSE2 + // step == 3 is pretty ugly on the final interleave, and i'm not convinced + // it's useful in practice (you wouldn't use it for textures, for example). + // so just accelerate step == 4 case. + if (step == 4) { + // this is a fairly straightforward implementation and not super-optimized. + __m128i signflip = _mm_set1_epi8(-0x80); + __m128i cr_const0 = _mm_set1_epi16( (short) ( 1.40200f*4096.0f+0.5f)); + __m128i cr_const1 = _mm_set1_epi16( - (short) ( 0.71414f*4096.0f+0.5f)); + __m128i cb_const0 = _mm_set1_epi16( - (short) ( 0.34414f*4096.0f+0.5f)); + __m128i cb_const1 = _mm_set1_epi16( (short) ( 1.77200f*4096.0f+0.5f)); + __m128i y_bias = _mm_set1_epi8((char) (unsigned char) 128); + __m128i xw = _mm_set1_epi16(255); // alpha channel + + for (; i+7 < count; i += 8) { + // load + __m128i y_bytes = _mm_loadl_epi64((__m128i *) (y+i)); + __m128i cr_bytes = _mm_loadl_epi64((__m128i *) (pcr+i)); + __m128i cb_bytes = _mm_loadl_epi64((__m128i *) (pcb+i)); + __m128i cr_biased = _mm_xor_si128(cr_bytes, signflip); // -128 + __m128i cb_biased = _mm_xor_si128(cb_bytes, signflip); // -128 + + // unpack to short (and left-shift cr, cb by 8) + __m128i yw = _mm_unpacklo_epi8(y_bias, y_bytes); + __m128i crw = _mm_unpacklo_epi8(_mm_setzero_si128(), cr_biased); + __m128i cbw = _mm_unpacklo_epi8(_mm_setzero_si128(), cb_biased); + + // color transform + __m128i yws = _mm_srli_epi16(yw, 4); + __m128i cr0 = _mm_mulhi_epi16(cr_const0, crw); + __m128i cb0 = _mm_mulhi_epi16(cb_const0, cbw); + __m128i cb1 = _mm_mulhi_epi16(cbw, cb_const1); + __m128i cr1 = _mm_mulhi_epi16(crw, cr_const1); + __m128i rws = _mm_add_epi16(cr0, yws); + __m128i gwt = _mm_add_epi16(cb0, yws); + __m128i bws = _mm_add_epi16(yws, cb1); + __m128i gws = _mm_add_epi16(gwt, cr1); + + // descale + __m128i rw = _mm_srai_epi16(rws, 4); + __m128i bw = _mm_srai_epi16(bws, 4); + __m128i gw = _mm_srai_epi16(gws, 4); + + // back to byte, set up for transpose + __m128i brb = _mm_packus_epi16(rw, bw); + __m128i gxb = _mm_packus_epi16(gw, xw); + + // transpose to interleave channels + __m128i t0 = _mm_unpacklo_epi8(brb, gxb); + __m128i t1 = _mm_unpackhi_epi8(brb, gxb); + __m128i o0 = _mm_unpacklo_epi16(t0, t1); + __m128i o1 = _mm_unpackhi_epi16(t0, t1); + + // store + _mm_storeu_si128((__m128i *) (out + 0), o0); + _mm_storeu_si128((__m128i *) (out + 16), o1); + out += 32; + } + } +#endif + +#ifdef STBI_NEON + // in this version, step=3 support would be easy to add. but is there demand? + if (step == 4) { + // this is a fairly straightforward implementation and not super-optimized. + uint8x8_t signflip = vdup_n_u8(0x80); + int16x8_t cr_const0 = vdupq_n_s16( (short) ( 1.40200f*4096.0f+0.5f)); + int16x8_t cr_const1 = vdupq_n_s16( - (short) ( 0.71414f*4096.0f+0.5f)); + int16x8_t cb_const0 = vdupq_n_s16( - (short) ( 0.34414f*4096.0f+0.5f)); + int16x8_t cb_const1 = vdupq_n_s16( (short) ( 1.77200f*4096.0f+0.5f)); + + for (; i+7 < count; i += 8) { + // load + uint8x8_t y_bytes = vld1_u8(y + i); + uint8x8_t cr_bytes = vld1_u8(pcr + i); + uint8x8_t cb_bytes = vld1_u8(pcb + i); + int8x8_t cr_biased = vreinterpret_s8_u8(vsub_u8(cr_bytes, signflip)); + int8x8_t cb_biased = vreinterpret_s8_u8(vsub_u8(cb_bytes, signflip)); + + // expand to s16 + int16x8_t yws = vreinterpretq_s16_u16(vshll_n_u8(y_bytes, 4)); + int16x8_t crw = vshll_n_s8(cr_biased, 7); + int16x8_t cbw = vshll_n_s8(cb_biased, 7); + + // color transform + int16x8_t cr0 = vqdmulhq_s16(crw, cr_const0); + int16x8_t cb0 = vqdmulhq_s16(cbw, cb_const0); + int16x8_t cr1 = vqdmulhq_s16(crw, cr_const1); + int16x8_t cb1 = vqdmulhq_s16(cbw, cb_const1); + int16x8_t rws = vaddq_s16(yws, cr0); + int16x8_t gws = vaddq_s16(vaddq_s16(yws, cb0), cr1); + int16x8_t bws = vaddq_s16(yws, cb1); + + // undo scaling, round, convert to byte + uint8x8x4_t o; + o.val[0] = vqrshrun_n_s16(rws, 4); + o.val[1] = vqrshrun_n_s16(gws, 4); + o.val[2] = vqrshrun_n_s16(bws, 4); + o.val[3] = vdup_n_u8(255); + + // store, interleaving r/g/b/a + vst4_u8(out, o); + out += 8*4; + } + } +#endif + + for (; i < count; ++i) { + int y_fixed = (y[i] << 20) + (1<<19); // rounding + int r,g,b; + int cr = pcr[i] - 128; + int cb = pcb[i] - 128; + r = y_fixed + cr* stbi__float2fixed(1.40200f); + g = y_fixed + cr*-stbi__float2fixed(0.71414f) + ((cb*-stbi__float2fixed(0.34414f)) & 0xffff0000); + b = y_fixed + cb* stbi__float2fixed(1.77200f); + r >>= 20; + g >>= 20; + b >>= 20; + if ((unsigned) r > 255) { if (r < 0) r = 0; else r = 255; } + if ((unsigned) g > 255) { if (g < 0) g = 0; else g = 255; } + if ((unsigned) b > 255) { if (b < 0) b = 0; else b = 255; } + out[0] = (stbi_uc)r; + out[1] = (stbi_uc)g; + out[2] = (stbi_uc)b; + out[3] = 255; + out += step; + } +} +#endif + +// set up the kernels +static void stbi__setup_jpeg(stbi__jpeg *j) +{ + j->idct_block_kernel = stbi__idct_block; + j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_row; + j->resample_row_hv_2_kernel = stbi__resample_row_hv_2; + +#ifdef STBI_SSE2 + if (stbi__sse2_available()) { + j->idct_block_kernel = stbi__idct_simd; + j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_simd; + j->resample_row_hv_2_kernel = stbi__resample_row_hv_2_simd; + } +#endif + +#ifdef STBI_NEON + j->idct_block_kernel = stbi__idct_simd; + j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_simd; + j->resample_row_hv_2_kernel = stbi__resample_row_hv_2_simd; +#endif +} + +// clean up the temporary component buffers +static void stbi__cleanup_jpeg(stbi__jpeg *j) +{ + stbi__free_jpeg_components(j, j->s->img_n, 0); +} + +typedef struct +{ + resample_row_func resample; + stbi_uc *line0,*line1; + int hs,vs; // expansion factor in each axis + int w_lores; // horizontal pixels pre-expansion + int ystep; // how far through vertical expansion we are + int ypos; // which pre-expansion row we're on +} stbi__resample; + +// fast 0..255 * 0..255 => 0..255 rounded multiplication +static stbi_uc stbi__blinn_8x8(stbi_uc x, stbi_uc y) +{ + unsigned int t = x*y + 128; + return (stbi_uc) ((t + (t >>8)) >> 8); +} + +static stbi_uc *load_jpeg_image(stbi__jpeg *z, int *out_x, int *out_y, int *comp, int req_comp) +{ + int n, decode_n, is_rgb; + z->s->img_n = 0; // make stbi__cleanup_jpeg safe + + // validate req_comp + if (req_comp < 0 || req_comp > 4) return stbi__errpuc("bad req_comp", "Internal error"); + + // load a jpeg image from whichever source, but leave in YCbCr format + if (!stbi__decode_jpeg_image(z)) { stbi__cleanup_jpeg(z); return NULL; } + + // determine actual number of components to generate + n = req_comp ? req_comp : z->s->img_n >= 3 ? 3 : 1; + + is_rgb = z->s->img_n == 3 && (z->rgb == 3 || (z->app14_color_transform == 0 && !z->jfif)); + + if (z->s->img_n == 3 && n < 3 && !is_rgb) + decode_n = 1; + else + decode_n = z->s->img_n; + + // resample and color-convert + { + int k; + unsigned int i,j; + stbi_uc *output; + stbi_uc *coutput[4]; + + stbi__resample res_comp[4]; + + for (k=0; k < decode_n; ++k) { + stbi__resample *r = &res_comp[k]; + + // allocate line buffer big enough for upsampling off the edges + // with upsample factor of 4 + z->img_comp[k].linebuf = (stbi_uc *) stbi__malloc(z->s->img_x + 3); + if (!z->img_comp[k].linebuf) { stbi__cleanup_jpeg(z); return stbi__errpuc("outofmem", "Out of memory"); } + + r->hs = z->img_h_max / z->img_comp[k].h; + r->vs = z->img_v_max / z->img_comp[k].v; + r->ystep = r->vs >> 1; + r->w_lores = (z->s->img_x + r->hs-1) / r->hs; + r->ypos = 0; + r->line0 = r->line1 = z->img_comp[k].data; + + if (r->hs == 1 && r->vs == 1) r->resample = resample_row_1; + else if (r->hs == 1 && r->vs == 2) r->resample = stbi__resample_row_v_2; + else if (r->hs == 2 && r->vs == 1) r->resample = stbi__resample_row_h_2; + else if (r->hs == 2 && r->vs == 2) r->resample = z->resample_row_hv_2_kernel; + else r->resample = stbi__resample_row_generic; + } + + // can't error after this so, this is safe + output = (stbi_uc *) stbi__malloc_mad3(n, z->s->img_x, z->s->img_y, 1); + if (!output) { stbi__cleanup_jpeg(z); return stbi__errpuc("outofmem", "Out of memory"); } + + // now go ahead and resample + for (j=0; j < z->s->img_y; ++j) { + stbi_uc *out = output + n * z->s->img_x * j; + for (k=0; k < decode_n; ++k) { + stbi__resample *r = &res_comp[k]; + int y_bot = r->ystep >= (r->vs >> 1); + coutput[k] = r->resample(z->img_comp[k].linebuf, + y_bot ? r->line1 : r->line0, + y_bot ? r->line0 : r->line1, + r->w_lores, r->hs); + if (++r->ystep >= r->vs) { + r->ystep = 0; + r->line0 = r->line1; + if (++r->ypos < z->img_comp[k].y) + r->line1 += z->img_comp[k].w2; + } + } + if (n >= 3) { + stbi_uc *y = coutput[0]; + if (z->s->img_n == 3) { + if (is_rgb) { + for (i=0; i < z->s->img_x; ++i) { + out[0] = y[i]; + out[1] = coutput[1][i]; + out[2] = coutput[2][i]; + out[3] = 255; + out += n; + } + } else { + z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); + } + } else if (z->s->img_n == 4) { + if (z->app14_color_transform == 0) { // CMYK + for (i=0; i < z->s->img_x; ++i) { + stbi_uc m = coutput[3][i]; + out[0] = stbi__blinn_8x8(coutput[0][i], m); + out[1] = stbi__blinn_8x8(coutput[1][i], m); + out[2] = stbi__blinn_8x8(coutput[2][i], m); + out[3] = 255; + out += n; + } + } else if (z->app14_color_transform == 2) { // YCCK + z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); + for (i=0; i < z->s->img_x; ++i) { + stbi_uc m = coutput[3][i]; + out[0] = stbi__blinn_8x8(255 - out[0], m); + out[1] = stbi__blinn_8x8(255 - out[1], m); + out[2] = stbi__blinn_8x8(255 - out[2], m); + out += n; + } + } else { // YCbCr + alpha? Ignore the fourth channel for now + z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); + } + } else + for (i=0; i < z->s->img_x; ++i) { + out[0] = out[1] = out[2] = y[i]; + out[3] = 255; // not used if n==3 + out += n; + } + } else { + if (is_rgb) { + if (n == 1) + for (i=0; i < z->s->img_x; ++i) + *out++ = stbi__compute_y(coutput[0][i], coutput[1][i], coutput[2][i]); + else { + for (i=0; i < z->s->img_x; ++i, out += 2) { + out[0] = stbi__compute_y(coutput[0][i], coutput[1][i], coutput[2][i]); + out[1] = 255; + } + } + } else if (z->s->img_n == 4 && z->app14_color_transform == 0) { + for (i=0; i < z->s->img_x; ++i) { + stbi_uc m = coutput[3][i]; + stbi_uc r = stbi__blinn_8x8(coutput[0][i], m); + stbi_uc g = stbi__blinn_8x8(coutput[1][i], m); + stbi_uc b = stbi__blinn_8x8(coutput[2][i], m); + out[0] = stbi__compute_y(r, g, b); + out[1] = 255; + out += n; + } + } else if (z->s->img_n == 4 && z->app14_color_transform == 2) { + for (i=0; i < z->s->img_x; ++i) { + out[0] = stbi__blinn_8x8(255 - coutput[0][i], coutput[3][i]); + out[1] = 255; + out += n; + } + } else { + stbi_uc *y = coutput[0]; + if (n == 1) + for (i=0; i < z->s->img_x; ++i) out[i] = y[i]; + else + for (i=0; i < z->s->img_x; ++i) { *out++ = y[i]; *out++ = 255; } + } + } + } + stbi__cleanup_jpeg(z); + *out_x = z->s->img_x; + *out_y = z->s->img_y; + if (comp) *comp = z->s->img_n >= 3 ? 3 : 1; // report original components, not output + return output; + } +} + +static void *stbi__jpeg_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + unsigned char* result; + stbi__jpeg* j = (stbi__jpeg*) stbi__malloc(sizeof(stbi__jpeg)); + STBI_NOTUSED(ri); + j->s = s; + stbi__setup_jpeg(j); + result = load_jpeg_image(j, x,y,comp,req_comp); + STBI_FREE(j); + return result; +} + +static int stbi__jpeg_test(stbi__context *s) +{ + int r; + stbi__jpeg* j = (stbi__jpeg*)stbi__malloc(sizeof(stbi__jpeg)); + j->s = s; + stbi__setup_jpeg(j); + r = stbi__decode_jpeg_header(j, STBI__SCAN_type); + stbi__rewind(s); + STBI_FREE(j); + return r; +} + +static int stbi__jpeg_info_raw(stbi__jpeg *j, int *x, int *y, int *comp) +{ + if (!stbi__decode_jpeg_header(j, STBI__SCAN_header)) { + stbi__rewind( j->s ); + return 0; + } + if (x) *x = j->s->img_x; + if (y) *y = j->s->img_y; + if (comp) *comp = j->s->img_n >= 3 ? 3 : 1; + return 1; +} + +static int stbi__jpeg_info(stbi__context *s, int *x, int *y, int *comp) +{ + int result; + stbi__jpeg* j = (stbi__jpeg*) (stbi__malloc(sizeof(stbi__jpeg))); + j->s = s; + result = stbi__jpeg_info_raw(j, x, y, comp); + STBI_FREE(j); + return result; +} +#endif + +// public domain zlib decode v0.2 Sean Barrett 2006-11-18 +// simple implementation +// - all input must be provided in an upfront buffer +// - all output is written to a single output buffer (can malloc/realloc) +// performance +// - fast huffman + +#ifndef STBI_NO_ZLIB + +// fast-way is faster to check than jpeg huffman, but slow way is slower +#define STBI__ZFAST_BITS 9 // accelerate all cases in default tables +#define STBI__ZFAST_MASK ((1 << STBI__ZFAST_BITS) - 1) + +// zlib-style huffman encoding +// (jpegs packs from left, zlib from right, so can't share code) +typedef struct +{ + stbi__uint16 fast[1 << STBI__ZFAST_BITS]; + stbi__uint16 firstcode[16]; + int maxcode[17]; + stbi__uint16 firstsymbol[16]; + stbi_uc size[288]; + stbi__uint16 value[288]; +} stbi__zhuffman; + +stbi_inline static int stbi__bitreverse16(int n) +{ + n = ((n & 0xAAAA) >> 1) | ((n & 0x5555) << 1); + n = ((n & 0xCCCC) >> 2) | ((n & 0x3333) << 2); + n = ((n & 0xF0F0) >> 4) | ((n & 0x0F0F) << 4); + n = ((n & 0xFF00) >> 8) | ((n & 0x00FF) << 8); + return n; +} + +stbi_inline static int stbi__bit_reverse(int v, int bits) +{ + STBI_ASSERT(bits <= 16); + // to bit reverse n bits, reverse 16 and shift + // e.g. 11 bits, bit reverse and shift away 5 + return stbi__bitreverse16(v) >> (16-bits); +} + +static int stbi__zbuild_huffman(stbi__zhuffman *z, const stbi_uc *sizelist, int num) +{ + int i,k=0; + int code, next_code[16], sizes[17]; + + // DEFLATE spec for generating codes + memset(sizes, 0, sizeof(sizes)); + memset(z->fast, 0, sizeof(z->fast)); + for (i=0; i < num; ++i) + ++sizes[sizelist[i]]; + sizes[0] = 0; + for (i=1; i < 16; ++i) + if (sizes[i] > (1 << i)) + return stbi__err("bad sizes", "Corrupt PNG"); + code = 0; + for (i=1; i < 16; ++i) { + next_code[i] = code; + z->firstcode[i] = (stbi__uint16) code; + z->firstsymbol[i] = (stbi__uint16) k; + code = (code + sizes[i]); + if (sizes[i]) + if (code-1 >= (1 << i)) return stbi__err("bad codelengths","Corrupt PNG"); + z->maxcode[i] = code << (16-i); // preshift for inner loop + code <<= 1; + k += sizes[i]; + } + z->maxcode[16] = 0x10000; // sentinel + for (i=0; i < num; ++i) { + int s = sizelist[i]; + if (s) { + int c = next_code[s] - z->firstcode[s] + z->firstsymbol[s]; + stbi__uint16 fastv = (stbi__uint16) ((s << 9) | i); + z->size [c] = (stbi_uc ) s; + z->value[c] = (stbi__uint16) i; + if (s <= STBI__ZFAST_BITS) { + int j = stbi__bit_reverse(next_code[s],s); + while (j < (1 << STBI__ZFAST_BITS)) { + z->fast[j] = fastv; + j += (1 << s); + } + } + ++next_code[s]; + } + } + return 1; +} + +// zlib-from-memory implementation for PNG reading +// because PNG allows splitting the zlib stream arbitrarily, +// and it's annoying structurally to have PNG call ZLIB call PNG, +// we require PNG read all the IDATs and combine them into a single +// memory buffer + +typedef struct +{ + stbi_uc *zbuffer, *zbuffer_end; + int num_bits; + stbi__uint32 code_buffer; + + char *zout; + char *zout_start; + char *zout_end; + int z_expandable; + + stbi__zhuffman z_length, z_distance; +} stbi__zbuf; + +stbi_inline static stbi_uc stbi__zget8(stbi__zbuf *z) +{ + if (z->zbuffer >= z->zbuffer_end) return 0; + return *z->zbuffer++; +} + +static void stbi__fill_bits(stbi__zbuf *z) +{ + do { + STBI_ASSERT(z->code_buffer < (1U << z->num_bits)); + z->code_buffer |= (unsigned int) stbi__zget8(z) << z->num_bits; + z->num_bits += 8; + } while (z->num_bits <= 24); +} + +stbi_inline static unsigned int stbi__zreceive(stbi__zbuf *z, int n) +{ + unsigned int k; + if (z->num_bits < n) stbi__fill_bits(z); + k = z->code_buffer & ((1 << n) - 1); + z->code_buffer >>= n; + z->num_bits -= n; + return k; +} + +static int stbi__zhuffman_decode_slowpath(stbi__zbuf *a, stbi__zhuffman *z) +{ + int b,s,k; + // not resolved by fast table, so compute it the slow way + // use jpeg approach, which requires MSbits at top + k = stbi__bit_reverse(a->code_buffer, 16); + for (s=STBI__ZFAST_BITS+1; ; ++s) + if (k < z->maxcode[s]) + break; + if (s == 16) return -1; // invalid code! + // code size is s, so: + b = (k >> (16-s)) - z->firstcode[s] + z->firstsymbol[s]; + STBI_ASSERT(z->size[b] == s); + a->code_buffer >>= s; + a->num_bits -= s; + return z->value[b]; +} + +stbi_inline static int stbi__zhuffman_decode(stbi__zbuf *a, stbi__zhuffman *z) +{ + int b,s; + if (a->num_bits < 16) stbi__fill_bits(a); + b = z->fast[a->code_buffer & STBI__ZFAST_MASK]; + if (b) { + s = b >> 9; + a->code_buffer >>= s; + a->num_bits -= s; + return b & 511; + } + return stbi__zhuffman_decode_slowpath(a, z); +} + +static int stbi__zexpand(stbi__zbuf *z, char *zout, int n) // need to make room for n bytes +{ + char *q; + int cur, limit, old_limit; + z->zout = zout; + if (!z->z_expandable) return stbi__err("output buffer limit","Corrupt PNG"); + cur = (int) (z->zout - z->zout_start); + limit = old_limit = (int) (z->zout_end - z->zout_start); + while (cur + n > limit) + limit *= 2; + q = (char *) STBI_REALLOC_SIZED(z->zout_start, old_limit, limit); + STBI_NOTUSED(old_limit); + if (q == NULL) return stbi__err("outofmem", "Out of memory"); + z->zout_start = q; + z->zout = q + cur; + z->zout_end = q + limit; + return 1; +} + +static const int stbi__zlength_base[31] = { + 3,4,5,6,7,8,9,10,11,13, + 15,17,19,23,27,31,35,43,51,59, + 67,83,99,115,131,163,195,227,258,0,0 }; + +static const int stbi__zlength_extra[31]= +{ 0,0,0,0,0,0,0,0,1,1,1,1,2,2,2,2,3,3,3,3,4,4,4,4,5,5,5,5,0,0,0 }; + +static const int stbi__zdist_base[32] = { 1,2,3,4,5,7,9,13,17,25,33,49,65,97,129,193, +257,385,513,769,1025,1537,2049,3073,4097,6145,8193,12289,16385,24577,0,0}; + +static const int stbi__zdist_extra[32] = +{ 0,0,0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,11,11,12,12,13,13}; + +static int stbi__parse_huffman_block(stbi__zbuf *a) +{ + char *zout = a->zout; + for(;;) { + int z = stbi__zhuffman_decode(a, &a->z_length); + if (z < 256) { + if (z < 0) return stbi__err("bad huffman code","Corrupt PNG"); // error in huffman codes + if (zout >= a->zout_end) { + if (!stbi__zexpand(a, zout, 1)) return 0; + zout = a->zout; + } + *zout++ = (char) z; + } else { + stbi_uc *p; + int len,dist; + if (z == 256) { + a->zout = zout; + return 1; + } + z -= 257; + len = stbi__zlength_base[z]; + if (stbi__zlength_extra[z]) len += stbi__zreceive(a, stbi__zlength_extra[z]); + z = stbi__zhuffman_decode(a, &a->z_distance); + if (z < 0) return stbi__err("bad huffman code","Corrupt PNG"); + dist = stbi__zdist_base[z]; + if (stbi__zdist_extra[z]) dist += stbi__zreceive(a, stbi__zdist_extra[z]); + if (zout - a->zout_start < dist) return stbi__err("bad dist","Corrupt PNG"); + if (zout + len > a->zout_end) { + if (!stbi__zexpand(a, zout, len)) return 0; + zout = a->zout; + } + p = (stbi_uc *) (zout - dist); + if (dist == 1) { // run of one byte; common in images. + stbi_uc v = *p; + if (len) { do *zout++ = v; while (--len); } + } else { + if (len) { do *zout++ = *p++; while (--len); } + } + } + } +} + +static int stbi__compute_huffman_codes(stbi__zbuf *a) +{ + static const stbi_uc length_dezigzag[19] = { 16,17,18,0,8,7,9,6,10,5,11,4,12,3,13,2,14,1,15 }; + stbi__zhuffman z_codelength; + stbi_uc lencodes[286+32+137];//padding for maximum single op + stbi_uc codelength_sizes[19]; + int i,n; + + int hlit = stbi__zreceive(a,5) + 257; + int hdist = stbi__zreceive(a,5) + 1; + int hclen = stbi__zreceive(a,4) + 4; + int ntot = hlit + hdist; + + memset(codelength_sizes, 0, sizeof(codelength_sizes)); + for (i=0; i < hclen; ++i) { + int s = stbi__zreceive(a,3); + codelength_sizes[length_dezigzag[i]] = (stbi_uc) s; + } + if (!stbi__zbuild_huffman(&z_codelength, codelength_sizes, 19)) return 0; + + n = 0; + while (n < ntot) { + int c = stbi__zhuffman_decode(a, &z_codelength); + if (c < 0 || c >= 19) return stbi__err("bad codelengths", "Corrupt PNG"); + if (c < 16) + lencodes[n++] = (stbi_uc) c; + else { + stbi_uc fill = 0; + if (c == 16) { + c = stbi__zreceive(a,2)+3; + if (n == 0) return stbi__err("bad codelengths", "Corrupt PNG"); + fill = lencodes[n-1]; + } else if (c == 17) + c = stbi__zreceive(a,3)+3; + else { + STBI_ASSERT(c == 18); + c = stbi__zreceive(a,7)+11; + } + if (ntot - n < c) return stbi__err("bad codelengths", "Corrupt PNG"); + memset(lencodes+n, fill, c); + n += c; + } + } + if (n != ntot) return stbi__err("bad codelengths","Corrupt PNG"); + if (!stbi__zbuild_huffman(&a->z_length, lencodes, hlit)) return 0; + if (!stbi__zbuild_huffman(&a->z_distance, lencodes+hlit, hdist)) return 0; + return 1; +} + +static int stbi__parse_uncompressed_block(stbi__zbuf *a) +{ + stbi_uc header[4]; + int len,nlen,k; + if (a->num_bits & 7) + stbi__zreceive(a, a->num_bits & 7); // discard + // drain the bit-packed data into header + k = 0; + while (a->num_bits > 0) { + header[k++] = (stbi_uc) (a->code_buffer & 255); // suppress MSVC run-time check + a->code_buffer >>= 8; + a->num_bits -= 8; + } + STBI_ASSERT(a->num_bits == 0); + // now fill header the normal way + while (k < 4) + header[k++] = stbi__zget8(a); + len = header[1] * 256 + header[0]; + nlen = header[3] * 256 + header[2]; + if (nlen != (len ^ 0xffff)) return stbi__err("zlib corrupt","Corrupt PNG"); + if (a->zbuffer + len > a->zbuffer_end) return stbi__err("read past buffer","Corrupt PNG"); + if (a->zout + len > a->zout_end) + if (!stbi__zexpand(a, a->zout, len)) return 0; + memcpy(a->zout, a->zbuffer, len); + a->zbuffer += len; + a->zout += len; + return 1; +} + +static int stbi__parse_zlib_header(stbi__zbuf *a) +{ + int cmf = stbi__zget8(a); + int cm = cmf & 15; + /* int cinfo = cmf >> 4; */ + int flg = stbi__zget8(a); + if ((cmf*256+flg) % 31 != 0) return stbi__err("bad zlib header","Corrupt PNG"); // zlib spec + if (flg & 32) return stbi__err("no preset dict","Corrupt PNG"); // preset dictionary not allowed in png + if (cm != 8) return stbi__err("bad compression","Corrupt PNG"); // DEFLATE required for png + // window = 1 << (8 + cinfo)... but who cares, we fully buffer output + return 1; +} + +static const stbi_uc stbi__zdefault_length[288] = +{ + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7, 7,7,7,7,7,7,7,7,8,8,8,8,8,8,8,8 +}; +static const stbi_uc stbi__zdefault_distance[32] = +{ + 5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5 +}; +/* +Init algorithm: +{ + int i; // use <= to match clearly with spec + for (i=0; i <= 143; ++i) stbi__zdefault_length[i] = 8; + for ( ; i <= 255; ++i) stbi__zdefault_length[i] = 9; + for ( ; i <= 279; ++i) stbi__zdefault_length[i] = 7; + for ( ; i <= 287; ++i) stbi__zdefault_length[i] = 8; + + for (i=0; i <= 31; ++i) stbi__zdefault_distance[i] = 5; +} +*/ + +static int stbi__parse_zlib(stbi__zbuf *a, int parse_header) +{ + int final, type; + if (parse_header) + if (!stbi__parse_zlib_header(a)) return 0; + a->num_bits = 0; + a->code_buffer = 0; + do { + final = stbi__zreceive(a,1); + type = stbi__zreceive(a,2); + if (type == 0) { + if (!stbi__parse_uncompressed_block(a)) return 0; + } else if (type == 3) { + return 0; + } else { + if (type == 1) { + // use fixed code lengths + if (!stbi__zbuild_huffman(&a->z_length , stbi__zdefault_length , 288)) return 0; + if (!stbi__zbuild_huffman(&a->z_distance, stbi__zdefault_distance, 32)) return 0; + } else { + if (!stbi__compute_huffman_codes(a)) return 0; + } + if (!stbi__parse_huffman_block(a)) return 0; + } + } while (!final); + return 1; +} + +static int stbi__do_zlib(stbi__zbuf *a, char *obuf, int olen, int exp, int parse_header) +{ + a->zout_start = obuf; + a->zout = obuf; + a->zout_end = obuf + olen; + a->z_expandable = exp; + + return stbi__parse_zlib(a, parse_header); +} + +STBIDEF char *stbi_zlib_decode_malloc_guesssize(const char *buffer, int len, int initial_size, int *outlen) +{ + stbi__zbuf a; + char *p = (char *) stbi__malloc(initial_size); + if (p == NULL) return NULL; + a.zbuffer = (stbi_uc *) buffer; + a.zbuffer_end = (stbi_uc *) buffer + len; + if (stbi__do_zlib(&a, p, initial_size, 1, 1)) { + if (outlen) *outlen = (int) (a.zout - a.zout_start); + return a.zout_start; + } else { + STBI_FREE(a.zout_start); + return NULL; + } +} + +STBIDEF char *stbi_zlib_decode_malloc(char const *buffer, int len, int *outlen) +{ + return stbi_zlib_decode_malloc_guesssize(buffer, len, 16384, outlen); +} + +STBIDEF char *stbi_zlib_decode_malloc_guesssize_headerflag(const char *buffer, int len, int initial_size, int *outlen, int parse_header) +{ + stbi__zbuf a; + char *p = (char *) stbi__malloc(initial_size); + if (p == NULL) return NULL; + a.zbuffer = (stbi_uc *) buffer; + a.zbuffer_end = (stbi_uc *) buffer + len; + if (stbi__do_zlib(&a, p, initial_size, 1, parse_header)) { + if (outlen) *outlen = (int) (a.zout - a.zout_start); + return a.zout_start; + } else { + STBI_FREE(a.zout_start); + return NULL; + } +} + +STBIDEF int stbi_zlib_decode_buffer(char *obuffer, int olen, char const *ibuffer, int ilen) +{ + stbi__zbuf a; + a.zbuffer = (stbi_uc *) ibuffer; + a.zbuffer_end = (stbi_uc *) ibuffer + ilen; + if (stbi__do_zlib(&a, obuffer, olen, 0, 1)) + return (int) (a.zout - a.zout_start); + else + return -1; +} + +STBIDEF char *stbi_zlib_decode_noheader_malloc(char const *buffer, int len, int *outlen) +{ + stbi__zbuf a; + char *p = (char *) stbi__malloc(16384); + if (p == NULL) return NULL; + a.zbuffer = (stbi_uc *) buffer; + a.zbuffer_end = (stbi_uc *) buffer+len; + if (stbi__do_zlib(&a, p, 16384, 1, 0)) { + if (outlen) *outlen = (int) (a.zout - a.zout_start); + return a.zout_start; + } else { + STBI_FREE(a.zout_start); + return NULL; + } +} + +STBIDEF int stbi_zlib_decode_noheader_buffer(char *obuffer, int olen, const char *ibuffer, int ilen) +{ + stbi__zbuf a; + a.zbuffer = (stbi_uc *) ibuffer; + a.zbuffer_end = (stbi_uc *) ibuffer + ilen; + if (stbi__do_zlib(&a, obuffer, olen, 0, 0)) + return (int) (a.zout - a.zout_start); + else + return -1; +} +#endif + +// public domain "baseline" PNG decoder v0.10 Sean Barrett 2006-11-18 +// simple implementation +// - only 8-bit samples +// - no CRC checking +// - allocates lots of intermediate memory +// - avoids problem of streaming data between subsystems +// - avoids explicit window management +// performance +// - uses stb_zlib, a PD zlib implementation with fast huffman decoding + +#ifndef STBI_NO_PNG +typedef struct +{ + stbi__uint32 length; + stbi__uint32 type; +} stbi__pngchunk; + +static stbi__pngchunk stbi__get_chunk_header(stbi__context *s) +{ + stbi__pngchunk c; + c.length = stbi__get32be(s); + c.type = stbi__get32be(s); + return c; +} + +static int stbi__check_png_header(stbi__context *s) +{ + static const stbi_uc png_sig[8] = { 137,80,78,71,13,10,26,10 }; + int i; + for (i=0; i < 8; ++i) + if (stbi__get8(s) != png_sig[i]) return stbi__err("bad png sig","Not a PNG"); + return 1; +} + +typedef struct +{ + stbi__context *s; + stbi_uc *idata, *expanded, *out; + int depth; +} stbi__png; + + +enum { + STBI__F_none=0, + STBI__F_sub=1, + STBI__F_up=2, + STBI__F_avg=3, + STBI__F_paeth=4, + // synthetic filters used for first scanline to avoid needing a dummy row of 0s + STBI__F_avg_first, + STBI__F_paeth_first +}; + +static stbi_uc first_row_filter[5] = +{ + STBI__F_none, + STBI__F_sub, + STBI__F_none, + STBI__F_avg_first, + STBI__F_paeth_first +}; + +static int stbi__paeth(int a, int b, int c) +{ + int p = a + b - c; + int pa = abs(p-a); + int pb = abs(p-b); + int pc = abs(p-c); + if (pa <= pb && pa <= pc) return a; + if (pb <= pc) return b; + return c; +} + +static const stbi_uc stbi__depth_scale_table[9] = { 0, 0xff, 0x55, 0, 0x11, 0,0,0, 0x01 }; + +// create the png data from post-deflated data +static int stbi__create_png_image_raw(stbi__png *a, stbi_uc *raw, stbi__uint32 raw_len, int out_n, stbi__uint32 x, stbi__uint32 y, int depth, int color) +{ + int bytes = (depth == 16? 2 : 1); + stbi__context *s = a->s; + stbi__uint32 i,j,stride = x*out_n*bytes; + stbi__uint32 img_len, img_width_bytes; + int k; + int img_n = s->img_n; // copy it into a local for later + + int output_bytes = out_n*bytes; + int filter_bytes = img_n*bytes; + int width = x; + + STBI_ASSERT(out_n == s->img_n || out_n == s->img_n+1); + a->out = (stbi_uc *) stbi__malloc_mad3(x, y, output_bytes, 0); // extra bytes to write off the end into + if (!a->out) return stbi__err("outofmem", "Out of memory"); + + if (!stbi__mad3sizes_valid(img_n, x, depth, 7)) return stbi__err("too large", "Corrupt PNG"); + img_width_bytes = (((img_n * x * depth) + 7) >> 3); + img_len = (img_width_bytes + 1) * y; + + // we used to check for exact match between raw_len and img_len on non-interlaced PNGs, + // but issue #276 reported a PNG in the wild that had extra data at the end (all zeros), + // so just check for raw_len < img_len always. + if (raw_len < img_len) return stbi__err("not enough pixels","Corrupt PNG"); + + for (j=0; j < y; ++j) { + stbi_uc *cur = a->out + stride*j; + stbi_uc *prior; + int filter = *raw++; + + if (filter > 4) + return stbi__err("invalid filter","Corrupt PNG"); + + if (depth < 8) { + STBI_ASSERT(img_width_bytes <= x); + cur += x*out_n - img_width_bytes; // store output to the rightmost img_len bytes, so we can decode in place + filter_bytes = 1; + width = img_width_bytes; + } + prior = cur - stride; // bugfix: need to compute this after 'cur +=' computation above + + // if first row, use special filter that doesn't sample previous row + if (j == 0) filter = first_row_filter[filter]; + + // handle first byte explicitly + for (k=0; k < filter_bytes; ++k) { + switch (filter) { + case STBI__F_none : cur[k] = raw[k]; break; + case STBI__F_sub : cur[k] = raw[k]; break; + case STBI__F_up : cur[k] = STBI__BYTECAST(raw[k] + prior[k]); break; + case STBI__F_avg : cur[k] = STBI__BYTECAST(raw[k] + (prior[k]>>1)); break; + case STBI__F_paeth : cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(0,prior[k],0)); break; + case STBI__F_avg_first : cur[k] = raw[k]; break; + case STBI__F_paeth_first: cur[k] = raw[k]; break; + } + } + + if (depth == 8) { + if (img_n != out_n) + cur[img_n] = 255; // first pixel + raw += img_n; + cur += out_n; + prior += out_n; + } else if (depth == 16) { + if (img_n != out_n) { + cur[filter_bytes] = 255; // first pixel top byte + cur[filter_bytes+1] = 255; // first pixel bottom byte + } + raw += filter_bytes; + cur += output_bytes; + prior += output_bytes; + } else { + raw += 1; + cur += 1; + prior += 1; + } + + // this is a little gross, so that we don't switch per-pixel or per-component + if (depth < 8 || img_n == out_n) { + int nk = (width - 1)*filter_bytes; + #define STBI__CASE(f) \ + case f: \ + for (k=0; k < nk; ++k) + switch (filter) { + // "none" filter turns into a memcpy here; make that explicit. + case STBI__F_none: memcpy(cur, raw, nk); break; + STBI__CASE(STBI__F_sub) { cur[k] = STBI__BYTECAST(raw[k] + cur[k-filter_bytes]); } break; + STBI__CASE(STBI__F_up) { cur[k] = STBI__BYTECAST(raw[k] + prior[k]); } break; + STBI__CASE(STBI__F_avg) { cur[k] = STBI__BYTECAST(raw[k] + ((prior[k] + cur[k-filter_bytes])>>1)); } break; + STBI__CASE(STBI__F_paeth) { cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k-filter_bytes],prior[k],prior[k-filter_bytes])); } break; + STBI__CASE(STBI__F_avg_first) { cur[k] = STBI__BYTECAST(raw[k] + (cur[k-filter_bytes] >> 1)); } break; + STBI__CASE(STBI__F_paeth_first) { cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k-filter_bytes],0,0)); } break; + } + #undef STBI__CASE + raw += nk; + } else { + STBI_ASSERT(img_n+1 == out_n); + #define STBI__CASE(f) \ + case f: \ + for (i=x-1; i >= 1; --i, cur[filter_bytes]=255,raw+=filter_bytes,cur+=output_bytes,prior+=output_bytes) \ + for (k=0; k < filter_bytes; ++k) + switch (filter) { + STBI__CASE(STBI__F_none) { cur[k] = raw[k]; } break; + STBI__CASE(STBI__F_sub) { cur[k] = STBI__BYTECAST(raw[k] + cur[k- output_bytes]); } break; + STBI__CASE(STBI__F_up) { cur[k] = STBI__BYTECAST(raw[k] + prior[k]); } break; + STBI__CASE(STBI__F_avg) { cur[k] = STBI__BYTECAST(raw[k] + ((prior[k] + cur[k- output_bytes])>>1)); } break; + STBI__CASE(STBI__F_paeth) { cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k- output_bytes],prior[k],prior[k- output_bytes])); } break; + STBI__CASE(STBI__F_avg_first) { cur[k] = STBI__BYTECAST(raw[k] + (cur[k- output_bytes] >> 1)); } break; + STBI__CASE(STBI__F_paeth_first) { cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k- output_bytes],0,0)); } break; + } + #undef STBI__CASE + + // the loop above sets the high byte of the pixels' alpha, but for + // 16 bit png files we also need the low byte set. we'll do that here. + if (depth == 16) { + cur = a->out + stride*j; // start at the beginning of the row again + for (i=0; i < x; ++i,cur+=output_bytes) { + cur[filter_bytes+1] = 255; + } + } + } + } + + // we make a separate pass to expand bits to pixels; for performance, + // this could run two scanlines behind the above code, so it won't + // intefere with filtering but will still be in the cache. + if (depth < 8) { + for (j=0; j < y; ++j) { + stbi_uc *cur = a->out + stride*j; + stbi_uc *in = a->out + stride*j + x*out_n - img_width_bytes; + // unpack 1/2/4-bit into a 8-bit buffer. allows us to keep the common 8-bit path optimal at minimal cost for 1/2/4-bit + // png guarante byte alignment, if width is not multiple of 8/4/2 we'll decode dummy trailing data that will be skipped in the later loop + stbi_uc scale = (color == 0) ? stbi__depth_scale_table[depth] : 1; // scale grayscale values to 0..255 range + + // note that the final byte might overshoot and write more data than desired. + // we can allocate enough data that this never writes out of memory, but it + // could also overwrite the next scanline. can it overwrite non-empty data + // on the next scanline? yes, consider 1-pixel-wide scanlines with 1-bit-per-pixel. + // so we need to explicitly clamp the final ones + + if (depth == 4) { + for (k=x*img_n; k >= 2; k-=2, ++in) { + *cur++ = scale * ((*in >> 4) ); + *cur++ = scale * ((*in ) & 0x0f); + } + if (k > 0) *cur++ = scale * ((*in >> 4) ); + } else if (depth == 2) { + for (k=x*img_n; k >= 4; k-=4, ++in) { + *cur++ = scale * ((*in >> 6) ); + *cur++ = scale * ((*in >> 4) & 0x03); + *cur++ = scale * ((*in >> 2) & 0x03); + *cur++ = scale * ((*in ) & 0x03); + } + if (k > 0) *cur++ = scale * ((*in >> 6) ); + if (k > 1) *cur++ = scale * ((*in >> 4) & 0x03); + if (k > 2) *cur++ = scale * ((*in >> 2) & 0x03); + } else if (depth == 1) { + for (k=x*img_n; k >= 8; k-=8, ++in) { + *cur++ = scale * ((*in >> 7) ); + *cur++ = scale * ((*in >> 6) & 0x01); + *cur++ = scale * ((*in >> 5) & 0x01); + *cur++ = scale * ((*in >> 4) & 0x01); + *cur++ = scale * ((*in >> 3) & 0x01); + *cur++ = scale * ((*in >> 2) & 0x01); + *cur++ = scale * ((*in >> 1) & 0x01); + *cur++ = scale * ((*in ) & 0x01); + } + if (k > 0) *cur++ = scale * ((*in >> 7) ); + if (k > 1) *cur++ = scale * ((*in >> 6) & 0x01); + if (k > 2) *cur++ = scale * ((*in >> 5) & 0x01); + if (k > 3) *cur++ = scale * ((*in >> 4) & 0x01); + if (k > 4) *cur++ = scale * ((*in >> 3) & 0x01); + if (k > 5) *cur++ = scale * ((*in >> 2) & 0x01); + if (k > 6) *cur++ = scale * ((*in >> 1) & 0x01); + } + if (img_n != out_n) { + int q; + // insert alpha = 255 + cur = a->out + stride*j; + if (img_n == 1) { + for (q=x-1; q >= 0; --q) { + cur[q*2+1] = 255; + cur[q*2+0] = cur[q]; + } + } else { + STBI_ASSERT(img_n == 3); + for (q=x-1; q >= 0; --q) { + cur[q*4+3] = 255; + cur[q*4+2] = cur[q*3+2]; + cur[q*4+1] = cur[q*3+1]; + cur[q*4+0] = cur[q*3+0]; + } + } + } + } + } else if (depth == 16) { + // force the image data from big-endian to platform-native. + // this is done in a separate pass due to the decoding relying + // on the data being untouched, but could probably be done + // per-line during decode if care is taken. + stbi_uc *cur = a->out; + stbi__uint16 *cur16 = (stbi__uint16*)cur; + + for(i=0; i < x*y*out_n; ++i,cur16++,cur+=2) { + *cur16 = (cur[0] << 8) | cur[1]; + } + } + + return 1; +} + +static int stbi__create_png_image(stbi__png *a, stbi_uc *image_data, stbi__uint32 image_data_len, int out_n, int depth, int color, int interlaced) +{ + int bytes = (depth == 16 ? 2 : 1); + int out_bytes = out_n * bytes; + stbi_uc *final; + int p; + if (!interlaced) + return stbi__create_png_image_raw(a, image_data, image_data_len, out_n, a->s->img_x, a->s->img_y, depth, color); + + // de-interlacing + final = (stbi_uc *) stbi__malloc_mad3(a->s->img_x, a->s->img_y, out_bytes, 0); + for (p=0; p < 7; ++p) { + int xorig[] = { 0,4,0,2,0,1,0 }; + int yorig[] = { 0,0,4,0,2,0,1 }; + int xspc[] = { 8,8,4,4,2,2,1 }; + int yspc[] = { 8,8,8,4,4,2,2 }; + int i,j,x,y; + // pass1_x[4] = 0, pass1_x[5] = 1, pass1_x[12] = 1 + x = (a->s->img_x - xorig[p] + xspc[p]-1) / xspc[p]; + y = (a->s->img_y - yorig[p] + yspc[p]-1) / yspc[p]; + if (x && y) { + stbi__uint32 img_len = ((((a->s->img_n * x * depth) + 7) >> 3) + 1) * y; + if (!stbi__create_png_image_raw(a, image_data, image_data_len, out_n, x, y, depth, color)) { + STBI_FREE(final); + return 0; + } + for (j=0; j < y; ++j) { + for (i=0; i < x; ++i) { + int out_y = j*yspc[p]+yorig[p]; + int out_x = i*xspc[p]+xorig[p]; + memcpy(final + out_y*a->s->img_x*out_bytes + out_x*out_bytes, + a->out + (j*x+i)*out_bytes, out_bytes); + } + } + STBI_FREE(a->out); + image_data += img_len; + image_data_len -= img_len; + } + } + a->out = final; + + return 1; +} + +static int stbi__compute_transparency(stbi__png *z, stbi_uc tc[3], int out_n) +{ + stbi__context *s = z->s; + stbi__uint32 i, pixel_count = s->img_x * s->img_y; + stbi_uc *p = z->out; + + // compute color-based transparency, assuming we've + // already got 255 as the alpha value in the output + STBI_ASSERT(out_n == 2 || out_n == 4); + + if (out_n == 2) { + for (i=0; i < pixel_count; ++i) { + p[1] = (p[0] == tc[0] ? 0 : 255); + p += 2; + } + } else { + for (i=0; i < pixel_count; ++i) { + if (p[0] == tc[0] && p[1] == tc[1] && p[2] == tc[2]) + p[3] = 0; + p += 4; + } + } + return 1; +} + +static int stbi__compute_transparency16(stbi__png *z, stbi__uint16 tc[3], int out_n) +{ + stbi__context *s = z->s; + stbi__uint32 i, pixel_count = s->img_x * s->img_y; + stbi__uint16 *p = (stbi__uint16*) z->out; + + // compute color-based transparency, assuming we've + // already got 65535 as the alpha value in the output + STBI_ASSERT(out_n == 2 || out_n == 4); + + if (out_n == 2) { + for (i = 0; i < pixel_count; ++i) { + p[1] = (p[0] == tc[0] ? 0 : 65535); + p += 2; + } + } else { + for (i = 0; i < pixel_count; ++i) { + if (p[0] == tc[0] && p[1] == tc[1] && p[2] == tc[2]) + p[3] = 0; + p += 4; + } + } + return 1; +} + +static int stbi__expand_png_palette(stbi__png *a, stbi_uc *palette, int len, int pal_img_n) +{ + stbi__uint32 i, pixel_count = a->s->img_x * a->s->img_y; + stbi_uc *p, *temp_out, *orig = a->out; + + p = (stbi_uc *) stbi__malloc_mad2(pixel_count, pal_img_n, 0); + if (p == NULL) return stbi__err("outofmem", "Out of memory"); + + // between here and free(out) below, exitting would leak + temp_out = p; + + if (pal_img_n == 3) { + for (i=0; i < pixel_count; ++i) { + int n = orig[i]*4; + p[0] = palette[n ]; + p[1] = palette[n+1]; + p[2] = palette[n+2]; + p += 3; + } + } else { + for (i=0; i < pixel_count; ++i) { + int n = orig[i]*4; + p[0] = palette[n ]; + p[1] = palette[n+1]; + p[2] = palette[n+2]; + p[3] = palette[n+3]; + p += 4; + } + } + STBI_FREE(a->out); + a->out = temp_out; + + STBI_NOTUSED(len); + + return 1; +} + +static int stbi__unpremultiply_on_load = 0; +static int stbi__de_iphone_flag = 0; + +STBIDEF void stbi_set_unpremultiply_on_load(int flag_true_if_should_unpremultiply) +{ + stbi__unpremultiply_on_load = flag_true_if_should_unpremultiply; +} + +STBIDEF void stbi_convert_iphone_png_to_rgb(int flag_true_if_should_convert) +{ + stbi__de_iphone_flag = flag_true_if_should_convert; +} + +static void stbi__de_iphone(stbi__png *z) +{ + stbi__context *s = z->s; + stbi__uint32 i, pixel_count = s->img_x * s->img_y; + stbi_uc *p = z->out; + + if (s->img_out_n == 3) { // convert bgr to rgb + for (i=0; i < pixel_count; ++i) { + stbi_uc t = p[0]; + p[0] = p[2]; + p[2] = t; + p += 3; + } + } else { + STBI_ASSERT(s->img_out_n == 4); + if (stbi__unpremultiply_on_load) { + // convert bgr to rgb and unpremultiply + for (i=0; i < pixel_count; ++i) { + stbi_uc a = p[3]; + stbi_uc t = p[0]; + if (a) { + stbi_uc half = a / 2; + p[0] = (p[2] * 255 + half) / a; + p[1] = (p[1] * 255 + half) / a; + p[2] = ( t * 255 + half) / a; + } else { + p[0] = p[2]; + p[2] = t; + } + p += 4; + } + } else { + // convert bgr to rgb + for (i=0; i < pixel_count; ++i) { + stbi_uc t = p[0]; + p[0] = p[2]; + p[2] = t; + p += 4; + } + } + } +} + +#define STBI__PNG_TYPE(a,b,c,d) (((unsigned) (a) << 24) + ((unsigned) (b) << 16) + ((unsigned) (c) << 8) + (unsigned) (d)) + +static int stbi__parse_png_file(stbi__png *z, int scan, int req_comp) +{ + stbi_uc palette[1024], pal_img_n=0; + stbi_uc has_trans=0, tc[3]={0}; + stbi__uint16 tc16[3]; + stbi__uint32 ioff=0, idata_limit=0, i, pal_len=0; + int first=1,k,interlace=0, color=0, is_iphone=0; + stbi__context *s = z->s; + + z->expanded = NULL; + z->idata = NULL; + z->out = NULL; + + if (!stbi__check_png_header(s)) return 0; + + if (scan == STBI__SCAN_type) return 1; + + for (;;) { + stbi__pngchunk c = stbi__get_chunk_header(s); + switch (c.type) { + case STBI__PNG_TYPE('C','g','B','I'): + is_iphone = 1; + stbi__skip(s, c.length); + break; + case STBI__PNG_TYPE('I','H','D','R'): { + int comp,filter; + if (!first) return stbi__err("multiple IHDR","Corrupt PNG"); + first = 0; + if (c.length != 13) return stbi__err("bad IHDR len","Corrupt PNG"); + s->img_x = stbi__get32be(s); if (s->img_x > (1 << 24)) return stbi__err("too large","Very large image (corrupt?)"); + s->img_y = stbi__get32be(s); if (s->img_y > (1 << 24)) return stbi__err("too large","Very large image (corrupt?)"); + z->depth = stbi__get8(s); if (z->depth != 1 && z->depth != 2 && z->depth != 4 && z->depth != 8 && z->depth != 16) return stbi__err("1/2/4/8/16-bit only","PNG not supported: 1/2/4/8/16-bit only"); + color = stbi__get8(s); if (color > 6) return stbi__err("bad ctype","Corrupt PNG"); + if (color == 3 && z->depth == 16) return stbi__err("bad ctype","Corrupt PNG"); + if (color == 3) pal_img_n = 3; else if (color & 1) return stbi__err("bad ctype","Corrupt PNG"); + comp = stbi__get8(s); if (comp) return stbi__err("bad comp method","Corrupt PNG"); + filter= stbi__get8(s); if (filter) return stbi__err("bad filter method","Corrupt PNG"); + interlace = stbi__get8(s); if (interlace>1) return stbi__err("bad interlace method","Corrupt PNG"); + if (!s->img_x || !s->img_y) return stbi__err("0-pixel image","Corrupt PNG"); + if (!pal_img_n) { + s->img_n = (color & 2 ? 3 : 1) + (color & 4 ? 1 : 0); + if ((1 << 30) / s->img_x / s->img_n < s->img_y) return stbi__err("too large", "Image too large to decode"); + if (scan == STBI__SCAN_header) return 1; + } else { + // if paletted, then pal_n is our final components, and + // img_n is # components to decompress/filter. + s->img_n = 1; + if ((1 << 30) / s->img_x / 4 < s->img_y) return stbi__err("too large","Corrupt PNG"); + // if SCAN_header, have to scan to see if we have a tRNS + } + break; + } + + case STBI__PNG_TYPE('P','L','T','E'): { + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (c.length > 256*3) return stbi__err("invalid PLTE","Corrupt PNG"); + pal_len = c.length / 3; + if (pal_len * 3 != c.length) return stbi__err("invalid PLTE","Corrupt PNG"); + for (i=0; i < pal_len; ++i) { + palette[i*4+0] = stbi__get8(s); + palette[i*4+1] = stbi__get8(s); + palette[i*4+2] = stbi__get8(s); + palette[i*4+3] = 255; + } + break; + } + + case STBI__PNG_TYPE('t','R','N','S'): { + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (z->idata) return stbi__err("tRNS after IDAT","Corrupt PNG"); + if (pal_img_n) { + if (scan == STBI__SCAN_header) { s->img_n = 4; return 1; } + if (pal_len == 0) return stbi__err("tRNS before PLTE","Corrupt PNG"); + if (c.length > pal_len) return stbi__err("bad tRNS len","Corrupt PNG"); + pal_img_n = 4; + for (i=0; i < c.length; ++i) + palette[i*4+3] = stbi__get8(s); + } else { + if (!(s->img_n & 1)) return stbi__err("tRNS with alpha","Corrupt PNG"); + if (c.length != (stbi__uint32) s->img_n*2) return stbi__err("bad tRNS len","Corrupt PNG"); + has_trans = 1; + if (z->depth == 16) { + for (k = 0; k < s->img_n; ++k) tc16[k] = (stbi__uint16)stbi__get16be(s); // copy the values as-is + } else { + for (k = 0; k < s->img_n; ++k) tc[k] = (stbi_uc)(stbi__get16be(s) & 255) * stbi__depth_scale_table[z->depth]; // non 8-bit images will be larger + } + } + break; + } + + case STBI__PNG_TYPE('I','D','A','T'): { + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (pal_img_n && !pal_len) return stbi__err("no PLTE","Corrupt PNG"); + if (scan == STBI__SCAN_header) { s->img_n = pal_img_n; return 1; } + if ((int)(ioff + c.length) < (int)ioff) return 0; + if (ioff + c.length > idata_limit) { + stbi__uint32 idata_limit_old = idata_limit; + stbi_uc *p; + if (idata_limit == 0) idata_limit = c.length > 4096 ? c.length : 4096; + while (ioff + c.length > idata_limit) + idata_limit *= 2; + STBI_NOTUSED(idata_limit_old); + p = (stbi_uc *) STBI_REALLOC_SIZED(z->idata, idata_limit_old, idata_limit); if (p == NULL) return stbi__err("outofmem", "Out of memory"); + z->idata = p; + } + if (!stbi__getn(s, z->idata+ioff,c.length)) return stbi__err("outofdata","Corrupt PNG"); + ioff += c.length; + break; + } + + case STBI__PNG_TYPE('I','E','N','D'): { + stbi__uint32 raw_len, bpl; + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (scan != STBI__SCAN_load) return 1; + if (z->idata == NULL) return stbi__err("no IDAT","Corrupt PNG"); + // initial guess for decoded data size to avoid unnecessary reallocs + bpl = (s->img_x * z->depth + 7) / 8; // bytes per line, per component + raw_len = bpl * s->img_y * s->img_n /* pixels */ + s->img_y /* filter mode per row */; + z->expanded = (stbi_uc *) stbi_zlib_decode_malloc_guesssize_headerflag((char *) z->idata, ioff, raw_len, (int *) &raw_len, !is_iphone); + if (z->expanded == NULL) return 0; // zlib should set error + STBI_FREE(z->idata); z->idata = NULL; + if ((req_comp == s->img_n+1 && req_comp != 3 && !pal_img_n) || has_trans) + s->img_out_n = s->img_n+1; + else + s->img_out_n = s->img_n; + if (!stbi__create_png_image(z, z->expanded, raw_len, s->img_out_n, z->depth, color, interlace)) return 0; + if (has_trans) { + if (z->depth == 16) { + if (!stbi__compute_transparency16(z, tc16, s->img_out_n)) return 0; + } else { + if (!stbi__compute_transparency(z, tc, s->img_out_n)) return 0; + } + } + if (is_iphone && stbi__de_iphone_flag && s->img_out_n > 2) + stbi__de_iphone(z); + if (pal_img_n) { + // pal_img_n == 3 or 4 + s->img_n = pal_img_n; // record the actual colors we had + s->img_out_n = pal_img_n; + if (req_comp >= 3) s->img_out_n = req_comp; + if (!stbi__expand_png_palette(z, palette, pal_len, s->img_out_n)) + return 0; + } else if (has_trans) { + // non-paletted image with tRNS -> source image has (constant) alpha + ++s->img_n; + } + STBI_FREE(z->expanded); z->expanded = NULL; + return 1; + } + + default: + // if critical, fail + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if ((c.type & (1 << 29)) == 0) { + #ifndef STBI_NO_FAILURE_STRINGS + // not threadsafe + static char invalid_chunk[] = "XXXX PNG chunk not known"; + invalid_chunk[0] = STBI__BYTECAST(c.type >> 24); + invalid_chunk[1] = STBI__BYTECAST(c.type >> 16); + invalid_chunk[2] = STBI__BYTECAST(c.type >> 8); + invalid_chunk[3] = STBI__BYTECAST(c.type >> 0); + #endif + return stbi__err(invalid_chunk, "PNG not supported: unknown PNG chunk type"); + } + stbi__skip(s, c.length); + break; + } + // end of PNG chunk, read and skip CRC + stbi__get32be(s); + } +} + +static void *stbi__do_png(stbi__png *p, int *x, int *y, int *n, int req_comp, stbi__result_info *ri) +{ + void *result=NULL; + if (req_comp < 0 || req_comp > 4) return stbi__errpuc("bad req_comp", "Internal error"); + if (stbi__parse_png_file(p, STBI__SCAN_load, req_comp)) { + if (p->depth < 8) + ri->bits_per_channel = 8; + else + ri->bits_per_channel = p->depth; + result = p->out; + p->out = NULL; + if (req_comp && req_comp != p->s->img_out_n) { + if (ri->bits_per_channel == 8) + result = stbi__convert_format((unsigned char *) result, p->s->img_out_n, req_comp, p->s->img_x, p->s->img_y); + else + result = stbi__convert_format16((stbi__uint16 *) result, p->s->img_out_n, req_comp, p->s->img_x, p->s->img_y); + p->s->img_out_n = req_comp; + if (result == NULL) return result; + } + *x = p->s->img_x; + *y = p->s->img_y; + if (n) *n = p->s->img_n; + } + STBI_FREE(p->out); p->out = NULL; + STBI_FREE(p->expanded); p->expanded = NULL; + STBI_FREE(p->idata); p->idata = NULL; + + return result; +} + +static void *stbi__png_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi__png p; + p.s = s; + return stbi__do_png(&p, x,y,comp,req_comp, ri); +} + +static int stbi__png_test(stbi__context *s) +{ + int r; + r = stbi__check_png_header(s); + stbi__rewind(s); + return r; +} + +static int stbi__png_info_raw(stbi__png *p, int *x, int *y, int *comp) +{ + if (!stbi__parse_png_file(p, STBI__SCAN_header, 0)) { + stbi__rewind( p->s ); + return 0; + } + if (x) *x = p->s->img_x; + if (y) *y = p->s->img_y; + if (comp) *comp = p->s->img_n; + return 1; +} + +static int stbi__png_info(stbi__context *s, int *x, int *y, int *comp) +{ + stbi__png p; + p.s = s; + return stbi__png_info_raw(&p, x, y, comp); +} + +static int stbi__png_is16(stbi__context *s) +{ + stbi__png p; + p.s = s; + if (!stbi__png_info_raw(&p, NULL, NULL, NULL)) + return 0; + if (p.depth != 16) { + stbi__rewind(p.s); + return 0; + } + return 1; +} +#endif + +// Microsoft/Windows BMP image + +#ifndef STBI_NO_BMP +static int stbi__bmp_test_raw(stbi__context *s) +{ + int r; + int sz; + if (stbi__get8(s) != 'B') return 0; + if (stbi__get8(s) != 'M') return 0; + stbi__get32le(s); // discard filesize + stbi__get16le(s); // discard reserved + stbi__get16le(s); // discard reserved + stbi__get32le(s); // discard data offset + sz = stbi__get32le(s); + r = (sz == 12 || sz == 40 || sz == 56 || sz == 108 || sz == 124); + return r; +} + +static int stbi__bmp_test(stbi__context *s) +{ + int r = stbi__bmp_test_raw(s); + stbi__rewind(s); + return r; +} + + +// returns 0..31 for the highest set bit +static int stbi__high_bit(unsigned int z) +{ + int n=0; + if (z == 0) return -1; + if (z >= 0x10000) { n += 16; z >>= 16; } + if (z >= 0x00100) { n += 8; z >>= 8; } + if (z >= 0x00010) { n += 4; z >>= 4; } + if (z >= 0x00004) { n += 2; z >>= 2; } + if (z >= 0x00002) { n += 1; z >>= 1; } + return n; +} + +static int stbi__bitcount(unsigned int a) +{ + a = (a & 0x55555555) + ((a >> 1) & 0x55555555); // max 2 + a = (a & 0x33333333) + ((a >> 2) & 0x33333333); // max 4 + a = (a + (a >> 4)) & 0x0f0f0f0f; // max 8 per 4, now 8 bits + a = (a + (a >> 8)); // max 16 per 8 bits + a = (a + (a >> 16)); // max 32 per 8 bits + return a & 0xff; +} + +// extract an arbitrarily-aligned N-bit value (N=bits) +// from v, and then make it 8-bits long and fractionally +// extend it to full full range. +static int stbi__shiftsigned(unsigned int v, int shift, int bits) +{ + static unsigned int mul_table[9] = { + 0, + 0xff/*0b11111111*/, 0x55/*0b01010101*/, 0x49/*0b01001001*/, 0x11/*0b00010001*/, + 0x21/*0b00100001*/, 0x41/*0b01000001*/, 0x81/*0b10000001*/, 0x01/*0b00000001*/, + }; + static unsigned int shift_table[9] = { + 0, 0,0,1,0,2,4,6,0, + }; + if (shift < 0) + v <<= -shift; + else + v >>= shift; + STBI_ASSERT(v >= 0 && v < 256); + v >>= (8-bits); + STBI_ASSERT(bits >= 0 && bits <= 8); + return (int) ((unsigned) v * mul_table[bits]) >> shift_table[bits]; +} + +typedef struct +{ + int bpp, offset, hsz; + unsigned int mr,mg,mb,ma, all_a; +} stbi__bmp_data; + +static void *stbi__bmp_parse_header(stbi__context *s, stbi__bmp_data *info) +{ + int hsz; + if (stbi__get8(s) != 'B' || stbi__get8(s) != 'M') return stbi__errpuc("not BMP", "Corrupt BMP"); + stbi__get32le(s); // discard filesize + stbi__get16le(s); // discard reserved + stbi__get16le(s); // discard reserved + info->offset = stbi__get32le(s); + info->hsz = hsz = stbi__get32le(s); + info->mr = info->mg = info->mb = info->ma = 0; + + if (hsz != 12 && hsz != 40 && hsz != 56 && hsz != 108 && hsz != 124) return stbi__errpuc("unknown BMP", "BMP type not supported: unknown"); + if (hsz == 12) { + s->img_x = stbi__get16le(s); + s->img_y = stbi__get16le(s); + } else { + s->img_x = stbi__get32le(s); + s->img_y = stbi__get32le(s); + } + if (stbi__get16le(s) != 1) return stbi__errpuc("bad BMP", "bad BMP"); + info->bpp = stbi__get16le(s); + if (hsz != 12) { + int compress = stbi__get32le(s); + if (compress == 1 || compress == 2) return stbi__errpuc("BMP RLE", "BMP type not supported: RLE"); + stbi__get32le(s); // discard sizeof + stbi__get32le(s); // discard hres + stbi__get32le(s); // discard vres + stbi__get32le(s); // discard colorsused + stbi__get32le(s); // discard max important + if (hsz == 40 || hsz == 56) { + if (hsz == 56) { + stbi__get32le(s); + stbi__get32le(s); + stbi__get32le(s); + stbi__get32le(s); + } + if (info->bpp == 16 || info->bpp == 32) { + if (compress == 0) { + if (info->bpp == 32) { + info->mr = 0xffu << 16; + info->mg = 0xffu << 8; + info->mb = 0xffu << 0; + info->ma = 0xffu << 24; + info->all_a = 0; // if all_a is 0 at end, then we loaded alpha channel but it was all 0 + } else { + info->mr = 31u << 10; + info->mg = 31u << 5; + info->mb = 31u << 0; + } + } else if (compress == 3) { + info->mr = stbi__get32le(s); + info->mg = stbi__get32le(s); + info->mb = stbi__get32le(s); + // not documented, but generated by photoshop and handled by mspaint + if (info->mr == info->mg && info->mg == info->mb) { + // ?!?!? + return stbi__errpuc("bad BMP", "bad BMP"); + } + } else + return stbi__errpuc("bad BMP", "bad BMP"); + } + } else { + int i; + if (hsz != 108 && hsz != 124) + return stbi__errpuc("bad BMP", "bad BMP"); + info->mr = stbi__get32le(s); + info->mg = stbi__get32le(s); + info->mb = stbi__get32le(s); + info->ma = stbi__get32le(s); + stbi__get32le(s); // discard color space + for (i=0; i < 12; ++i) + stbi__get32le(s); // discard color space parameters + if (hsz == 124) { + stbi__get32le(s); // discard rendering intent + stbi__get32le(s); // discard offset of profile data + stbi__get32le(s); // discard size of profile data + stbi__get32le(s); // discard reserved + } + } + } + return (void *) 1; +} + + +static void *stbi__bmp_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi_uc *out; + unsigned int mr=0,mg=0,mb=0,ma=0, all_a; + stbi_uc pal[256][4]; + int psize=0,i,j,width; + int flip_vertically, pad, target; + stbi__bmp_data info; + STBI_NOTUSED(ri); + + info.all_a = 255; + if (stbi__bmp_parse_header(s, &info) == NULL) + return NULL; // error code already set + + flip_vertically = ((int) s->img_y) > 0; + s->img_y = abs((int) s->img_y); + + mr = info.mr; + mg = info.mg; + mb = info.mb; + ma = info.ma; + all_a = info.all_a; + + if (info.hsz == 12) { + if (info.bpp < 24) + psize = (info.offset - 14 - 24) / 3; + } else { + if (info.bpp < 16) + psize = (info.offset - 14 - info.hsz) >> 2; + } + + s->img_n = ma ? 4 : 3; + if (req_comp && req_comp >= 3) // we can directly decode 3 or 4 + target = req_comp; + else + target = s->img_n; // if they want monochrome, we'll post-convert + + // sanity-check size + if (!stbi__mad3sizes_valid(target, s->img_x, s->img_y, 0)) + return stbi__errpuc("too large", "Corrupt BMP"); + + out = (stbi_uc *) stbi__malloc_mad3(target, s->img_x, s->img_y, 0); + if (!out) return stbi__errpuc("outofmem", "Out of memory"); + if (info.bpp < 16) { + int z=0; + if (psize == 0 || psize > 256) { STBI_FREE(out); return stbi__errpuc("invalid", "Corrupt BMP"); } + for (i=0; i < psize; ++i) { + pal[i][2] = stbi__get8(s); + pal[i][1] = stbi__get8(s); + pal[i][0] = stbi__get8(s); + if (info.hsz != 12) stbi__get8(s); + pal[i][3] = 255; + } + stbi__skip(s, info.offset - 14 - info.hsz - psize * (info.hsz == 12 ? 3 : 4)); + if (info.bpp == 1) width = (s->img_x + 7) >> 3; + else if (info.bpp == 4) width = (s->img_x + 1) >> 1; + else if (info.bpp == 8) width = s->img_x; + else { STBI_FREE(out); return stbi__errpuc("bad bpp", "Corrupt BMP"); } + pad = (-width)&3; + if (info.bpp == 1) { + for (j=0; j < (int) s->img_y; ++j) { + int bit_offset = 7, v = stbi__get8(s); + for (i=0; i < (int) s->img_x; ++i) { + int color = (v>>bit_offset)&0x1; + out[z++] = pal[color][0]; + out[z++] = pal[color][1]; + out[z++] = pal[color][2]; + if (target == 4) out[z++] = 255; + if (i+1 == (int) s->img_x) break; + if((--bit_offset) < 0) { + bit_offset = 7; + v = stbi__get8(s); + } + } + stbi__skip(s, pad); + } + } else { + for (j=0; j < (int) s->img_y; ++j) { + for (i=0; i < (int) s->img_x; i += 2) { + int v=stbi__get8(s),v2=0; + if (info.bpp == 4) { + v2 = v & 15; + v >>= 4; + } + out[z++] = pal[v][0]; + out[z++] = pal[v][1]; + out[z++] = pal[v][2]; + if (target == 4) out[z++] = 255; + if (i+1 == (int) s->img_x) break; + v = (info.bpp == 8) ? stbi__get8(s) : v2; + out[z++] = pal[v][0]; + out[z++] = pal[v][1]; + out[z++] = pal[v][2]; + if (target == 4) out[z++] = 255; + } + stbi__skip(s, pad); + } + } + } else { + int rshift=0,gshift=0,bshift=0,ashift=0,rcount=0,gcount=0,bcount=0,acount=0; + int z = 0; + int easy=0; + stbi__skip(s, info.offset - 14 - info.hsz); + if (info.bpp == 24) width = 3 * s->img_x; + else if (info.bpp == 16) width = 2*s->img_x; + else /* bpp = 32 and pad = 0 */ width=0; + pad = (-width) & 3; + if (info.bpp == 24) { + easy = 1; + } else if (info.bpp == 32) { + if (mb == 0xff && mg == 0xff00 && mr == 0x00ff0000 && ma == 0xff000000) + easy = 2; + } + if (!easy) { + if (!mr || !mg || !mb) { STBI_FREE(out); return stbi__errpuc("bad masks", "Corrupt BMP"); } + // right shift amt to put high bit in position #7 + rshift = stbi__high_bit(mr)-7; rcount = stbi__bitcount(mr); + gshift = stbi__high_bit(mg)-7; gcount = stbi__bitcount(mg); + bshift = stbi__high_bit(mb)-7; bcount = stbi__bitcount(mb); + ashift = stbi__high_bit(ma)-7; acount = stbi__bitcount(ma); + } + for (j=0; j < (int) s->img_y; ++j) { + if (easy) { + for (i=0; i < (int) s->img_x; ++i) { + unsigned char a; + out[z+2] = stbi__get8(s); + out[z+1] = stbi__get8(s); + out[z+0] = stbi__get8(s); + z += 3; + a = (easy == 2 ? stbi__get8(s) : 255); + all_a |= a; + if (target == 4) out[z++] = a; + } + } else { + int bpp = info.bpp; + for (i=0; i < (int) s->img_x; ++i) { + stbi__uint32 v = (bpp == 16 ? (stbi__uint32) stbi__get16le(s) : stbi__get32le(s)); + unsigned int a; + out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mr, rshift, rcount)); + out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mg, gshift, gcount)); + out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mb, bshift, bcount)); + a = (ma ? stbi__shiftsigned(v & ma, ashift, acount) : 255); + all_a |= a; + if (target == 4) out[z++] = STBI__BYTECAST(a); + } + } + stbi__skip(s, pad); + } + } + + // if alpha channel is all 0s, replace with all 255s + if (target == 4 && all_a == 0) + for (i=4*s->img_x*s->img_y-1; i >= 0; i -= 4) + out[i] = 255; + + if (flip_vertically) { + stbi_uc t; + for (j=0; j < (int) s->img_y>>1; ++j) { + stbi_uc *p1 = out + j *s->img_x*target; + stbi_uc *p2 = out + (s->img_y-1-j)*s->img_x*target; + for (i=0; i < (int) s->img_x*target; ++i) { + t = p1[i]; p1[i] = p2[i]; p2[i] = t; + } + } + } + + if (req_comp && req_comp != target) { + out = stbi__convert_format(out, target, req_comp, s->img_x, s->img_y); + if (out == NULL) return out; // stbi__convert_format frees input on failure + } + + *x = s->img_x; + *y = s->img_y; + if (comp) *comp = s->img_n; + return out; +} +#endif + +// Targa Truevision - TGA +// by Jonathan Dummer +#ifndef STBI_NO_TGA +// returns STBI_rgb or whatever, 0 on error +static int stbi__tga_get_comp(int bits_per_pixel, int is_grey, int* is_rgb16) +{ + // only RGB or RGBA (incl. 16bit) or grey allowed + if (is_rgb16) *is_rgb16 = 0; + switch(bits_per_pixel) { + case 8: return STBI_grey; + case 16: if(is_grey) return STBI_grey_alpha; + // fallthrough + case 15: if(is_rgb16) *is_rgb16 = 1; + return STBI_rgb; + case 24: // fallthrough + case 32: return bits_per_pixel/8; + default: return 0; + } +} + +static int stbi__tga_info(stbi__context *s, int *x, int *y, int *comp) +{ + int tga_w, tga_h, tga_comp, tga_image_type, tga_bits_per_pixel, tga_colormap_bpp; + int sz, tga_colormap_type; + stbi__get8(s); // discard Offset + tga_colormap_type = stbi__get8(s); // colormap type + if( tga_colormap_type > 1 ) { + stbi__rewind(s); + return 0; // only RGB or indexed allowed + } + tga_image_type = stbi__get8(s); // image type + if ( tga_colormap_type == 1 ) { // colormapped (paletted) image + if (tga_image_type != 1 && tga_image_type != 9) { + stbi__rewind(s); + return 0; + } + stbi__skip(s,4); // skip index of first colormap entry and number of entries + sz = stbi__get8(s); // check bits per palette color entry + if ( (sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32) ) { + stbi__rewind(s); + return 0; + } + stbi__skip(s,4); // skip image x and y origin + tga_colormap_bpp = sz; + } else { // "normal" image w/o colormap - only RGB or grey allowed, +/- RLE + if ( (tga_image_type != 2) && (tga_image_type != 3) && (tga_image_type != 10) && (tga_image_type != 11) ) { + stbi__rewind(s); + return 0; // only RGB or grey allowed, +/- RLE + } + stbi__skip(s,9); // skip colormap specification and image x/y origin + tga_colormap_bpp = 0; + } + tga_w = stbi__get16le(s); + if( tga_w < 1 ) { + stbi__rewind(s); + return 0; // test width + } + tga_h = stbi__get16le(s); + if( tga_h < 1 ) { + stbi__rewind(s); + return 0; // test height + } + tga_bits_per_pixel = stbi__get8(s); // bits per pixel + stbi__get8(s); // ignore alpha bits + if (tga_colormap_bpp != 0) { + if((tga_bits_per_pixel != 8) && (tga_bits_per_pixel != 16)) { + // when using a colormap, tga_bits_per_pixel is the size of the indexes + // I don't think anything but 8 or 16bit indexes makes sense + stbi__rewind(s); + return 0; + } + tga_comp = stbi__tga_get_comp(tga_colormap_bpp, 0, NULL); + } else { + tga_comp = stbi__tga_get_comp(tga_bits_per_pixel, (tga_image_type == 3) || (tga_image_type == 11), NULL); + } + if(!tga_comp) { + stbi__rewind(s); + return 0; + } + if (x) *x = tga_w; + if (y) *y = tga_h; + if (comp) *comp = tga_comp; + return 1; // seems to have passed everything +} + +static int stbi__tga_test(stbi__context *s) +{ + int res = 0; + int sz, tga_color_type; + stbi__get8(s); // discard Offset + tga_color_type = stbi__get8(s); // color type + if ( tga_color_type > 1 ) goto errorEnd; // only RGB or indexed allowed + sz = stbi__get8(s); // image type + if ( tga_color_type == 1 ) { // colormapped (paletted) image + if (sz != 1 && sz != 9) goto errorEnd; // colortype 1 demands image type 1 or 9 + stbi__skip(s,4); // skip index of first colormap entry and number of entries + sz = stbi__get8(s); // check bits per palette color entry + if ( (sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32) ) goto errorEnd; + stbi__skip(s,4); // skip image x and y origin + } else { // "normal" image w/o colormap + if ( (sz != 2) && (sz != 3) && (sz != 10) && (sz != 11) ) goto errorEnd; // only RGB or grey allowed, +/- RLE + stbi__skip(s,9); // skip colormap specification and image x/y origin + } + if ( stbi__get16le(s) < 1 ) goto errorEnd; // test width + if ( stbi__get16le(s) < 1 ) goto errorEnd; // test height + sz = stbi__get8(s); // bits per pixel + if ( (tga_color_type == 1) && (sz != 8) && (sz != 16) ) goto errorEnd; // for colormapped images, bpp is size of an index + if ( (sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32) ) goto errorEnd; + + res = 1; // if we got this far, everything's good and we can return 1 instead of 0 + +errorEnd: + stbi__rewind(s); + return res; +} + +// read 16bit value and convert to 24bit RGB +static void stbi__tga_read_rgb16(stbi__context *s, stbi_uc* out) +{ + stbi__uint16 px = (stbi__uint16)stbi__get16le(s); + stbi__uint16 fiveBitMask = 31; + // we have 3 channels with 5bits each + int r = (px >> 10) & fiveBitMask; + int g = (px >> 5) & fiveBitMask; + int b = px & fiveBitMask; + // Note that this saves the data in RGB(A) order, so it doesn't need to be swapped later + out[0] = (stbi_uc)((r * 255)/31); + out[1] = (stbi_uc)((g * 255)/31); + out[2] = (stbi_uc)((b * 255)/31); + + // some people claim that the most significant bit might be used for alpha + // (possibly if an alpha-bit is set in the "image descriptor byte") + // but that only made 16bit test images completely translucent.. + // so let's treat all 15 and 16bit TGAs as RGB with no alpha. +} + +static void *stbi__tga_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + // read in the TGA header stuff + int tga_offset = stbi__get8(s); + int tga_indexed = stbi__get8(s); + int tga_image_type = stbi__get8(s); + int tga_is_RLE = 0; + int tga_palette_start = stbi__get16le(s); + int tga_palette_len = stbi__get16le(s); + int tga_palette_bits = stbi__get8(s); + int tga_x_origin = stbi__get16le(s); + int tga_y_origin = stbi__get16le(s); + int tga_width = stbi__get16le(s); + int tga_height = stbi__get16le(s); + int tga_bits_per_pixel = stbi__get8(s); + int tga_comp, tga_rgb16=0; + int tga_inverted = stbi__get8(s); + // int tga_alpha_bits = tga_inverted & 15; // the 4 lowest bits - unused (useless?) + // image data + unsigned char *tga_data; + unsigned char *tga_palette = NULL; + int i, j; + unsigned char raw_data[4] = {0}; + int RLE_count = 0; + int RLE_repeating = 0; + int read_next_pixel = 1; + STBI_NOTUSED(ri); + + // do a tiny bit of precessing + if ( tga_image_type >= 8 ) + { + tga_image_type -= 8; + tga_is_RLE = 1; + } + tga_inverted = 1 - ((tga_inverted >> 5) & 1); + + // If I'm paletted, then I'll use the number of bits from the palette + if ( tga_indexed ) tga_comp = stbi__tga_get_comp(tga_palette_bits, 0, &tga_rgb16); + else tga_comp = stbi__tga_get_comp(tga_bits_per_pixel, (tga_image_type == 3), &tga_rgb16); + + if(!tga_comp) // shouldn't really happen, stbi__tga_test() should have ensured basic consistency + return stbi__errpuc("bad format", "Can't find out TGA pixelformat"); + + // tga info + *x = tga_width; + *y = tga_height; + if (comp) *comp = tga_comp; + + if (!stbi__mad3sizes_valid(tga_width, tga_height, tga_comp, 0)) + return stbi__errpuc("too large", "Corrupt TGA"); + + tga_data = (unsigned char*)stbi__malloc_mad3(tga_width, tga_height, tga_comp, 0); + if (!tga_data) return stbi__errpuc("outofmem", "Out of memory"); + + // skip to the data's starting position (offset usually = 0) + stbi__skip(s, tga_offset ); + + if ( !tga_indexed && !tga_is_RLE && !tga_rgb16 ) { + for (i=0; i < tga_height; ++i) { + int row = tga_inverted ? tga_height -i - 1 : i; + stbi_uc *tga_row = tga_data + row*tga_width*tga_comp; + stbi__getn(s, tga_row, tga_width * tga_comp); + } + } else { + // do I need to load a palette? + if ( tga_indexed) + { + // any data to skip? (offset usually = 0) + stbi__skip(s, tga_palette_start ); + // load the palette + tga_palette = (unsigned char*)stbi__malloc_mad2(tga_palette_len, tga_comp, 0); + if (!tga_palette) { + STBI_FREE(tga_data); + return stbi__errpuc("outofmem", "Out of memory"); + } + if (tga_rgb16) { + stbi_uc *pal_entry = tga_palette; + STBI_ASSERT(tga_comp == STBI_rgb); + for (i=0; i < tga_palette_len; ++i) { + stbi__tga_read_rgb16(s, pal_entry); + pal_entry += tga_comp; + } + } else if (!stbi__getn(s, tga_palette, tga_palette_len * tga_comp)) { + STBI_FREE(tga_data); + STBI_FREE(tga_palette); + return stbi__errpuc("bad palette", "Corrupt TGA"); + } + } + // load the data + for (i=0; i < tga_width * tga_height; ++i) + { + // if I'm in RLE mode, do I need to get a RLE stbi__pngchunk? + if ( tga_is_RLE ) + { + if ( RLE_count == 0 ) + { + // yep, get the next byte as a RLE command + int RLE_cmd = stbi__get8(s); + RLE_count = 1 + (RLE_cmd & 127); + RLE_repeating = RLE_cmd >> 7; + read_next_pixel = 1; + } else if ( !RLE_repeating ) + { + read_next_pixel = 1; + } + } else + { + read_next_pixel = 1; + } + // OK, if I need to read a pixel, do it now + if ( read_next_pixel ) + { + // load however much data we did have + if ( tga_indexed ) + { + // read in index, then perform the lookup + int pal_idx = (tga_bits_per_pixel == 8) ? stbi__get8(s) : stbi__get16le(s); + if ( pal_idx >= tga_palette_len ) { + // invalid index + pal_idx = 0; + } + pal_idx *= tga_comp; + for (j = 0; j < tga_comp; ++j) { + raw_data[j] = tga_palette[pal_idx+j]; + } + } else if(tga_rgb16) { + STBI_ASSERT(tga_comp == STBI_rgb); + stbi__tga_read_rgb16(s, raw_data); + } else { + // read in the data raw + for (j = 0; j < tga_comp; ++j) { + raw_data[j] = stbi__get8(s); + } + } + // clear the reading flag for the next pixel + read_next_pixel = 0; + } // end of reading a pixel + + // copy data + for (j = 0; j < tga_comp; ++j) + tga_data[i*tga_comp+j] = raw_data[j]; + + // in case we're in RLE mode, keep counting down + --RLE_count; + } + // do I need to invert the image? + if ( tga_inverted ) + { + for (j = 0; j*2 < tga_height; ++j) + { + int index1 = j * tga_width * tga_comp; + int index2 = (tga_height - 1 - j) * tga_width * tga_comp; + for (i = tga_width * tga_comp; i > 0; --i) + { + unsigned char temp = tga_data[index1]; + tga_data[index1] = tga_data[index2]; + tga_data[index2] = temp; + ++index1; + ++index2; + } + } + } + // clear my palette, if I had one + if ( tga_palette != NULL ) + { + STBI_FREE( tga_palette ); + } + } + + // swap RGB - if the source data was RGB16, it already is in the right order + if (tga_comp >= 3 && !tga_rgb16) + { + unsigned char* tga_pixel = tga_data; + for (i=0; i < tga_width * tga_height; ++i) + { + unsigned char temp = tga_pixel[0]; + tga_pixel[0] = tga_pixel[2]; + tga_pixel[2] = temp; + tga_pixel += tga_comp; + } + } + + // convert to target component count + if (req_comp && req_comp != tga_comp) + tga_data = stbi__convert_format(tga_data, tga_comp, req_comp, tga_width, tga_height); + + // the things I do to get rid of an error message, and yet keep + // Microsoft's C compilers happy... [8^( + tga_palette_start = tga_palette_len = tga_palette_bits = + tga_x_origin = tga_y_origin = 0; + // OK, done + return tga_data; +} +#endif + +// ************************************************************************************************* +// Photoshop PSD loader -- PD by Thatcher Ulrich, integration by Nicolas Schulz, tweaked by STB + +#ifndef STBI_NO_PSD +static int stbi__psd_test(stbi__context *s) +{ + int r = (stbi__get32be(s) == 0x38425053); + stbi__rewind(s); + return r; +} + +static int stbi__psd_decode_rle(stbi__context *s, stbi_uc *p, int pixelCount) +{ + int count, nleft, len; + + count = 0; + while ((nleft = pixelCount - count) > 0) { + len = stbi__get8(s); + if (len == 128) { + // No-op. + } else if (len < 128) { + // Copy next len+1 bytes literally. + len++; + if (len > nleft) return 0; // corrupt data + count += len; + while (len) { + *p = stbi__get8(s); + p += 4; + len--; + } + } else if (len > 128) { + stbi_uc val; + // Next -len+1 bytes in the dest are replicated from next source byte. + // (Interpret len as a negative 8-bit int.) + len = 257 - len; + if (len > nleft) return 0; // corrupt data + val = stbi__get8(s); + count += len; + while (len) { + *p = val; + p += 4; + len--; + } + } + } + + return 1; +} + +static void *stbi__psd_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri, int bpc) +{ + int pixelCount; + int channelCount, compression; + int channel, i; + int bitdepth; + int w,h; + stbi_uc *out; + STBI_NOTUSED(ri); + + // Check identifier + if (stbi__get32be(s) != 0x38425053) // "8BPS" + return stbi__errpuc("not PSD", "Corrupt PSD image"); + + // Check file type version. + if (stbi__get16be(s) != 1) + return stbi__errpuc("wrong version", "Unsupported version of PSD image"); + + // Skip 6 reserved bytes. + stbi__skip(s, 6 ); + + // Read the number of channels (R, G, B, A, etc). + channelCount = stbi__get16be(s); + if (channelCount < 0 || channelCount > 16) + return stbi__errpuc("wrong channel count", "Unsupported number of channels in PSD image"); + + // Read the rows and columns of the image. + h = stbi__get32be(s); + w = stbi__get32be(s); + + // Make sure the depth is 8 bits. + bitdepth = stbi__get16be(s); + if (bitdepth != 8 && bitdepth != 16) + return stbi__errpuc("unsupported bit depth", "PSD bit depth is not 8 or 16 bit"); + + // Make sure the color mode is RGB. + // Valid options are: + // 0: Bitmap + // 1: Grayscale + // 2: Indexed color + // 3: RGB color + // 4: CMYK color + // 7: Multichannel + // 8: Duotone + // 9: Lab color + if (stbi__get16be(s) != 3) + return stbi__errpuc("wrong color format", "PSD is not in RGB color format"); + + // Skip the Mode Data. (It's the palette for indexed color; other info for other modes.) + stbi__skip(s,stbi__get32be(s) ); + + // Skip the image resources. (resolution, pen tool paths, etc) + stbi__skip(s, stbi__get32be(s) ); + + // Skip the reserved data. + stbi__skip(s, stbi__get32be(s) ); + + // Find out if the data is compressed. + // Known values: + // 0: no compression + // 1: RLE compressed + compression = stbi__get16be(s); + if (compression > 1) + return stbi__errpuc("bad compression", "PSD has an unknown compression format"); + + // Check size + if (!stbi__mad3sizes_valid(4, w, h, 0)) + return stbi__errpuc("too large", "Corrupt PSD"); + + // Create the destination image. + + if (!compression && bitdepth == 16 && bpc == 16) { + out = (stbi_uc *) stbi__malloc_mad3(8, w, h, 0); + ri->bits_per_channel = 16; + } else + out = (stbi_uc *) stbi__malloc(4 * w*h); + + if (!out) return stbi__errpuc("outofmem", "Out of memory"); + pixelCount = w*h; + + // Initialize the data to zero. + //memset( out, 0, pixelCount * 4 ); + + // Finally, the image data. + if (compression) { + // RLE as used by .PSD and .TIFF + // Loop until you get the number of unpacked bytes you are expecting: + // Read the next source byte into n. + // If n is between 0 and 127 inclusive, copy the next n+1 bytes literally. + // Else if n is between -127 and -1 inclusive, copy the next byte -n+1 times. + // Else if n is 128, noop. + // Endloop + + // The RLE-compressed data is preceded by a 2-byte data count for each row in the data, + // which we're going to just skip. + stbi__skip(s, h * channelCount * 2 ); + + // Read the RLE data by channel. + for (channel = 0; channel < 4; channel++) { + stbi_uc *p; + + p = out+channel; + if (channel >= channelCount) { + // Fill this channel with default data. + for (i = 0; i < pixelCount; i++, p += 4) + *p = (channel == 3 ? 255 : 0); + } else { + // Read the RLE data. + if (!stbi__psd_decode_rle(s, p, pixelCount)) { + STBI_FREE(out); + return stbi__errpuc("corrupt", "bad RLE data"); + } + } + } + + } else { + // We're at the raw image data. It's each channel in order (Red, Green, Blue, Alpha, ...) + // where each channel consists of an 8-bit (or 16-bit) value for each pixel in the image. + + // Read the data by channel. + for (channel = 0; channel < 4; channel++) { + if (channel >= channelCount) { + // Fill this channel with default data. + if (bitdepth == 16 && bpc == 16) { + stbi__uint16 *q = ((stbi__uint16 *) out) + channel; + stbi__uint16 val = channel == 3 ? 65535 : 0; + for (i = 0; i < pixelCount; i++, q += 4) + *q = val; + } else { + stbi_uc *p = out+channel; + stbi_uc val = channel == 3 ? 255 : 0; + for (i = 0; i < pixelCount; i++, p += 4) + *p = val; + } + } else { + if (ri->bits_per_channel == 16) { // output bpc + stbi__uint16 *q = ((stbi__uint16 *) out) + channel; + for (i = 0; i < pixelCount; i++, q += 4) + *q = (stbi__uint16) stbi__get16be(s); + } else { + stbi_uc *p = out+channel; + if (bitdepth == 16) { // input bpc + for (i = 0; i < pixelCount; i++, p += 4) + *p = (stbi_uc) (stbi__get16be(s) >> 8); + } else { + for (i = 0; i < pixelCount; i++, p += 4) + *p = stbi__get8(s); + } + } + } + } + } + + // remove weird white matte from PSD + if (channelCount >= 4) { + if (ri->bits_per_channel == 16) { + for (i=0; i < w*h; ++i) { + stbi__uint16 *pixel = (stbi__uint16 *) out + 4*i; + if (pixel[3] != 0 && pixel[3] != 65535) { + float a = pixel[3] / 65535.0f; + float ra = 1.0f / a; + float inv_a = 65535.0f * (1 - ra); + pixel[0] = (stbi__uint16) (pixel[0]*ra + inv_a); + pixel[1] = (stbi__uint16) (pixel[1]*ra + inv_a); + pixel[2] = (stbi__uint16) (pixel[2]*ra + inv_a); + } + } + } else { + for (i=0; i < w*h; ++i) { + unsigned char *pixel = out + 4*i; + if (pixel[3] != 0 && pixel[3] != 255) { + float a = pixel[3] / 255.0f; + float ra = 1.0f / a; + float inv_a = 255.0f * (1 - ra); + pixel[0] = (unsigned char) (pixel[0]*ra + inv_a); + pixel[1] = (unsigned char) (pixel[1]*ra + inv_a); + pixel[2] = (unsigned char) (pixel[2]*ra + inv_a); + } + } + } + } + + // convert to desired output format + if (req_comp && req_comp != 4) { + if (ri->bits_per_channel == 16) + out = (stbi_uc *) stbi__convert_format16((stbi__uint16 *) out, 4, req_comp, w, h); + else + out = stbi__convert_format(out, 4, req_comp, w, h); + if (out == NULL) return out; // stbi__convert_format frees input on failure + } + + if (comp) *comp = 4; + *y = h; + *x = w; + + return out; +} +#endif + +// ************************************************************************************************* +// Softimage PIC loader +// by Tom Seddon +// +// See http://softimage.wiki.softimage.com/index.php/INFO:_PIC_file_format +// See http://ozviz.wasp.uwa.edu.au/~pbourke/dataformats/softimagepic/ + +#ifndef STBI_NO_PIC +static int stbi__pic_is4(stbi__context *s,const char *str) +{ + int i; + for (i=0; i<4; ++i) + if (stbi__get8(s) != (stbi_uc)str[i]) + return 0; + + return 1; +} + +static int stbi__pic_test_core(stbi__context *s) +{ + int i; + + if (!stbi__pic_is4(s,"\x53\x80\xF6\x34")) + return 0; + + for(i=0;i<84;++i) + stbi__get8(s); + + if (!stbi__pic_is4(s,"PICT")) + return 0; + + return 1; +} + +typedef struct +{ + stbi_uc size,type,channel; +} stbi__pic_packet; + +static stbi_uc *stbi__readval(stbi__context *s, int channel, stbi_uc *dest) +{ + int mask=0x80, i; + + for (i=0; i<4; ++i, mask>>=1) { + if (channel & mask) { + if (stbi__at_eof(s)) return stbi__errpuc("bad file","PIC file too short"); + dest[i]=stbi__get8(s); + } + } + + return dest; +} + +static void stbi__copyval(int channel,stbi_uc *dest,const stbi_uc *src) +{ + int mask=0x80,i; + + for (i=0;i<4; ++i, mask>>=1) + if (channel&mask) + dest[i]=src[i]; +} + +static stbi_uc *stbi__pic_load_core(stbi__context *s,int width,int height,int *comp, stbi_uc *result) +{ + int act_comp=0,num_packets=0,y,chained; + stbi__pic_packet packets[10]; + + // this will (should...) cater for even some bizarre stuff like having data + // for the same channel in multiple packets. + do { + stbi__pic_packet *packet; + + if (num_packets==sizeof(packets)/sizeof(packets[0])) + return stbi__errpuc("bad format","too many packets"); + + packet = &packets[num_packets++]; + + chained = stbi__get8(s); + packet->size = stbi__get8(s); + packet->type = stbi__get8(s); + packet->channel = stbi__get8(s); + + act_comp |= packet->channel; + + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (reading packets)"); + if (packet->size != 8) return stbi__errpuc("bad format","packet isn't 8bpp"); + } while (chained); + + *comp = (act_comp & 0x10 ? 4 : 3); // has alpha channel? + + for(y=0; ytype) { + default: + return stbi__errpuc("bad format","packet has bad compression type"); + + case 0: {//uncompressed + int x; + + for(x=0;xchannel,dest)) + return 0; + break; + } + + case 1://Pure RLE + { + int left=width, i; + + while (left>0) { + stbi_uc count,value[4]; + + count=stbi__get8(s); + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (pure read count)"); + + if (count > left) + count = (stbi_uc) left; + + if (!stbi__readval(s,packet->channel,value)) return 0; + + for(i=0; ichannel,dest,value); + left -= count; + } + } + break; + + case 2: {//Mixed RLE + int left=width; + while (left>0) { + int count = stbi__get8(s), i; + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (mixed read count)"); + + if (count >= 128) { // Repeated + stbi_uc value[4]; + + if (count==128) + count = stbi__get16be(s); + else + count -= 127; + if (count > left) + return stbi__errpuc("bad file","scanline overrun"); + + if (!stbi__readval(s,packet->channel,value)) + return 0; + + for(i=0;ichannel,dest,value); + } else { // Raw + ++count; + if (count>left) return stbi__errpuc("bad file","scanline overrun"); + + for(i=0;ichannel,dest)) + return 0; + } + left-=count; + } + break; + } + } + } + } + + return result; +} + +static void *stbi__pic_load(stbi__context *s,int *px,int *py,int *comp,int req_comp, stbi__result_info *ri) +{ + stbi_uc *result; + int i, x,y, internal_comp; + STBI_NOTUSED(ri); + + if (!comp) comp = &internal_comp; + + for (i=0; i<92; ++i) + stbi__get8(s); + + x = stbi__get16be(s); + y = stbi__get16be(s); + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (pic header)"); + if (!stbi__mad3sizes_valid(x, y, 4, 0)) return stbi__errpuc("too large", "PIC image too large to decode"); + + stbi__get32be(s); //skip `ratio' + stbi__get16be(s); //skip `fields' + stbi__get16be(s); //skip `pad' + + // intermediate buffer is RGBA + result = (stbi_uc *) stbi__malloc_mad3(x, y, 4, 0); + memset(result, 0xff, x*y*4); + + if (!stbi__pic_load_core(s,x,y,comp, result)) { + STBI_FREE(result); + result=0; + } + *px = x; + *py = y; + if (req_comp == 0) req_comp = *comp; + result=stbi__convert_format(result,4,req_comp,x,y); + + return result; +} + +static int stbi__pic_test(stbi__context *s) +{ + int r = stbi__pic_test_core(s); + stbi__rewind(s); + return r; +} +#endif + +// ************************************************************************************************* +// GIF loader -- public domain by Jean-Marc Lienher -- simplified/shrunk by stb + +#ifndef STBI_NO_GIF +typedef struct +{ + stbi__int16 prefix; + stbi_uc first; + stbi_uc suffix; +} stbi__gif_lzw; + +typedef struct +{ + int w,h; + stbi_uc *out; // output buffer (always 4 components) + stbi_uc *background; // The current "background" as far as a gif is concerned + stbi_uc *history; + int flags, bgindex, ratio, transparent, eflags; + stbi_uc pal[256][4]; + stbi_uc lpal[256][4]; + stbi__gif_lzw codes[8192]; + stbi_uc *color_table; + int parse, step; + int lflags; + int start_x, start_y; + int max_x, max_y; + int cur_x, cur_y; + int line_size; + int delay; +} stbi__gif; + +static int stbi__gif_test_raw(stbi__context *s) +{ + int sz; + if (stbi__get8(s) != 'G' || stbi__get8(s) != 'I' || stbi__get8(s) != 'F' || stbi__get8(s) != '8') return 0; + sz = stbi__get8(s); + if (sz != '9' && sz != '7') return 0; + if (stbi__get8(s) != 'a') return 0; + return 1; +} + +static int stbi__gif_test(stbi__context *s) +{ + int r = stbi__gif_test_raw(s); + stbi__rewind(s); + return r; +} + +static void stbi__gif_parse_colortable(stbi__context *s, stbi_uc pal[256][4], int num_entries, int transp) +{ + int i; + for (i=0; i < num_entries; ++i) { + pal[i][2] = stbi__get8(s); + pal[i][1] = stbi__get8(s); + pal[i][0] = stbi__get8(s); + pal[i][3] = transp == i ? 0 : 255; + } +} + +static int stbi__gif_header(stbi__context *s, stbi__gif *g, int *comp, int is_info) +{ + stbi_uc version; + if (stbi__get8(s) != 'G' || stbi__get8(s) != 'I' || stbi__get8(s) != 'F' || stbi__get8(s) != '8') + return stbi__err("not GIF", "Corrupt GIF"); + + version = stbi__get8(s); + if (version != '7' && version != '9') return stbi__err("not GIF", "Corrupt GIF"); + if (stbi__get8(s) != 'a') return stbi__err("not GIF", "Corrupt GIF"); + + stbi__g_failure_reason = ""; + g->w = stbi__get16le(s); + g->h = stbi__get16le(s); + g->flags = stbi__get8(s); + g->bgindex = stbi__get8(s); + g->ratio = stbi__get8(s); + g->transparent = -1; + + if (comp != 0) *comp = 4; // can't actually tell whether it's 3 or 4 until we parse the comments + + if (is_info) return 1; + + if (g->flags & 0x80) + stbi__gif_parse_colortable(s,g->pal, 2 << (g->flags & 7), -1); + + return 1; +} + +static int stbi__gif_info_raw(stbi__context *s, int *x, int *y, int *comp) +{ + stbi__gif* g = (stbi__gif*) stbi__malloc(sizeof(stbi__gif)); + if (!stbi__gif_header(s, g, comp, 1)) { + STBI_FREE(g); + stbi__rewind( s ); + return 0; + } + if (x) *x = g->w; + if (y) *y = g->h; + STBI_FREE(g); + return 1; +} + +static void stbi__out_gif_code(stbi__gif *g, stbi__uint16 code) +{ + stbi_uc *p, *c; + int idx; + + // recurse to decode the prefixes, since the linked-list is backwards, + // and working backwards through an interleaved image would be nasty + if (g->codes[code].prefix >= 0) + stbi__out_gif_code(g, g->codes[code].prefix); + + if (g->cur_y >= g->max_y) return; + + idx = g->cur_x + g->cur_y; + p = &g->out[idx]; + g->history[idx / 4] = 1; + + c = &g->color_table[g->codes[code].suffix * 4]; + if (c[3] > 128) { // don't render transparent pixels; + p[0] = c[2]; + p[1] = c[1]; + p[2] = c[0]; + p[3] = c[3]; + } + g->cur_x += 4; + + if (g->cur_x >= g->max_x) { + g->cur_x = g->start_x; + g->cur_y += g->step; + + while (g->cur_y >= g->max_y && g->parse > 0) { + g->step = (1 << g->parse) * g->line_size; + g->cur_y = g->start_y + (g->step >> 1); + --g->parse; + } + } +} + +static stbi_uc *stbi__process_gif_raster(stbi__context *s, stbi__gif *g) +{ + stbi_uc lzw_cs; + stbi__int32 len, init_code; + stbi__uint32 first; + stbi__int32 codesize, codemask, avail, oldcode, bits, valid_bits, clear; + stbi__gif_lzw *p; + + lzw_cs = stbi__get8(s); + if (lzw_cs > 12) return NULL; + clear = 1 << lzw_cs; + first = 1; + codesize = lzw_cs + 1; + codemask = (1 << codesize) - 1; + bits = 0; + valid_bits = 0; + for (init_code = 0; init_code < clear; init_code++) { + g->codes[init_code].prefix = -1; + g->codes[init_code].first = (stbi_uc) init_code; + g->codes[init_code].suffix = (stbi_uc) init_code; + } + + // support no starting clear code + avail = clear+2; + oldcode = -1; + + len = 0; + for(;;) { + if (valid_bits < codesize) { + if (len == 0) { + len = stbi__get8(s); // start new block + if (len == 0) + return g->out; + } + --len; + bits |= (stbi__int32) stbi__get8(s) << valid_bits; + valid_bits += 8; + } else { + stbi__int32 code = bits & codemask; + bits >>= codesize; + valid_bits -= codesize; + // @OPTIMIZE: is there some way we can accelerate the non-clear path? + if (code == clear) { // clear code + codesize = lzw_cs + 1; + codemask = (1 << codesize) - 1; + avail = clear + 2; + oldcode = -1; + first = 0; + } else if (code == clear + 1) { // end of stream code + stbi__skip(s, len); + while ((len = stbi__get8(s)) > 0) + stbi__skip(s,len); + return g->out; + } else if (code <= avail) { + if (first) { + return stbi__errpuc("no clear code", "Corrupt GIF"); + } + + if (oldcode >= 0) { + p = &g->codes[avail++]; + if (avail > 8192) { + return stbi__errpuc("too many codes", "Corrupt GIF"); + } + + p->prefix = (stbi__int16) oldcode; + p->first = g->codes[oldcode].first; + p->suffix = (code == avail) ? p->first : g->codes[code].first; + } else if (code == avail) + return stbi__errpuc("illegal code in raster", "Corrupt GIF"); + + stbi__out_gif_code(g, (stbi__uint16) code); + + if ((avail & codemask) == 0 && avail <= 0x0FFF) { + codesize++; + codemask = (1 << codesize) - 1; + } + + oldcode = code; + } else { + return stbi__errpuc("illegal code in raster", "Corrupt GIF"); + } + } + } +} + +// this function is designed to support animated gifs, although stb_image doesn't support it +// two back is the image from two frames ago, used for a very specific disposal format +static stbi_uc *stbi__gif_load_next(stbi__context *s, stbi__gif *g, int *comp, int req_comp, stbi_uc *two_back) +{ + int dispose; + int first_frame; + int pi; + int pcount; + STBI_NOTUSED(req_comp); + + // on first frame, any non-written pixels get the background colour (non-transparent) + first_frame = 0; + if (g->out == 0) { + if (!stbi__gif_header(s, g, comp,0)) return 0; // stbi__g_failure_reason set by stbi__gif_header + g->out = (stbi_uc *) stbi__malloc(4 * g->w * g->h); + g->background = (stbi_uc *) stbi__malloc(4 * g->w * g->h); + g->history = (stbi_uc *) stbi__malloc(g->w * g->h); + if (g->out == 0) return stbi__errpuc("outofmem", "Out of memory"); + + // image is treated as "transparent" at the start - ie, nothing overwrites the current background; + // background colour is only used for pixels that are not rendered first frame, after that "background" + // color refers to the color that was there the previous frame. + memset( g->out, 0x00, 4 * g->w * g->h ); + memset( g->background, 0x00, 4 * g->w * g->h ); // state of the background (starts transparent) + memset( g->history, 0x00, g->w * g->h ); // pixels that were affected previous frame + first_frame = 1; + } else { + // second frame - how do we dispoase of the previous one? + dispose = (g->eflags & 0x1C) >> 2; + pcount = g->w * g->h; + + if ((dispose == 3) && (two_back == 0)) { + dispose = 2; // if I don't have an image to revert back to, default to the old background + } + + if (dispose == 3) { // use previous graphic + for (pi = 0; pi < pcount; ++pi) { + if (g->history[pi]) { + memcpy( &g->out[pi * 4], &two_back[pi * 4], 4 ); + } + } + } else if (dispose == 2) { + // restore what was changed last frame to background before that frame; + for (pi = 0; pi < pcount; ++pi) { + if (g->history[pi]) { + memcpy( &g->out[pi * 4], &g->background[pi * 4], 4 ); + } + } + } else { + // This is a non-disposal case eithe way, so just + // leave the pixels as is, and they will become the new background + // 1: do not dispose + // 0: not specified. + } + + // background is what out is after the undoing of the previou frame; + memcpy( g->background, g->out, 4 * g->w * g->h ); + } + + // clear my history; + memset( g->history, 0x00, g->w * g->h ); // pixels that were affected previous frame + + for (;;) { + int tag = stbi__get8(s); + switch (tag) { + case 0x2C: /* Image Descriptor */ + { + stbi__int32 x, y, w, h; + stbi_uc *o; + + x = stbi__get16le(s); + y = stbi__get16le(s); + w = stbi__get16le(s); + h = stbi__get16le(s); + if (((x + w) > (g->w)) || ((y + h) > (g->h))) + return stbi__errpuc("bad Image Descriptor", "Corrupt GIF"); + + g->line_size = g->w * 4; + g->start_x = x * 4; + g->start_y = y * g->line_size; + g->max_x = g->start_x + w * 4; + g->max_y = g->start_y + h * g->line_size; + g->cur_x = g->start_x; + g->cur_y = g->start_y; + + g->lflags = stbi__get8(s); + + if (g->lflags & 0x40) { + g->step = 8 * g->line_size; // first interlaced spacing + g->parse = 3; + } else { + g->step = g->line_size; + g->parse = 0; + } + + if (g->lflags & 0x80) { + stbi__gif_parse_colortable(s,g->lpal, 2 << (g->lflags & 7), g->eflags & 0x01 ? g->transparent : -1); + g->color_table = (stbi_uc *) g->lpal; + } else if (g->flags & 0x80) { + g->color_table = (stbi_uc *) g->pal; + } else + return stbi__errpuc("missing color table", "Corrupt GIF"); + + o = stbi__process_gif_raster(s, g); + if (o == NULL) return NULL; + + // if this was the first frame, + pcount = g->w * g->h; + if (first_frame && (g->bgindex > 0)) { + // if first frame, any pixel not drawn to gets the background color + for (pi = 0; pi < pcount; ++pi) { + if (g->history[pi] == 0) { + g->pal[g->bgindex][3] = 255; // just in case it was made transparent, undo that; It will be reset next frame if need be; + memcpy( &g->out[pi * 4], &g->pal[g->bgindex], 4 ); + } + } + } + + return o; + } + + case 0x21: // Comment Extension. + { + int len; + int ext = stbi__get8(s); + if (ext == 0xF9) { // Graphic Control Extension. + len = stbi__get8(s); + if (len == 4) { + g->eflags = stbi__get8(s); + g->delay = 10 * stbi__get16le(s); // delay - 1/100th of a second, saving as 1/1000ths. + + // unset old transparent + if (g->transparent >= 0) { + g->pal[g->transparent][3] = 255; + } + if (g->eflags & 0x01) { + g->transparent = stbi__get8(s); + if (g->transparent >= 0) { + g->pal[g->transparent][3] = 0; + } + } else { + // don't need transparent + stbi__skip(s, 1); + g->transparent = -1; + } + } else { + stbi__skip(s, len); + break; + } + } + while ((len = stbi__get8(s)) != 0) { + stbi__skip(s, len); + } + break; + } + + case 0x3B: // gif stream termination code + return (stbi_uc *) s; // using '1' causes warning on some compilers + + default: + return stbi__errpuc("unknown code", "Corrupt GIF"); + } + } +} + +static void *stbi__load_gif_main(stbi__context *s, int **delays, int *x, int *y, int *z, int *comp, int req_comp) +{ + if (stbi__gif_test(s)) { + int layers = 0; + stbi_uc *u = 0; + stbi_uc *out = 0; + stbi_uc *two_back = 0; + stbi__gif g; + int stride; + memset(&g, 0, sizeof(g)); + if (delays) { + *delays = 0; + } + + do { + u = stbi__gif_load_next(s, &g, comp, req_comp, two_back); + if (u == (stbi_uc *) s) u = 0; // end of animated gif marker + + if (u) { + *x = g.w; + *y = g.h; + ++layers; + stride = g.w * g.h * 4; + + if (out) { + out = (stbi_uc*) STBI_REALLOC( out, layers * stride ); + if (delays) { + *delays = (int*) STBI_REALLOC( *delays, sizeof(int) * layers ); + } + } else { + out = (stbi_uc*)stbi__malloc( layers * stride ); + if (delays) { + *delays = (int*) stbi__malloc( layers * sizeof(int) ); + } + } + memcpy( out + ((layers - 1) * stride), u, stride ); + if (layers >= 2) { + two_back = out - 2 * stride; + } + + if (delays) { + (*delays)[layers - 1U] = g.delay; + } + } + } while (u != 0); + + // free temp buffer; + STBI_FREE(g.out); + STBI_FREE(g.history); + STBI_FREE(g.background); + + // do the final conversion after loading everything; + if (req_comp && req_comp != 4) + out = stbi__convert_format(out, 4, req_comp, layers * g.w, g.h); + + *z = layers; + return out; + } else { + return stbi__errpuc("not GIF", "Image was not as a gif type."); + } +} + +static void *stbi__gif_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi_uc *u = 0; + stbi__gif g; + memset(&g, 0, sizeof(g)); + STBI_NOTUSED(ri); + + u = stbi__gif_load_next(s, &g, comp, req_comp, 0); + if (u == (stbi_uc *) s) u = 0; // end of animated gif marker + if (u) { + *x = g.w; + *y = g.h; + + // moved conversion to after successful load so that the same + // can be done for multiple frames. + if (req_comp && req_comp != 4) + u = stbi__convert_format(u, 4, req_comp, g.w, g.h); + } + + // free buffers needed for multiple frame loading; + STBI_FREE(g.history); + STBI_FREE(g.background); + + return u; +} + +static int stbi__gif_info(stbi__context *s, int *x, int *y, int *comp) +{ + return stbi__gif_info_raw(s,x,y,comp); +} +#endif + +// ************************************************************************************************* +// Radiance RGBE HDR loader +// originally by Nicolas Schulz +#ifndef STBI_NO_HDR +static int stbi__hdr_test_core(stbi__context *s, const char *signature) +{ + int i; + for (i=0; signature[i]; ++i) + if (stbi__get8(s) != signature[i]) + return 0; + stbi__rewind(s); + return 1; +} + +static int stbi__hdr_test(stbi__context* s) +{ + int r = stbi__hdr_test_core(s, "#?RADIANCE\n"); + stbi__rewind(s); + if(!r) { + r = stbi__hdr_test_core(s, "#?RGBE\n"); + stbi__rewind(s); + } + return r; +} + +#define STBI__HDR_BUFLEN 1024 +static char *stbi__hdr_gettoken(stbi__context *z, char *buffer) +{ + int len=0; + char c = '\0'; + + c = (char) stbi__get8(z); + + while (!stbi__at_eof(z) && c != '\n') { + buffer[len++] = c; + if (len == STBI__HDR_BUFLEN-1) { + // flush to end of line + while (!stbi__at_eof(z) && stbi__get8(z) != '\n') + ; + break; + } + c = (char) stbi__get8(z); + } + + buffer[len] = 0; + return buffer; +} + +static void stbi__hdr_convert(float *output, stbi_uc *input, int req_comp) +{ + if ( input[3] != 0 ) { + float f1; + // Exponent + f1 = (float) ldexp(1.0f, input[3] - (int)(128 + 8)); + if (req_comp <= 2) + output[0] = (input[0] + input[1] + input[2]) * f1 / 3; + else { + output[0] = input[0] * f1; + output[1] = input[1] * f1; + output[2] = input[2] * f1; + } + if (req_comp == 2) output[1] = 1; + if (req_comp == 4) output[3] = 1; + } else { + switch (req_comp) { + case 4: output[3] = 1; /* fallthrough */ + case 3: output[0] = output[1] = output[2] = 0; + break; + case 2: output[1] = 1; /* fallthrough */ + case 1: output[0] = 0; + break; + } + } +} + +static float *stbi__hdr_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + char buffer[STBI__HDR_BUFLEN]; + char *token; + int valid = 0; + int width, height; + stbi_uc *scanline; + float *hdr_data; + int len; + unsigned char count, value; + int i, j, k, c1,c2, z; + const char *headerToken; + STBI_NOTUSED(ri); + + // Check identifier + headerToken = stbi__hdr_gettoken(s,buffer); + if (strcmp(headerToken, "#?RADIANCE") != 0 && strcmp(headerToken, "#?RGBE") != 0) + return stbi__errpf("not HDR", "Corrupt HDR image"); + + // Parse header + for(;;) { + token = stbi__hdr_gettoken(s,buffer); + if (token[0] == 0) break; + if (strcmp(token, "FORMAT=32-bit_rle_rgbe") == 0) valid = 1; + } + + if (!valid) return stbi__errpf("unsupported format", "Unsupported HDR format"); + + // Parse width and height + // can't use sscanf() if we're not using stdio! + token = stbi__hdr_gettoken(s,buffer); + if (strncmp(token, "-Y ", 3)) return stbi__errpf("unsupported data layout", "Unsupported HDR format"); + token += 3; + height = (int) strtol(token, &token, 10); + while (*token == ' ') ++token; + if (strncmp(token, "+X ", 3)) return stbi__errpf("unsupported data layout", "Unsupported HDR format"); + token += 3; + width = (int) strtol(token, NULL, 10); + + *x = width; + *y = height; + + if (comp) *comp = 3; + if (req_comp == 0) req_comp = 3; + + if (!stbi__mad4sizes_valid(width, height, req_comp, sizeof(float), 0)) + return stbi__errpf("too large", "HDR image is too large"); + + // Read data + hdr_data = (float *) stbi__malloc_mad4(width, height, req_comp, sizeof(float), 0); + if (!hdr_data) + return stbi__errpf("outofmem", "Out of memory"); + + // Load image data + // image data is stored as some number of sca + if ( width < 8 || width >= 32768) { + // Read flat data + for (j=0; j < height; ++j) { + for (i=0; i < width; ++i) { + stbi_uc rgbe[4]; + main_decode_loop: + stbi__getn(s, rgbe, 4); + stbi__hdr_convert(hdr_data + j * width * req_comp + i * req_comp, rgbe, req_comp); + } + } + } else { + // Read RLE-encoded data + scanline = NULL; + + for (j = 0; j < height; ++j) { + c1 = stbi__get8(s); + c2 = stbi__get8(s); + len = stbi__get8(s); + if (c1 != 2 || c2 != 2 || (len & 0x80)) { + // not run-length encoded, so we have to actually use THIS data as a decoded + // pixel (note this can't be a valid pixel--one of RGB must be >= 128) + stbi_uc rgbe[4]; + rgbe[0] = (stbi_uc) c1; + rgbe[1] = (stbi_uc) c2; + rgbe[2] = (stbi_uc) len; + rgbe[3] = (stbi_uc) stbi__get8(s); + stbi__hdr_convert(hdr_data, rgbe, req_comp); + i = 1; + j = 0; + STBI_FREE(scanline); + goto main_decode_loop; // yes, this makes no sense + } + len <<= 8; + len |= stbi__get8(s); + if (len != width) { STBI_FREE(hdr_data); STBI_FREE(scanline); return stbi__errpf("invalid decoded scanline length", "corrupt HDR"); } + if (scanline == NULL) { + scanline = (stbi_uc *) stbi__malloc_mad2(width, 4, 0); + if (!scanline) { + STBI_FREE(hdr_data); + return stbi__errpf("outofmem", "Out of memory"); + } + } + + for (k = 0; k < 4; ++k) { + int nleft; + i = 0; + while ((nleft = width - i) > 0) { + count = stbi__get8(s); + if (count > 128) { + // Run + value = stbi__get8(s); + count -= 128; + if (count > nleft) { STBI_FREE(hdr_data); STBI_FREE(scanline); return stbi__errpf("corrupt", "bad RLE data in HDR"); } + for (z = 0; z < count; ++z) + scanline[i++ * 4 + k] = value; + } else { + // Dump + if (count > nleft) { STBI_FREE(hdr_data); STBI_FREE(scanline); return stbi__errpf("corrupt", "bad RLE data in HDR"); } + for (z = 0; z < count; ++z) + scanline[i++ * 4 + k] = stbi__get8(s); + } + } + } + for (i=0; i < width; ++i) + stbi__hdr_convert(hdr_data+(j*width + i)*req_comp, scanline + i*4, req_comp); + } + if (scanline) + STBI_FREE(scanline); + } + + return hdr_data; +} + +static int stbi__hdr_info(stbi__context *s, int *x, int *y, int *comp) +{ + char buffer[STBI__HDR_BUFLEN]; + char *token; + int valid = 0; + int dummy; + + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + + if (stbi__hdr_test(s) == 0) { + stbi__rewind( s ); + return 0; + } + + for(;;) { + token = stbi__hdr_gettoken(s,buffer); + if (token[0] == 0) break; + if (strcmp(token, "FORMAT=32-bit_rle_rgbe") == 0) valid = 1; + } + + if (!valid) { + stbi__rewind( s ); + return 0; + } + token = stbi__hdr_gettoken(s,buffer); + if (strncmp(token, "-Y ", 3)) { + stbi__rewind( s ); + return 0; + } + token += 3; + *y = (int) strtol(token, &token, 10); + while (*token == ' ') ++token; + if (strncmp(token, "+X ", 3)) { + stbi__rewind( s ); + return 0; + } + token += 3; + *x = (int) strtol(token, NULL, 10); + *comp = 3; + return 1; +} +#endif // STBI_NO_HDR + +#ifndef STBI_NO_BMP +static int stbi__bmp_info(stbi__context *s, int *x, int *y, int *comp) +{ + void *p; + stbi__bmp_data info; + + info.all_a = 255; + p = stbi__bmp_parse_header(s, &info); + stbi__rewind( s ); + if (p == NULL) + return 0; + if (x) *x = s->img_x; + if (y) *y = s->img_y; + if (comp) *comp = info.ma ? 4 : 3; + return 1; +} +#endif + +#ifndef STBI_NO_PSD +static int stbi__psd_info(stbi__context *s, int *x, int *y, int *comp) +{ + int channelCount, dummy, depth; + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + if (stbi__get32be(s) != 0x38425053) { + stbi__rewind( s ); + return 0; + } + if (stbi__get16be(s) != 1) { + stbi__rewind( s ); + return 0; + } + stbi__skip(s, 6); + channelCount = stbi__get16be(s); + if (channelCount < 0 || channelCount > 16) { + stbi__rewind( s ); + return 0; + } + *y = stbi__get32be(s); + *x = stbi__get32be(s); + depth = stbi__get16be(s); + if (depth != 8 && depth != 16) { + stbi__rewind( s ); + return 0; + } + if (stbi__get16be(s) != 3) { + stbi__rewind( s ); + return 0; + } + *comp = 4; + return 1; +} + +static int stbi__psd_is16(stbi__context *s) +{ + int channelCount, depth; + if (stbi__get32be(s) != 0x38425053) { + stbi__rewind( s ); + return 0; + } + if (stbi__get16be(s) != 1) { + stbi__rewind( s ); + return 0; + } + stbi__skip(s, 6); + channelCount = stbi__get16be(s); + if (channelCount < 0 || channelCount > 16) { + stbi__rewind( s ); + return 0; + } + (void) stbi__get32be(s); + (void) stbi__get32be(s); + depth = stbi__get16be(s); + if (depth != 16) { + stbi__rewind( s ); + return 0; + } + return 1; +} +#endif + +#ifndef STBI_NO_PIC +static int stbi__pic_info(stbi__context *s, int *x, int *y, int *comp) +{ + int act_comp=0,num_packets=0,chained,dummy; + stbi__pic_packet packets[10]; + + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + + if (!stbi__pic_is4(s,"\x53\x80\xF6\x34")) { + stbi__rewind(s); + return 0; + } + + stbi__skip(s, 88); + + *x = stbi__get16be(s); + *y = stbi__get16be(s); + if (stbi__at_eof(s)) { + stbi__rewind( s); + return 0; + } + if ( (*x) != 0 && (1 << 28) / (*x) < (*y)) { + stbi__rewind( s ); + return 0; + } + + stbi__skip(s, 8); + + do { + stbi__pic_packet *packet; + + if (num_packets==sizeof(packets)/sizeof(packets[0])) + return 0; + + packet = &packets[num_packets++]; + chained = stbi__get8(s); + packet->size = stbi__get8(s); + packet->type = stbi__get8(s); + packet->channel = stbi__get8(s); + act_comp |= packet->channel; + + if (stbi__at_eof(s)) { + stbi__rewind( s ); + return 0; + } + if (packet->size != 8) { + stbi__rewind( s ); + return 0; + } + } while (chained); + + *comp = (act_comp & 0x10 ? 4 : 3); + + return 1; +} +#endif + +// ************************************************************************************************* +// Portable Gray Map and Portable Pixel Map loader +// by Ken Miller +// +// PGM: http://netpbm.sourceforge.net/doc/pgm.html +// PPM: http://netpbm.sourceforge.net/doc/ppm.html +// +// Known limitations: +// Does not support comments in the header section +// Does not support ASCII image data (formats P2 and P3) +// Does not support 16-bit-per-channel + +#ifndef STBI_NO_PNM + +static int stbi__pnm_test(stbi__context *s) +{ + char p, t; + p = (char) stbi__get8(s); + t = (char) stbi__get8(s); + if (p != 'P' || (t != '5' && t != '6')) { + stbi__rewind( s ); + return 0; + } + return 1; +} + +static void *stbi__pnm_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi_uc *out; + STBI_NOTUSED(ri); + + if (!stbi__pnm_info(s, (int *)&s->img_x, (int *)&s->img_y, (int *)&s->img_n)) + return 0; + + *x = s->img_x; + *y = s->img_y; + if (comp) *comp = s->img_n; + + if (!stbi__mad3sizes_valid(s->img_n, s->img_x, s->img_y, 0)) + return stbi__errpuc("too large", "PNM too large"); + + out = (stbi_uc *) stbi__malloc_mad3(s->img_n, s->img_x, s->img_y, 0); + if (!out) return stbi__errpuc("outofmem", "Out of memory"); + stbi__getn(s, out, s->img_n * s->img_x * s->img_y); + + if (req_comp && req_comp != s->img_n) { + out = stbi__convert_format(out, s->img_n, req_comp, s->img_x, s->img_y); + if (out == NULL) return out; // stbi__convert_format frees input on failure + } + return out; +} + +static int stbi__pnm_isspace(char c) +{ + return c == ' ' || c == '\t' || c == '\n' || c == '\v' || c == '\f' || c == '\r'; +} + +static void stbi__pnm_skip_whitespace(stbi__context *s, char *c) +{ + for (;;) { + while (!stbi__at_eof(s) && stbi__pnm_isspace(*c)) + *c = (char) stbi__get8(s); + + if (stbi__at_eof(s) || *c != '#') + break; + + while (!stbi__at_eof(s) && *c != '\n' && *c != '\r' ) + *c = (char) stbi__get8(s); + } +} + +static int stbi__pnm_isdigit(char c) +{ + return c >= '0' && c <= '9'; +} + +static int stbi__pnm_getinteger(stbi__context *s, char *c) +{ + int value = 0; + + while (!stbi__at_eof(s) && stbi__pnm_isdigit(*c)) { + value = value*10 + (*c - '0'); + *c = (char) stbi__get8(s); + } + + return value; +} + +static int stbi__pnm_info(stbi__context *s, int *x, int *y, int *comp) +{ + int maxv, dummy; + char c, p, t; + + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + + stbi__rewind(s); + + // Get identifier + p = (char) stbi__get8(s); + t = (char) stbi__get8(s); + if (p != 'P' || (t != '5' && t != '6')) { + stbi__rewind(s); + return 0; + } + + *comp = (t == '6') ? 3 : 1; // '5' is 1-component .pgm; '6' is 3-component .ppm + + c = (char) stbi__get8(s); + stbi__pnm_skip_whitespace(s, &c); + + *x = stbi__pnm_getinteger(s, &c); // read width + stbi__pnm_skip_whitespace(s, &c); + + *y = stbi__pnm_getinteger(s, &c); // read height + stbi__pnm_skip_whitespace(s, &c); + + maxv = stbi__pnm_getinteger(s, &c); // read max value + + if (maxv > 255) + return stbi__err("max value > 255", "PPM image not 8-bit"); + else + return 1; +} +#endif + +static int stbi__info_main(stbi__context *s, int *x, int *y, int *comp) +{ + #ifndef STBI_NO_JPEG + if (stbi__jpeg_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_PNG + if (stbi__png_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_GIF + if (stbi__gif_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_BMP + if (stbi__bmp_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_PSD + if (stbi__psd_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_PIC + if (stbi__pic_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_PNM + if (stbi__pnm_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_HDR + if (stbi__hdr_info(s, x, y, comp)) return 1; + #endif + + // test tga last because it's a crappy test! + #ifndef STBI_NO_TGA + if (stbi__tga_info(s, x, y, comp)) + return 1; + #endif + return stbi__err("unknown image type", "Image not of any known type, or corrupt"); +} + +static int stbi__is_16_main(stbi__context *s) +{ + #ifndef STBI_NO_PNG + if (stbi__png_is16(s)) return 1; + #endif + + #ifndef STBI_NO_PSD + if (stbi__psd_is16(s)) return 1; + #endif + + return 0; +} + +#ifndef STBI_NO_STDIO +STBIDEF int stbi_info(char const *filename, int *x, int *y, int *comp) +{ + FILE *f = stbi__fopen(filename, "rb"); + int result; + if (!f) return stbi__err("can't fopen", "Unable to open file"); + result = stbi_info_from_file(f, x, y, comp); + fclose(f); + return result; +} + +STBIDEF int stbi_info_from_file(FILE *f, int *x, int *y, int *comp) +{ + int r; + stbi__context s; + long pos = ftell(f); + stbi__start_file(&s, f); + r = stbi__info_main(&s,x,y,comp); + fseek(f,pos,SEEK_SET); + return r; +} + +STBIDEF int stbi_is_16_bit(char const *filename) +{ + FILE *f = stbi__fopen(filename, "rb"); + int result; + if (!f) return stbi__err("can't fopen", "Unable to open file"); + result = stbi_is_16_bit_from_file(f); + fclose(f); + return result; +} + +STBIDEF int stbi_is_16_bit_from_file(FILE *f) +{ + int r; + stbi__context s; + long pos = ftell(f); + stbi__start_file(&s, f); + r = stbi__is_16_main(&s); + fseek(f,pos,SEEK_SET); + return r; +} +#endif // !STBI_NO_STDIO + +STBIDEF int stbi_info_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__info_main(&s,x,y,comp); +} + +STBIDEF int stbi_info_from_callbacks(stbi_io_callbacks const *c, void *user, int *x, int *y, int *comp) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) c, user); + return stbi__info_main(&s,x,y,comp); +} + +STBIDEF int stbi_is_16_bit_from_memory(stbi_uc const *buffer, int len) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__is_16_main(&s); +} + +STBIDEF int stbi_is_16_bit_from_callbacks(stbi_io_callbacks const *c, void *user) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) c, user); + return stbi__is_16_main(&s); +} + +#endif // STB_IMAGE_IMPLEMENTATION + +/* + revision history: + 2.20 (2019-02-07) support utf8 filenames in Windows; fix warnings and platform ifdefs + 2.19 (2018-02-11) fix warning + 2.18 (2018-01-30) fix warnings + 2.17 (2018-01-29) change sbti__shiftsigned to avoid clang -O2 bug + 1-bit BMP + *_is_16_bit api + avoid warnings + 2.16 (2017-07-23) all functions have 16-bit variants; + STBI_NO_STDIO works again; + compilation fixes; + fix rounding in unpremultiply; + optimize vertical flip; + disable raw_len validation; + documentation fixes + 2.15 (2017-03-18) fix png-1,2,4 bug; now all Imagenet JPGs decode; + warning fixes; disable run-time SSE detection on gcc; + uniform handling of optional "return" values; + thread-safe initialization of zlib tables + 2.14 (2017-03-03) remove deprecated STBI_JPEG_OLD; fixes for Imagenet JPGs + 2.13 (2016-11-29) add 16-bit API, only supported for PNG right now + 2.12 (2016-04-02) fix typo in 2.11 PSD fix that caused crashes + 2.11 (2016-04-02) allocate large structures on the stack + remove white matting for transparent PSD + fix reported channel count for PNG & BMP + re-enable SSE2 in non-gcc 64-bit + support RGB-formatted JPEG + read 16-bit PNGs (only as 8-bit) + 2.10 (2016-01-22) avoid warning introduced in 2.09 by STBI_REALLOC_SIZED + 2.09 (2016-01-16) allow comments in PNM files + 16-bit-per-pixel TGA (not bit-per-component) + info() for TGA could break due to .hdr handling + info() for BMP to shares code instead of sloppy parse + can use STBI_REALLOC_SIZED if allocator doesn't support realloc + code cleanup + 2.08 (2015-09-13) fix to 2.07 cleanup, reading RGB PSD as RGBA + 2.07 (2015-09-13) fix compiler warnings + partial animated GIF support + limited 16-bpc PSD support + #ifdef unused functions + bug with < 92 byte PIC,PNM,HDR,TGA + 2.06 (2015-04-19) fix bug where PSD returns wrong '*comp' value + 2.05 (2015-04-19) fix bug in progressive JPEG handling, fix warning + 2.04 (2015-04-15) try to re-enable SIMD on MinGW 64-bit + 2.03 (2015-04-12) extra corruption checking (mmozeiko) + stbi_set_flip_vertically_on_load (nguillemot) + fix NEON support; fix mingw support + 2.02 (2015-01-19) fix incorrect assert, fix warning + 2.01 (2015-01-17) fix various warnings; suppress SIMD on gcc 32-bit without -msse2 + 2.00b (2014-12-25) fix STBI_MALLOC in progressive JPEG + 2.00 (2014-12-25) optimize JPG, including x86 SSE2 & NEON SIMD (ryg) + progressive JPEG (stb) + PGM/PPM support (Ken Miller) + STBI_MALLOC,STBI_REALLOC,STBI_FREE + GIF bugfix -- seemingly never worked + STBI_NO_*, STBI_ONLY_* + 1.48 (2014-12-14) fix incorrectly-named assert() + 1.47 (2014-12-14) 1/2/4-bit PNG support, both direct and paletted (Omar Cornut & stb) + optimize PNG (ryg) + fix bug in interlaced PNG with user-specified channel count (stb) + 1.46 (2014-08-26) + fix broken tRNS chunk (colorkey-style transparency) in non-paletted PNG + 1.45 (2014-08-16) + fix MSVC-ARM internal compiler error by wrapping malloc + 1.44 (2014-08-07) + various warning fixes from Ronny Chevalier + 1.43 (2014-07-15) + fix MSVC-only compiler problem in code changed in 1.42 + 1.42 (2014-07-09) + don't define _CRT_SECURE_NO_WARNINGS (affects user code) + fixes to stbi__cleanup_jpeg path + added STBI_ASSERT to avoid requiring assert.h + 1.41 (2014-06-25) + fix search&replace from 1.36 that messed up comments/error messages + 1.40 (2014-06-22) + fix gcc struct-initialization warning + 1.39 (2014-06-15) + fix to TGA optimization when req_comp != number of components in TGA; + fix to GIF loading because BMP wasn't rewinding (whoops, no GIFs in my test suite) + add support for BMP version 5 (more ignored fields) + 1.38 (2014-06-06) + suppress MSVC warnings on integer casts truncating values + fix accidental rename of 'skip' field of I/O + 1.37 (2014-06-04) + remove duplicate typedef + 1.36 (2014-06-03) + convert to header file single-file library + if de-iphone isn't set, load iphone images color-swapped instead of returning NULL + 1.35 (2014-05-27) + various warnings + fix broken STBI_SIMD path + fix bug where stbi_load_from_file no longer left file pointer in correct place + fix broken non-easy path for 32-bit BMP (possibly never used) + TGA optimization by Arseny Kapoulkine + 1.34 (unknown) + use STBI_NOTUSED in stbi__resample_row_generic(), fix one more leak in tga failure case + 1.33 (2011-07-14) + make stbi_is_hdr work in STBI_NO_HDR (as specified), minor compiler-friendly improvements + 1.32 (2011-07-13) + support for "info" function for all supported filetypes (SpartanJ) + 1.31 (2011-06-20) + a few more leak fixes, bug in PNG handling (SpartanJ) + 1.30 (2011-06-11) + added ability to load files via callbacks to accomidate custom input streams (Ben Wenger) + removed deprecated format-specific test/load functions + removed support for installable file formats (stbi_loader) -- would have been broken for IO callbacks anyway + error cases in bmp and tga give messages and don't leak (Raymond Barbiero, grisha) + fix inefficiency in decoding 32-bit BMP (David Woo) + 1.29 (2010-08-16) + various warning fixes from Aurelien Pocheville + 1.28 (2010-08-01) + fix bug in GIF palette transparency (SpartanJ) + 1.27 (2010-08-01) + cast-to-stbi_uc to fix warnings + 1.26 (2010-07-24) + fix bug in file buffering for PNG reported by SpartanJ + 1.25 (2010-07-17) + refix trans_data warning (Won Chun) + 1.24 (2010-07-12) + perf improvements reading from files on platforms with lock-heavy fgetc() + minor perf improvements for jpeg + deprecated type-specific functions so we'll get feedback if they're needed + attempt to fix trans_data warning (Won Chun) + 1.23 fixed bug in iPhone support + 1.22 (2010-07-10) + removed image *writing* support + stbi_info support from Jetro Lauha + GIF support from Jean-Marc Lienher + iPhone PNG-extensions from James Brown + warning-fixes from Nicolas Schulz and Janez Zemva (i.stbi__err. Janez (U+017D)emva) + 1.21 fix use of 'stbi_uc' in header (reported by jon blow) + 1.20 added support for Softimage PIC, by Tom Seddon + 1.19 bug in interlaced PNG corruption check (found by ryg) + 1.18 (2008-08-02) + fix a threading bug (local mutable static) + 1.17 support interlaced PNG + 1.16 major bugfix - stbi__convert_format converted one too many pixels + 1.15 initialize some fields for thread safety + 1.14 fix threadsafe conversion bug + header-file-only version (#define STBI_HEADER_FILE_ONLY before including) + 1.13 threadsafe + 1.12 const qualifiers in the API + 1.11 Support installable IDCT, colorspace conversion routines + 1.10 Fixes for 64-bit (don't use "unsigned long") + optimized upsampling by Fabian "ryg" Giesen + 1.09 Fix format-conversion for PSD code (bad global variables!) + 1.08 Thatcher Ulrich's PSD code integrated by Nicolas Schulz + 1.07 attempt to fix C++ warning/errors again + 1.06 attempt to fix C++ warning/errors again + 1.05 fix TGA loading to return correct *comp and use good luminance calc + 1.04 default float alpha is 1, not 255; use 'void *' for stbi_image_free + 1.03 bugfixes to STBI_NO_STDIO, STBI_NO_HDR + 1.02 support for (subset of) HDR files, float interface for preferred access to them + 1.01 fix bug: possible bug in handling right-side up bmps... not sure + fix bug: the stbi__bmp_load() and stbi__tga_load() functions didn't work at all + 1.00 interface to zlib that skips zlib header + 0.99 correct handling of alpha in palette + 0.98 TGA loader by lonesock; dynamically add loaders (untested) + 0.97 jpeg errors on too large a file; also catch another malloc failure + 0.96 fix detection of invalid v value - particleman@mollyrocket forum + 0.95 during header scan, seek to markers in case of padding + 0.94 STBI_NO_STDIO to disable stdio usage; rename all #defines the same + 0.93 handle jpegtran output; verbose errors + 0.92 read 4,8,16,24,32-bit BMP files of several formats + 0.91 output 24-bit Windows 3.0 BMP files + 0.90 fix a few more warnings; bump version number to approach 1.0 + 0.61 bugfixes due to Marc LeBlanc, Christopher Lloyd + 0.60 fix compiling as c++ + 0.59 fix warnings: merge Dave Moore's -Wall fixes + 0.58 fix bug: zlib uncompressed mode len/nlen was wrong endian + 0.57 fix bug: jpg last huffman symbol before marker was >9 bits but less than 16 available + 0.56 fix bug: zlib uncompressed mode len vs. nlen + 0.55 fix bug: restart_interval not initialized to 0 + 0.54 allow NULL for 'int *comp' + 0.53 fix bug in png 3->4; speedup png decoding + 0.52 png handles req_comp=3,4 directly; minor cleanup; jpeg comments + 0.51 obey req_comp requests, 1-component jpegs return as 1-component, + on 'test' only check type, not whether we support this variant + 0.50 (2006-11-19) + first released version +*/ + + +/* +------------------------------------------------------------------------------ +This software is available under 2 licenses -- choose whichever you prefer. +------------------------------------------------------------------------------ +ALTERNATIVE A - MIT License +Copyright (c) 2017 Sean Barrett +Permission is hereby granted, free of charge, to any person obtaining a copy of +this software and associated documentation files (the "Software"), to deal in +the Software without restriction, including without limitation the rights to +use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies +of the Software, and to permit persons to whom the Software is furnished to do +so, subject to the following conditions: +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. +------------------------------------------------------------------------------ +ALTERNATIVE B - Public Domain (www.unlicense.org) +This is free and unencumbered software released into the public domain. +Anyone is free to copy, modify, publish, use, compile, sell, or distribute this +software, either in source code form or as a compiled binary, for any purpose, +commercial or non-commercial, and by any means. +In jurisdictions that recognize copyright laws, the author or authors of this +software dedicate any and all copyright interest in the software to the public +domain. We make this dedication for the benefit of the public at large and to +the detriment of our heirs and successors. We intend this dedication to be an +overt act of relinquishment in perpetuity of all present and future rights to +this software under copyright law. +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN +ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION +WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +------------------------------------------------------------------------------ +*/ \ No newline at end of file diff --git a/third-party/stb_image/stb_image_write.h b/third-party/stb_image/stb_image_write.h new file mode 100644 index 0000000000..c1e4f3479a --- /dev/null +++ b/third-party/stb_image/stb_image_write.h @@ -0,0 +1,1621 @@ +/* stb_image_write - v1.11 - public domain - http://nothings.org/stb/stb_image_write.h + writes out PNG/BMP/TGA/JPEG/HDR images to C stdio - Sean Barrett 2010-2015 + no warranty implied; use at your own risk + + Before #including, + + #define STB_IMAGE_WRITE_IMPLEMENTATION + + in the file that you want to have the implementation. + + Will probably not work correctly with strict-aliasing optimizations. + + If using a modern Microsoft Compiler, non-safe versions of CRT calls may cause + compilation warnings or even errors. To avoid this, also before #including, + + #define STBI_MSC_SECURE_CRT + +ABOUT: + + This header file is a library for writing images to C stdio or a callback. + + The PNG output is not optimal; it is 20-50% larger than the file + written by a decent optimizing implementation; though providing a custom + zlib compress function (see STBIW_ZLIB_COMPRESS) can mitigate that. + This library is designed for source code compactness and simplicity, + not optimal image file size or run-time performance. + +BUILDING: + + You can #define STBIW_ASSERT(x) before the #include to avoid using assert.h. + You can #define STBIW_MALLOC(), STBIW_REALLOC(), and STBIW_FREE() to replace + malloc,realloc,free. + You can #define STBIW_MEMMOVE() to replace memmove() + You can #define STBIW_ZLIB_COMPRESS to use a custom zlib-style compress function + for PNG compression (instead of the builtin one), it must have the following signature: + unsigned char * my_compress(unsigned char *data, int data_len, int *out_len, int quality); + The returned data will be freed with STBIW_FREE() (free() by default), + so it must be heap allocated with STBIW_MALLOC() (malloc() by default), + +UNICODE: + + If compiling for Windows and you wish to use Unicode filenames, compile + with + #define STBIW_WINDOWS_UTF8 + and pass utf8-encoded filenames. Call stbiw_convert_wchar_to_utf8 to convert + Windows wchar_t filenames to utf8. + +USAGE: + + There are five functions, one for each image file format: + + int stbi_write_png(char const *filename, int w, int h, int comp, const void *data, int stride_in_bytes); + int stbi_write_bmp(char const *filename, int w, int h, int comp, const void *data); + int stbi_write_tga(char const *filename, int w, int h, int comp, const void *data); + int stbi_write_jpg(char const *filename, int w, int h, int comp, const void *data, int quality); + int stbi_write_hdr(char const *filename, int w, int h, int comp, const float *data); + + void stbi_flip_vertically_on_write(int flag); // flag is non-zero to flip data vertically + + There are also five equivalent functions that use an arbitrary write function. You are + expected to open/close your file-equivalent before and after calling these: + + int stbi_write_png_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const void *data, int stride_in_bytes); + int stbi_write_bmp_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const void *data); + int stbi_write_tga_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const void *data); + int stbi_write_hdr_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const float *data); + int stbi_write_jpg_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data, int quality); + + where the callback is: + void stbi_write_func(void *context, void *data, int size); + + You can configure it with these global variables: + int stbi_write_tga_with_rle; // defaults to true; set to 0 to disable RLE + int stbi_write_png_compression_level; // defaults to 8; set to higher for more compression + int stbi_write_force_png_filter; // defaults to -1; set to 0..5 to force a filter mode + + + You can define STBI_WRITE_NO_STDIO to disable the file variant of these + functions, so the library will not use stdio.h at all. However, this will + also disable HDR writing, because it requires stdio for formatted output. + + Each function returns 0 on failure and non-0 on success. + + The functions create an image file defined by the parameters. The image + is a rectangle of pixels stored from left-to-right, top-to-bottom. + Each pixel contains 'comp' channels of data stored interleaved with 8-bits + per channel, in the following order: 1=Y, 2=YA, 3=RGB, 4=RGBA. (Y is + monochrome color.) The rectangle is 'w' pixels wide and 'h' pixels tall. + The *data pointer points to the first byte of the top-left-most pixel. + For PNG, "stride_in_bytes" is the distance in bytes from the first byte of + a row of pixels to the first byte of the next row of pixels. + + PNG creates output files with the same number of components as the input. + The BMP format expands Y to RGB in the file format and does not + output alpha. + + PNG supports writing rectangles of data even when the bytes storing rows of + data are not consecutive in memory (e.g. sub-rectangles of a larger image), + by supplying the stride between the beginning of adjacent rows. The other + formats do not. (Thus you cannot write a native-format BMP through the BMP + writer, both because it is in BGR order and because it may have padding + at the end of the line.) + + PNG allows you to set the deflate compression level by setting the global + variable 'stbi_write_png_compression_level' (it defaults to 8). + + HDR expects linear float data. Since the format is always 32-bit rgb(e) + data, alpha (if provided) is discarded, and for monochrome data it is + replicated across all three channels. + + TGA supports RLE or non-RLE compressed data. To use non-RLE-compressed + data, set the global variable 'stbi_write_tga_with_rle' to 0. + + JPEG does ignore alpha channels in input data; quality is between 1 and 100. + Higher quality looks better but results in a bigger image. + JPEG baseline (no JPEG progressive). + +CREDITS: + + + Sean Barrett - PNG/BMP/TGA + Baldur Karlsson - HDR + Jean-Sebastien Guay - TGA monochrome + Tim Kelsey - misc enhancements + Alan Hickman - TGA RLE + Emmanuel Julien - initial file IO callback implementation + Jon Olick - original jo_jpeg.cpp code + Daniel Gibson - integrate JPEG, allow external zlib + Aarni Koskela - allow choosing PNG filter + + bugfixes: + github:Chribba + Guillaume Chereau + github:jry2 + github:romigrou + Sergio Gonzalez + Jonas Karlsson + Filip Wasil + Thatcher Ulrich + github:poppolopoppo + Patrick Boettcher + github:xeekworx + Cap Petschulat + Simon Rodriguez + Ivan Tikhonov + github:ignotion + Adam Schackart + +LICENSE + + See end of file for license information. + +*/ + +#ifndef INCLUDE_STB_IMAGE_WRITE_H +#define INCLUDE_STB_IMAGE_WRITE_H + +#include + +// if STB_IMAGE_WRITE_STATIC causes problems, try defining STBIWDEF to 'inline' or 'static inline' +#ifndef STBIWDEF +#ifdef STB_IMAGE_WRITE_STATIC +#define STBIWDEF static +#else +#ifdef __cplusplus +#define STBIWDEF extern "C" +#else +#define STBIWDEF extern +#endif +#endif +#endif + +#ifndef STB_IMAGE_WRITE_STATIC // C++ forbids static forward declarations +extern int stbi_write_tga_with_rle; +extern int stbi_write_png_compression_level; +extern int stbi_write_force_png_filter; +#endif + +#ifndef STBI_WRITE_NO_STDIO +STBIWDEF int stbi_write_png(char const *filename, int w, int h, int comp, const void *data, int stride_in_bytes); +STBIWDEF int stbi_write_bmp(char const *filename, int w, int h, int comp, const void *data); +STBIWDEF int stbi_write_tga(char const *filename, int w, int h, int comp, const void *data); +STBIWDEF int stbi_write_hdr(char const *filename, int w, int h, int comp, const float *data); +STBIWDEF int stbi_write_jpg(char const *filename, int x, int y, int comp, const void *data, int quality); + +#ifdef STBI_WINDOWS_UTF8 +STBIWDEF int stbiw_convert_wchar_to_utf8(char *buffer, size_t bufferlen, const wchar_t* input); +#endif +#endif + +typedef void stbi_write_func(void *context, void *data, int size); + +STBIWDEF int stbi_write_png_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const void *data, int stride_in_bytes); +STBIWDEF int stbi_write_bmp_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const void *data); +STBIWDEF int stbi_write_tga_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const void *data); +STBIWDEF int stbi_write_hdr_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const float *data); +STBIWDEF int stbi_write_jpg_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data, int quality); + +STBIWDEF void stbi_flip_vertically_on_write(int flip_boolean); + +#endif//INCLUDE_STB_IMAGE_WRITE_H + +#ifdef STB_IMAGE_WRITE_IMPLEMENTATION + +#ifdef _WIN32 + #ifndef _CRT_SECURE_NO_WARNINGS + #define _CRT_SECURE_NO_WARNINGS + #endif + #ifndef _CRT_NONSTDC_NO_DEPRECATE + #define _CRT_NONSTDC_NO_DEPRECATE + #endif +#endif + +#ifndef STBI_WRITE_NO_STDIO +#include +#endif // STBI_WRITE_NO_STDIO + +#include +#include +#include +#include + +#if defined(STBIW_MALLOC) && defined(STBIW_FREE) && (defined(STBIW_REALLOC) || defined(STBIW_REALLOC_SIZED)) +// ok +#elif !defined(STBIW_MALLOC) && !defined(STBIW_FREE) && !defined(STBIW_REALLOC) && !defined(STBIW_REALLOC_SIZED) +// ok +#else +#error "Must define all or none of STBIW_MALLOC, STBIW_FREE, and STBIW_REALLOC (or STBIW_REALLOC_SIZED)." +#endif + +#ifndef STBIW_MALLOC +#define STBIW_MALLOC(sz) malloc(sz) +#define STBIW_REALLOC(p,newsz) realloc(p,newsz) +#define STBIW_FREE(p) free(p) +#endif + +#ifndef STBIW_REALLOC_SIZED +#define STBIW_REALLOC_SIZED(p,oldsz,newsz) STBIW_REALLOC(p,newsz) +#endif + + +#ifndef STBIW_MEMMOVE +#define STBIW_MEMMOVE(a,b,sz) memmove(a,b,sz) +#endif + + +#ifndef STBIW_ASSERT +#include +#define STBIW_ASSERT(x) assert(x) +#endif + +#define STBIW_UCHAR(x) (unsigned char) ((x) & 0xff) + +#ifdef STB_IMAGE_WRITE_STATIC +static int stbi__flip_vertically_on_write=0; +static int stbi_write_png_compression_level = 8; +static int stbi_write_tga_with_rle = 1; +static int stbi_write_force_png_filter = -1; +#else +int stbi_write_png_compression_level = 8; +int stbi__flip_vertically_on_write=0; +int stbi_write_tga_with_rle = 1; +int stbi_write_force_png_filter = -1; +#endif + +STBIWDEF void stbi_flip_vertically_on_write(int flag) +{ + stbi__flip_vertically_on_write = flag; +} + +typedef struct +{ + stbi_write_func *func; + void *context; +} stbi__write_context; + +// initialize a callback-based context +static void stbi__start_write_callbacks(stbi__write_context *s, stbi_write_func *c, void *context) +{ + s->func = c; + s->context = context; +} + +#ifndef STBI_WRITE_NO_STDIO + +static void stbi__stdio_write(void *context, void *data, int size) +{ + fwrite(data,1,size,(FILE*) context); +} + +#if defined(_MSC_VER) && defined(STBI_WINDOWS_UTF8) +#ifdef __cplusplus +#define STBIW_EXTERN extern "C" +#else +#define STBIW_EXTERN extern +#endif +STBIW_EXTERN __declspec(dllimport) int __stdcall MultiByteToWideChar(unsigned int cp, unsigned long flags, const char *str, int cbmb, wchar_t *widestr, int cchwide); +STBIW_EXTERN __declspec(dllimport) int __stdcall WideCharToMultiByte(unsigned int cp, unsigned long flags, const wchar_t *widestr, int cchwide, char *str, int cbmb, const char *defchar, int *used_default); + +STBIWDEF int stbiw_convert_wchar_to_utf8(char *buffer, size_t bufferlen, const wchar_t* input) +{ + return WideCharToMultiByte(65001 /* UTF8 */, 0, input, -1, buffer, bufferlen, NULL, NULL); +} +#endif + +static FILE *stbiw__fopen(char const *filename, char const *mode) +{ + FILE *f; +#if defined(_MSC_VER) && defined(STBI_WINDOWS_UTF8) + wchar_t wMode[64]; + wchar_t wFilename[1024]; + if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, filename, -1, wFilename, sizeof(wFilename))) + return 0; + + if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, mode, -1, wMode, sizeof(wMode))) + return 0; + +#if _MSC_VER >= 1400 + if (0 != _wfopen_s(&f, wFilename, wMode)) + f = 0; +#else + f = _wfopen(wFilename, wMode); +#endif + +#elif defined(_MSC_VER) && _MSC_VER >= 1400 + if (0 != fopen_s(&f, filename, mode)) + f=0; +#else + f = fopen(filename, mode); +#endif + return f; +} + +static int stbi__start_write_file(stbi__write_context *s, const char *filename) +{ + FILE *f = stbiw__fopen(filename, "wb"); + stbi__start_write_callbacks(s, stbi__stdio_write, (void *) f); + return f != NULL; +} + +static void stbi__end_write_file(stbi__write_context *s) +{ + fclose((FILE *)s->context); +} + +#endif // !STBI_WRITE_NO_STDIO + +typedef unsigned int stbiw_uint32; +typedef int stb_image_write_test[sizeof(stbiw_uint32)==4 ? 1 : -1]; + +static void stbiw__writefv(stbi__write_context *s, const char *fmt, va_list v) +{ + while (*fmt) { + switch (*fmt++) { + case ' ': break; + case '1': { unsigned char x = STBIW_UCHAR(va_arg(v, int)); + s->func(s->context,&x,1); + break; } + case '2': { int x = va_arg(v,int); + unsigned char b[2]; + b[0] = STBIW_UCHAR(x); + b[1] = STBIW_UCHAR(x>>8); + s->func(s->context,b,2); + break; } + case '4': { stbiw_uint32 x = va_arg(v,int); + unsigned char b[4]; + b[0]=STBIW_UCHAR(x); + b[1]=STBIW_UCHAR(x>>8); + b[2]=STBIW_UCHAR(x>>16); + b[3]=STBIW_UCHAR(x>>24); + s->func(s->context,b,4); + break; } + default: + STBIW_ASSERT(0); + return; + } + } +} + +static void stbiw__writef(stbi__write_context *s, const char *fmt, ...) +{ + va_list v; + va_start(v, fmt); + stbiw__writefv(s, fmt, v); + va_end(v); +} + +static void stbiw__putc(stbi__write_context *s, unsigned char c) +{ + s->func(s->context, &c, 1); +} + +static void stbiw__write3(stbi__write_context *s, unsigned char a, unsigned char b, unsigned char c) +{ + unsigned char arr[3]; + arr[0] = a, arr[1] = b, arr[2] = c; + s->func(s->context, arr, 3); +} + +static void stbiw__write_pixel(stbi__write_context *s, int rgb_dir, int comp, int write_alpha, int expand_mono, unsigned char *d) +{ + unsigned char bg[3] = { 255, 0, 255}, px[3]; + int k; + + if (write_alpha < 0) + s->func(s->context, &d[comp - 1], 1); + + switch (comp) { + case 2: // 2 pixels = mono + alpha, alpha is written separately, so same as 1-channel case + case 1: + if (expand_mono) + stbiw__write3(s, d[0], d[0], d[0]); // monochrome bmp + else + s->func(s->context, d, 1); // monochrome TGA + break; + case 4: + if (!write_alpha) { + // composite against pink background + for (k = 0; k < 3; ++k) + px[k] = bg[k] + ((d[k] - bg[k]) * d[3]) / 255; + stbiw__write3(s, px[1 - rgb_dir], px[1], px[1 + rgb_dir]); + break; + } + /* FALLTHROUGH */ + case 3: + stbiw__write3(s, d[1 - rgb_dir], d[1], d[1 + rgb_dir]); + break; + } + if (write_alpha > 0) + s->func(s->context, &d[comp - 1], 1); +} + +static void stbiw__write_pixels(stbi__write_context *s, int rgb_dir, int vdir, int x, int y, int comp, void *data, int write_alpha, int scanline_pad, int expand_mono) +{ + stbiw_uint32 zero = 0; + int i,j, j_end; + + if (y <= 0) + return; + + if (stbi__flip_vertically_on_write) + vdir *= -1; + + if (vdir < 0) + j_end = -1, j = y-1; + else + j_end = y, j = 0; + + for (; j != j_end; j += vdir) { + for (i=0; i < x; ++i) { + unsigned char *d = (unsigned char *) data + (j*x+i)*comp; + stbiw__write_pixel(s, rgb_dir, comp, write_alpha, expand_mono, d); + } + s->func(s->context, &zero, scanline_pad); + } +} + +static int stbiw__outfile(stbi__write_context *s, int rgb_dir, int vdir, int x, int y, int comp, int expand_mono, void *data, int alpha, int pad, const char *fmt, ...) +{ + if (y < 0 || x < 0) { + return 0; + } else { + va_list v; + va_start(v, fmt); + stbiw__writefv(s, fmt, v); + va_end(v); + stbiw__write_pixels(s,rgb_dir,vdir,x,y,comp,data,alpha,pad, expand_mono); + return 1; + } +} + +static int stbi_write_bmp_core(stbi__write_context *s, int x, int y, int comp, const void *data) +{ + int pad = (-x*3) & 3; + return stbiw__outfile(s,-1,-1,x,y,comp,1,(void *) data,0,pad, + "11 4 22 4" "4 44 22 444444", + 'B', 'M', 14+40+(x*3+pad)*y, 0,0, 14+40, // file header + 40, x,y, 1,24, 0,0,0,0,0,0); // bitmap header +} + +STBIWDEF int stbi_write_bmp_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data) +{ + stbi__write_context s; + stbi__start_write_callbacks(&s, func, context); + return stbi_write_bmp_core(&s, x, y, comp, data); +} + +#ifndef STBI_WRITE_NO_STDIO +STBIWDEF int stbi_write_bmp(char const *filename, int x, int y, int comp, const void *data) +{ + stbi__write_context s; + if (stbi__start_write_file(&s,filename)) { + int r = stbi_write_bmp_core(&s, x, y, comp, data); + stbi__end_write_file(&s); + return r; + } else + return 0; +} +#endif //!STBI_WRITE_NO_STDIO + +static int stbi_write_tga_core(stbi__write_context *s, int x, int y, int comp, void *data) +{ + int has_alpha = (comp == 2 || comp == 4); + int colorbytes = has_alpha ? comp-1 : comp; + int format = colorbytes < 2 ? 3 : 2; // 3 color channels (RGB/RGBA) = 2, 1 color channel (Y/YA) = 3 + + if (y < 0 || x < 0) + return 0; + + if (!stbi_write_tga_with_rle) { + return stbiw__outfile(s, -1, -1, x, y, comp, 0, (void *) data, has_alpha, 0, + "111 221 2222 11", 0, 0, format, 0, 0, 0, 0, 0, x, y, (colorbytes + has_alpha) * 8, has_alpha * 8); + } else { + int i,j,k; + int jend, jdir; + + stbiw__writef(s, "111 221 2222 11", 0,0,format+8, 0,0,0, 0,0,x,y, (colorbytes + has_alpha) * 8, has_alpha * 8); + + if (stbi__flip_vertically_on_write) { + j = 0; + jend = y; + jdir = 1; + } else { + j = y-1; + jend = -1; + jdir = -1; + } + for (; j != jend; j += jdir) { + unsigned char *row = (unsigned char *) data + j * x * comp; + int len; + + for (i = 0; i < x; i += len) { + unsigned char *begin = row + i * comp; + int diff = 1; + len = 1; + + if (i < x - 1) { + ++len; + diff = memcmp(begin, row + (i + 1) * comp, comp); + if (diff) { + const unsigned char *prev = begin; + for (k = i + 2; k < x && len < 128; ++k) { + if (memcmp(prev, row + k * comp, comp)) { + prev += comp; + ++len; + } else { + --len; + break; + } + } + } else { + for (k = i + 2; k < x && len < 128; ++k) { + if (!memcmp(begin, row + k * comp, comp)) { + ++len; + } else { + break; + } + } + } + } + + if (diff) { + unsigned char header = STBIW_UCHAR(len - 1); + s->func(s->context, &header, 1); + for (k = 0; k < len; ++k) { + stbiw__write_pixel(s, -1, comp, has_alpha, 0, begin + k * comp); + } + } else { + unsigned char header = STBIW_UCHAR(len - 129); + s->func(s->context, &header, 1); + stbiw__write_pixel(s, -1, comp, has_alpha, 0, begin); + } + } + } + } + return 1; +} + +STBIWDEF int stbi_write_tga_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data) +{ + stbi__write_context s; + stbi__start_write_callbacks(&s, func, context); + return stbi_write_tga_core(&s, x, y, comp, (void *) data); +} + +#ifndef STBI_WRITE_NO_STDIO +STBIWDEF int stbi_write_tga(char const *filename, int x, int y, int comp, const void *data) +{ + stbi__write_context s; + if (stbi__start_write_file(&s,filename)) { + int r = stbi_write_tga_core(&s, x, y, comp, (void *) data); + stbi__end_write_file(&s); + return r; + } else + return 0; +} +#endif + +// ************************************************************************************************* +// Radiance RGBE HDR writer +// by Baldur Karlsson + +#define stbiw__max(a, b) ((a) > (b) ? (a) : (b)) + +static void stbiw__linear_to_rgbe(unsigned char *rgbe, float *linear) +{ + int exponent; + float maxcomp = stbiw__max(linear[0], stbiw__max(linear[1], linear[2])); + + if (maxcomp < 1e-32f) { + rgbe[0] = rgbe[1] = rgbe[2] = rgbe[3] = 0; + } else { + float normalize = (float) frexp(maxcomp, &exponent) * 256.0f/maxcomp; + + rgbe[0] = (unsigned char)(linear[0] * normalize); + rgbe[1] = (unsigned char)(linear[1] * normalize); + rgbe[2] = (unsigned char)(linear[2] * normalize); + rgbe[3] = (unsigned char)(exponent + 128); + } +} + +static void stbiw__write_run_data(stbi__write_context *s, int length, unsigned char databyte) +{ + unsigned char lengthbyte = STBIW_UCHAR(length+128); + STBIW_ASSERT(length+128 <= 255); + s->func(s->context, &lengthbyte, 1); + s->func(s->context, &databyte, 1); +} + +static void stbiw__write_dump_data(stbi__write_context *s, int length, unsigned char *data) +{ + unsigned char lengthbyte = STBIW_UCHAR(length); + STBIW_ASSERT(length <= 128); // inconsistent with spec but consistent with official code + s->func(s->context, &lengthbyte, 1); + s->func(s->context, data, length); +} + +static void stbiw__write_hdr_scanline(stbi__write_context *s, int width, int ncomp, unsigned char *scratch, float *scanline) +{ + unsigned char scanlineheader[4] = { 2, 2, 0, 0 }; + unsigned char rgbe[4]; + float linear[3]; + int x; + + scanlineheader[2] = (width&0xff00)>>8; + scanlineheader[3] = (width&0x00ff); + + /* skip RLE for images too small or large */ + if (width < 8 || width >= 32768) { + for (x=0; x < width; x++) { + switch (ncomp) { + case 4: /* fallthrough */ + case 3: linear[2] = scanline[x*ncomp + 2]; + linear[1] = scanline[x*ncomp + 1]; + linear[0] = scanline[x*ncomp + 0]; + break; + default: + linear[0] = linear[1] = linear[2] = scanline[x*ncomp + 0]; + break; + } + stbiw__linear_to_rgbe(rgbe, linear); + s->func(s->context, rgbe, 4); + } + } else { + int c,r; + /* encode into scratch buffer */ + for (x=0; x < width; x++) { + switch(ncomp) { + case 4: /* fallthrough */ + case 3: linear[2] = scanline[x*ncomp + 2]; + linear[1] = scanline[x*ncomp + 1]; + linear[0] = scanline[x*ncomp + 0]; + break; + default: + linear[0] = linear[1] = linear[2] = scanline[x*ncomp + 0]; + break; + } + stbiw__linear_to_rgbe(rgbe, linear); + scratch[x + width*0] = rgbe[0]; + scratch[x + width*1] = rgbe[1]; + scratch[x + width*2] = rgbe[2]; + scratch[x + width*3] = rgbe[3]; + } + + s->func(s->context, scanlineheader, 4); + + /* RLE each component separately */ + for (c=0; c < 4; c++) { + unsigned char *comp = &scratch[width*c]; + + x = 0; + while (x < width) { + // find first run + r = x; + while (r+2 < width) { + if (comp[r] == comp[r+1] && comp[r] == comp[r+2]) + break; + ++r; + } + if (r+2 >= width) + r = width; + // dump up to first run + while (x < r) { + int len = r-x; + if (len > 128) len = 128; + stbiw__write_dump_data(s, len, &comp[x]); + x += len; + } + // if there's a run, output it + if (r+2 < width) { // same test as what we break out of in search loop, so only true if we break'd + // find next byte after run + while (r < width && comp[r] == comp[x]) + ++r; + // output run up to r + while (x < r) { + int len = r-x; + if (len > 127) len = 127; + stbiw__write_run_data(s, len, comp[x]); + x += len; + } + } + } + } + } +} + +static int stbi_write_hdr_core(stbi__write_context *s, int x, int y, int comp, float *data) +{ + if (y <= 0 || x <= 0 || data == NULL) + return 0; + else { + // Each component is stored separately. Allocate scratch space for full output scanline. + unsigned char *scratch = (unsigned char *) STBIW_MALLOC(x*4); + int i, len; + char buffer[128]; + char header[] = "#?RADIANCE\n# Written by stb_image_write.h\nFORMAT=32-bit_rle_rgbe\n"; + s->func(s->context, header, sizeof(header)-1); + +#ifdef STBI_MSC_SECURE_CRT + len = sprintf_s(buffer, "EXPOSURE= 1.0000000000000\n\n-Y %d +X %d\n", y, x); +#else + len = sprintf(buffer, "EXPOSURE= 1.0000000000000\n\n-Y %d +X %d\n", y, x); +#endif + s->func(s->context, buffer, len); + + for(i=0; i < y; i++) + stbiw__write_hdr_scanline(s, x, comp, scratch, data + comp*x*(stbi__flip_vertically_on_write ? y-1-i : i)); + STBIW_FREE(scratch); + return 1; + } +} + +STBIWDEF int stbi_write_hdr_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const float *data) +{ + stbi__write_context s; + stbi__start_write_callbacks(&s, func, context); + return stbi_write_hdr_core(&s, x, y, comp, (float *) data); +} + +#ifndef STBI_WRITE_NO_STDIO +STBIWDEF int stbi_write_hdr(char const *filename, int x, int y, int comp, const float *data) +{ + stbi__write_context s; + if (stbi__start_write_file(&s,filename)) { + int r = stbi_write_hdr_core(&s, x, y, comp, (float *) data); + stbi__end_write_file(&s); + return r; + } else + return 0; +} +#endif // STBI_WRITE_NO_STDIO + + +////////////////////////////////////////////////////////////////////////////// +// +// PNG writer +// + +#ifndef STBIW_ZLIB_COMPRESS +// stretchy buffer; stbiw__sbpush() == vector<>::push_back() -- stbiw__sbcount() == vector<>::size() +#define stbiw__sbraw(a) ((int *) (a) - 2) +#define stbiw__sbm(a) stbiw__sbraw(a)[0] +#define stbiw__sbn(a) stbiw__sbraw(a)[1] + +#define stbiw__sbneedgrow(a,n) ((a)==0 || stbiw__sbn(a)+n >= stbiw__sbm(a)) +#define stbiw__sbmaybegrow(a,n) (stbiw__sbneedgrow(a,(n)) ? stbiw__sbgrow(a,n) : 0) +#define stbiw__sbgrow(a,n) stbiw__sbgrowf((void **) &(a), (n), sizeof(*(a))) + +#define stbiw__sbpush(a, v) (stbiw__sbmaybegrow(a,1), (a)[stbiw__sbn(a)++] = (v)) +#define stbiw__sbcount(a) ((a) ? stbiw__sbn(a) : 0) +#define stbiw__sbfree(a) ((a) ? STBIW_FREE(stbiw__sbraw(a)),0 : 0) + +static void *stbiw__sbgrowf(void **arr, int increment, int itemsize) +{ + int m = *arr ? 2*stbiw__sbm(*arr)+increment : increment+1; + void *p = STBIW_REALLOC_SIZED(*arr ? stbiw__sbraw(*arr) : 0, *arr ? (stbiw__sbm(*arr)*itemsize + sizeof(int)*2) : 0, itemsize * m + sizeof(int)*2); + STBIW_ASSERT(p); + if (p) { + if (!*arr) ((int *) p)[1] = 0; + *arr = (void *) ((int *) p + 2); + stbiw__sbm(*arr) = m; + } + return *arr; +} + +static unsigned char *stbiw__zlib_flushf(unsigned char *data, unsigned int *bitbuffer, int *bitcount) +{ + while (*bitcount >= 8) { + stbiw__sbpush(data, STBIW_UCHAR(*bitbuffer)); + *bitbuffer >>= 8; + *bitcount -= 8; + } + return data; +} + +static int stbiw__zlib_bitrev(int code, int codebits) +{ + int res=0; + while (codebits--) { + res = (res << 1) | (code & 1); + code >>= 1; + } + return res; +} + +static unsigned int stbiw__zlib_countm(unsigned char *a, unsigned char *b, int limit) +{ + int i; + for (i=0; i < limit && i < 258; ++i) + if (a[i] != b[i]) break; + return i; +} + +static unsigned int stbiw__zhash(unsigned char *data) +{ + stbiw_uint32 hash = data[0] + (data[1] << 8) + (data[2] << 16); + hash ^= hash << 3; + hash += hash >> 5; + hash ^= hash << 4; + hash += hash >> 17; + hash ^= hash << 25; + hash += hash >> 6; + return hash; +} + +#define stbiw__zlib_flush() (out = stbiw__zlib_flushf(out, &bitbuf, &bitcount)) +#define stbiw__zlib_add(code,codebits) \ + (bitbuf |= (code) << bitcount, bitcount += (codebits), stbiw__zlib_flush()) +#define stbiw__zlib_huffa(b,c) stbiw__zlib_add(stbiw__zlib_bitrev(b,c),c) +// default huffman tables +#define stbiw__zlib_huff1(n) stbiw__zlib_huffa(0x30 + (n), 8) +#define stbiw__zlib_huff2(n) stbiw__zlib_huffa(0x190 + (n)-144, 9) +#define stbiw__zlib_huff3(n) stbiw__zlib_huffa(0 + (n)-256,7) +#define stbiw__zlib_huff4(n) stbiw__zlib_huffa(0xc0 + (n)-280,8) +#define stbiw__zlib_huff(n) ((n) <= 143 ? stbiw__zlib_huff1(n) : (n) <= 255 ? stbiw__zlib_huff2(n) : (n) <= 279 ? stbiw__zlib_huff3(n) : stbiw__zlib_huff4(n)) +#define stbiw__zlib_huffb(n) ((n) <= 143 ? stbiw__zlib_huff1(n) : stbiw__zlib_huff2(n)) + +#define stbiw__ZHASH 16384 + +#endif // STBIW_ZLIB_COMPRESS + +STBIWDEF unsigned char * stbi_zlib_compress(unsigned char *data, int data_len, int *out_len, int quality) +{ +#ifdef STBIW_ZLIB_COMPRESS + // user provided a zlib compress implementation, use that + return STBIW_ZLIB_COMPRESS(data, data_len, out_len, quality); +#else // use builtin + static unsigned short lengthc[] = { 3,4,5,6,7,8,9,10,11,13,15,17,19,23,27,31,35,43,51,59,67,83,99,115,131,163,195,227,258, 259 }; + static unsigned char lengtheb[]= { 0,0,0,0,0,0,0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0 }; + static unsigned short distc[] = { 1,2,3,4,5,7,9,13,17,25,33,49,65,97,129,193,257,385,513,769,1025,1537,2049,3073,4097,6145,8193,12289,16385,24577, 32768 }; + static unsigned char disteb[] = { 0,0,0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,11,11,12,12,13,13 }; + unsigned int bitbuf=0; + int i,j, bitcount=0; + unsigned char *out = NULL; + unsigned char ***hash_table = (unsigned char***) STBIW_MALLOC(stbiw__ZHASH * sizeof(char**)); + if (hash_table == NULL) + return NULL; + if (quality < 5) quality = 5; + + stbiw__sbpush(out, 0x78); // DEFLATE 32K window + stbiw__sbpush(out, 0x5e); // FLEVEL = 1 + stbiw__zlib_add(1,1); // BFINAL = 1 + stbiw__zlib_add(1,2); // BTYPE = 1 -- fixed huffman + + for (i=0; i < stbiw__ZHASH; ++i) + hash_table[i] = NULL; + + i=0; + while (i < data_len-3) { + // hash next 3 bytes of data to be compressed + int h = stbiw__zhash(data+i)&(stbiw__ZHASH-1), best=3; + unsigned char *bestloc = 0; + unsigned char **hlist = hash_table[h]; + int n = stbiw__sbcount(hlist); + for (j=0; j < n; ++j) { + if (hlist[j]-data > i-32768) { // if entry lies within window + int d = stbiw__zlib_countm(hlist[j], data+i, data_len-i); + if (d >= best) best=d,bestloc=hlist[j]; + } + } + // when hash table entry is too long, delete half the entries + if (hash_table[h] && stbiw__sbn(hash_table[h]) == 2*quality) { + STBIW_MEMMOVE(hash_table[h], hash_table[h]+quality, sizeof(hash_table[h][0])*quality); + stbiw__sbn(hash_table[h]) = quality; + } + stbiw__sbpush(hash_table[h],data+i); + + if (bestloc) { + // "lazy matching" - check match at *next* byte, and if it's better, do cur byte as literal + h = stbiw__zhash(data+i+1)&(stbiw__ZHASH-1); + hlist = hash_table[h]; + n = stbiw__sbcount(hlist); + for (j=0; j < n; ++j) { + if (hlist[j]-data > i-32767) { + int e = stbiw__zlib_countm(hlist[j], data+i+1, data_len-i-1); + if (e > best) { // if next match is better, bail on current match + bestloc = NULL; + break; + } + } + } + } + + if (bestloc) { + int d = (int) (data+i - bestloc); // distance back + STBIW_ASSERT(d <= 32767 && best <= 258); + for (j=0; best > lengthc[j+1]-1; ++j); + stbiw__zlib_huff(j+257); + if (lengtheb[j]) stbiw__zlib_add(best - lengthc[j], lengtheb[j]); + for (j=0; d > distc[j+1]-1; ++j); + stbiw__zlib_add(stbiw__zlib_bitrev(j,5),5); + if (disteb[j]) stbiw__zlib_add(d - distc[j], disteb[j]); + i += best; + } else { + stbiw__zlib_huffb(data[i]); + ++i; + } + } + // write out final bytes + for (;i < data_len; ++i) + stbiw__zlib_huffb(data[i]); + stbiw__zlib_huff(256); // end of block + // pad with 0 bits to byte boundary + while (bitcount) + stbiw__zlib_add(0,1); + + for (i=0; i < stbiw__ZHASH; ++i) + (void) stbiw__sbfree(hash_table[i]); + STBIW_FREE(hash_table); + + { + // compute adler32 on input + unsigned int s1=1, s2=0; + int blocklen = (int) (data_len % 5552); + j=0; + while (j < data_len) { + for (i=0; i < blocklen; ++i) s1 += data[j+i], s2 += s1; + s1 %= 65521, s2 %= 65521; + j += blocklen; + blocklen = 5552; + } + stbiw__sbpush(out, STBIW_UCHAR(s2 >> 8)); + stbiw__sbpush(out, STBIW_UCHAR(s2)); + stbiw__sbpush(out, STBIW_UCHAR(s1 >> 8)); + stbiw__sbpush(out, STBIW_UCHAR(s1)); + } + *out_len = stbiw__sbn(out); + // make returned pointer freeable + STBIW_MEMMOVE(stbiw__sbraw(out), out, *out_len); + return (unsigned char *) stbiw__sbraw(out); +#endif // STBIW_ZLIB_COMPRESS +} + +static unsigned int stbiw__crc32(unsigned char *buffer, int len) +{ +#ifdef STBIW_CRC32 + return STBIW_CRC32(buffer, len); +#else + static unsigned int crc_table[256] = + { + 0x00000000, 0x77073096, 0xEE0E612C, 0x990951BA, 0x076DC419, 0x706AF48F, 0xE963A535, 0x9E6495A3, + 0x0eDB8832, 0x79DCB8A4, 0xE0D5E91E, 0x97D2D988, 0x09B64C2B, 0x7EB17CBD, 0xE7B82D07, 0x90BF1D91, + 0x1DB71064, 0x6AB020F2, 0xF3B97148, 0x84BE41DE, 0x1ADAD47D, 0x6DDDE4EB, 0xF4D4B551, 0x83D385C7, + 0x136C9856, 0x646BA8C0, 0xFD62F97A, 0x8A65C9EC, 0x14015C4F, 0x63066CD9, 0xFA0F3D63, 0x8D080DF5, + 0x3B6E20C8, 0x4C69105E, 0xD56041E4, 0xA2677172, 0x3C03E4D1, 0x4B04D447, 0xD20D85FD, 0xA50AB56B, + 0x35B5A8FA, 0x42B2986C, 0xDBBBC9D6, 0xACBCF940, 0x32D86CE3, 0x45DF5C75, 0xDCD60DCF, 0xABD13D59, + 0x26D930AC, 0x51DE003A, 0xC8D75180, 0xBFD06116, 0x21B4F4B5, 0x56B3C423, 0xCFBA9599, 0xB8BDA50F, + 0x2802B89E, 0x5F058808, 0xC60CD9B2, 0xB10BE924, 0x2F6F7C87, 0x58684C11, 0xC1611DAB, 0xB6662D3D, + 0x76DC4190, 0x01DB7106, 0x98D220BC, 0xEFD5102A, 0x71B18589, 0x06B6B51F, 0x9FBFE4A5, 0xE8B8D433, + 0x7807C9A2, 0x0F00F934, 0x9609A88E, 0xE10E9818, 0x7F6A0DBB, 0x086D3D2D, 0x91646C97, 0xE6635C01, + 0x6B6B51F4, 0x1C6C6162, 0x856530D8, 0xF262004E, 0x6C0695ED, 0x1B01A57B, 0x8208F4C1, 0xF50FC457, + 0x65B0D9C6, 0x12B7E950, 0x8BBEB8EA, 0xFCB9887C, 0x62DD1DDF, 0x15DA2D49, 0x8CD37CF3, 0xFBD44C65, + 0x4DB26158, 0x3AB551CE, 0xA3BC0074, 0xD4BB30E2, 0x4ADFA541, 0x3DD895D7, 0xA4D1C46D, 0xD3D6F4FB, + 0x4369E96A, 0x346ED9FC, 0xAD678846, 0xDA60B8D0, 0x44042D73, 0x33031DE5, 0xAA0A4C5F, 0xDD0D7CC9, + 0x5005713C, 0x270241AA, 0xBE0B1010, 0xC90C2086, 0x5768B525, 0x206F85B3, 0xB966D409, 0xCE61E49F, + 0x5EDEF90E, 0x29D9C998, 0xB0D09822, 0xC7D7A8B4, 0x59B33D17, 0x2EB40D81, 0xB7BD5C3B, 0xC0BA6CAD, + 0xEDB88320, 0x9ABFB3B6, 0x03B6E20C, 0x74B1D29A, 0xEAD54739, 0x9DD277AF, 0x04DB2615, 0x73DC1683, + 0xE3630B12, 0x94643B84, 0x0D6D6A3E, 0x7A6A5AA8, 0xE40ECF0B, 0x9309FF9D, 0x0A00AE27, 0x7D079EB1, + 0xF00F9344, 0x8708A3D2, 0x1E01F268, 0x6906C2FE, 0xF762575D, 0x806567CB, 0x196C3671, 0x6E6B06E7, + 0xFED41B76, 0x89D32BE0, 0x10DA7A5A, 0x67DD4ACC, 0xF9B9DF6F, 0x8EBEEFF9, 0x17B7BE43, 0x60B08ED5, + 0xD6D6A3E8, 0xA1D1937E, 0x38D8C2C4, 0x4FDFF252, 0xD1BB67F1, 0xA6BC5767, 0x3FB506DD, 0x48B2364B, + 0xD80D2BDA, 0xAF0A1B4C, 0x36034AF6, 0x41047A60, 0xDF60EFC3, 0xA867DF55, 0x316E8EEF, 0x4669BE79, + 0xCB61B38C, 0xBC66831A, 0x256FD2A0, 0x5268E236, 0xCC0C7795, 0xBB0B4703, 0x220216B9, 0x5505262F, + 0xC5BA3BBE, 0xB2BD0B28, 0x2BB45A92, 0x5CB36A04, 0xC2D7FFA7, 0xB5D0CF31, 0x2CD99E8B, 0x5BDEAE1D, + 0x9B64C2B0, 0xEC63F226, 0x756AA39C, 0x026D930A, 0x9C0906A9, 0xEB0E363F, 0x72076785, 0x05005713, + 0x95BF4A82, 0xE2B87A14, 0x7BB12BAE, 0x0CB61B38, 0x92D28E9B, 0xE5D5BE0D, 0x7CDCEFB7, 0x0BDBDF21, + 0x86D3D2D4, 0xF1D4E242, 0x68DDB3F8, 0x1FDA836E, 0x81BE16CD, 0xF6B9265B, 0x6FB077E1, 0x18B74777, + 0x88085AE6, 0xFF0F6A70, 0x66063BCA, 0x11010B5C, 0x8F659EFF, 0xF862AE69, 0x616BFFD3, 0x166CCF45, + 0xA00AE278, 0xD70DD2EE, 0x4E048354, 0x3903B3C2, 0xA7672661, 0xD06016F7, 0x4969474D, 0x3E6E77DB, + 0xAED16A4A, 0xD9D65ADC, 0x40DF0B66, 0x37D83BF0, 0xA9BCAE53, 0xDEBB9EC5, 0x47B2CF7F, 0x30B5FFE9, + 0xBDBDF21C, 0xCABAC28A, 0x53B39330, 0x24B4A3A6, 0xBAD03605, 0xCDD70693, 0x54DE5729, 0x23D967BF, + 0xB3667A2E, 0xC4614AB8, 0x5D681B02, 0x2A6F2B94, 0xB40BBE37, 0xC30C8EA1, 0x5A05DF1B, 0x2D02EF8D + }; + + unsigned int crc = ~0u; + int i; + for (i=0; i < len; ++i) + crc = (crc >> 8) ^ crc_table[buffer[i] ^ (crc & 0xff)]; + return ~crc; +#endif +} + +#define stbiw__wpng4(o,a,b,c,d) ((o)[0]=STBIW_UCHAR(a),(o)[1]=STBIW_UCHAR(b),(o)[2]=STBIW_UCHAR(c),(o)[3]=STBIW_UCHAR(d),(o)+=4) +#define stbiw__wp32(data,v) stbiw__wpng4(data, (v)>>24,(v)>>16,(v)>>8,(v)); +#define stbiw__wptag(data,s) stbiw__wpng4(data, s[0],s[1],s[2],s[3]) + +static void stbiw__wpcrc(unsigned char **data, int len) +{ + unsigned int crc = stbiw__crc32(*data - len - 4, len+4); + stbiw__wp32(*data, crc); +} + +static unsigned char stbiw__paeth(int a, int b, int c) +{ + int p = a + b - c, pa = abs(p-a), pb = abs(p-b), pc = abs(p-c); + if (pa <= pb && pa <= pc) return STBIW_UCHAR(a); + if (pb <= pc) return STBIW_UCHAR(b); + return STBIW_UCHAR(c); +} + +// @OPTIMIZE: provide an option that always forces left-predict or paeth predict +static void stbiw__encode_png_line(unsigned char *pixels, int stride_bytes, int width, int height, int y, int n, int filter_type, signed char *line_buffer) +{ + static int mapping[] = { 0,1,2,3,4 }; + static int firstmap[] = { 0,1,0,5,6 }; + int *mymap = (y != 0) ? mapping : firstmap; + int i; + int type = mymap[filter_type]; + unsigned char *z = pixels + stride_bytes * (stbi__flip_vertically_on_write ? height-1-y : y); + int signed_stride = stbi__flip_vertically_on_write ? -stride_bytes : stride_bytes; + + if (type==0) { + memcpy(line_buffer, z, width*n); + return; + } + + // first loop isn't optimized since it's just one pixel + for (i = 0; i < n; ++i) { + switch (type) { + case 1: line_buffer[i] = z[i]; break; + case 2: line_buffer[i] = z[i] - z[i-signed_stride]; break; + case 3: line_buffer[i] = z[i] - (z[i-signed_stride]>>1); break; + case 4: line_buffer[i] = (signed char) (z[i] - stbiw__paeth(0,z[i-signed_stride],0)); break; + case 5: line_buffer[i] = z[i]; break; + case 6: line_buffer[i] = z[i]; break; + } + } + switch (type) { + case 1: for (i=n; i < width*n; ++i) line_buffer[i] = z[i] - z[i-n]; break; + case 2: for (i=n; i < width*n; ++i) line_buffer[i] = z[i] - z[i-signed_stride]; break; + case 3: for (i=n; i < width*n; ++i) line_buffer[i] = z[i] - ((z[i-n] + z[i-signed_stride])>>1); break; + case 4: for (i=n; i < width*n; ++i) line_buffer[i] = z[i] - stbiw__paeth(z[i-n], z[i-signed_stride], z[i-signed_stride-n]); break; + case 5: for (i=n; i < width*n; ++i) line_buffer[i] = z[i] - (z[i-n]>>1); break; + case 6: for (i=n; i < width*n; ++i) line_buffer[i] = z[i] - stbiw__paeth(z[i-n], 0,0); break; + } +} + +STBIWDEF unsigned char *stbi_write_png_to_mem(const unsigned char *pixels, int stride_bytes, int x, int y, int n, int *out_len) +{ + int force_filter = stbi_write_force_png_filter; + int ctype[5] = { -1, 0, 4, 2, 6 }; + unsigned char sig[8] = { 137,80,78,71,13,10,26,10 }; + unsigned char *out,*o, *filt, *zlib; + signed char *line_buffer; + int j,zlen; + + if (stride_bytes == 0) + stride_bytes = x * n; + + if (force_filter >= 5) { + force_filter = -1; + } + + filt = (unsigned char *) STBIW_MALLOC((x*n+1) * y); if (!filt) return 0; + line_buffer = (signed char *) STBIW_MALLOC(x * n); if (!line_buffer) { STBIW_FREE(filt); return 0; } + for (j=0; j < y; ++j) { + int filter_type; + if (force_filter > -1) { + filter_type = force_filter; + stbiw__encode_png_line((unsigned char*)(pixels), stride_bytes, x, y, j, n, force_filter, line_buffer); + } else { // Estimate the best filter by running through all of them: + int best_filter = 0, best_filter_val = 0x7fffffff, est, i; + for (filter_type = 0; filter_type < 5; filter_type++) { + stbiw__encode_png_line((unsigned char*)(pixels), stride_bytes, x, y, j, n, filter_type, line_buffer); + + // Estimate the entropy of the line using this filter; the less, the better. + est = 0; + for (i = 0; i < x*n; ++i) { + est += abs((signed char) line_buffer[i]); + } + if (est < best_filter_val) { + best_filter_val = est; + best_filter = filter_type; + } + } + if (filter_type != best_filter) { // If the last iteration already got us the best filter, don't redo it + stbiw__encode_png_line((unsigned char*)(pixels), stride_bytes, x, y, j, n, best_filter, line_buffer); + filter_type = best_filter; + } + } + // when we get here, filter_type contains the filter type, and line_buffer contains the data + filt[j*(x*n+1)] = (unsigned char) filter_type; + STBIW_MEMMOVE(filt+j*(x*n+1)+1, line_buffer, x*n); + } + STBIW_FREE(line_buffer); + zlib = stbi_zlib_compress(filt, y*( x*n+1), &zlen, stbi_write_png_compression_level); + STBIW_FREE(filt); + if (!zlib) return 0; + + // each tag requires 12 bytes of overhead + out = (unsigned char *) STBIW_MALLOC(8 + 12+13 + 12+zlen + 12); + if (!out) return 0; + *out_len = 8 + 12+13 + 12+zlen + 12; + + o=out; + STBIW_MEMMOVE(o,sig,8); o+= 8; + stbiw__wp32(o, 13); // header length + stbiw__wptag(o, "IHDR"); + stbiw__wp32(o, x); + stbiw__wp32(o, y); + *o++ = 8; + *o++ = STBIW_UCHAR(ctype[n]); + *o++ = 0; + *o++ = 0; + *o++ = 0; + stbiw__wpcrc(&o,13); + + stbiw__wp32(o, zlen); + stbiw__wptag(o, "IDAT"); + STBIW_MEMMOVE(o, zlib, zlen); + o += zlen; + STBIW_FREE(zlib); + stbiw__wpcrc(&o, zlen); + + stbiw__wp32(o,0); + stbiw__wptag(o, "IEND"); + stbiw__wpcrc(&o,0); + + STBIW_ASSERT(o == out + *out_len); + + return out; +} + +#ifndef STBI_WRITE_NO_STDIO +STBIWDEF int stbi_write_png(char const *filename, int x, int y, int comp, const void *data, int stride_bytes) +{ + FILE *f; + int len; + unsigned char *png = stbi_write_png_to_mem((const unsigned char *) data, stride_bytes, x, y, comp, &len); + if (png == NULL) return 0; + + f = stbiw__fopen(filename, "wb"); + if (!f) { STBIW_FREE(png); return 0; } + fwrite(png, 1, len, f); + fclose(f); + STBIW_FREE(png); + return 1; +} +#endif + +STBIWDEF int stbi_write_png_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data, int stride_bytes) +{ + int len; + unsigned char *png = stbi_write_png_to_mem((const unsigned char *) data, stride_bytes, x, y, comp, &len); + if (png == NULL) return 0; + func(context, png, len); + STBIW_FREE(png); + return 1; +} + + +/* *************************************************************************** + * + * JPEG writer + * + * This is based on Jon Olick's jo_jpeg.cpp: + * public domain Simple, Minimalistic JPEG writer - http://www.jonolick.com/code.html + */ + +static const unsigned char stbiw__jpg_ZigZag[] = { 0,1,5,6,14,15,27,28,2,4,7,13,16,26,29,42,3,8,12,17,25,30,41,43,9,11,18, + 24,31,40,44,53,10,19,23,32,39,45,52,54,20,22,33,38,46,51,55,60,21,34,37,47,50,56,59,61,35,36,48,49,57,58,62,63 }; + +static void stbiw__jpg_writeBits(stbi__write_context *s, int *bitBufP, int *bitCntP, const unsigned short *bs) { + int bitBuf = *bitBufP, bitCnt = *bitCntP; + bitCnt += bs[1]; + bitBuf |= bs[0] << (24 - bitCnt); + while(bitCnt >= 8) { + unsigned char c = (bitBuf >> 16) & 255; + stbiw__putc(s, c); + if(c == 255) { + stbiw__putc(s, 0); + } + bitBuf <<= 8; + bitCnt -= 8; + } + *bitBufP = bitBuf; + *bitCntP = bitCnt; +} + +static void stbiw__jpg_DCT(float *d0p, float *d1p, float *d2p, float *d3p, float *d4p, float *d5p, float *d6p, float *d7p) { + float d0 = *d0p, d1 = *d1p, d2 = *d2p, d3 = *d3p, d4 = *d4p, d5 = *d5p, d6 = *d6p, d7 = *d7p; + float z1, z2, z3, z4, z5, z11, z13; + + float tmp0 = d0 + d7; + float tmp7 = d0 - d7; + float tmp1 = d1 + d6; + float tmp6 = d1 - d6; + float tmp2 = d2 + d5; + float tmp5 = d2 - d5; + float tmp3 = d3 + d4; + float tmp4 = d3 - d4; + + // Even part + float tmp10 = tmp0 + tmp3; // phase 2 + float tmp13 = tmp0 - tmp3; + float tmp11 = tmp1 + tmp2; + float tmp12 = tmp1 - tmp2; + + d0 = tmp10 + tmp11; // phase 3 + d4 = tmp10 - tmp11; + + z1 = (tmp12 + tmp13) * 0.707106781f; // c4 + d2 = tmp13 + z1; // phase 5 + d6 = tmp13 - z1; + + // Odd part + tmp10 = tmp4 + tmp5; // phase 2 + tmp11 = tmp5 + tmp6; + tmp12 = tmp6 + tmp7; + + // The rotator is modified from fig 4-8 to avoid extra negations. + z5 = (tmp10 - tmp12) * 0.382683433f; // c6 + z2 = tmp10 * 0.541196100f + z5; // c2-c6 + z4 = tmp12 * 1.306562965f + z5; // c2+c6 + z3 = tmp11 * 0.707106781f; // c4 + + z11 = tmp7 + z3; // phase 5 + z13 = tmp7 - z3; + + *d5p = z13 + z2; // phase 6 + *d3p = z13 - z2; + *d1p = z11 + z4; + *d7p = z11 - z4; + + *d0p = d0; *d2p = d2; *d4p = d4; *d6p = d6; +} + +static void stbiw__jpg_calcBits(int val, unsigned short bits[2]) { + int tmp1 = val < 0 ? -val : val; + val = val < 0 ? val-1 : val; + bits[1] = 1; + while(tmp1 >>= 1) { + ++bits[1]; + } + bits[0] = val & ((1<0)&&(DU[end0pos]==0); --end0pos) { + } + // end0pos = first element in reverse order !=0 + if(end0pos == 0) { + stbiw__jpg_writeBits(s, bitBuf, bitCnt, EOB); + return DU[0]; + } + for(i = 1; i <= end0pos; ++i) { + int startpos = i; + int nrzeroes; + unsigned short bits[2]; + for (; DU[i]==0 && i<=end0pos; ++i) { + } + nrzeroes = i-startpos; + if ( nrzeroes >= 16 ) { + int lng = nrzeroes>>4; + int nrmarker; + for (nrmarker=1; nrmarker <= lng; ++nrmarker) + stbiw__jpg_writeBits(s, bitBuf, bitCnt, M16zeroes); + nrzeroes &= 15; + } + stbiw__jpg_calcBits(DU[i], bits); + stbiw__jpg_writeBits(s, bitBuf, bitCnt, HTAC[(nrzeroes<<4)+bits[1]]); + stbiw__jpg_writeBits(s, bitBuf, bitCnt, bits); + } + if(end0pos != 63) { + stbiw__jpg_writeBits(s, bitBuf, bitCnt, EOB); + } + return DU[0]; +} + +static int stbi_write_jpg_core(stbi__write_context *s, int width, int height, int comp, const void* data, int quality) { + // Constants that don't pollute global namespace + static const unsigned char std_dc_luminance_nrcodes[] = {0,0,1,5,1,1,1,1,1,1,0,0,0,0,0,0,0}; + static const unsigned char std_dc_luminance_values[] = {0,1,2,3,4,5,6,7,8,9,10,11}; + static const unsigned char std_ac_luminance_nrcodes[] = {0,0,2,1,3,3,2,4,3,5,5,4,4,0,0,1,0x7d}; + static const unsigned char std_ac_luminance_values[] = { + 0x01,0x02,0x03,0x00,0x04,0x11,0x05,0x12,0x21,0x31,0x41,0x06,0x13,0x51,0x61,0x07,0x22,0x71,0x14,0x32,0x81,0x91,0xa1,0x08, + 0x23,0x42,0xb1,0xc1,0x15,0x52,0xd1,0xf0,0x24,0x33,0x62,0x72,0x82,0x09,0x0a,0x16,0x17,0x18,0x19,0x1a,0x25,0x26,0x27,0x28, + 0x29,0x2a,0x34,0x35,0x36,0x37,0x38,0x39,0x3a,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0x4a,0x53,0x54,0x55,0x56,0x57,0x58,0x59, + 0x5a,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0x6a,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7a,0x83,0x84,0x85,0x86,0x87,0x88,0x89, + 0x8a,0x92,0x93,0x94,0x95,0x96,0x97,0x98,0x99,0x9a,0xa2,0xa3,0xa4,0xa5,0xa6,0xa7,0xa8,0xa9,0xaa,0xb2,0xb3,0xb4,0xb5,0xb6, + 0xb7,0xb8,0xb9,0xba,0xc2,0xc3,0xc4,0xc5,0xc6,0xc7,0xc8,0xc9,0xca,0xd2,0xd3,0xd4,0xd5,0xd6,0xd7,0xd8,0xd9,0xda,0xe1,0xe2, + 0xe3,0xe4,0xe5,0xe6,0xe7,0xe8,0xe9,0xea,0xf1,0xf2,0xf3,0xf4,0xf5,0xf6,0xf7,0xf8,0xf9,0xfa + }; + static const unsigned char std_dc_chrominance_nrcodes[] = {0,0,3,1,1,1,1,1,1,1,1,1,0,0,0,0,0}; + static const unsigned char std_dc_chrominance_values[] = {0,1,2,3,4,5,6,7,8,9,10,11}; + static const unsigned char std_ac_chrominance_nrcodes[] = {0,0,2,1,2,4,4,3,4,7,5,4,4,0,1,2,0x77}; + static const unsigned char std_ac_chrominance_values[] = { + 0x00,0x01,0x02,0x03,0x11,0x04,0x05,0x21,0x31,0x06,0x12,0x41,0x51,0x07,0x61,0x71,0x13,0x22,0x32,0x81,0x08,0x14,0x42,0x91, + 0xa1,0xb1,0xc1,0x09,0x23,0x33,0x52,0xf0,0x15,0x62,0x72,0xd1,0x0a,0x16,0x24,0x34,0xe1,0x25,0xf1,0x17,0x18,0x19,0x1a,0x26, + 0x27,0x28,0x29,0x2a,0x35,0x36,0x37,0x38,0x39,0x3a,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0x4a,0x53,0x54,0x55,0x56,0x57,0x58, + 0x59,0x5a,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0x6a,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7a,0x82,0x83,0x84,0x85,0x86,0x87, + 0x88,0x89,0x8a,0x92,0x93,0x94,0x95,0x96,0x97,0x98,0x99,0x9a,0xa2,0xa3,0xa4,0xa5,0xa6,0xa7,0xa8,0xa9,0xaa,0xb2,0xb3,0xb4, + 0xb5,0xb6,0xb7,0xb8,0xb9,0xba,0xc2,0xc3,0xc4,0xc5,0xc6,0xc7,0xc8,0xc9,0xca,0xd2,0xd3,0xd4,0xd5,0xd6,0xd7,0xd8,0xd9,0xda, + 0xe2,0xe3,0xe4,0xe5,0xe6,0xe7,0xe8,0xe9,0xea,0xf2,0xf3,0xf4,0xf5,0xf6,0xf7,0xf8,0xf9,0xfa + }; + // Huffman tables + static const unsigned short YDC_HT[256][2] = { {0,2},{2,3},{3,3},{4,3},{5,3},{6,3},{14,4},{30,5},{62,6},{126,7},{254,8},{510,9}}; + static const unsigned short UVDC_HT[256][2] = { {0,2},{1,2},{2,2},{6,3},{14,4},{30,5},{62,6},{126,7},{254,8},{510,9},{1022,10},{2046,11}}; + static const unsigned short YAC_HT[256][2] = { + {10,4},{0,2},{1,2},{4,3},{11,4},{26,5},{120,7},{248,8},{1014,10},{65410,16},{65411,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {12,4},{27,5},{121,7},{502,9},{2038,11},{65412,16},{65413,16},{65414,16},{65415,16},{65416,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {28,5},{249,8},{1015,10},{4084,12},{65417,16},{65418,16},{65419,16},{65420,16},{65421,16},{65422,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {58,6},{503,9},{4085,12},{65423,16},{65424,16},{65425,16},{65426,16},{65427,16},{65428,16},{65429,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {59,6},{1016,10},{65430,16},{65431,16},{65432,16},{65433,16},{65434,16},{65435,16},{65436,16},{65437,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {122,7},{2039,11},{65438,16},{65439,16},{65440,16},{65441,16},{65442,16},{65443,16},{65444,16},{65445,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {123,7},{4086,12},{65446,16},{65447,16},{65448,16},{65449,16},{65450,16},{65451,16},{65452,16},{65453,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {250,8},{4087,12},{65454,16},{65455,16},{65456,16},{65457,16},{65458,16},{65459,16},{65460,16},{65461,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {504,9},{32704,15},{65462,16},{65463,16},{65464,16},{65465,16},{65466,16},{65467,16},{65468,16},{65469,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {505,9},{65470,16},{65471,16},{65472,16},{65473,16},{65474,16},{65475,16},{65476,16},{65477,16},{65478,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {506,9},{65479,16},{65480,16},{65481,16},{65482,16},{65483,16},{65484,16},{65485,16},{65486,16},{65487,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {1017,10},{65488,16},{65489,16},{65490,16},{65491,16},{65492,16},{65493,16},{65494,16},{65495,16},{65496,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {1018,10},{65497,16},{65498,16},{65499,16},{65500,16},{65501,16},{65502,16},{65503,16},{65504,16},{65505,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {2040,11},{65506,16},{65507,16},{65508,16},{65509,16},{65510,16},{65511,16},{65512,16},{65513,16},{65514,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {65515,16},{65516,16},{65517,16},{65518,16},{65519,16},{65520,16},{65521,16},{65522,16},{65523,16},{65524,16},{0,0},{0,0},{0,0},{0,0},{0,0}, + {2041,11},{65525,16},{65526,16},{65527,16},{65528,16},{65529,16},{65530,16},{65531,16},{65532,16},{65533,16},{65534,16},{0,0},{0,0},{0,0},{0,0},{0,0} + }; + static const unsigned short UVAC_HT[256][2] = { + {0,2},{1,2},{4,3},{10,4},{24,5},{25,5},{56,6},{120,7},{500,9},{1014,10},{4084,12},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {11,4},{57,6},{246,8},{501,9},{2038,11},{4085,12},{65416,16},{65417,16},{65418,16},{65419,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {26,5},{247,8},{1015,10},{4086,12},{32706,15},{65420,16},{65421,16},{65422,16},{65423,16},{65424,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {27,5},{248,8},{1016,10},{4087,12},{65425,16},{65426,16},{65427,16},{65428,16},{65429,16},{65430,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {58,6},{502,9},{65431,16},{65432,16},{65433,16},{65434,16},{65435,16},{65436,16},{65437,16},{65438,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {59,6},{1017,10},{65439,16},{65440,16},{65441,16},{65442,16},{65443,16},{65444,16},{65445,16},{65446,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {121,7},{2039,11},{65447,16},{65448,16},{65449,16},{65450,16},{65451,16},{65452,16},{65453,16},{65454,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {122,7},{2040,11},{65455,16},{65456,16},{65457,16},{65458,16},{65459,16},{65460,16},{65461,16},{65462,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {249,8},{65463,16},{65464,16},{65465,16},{65466,16},{65467,16},{65468,16},{65469,16},{65470,16},{65471,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {503,9},{65472,16},{65473,16},{65474,16},{65475,16},{65476,16},{65477,16},{65478,16},{65479,16},{65480,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {504,9},{65481,16},{65482,16},{65483,16},{65484,16},{65485,16},{65486,16},{65487,16},{65488,16},{65489,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {505,9},{65490,16},{65491,16},{65492,16},{65493,16},{65494,16},{65495,16},{65496,16},{65497,16},{65498,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {506,9},{65499,16},{65500,16},{65501,16},{65502,16},{65503,16},{65504,16},{65505,16},{65506,16},{65507,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {2041,11},{65508,16},{65509,16},{65510,16},{65511,16},{65512,16},{65513,16},{65514,16},{65515,16},{65516,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {16352,14},{65517,16},{65518,16},{65519,16},{65520,16},{65521,16},{65522,16},{65523,16},{65524,16},{65525,16},{0,0},{0,0},{0,0},{0,0},{0,0}, + {1018,10},{32707,15},{65526,16},{65527,16},{65528,16},{65529,16},{65530,16},{65531,16},{65532,16},{65533,16},{65534,16},{0,0},{0,0},{0,0},{0,0},{0,0} + }; + static const int YQT[] = {16,11,10,16,24,40,51,61,12,12,14,19,26,58,60,55,14,13,16,24,40,57,69,56,14,17,22,29,51,87,80,62,18,22, + 37,56,68,109,103,77,24,35,55,64,81,104,113,92,49,64,78,87,103,121,120,101,72,92,95,98,112,100,103,99}; + static const int UVQT[] = {17,18,24,47,99,99,99,99,18,21,26,66,99,99,99,99,24,26,56,99,99,99,99,99,47,66,99,99,99,99,99,99, + 99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99}; + static const float aasf[] = { 1.0f * 2.828427125f, 1.387039845f * 2.828427125f, 1.306562965f * 2.828427125f, 1.175875602f * 2.828427125f, + 1.0f * 2.828427125f, 0.785694958f * 2.828427125f, 0.541196100f * 2.828427125f, 0.275899379f * 2.828427125f }; + + int row, col, i, k; + float fdtbl_Y[64], fdtbl_UV[64]; + unsigned char YTable[64], UVTable[64]; + + if(!data || !width || !height || comp > 4 || comp < 1) { + return 0; + } + + quality = quality ? quality : 90; + quality = quality < 1 ? 1 : quality > 100 ? 100 : quality; + quality = quality < 50 ? 5000 / quality : 200 - quality * 2; + + for(i = 0; i < 64; ++i) { + int uvti, yti = (YQT[i]*quality+50)/100; + YTable[stbiw__jpg_ZigZag[i]] = (unsigned char) (yti < 1 ? 1 : yti > 255 ? 255 : yti); + uvti = (UVQT[i]*quality+50)/100; + UVTable[stbiw__jpg_ZigZag[i]] = (unsigned char) (uvti < 1 ? 1 : uvti > 255 ? 255 : uvti); + } + + for(row = 0, k = 0; row < 8; ++row) { + for(col = 0; col < 8; ++col, ++k) { + fdtbl_Y[k] = 1 / (YTable [stbiw__jpg_ZigZag[k]] * aasf[row] * aasf[col]); + fdtbl_UV[k] = 1 / (UVTable[stbiw__jpg_ZigZag[k]] * aasf[row] * aasf[col]); + } + } + + // Write Headers + { + static const unsigned char head0[] = { 0xFF,0xD8,0xFF,0xE0,0,0x10,'J','F','I','F',0,1,1,0,0,1,0,1,0,0,0xFF,0xDB,0,0x84,0 }; + static const unsigned char head2[] = { 0xFF,0xDA,0,0xC,3,1,0,2,0x11,3,0x11,0,0x3F,0 }; + const unsigned char head1[] = { 0xFF,0xC0,0,0x11,8,(unsigned char)(height>>8),STBIW_UCHAR(height),(unsigned char)(width>>8),STBIW_UCHAR(width), + 3,1,0x11,0,2,0x11,1,3,0x11,1,0xFF,0xC4,0x01,0xA2,0 }; + s->func(s->context, (void*)head0, sizeof(head0)); + s->func(s->context, (void*)YTable, sizeof(YTable)); + stbiw__putc(s, 1); + s->func(s->context, UVTable, sizeof(UVTable)); + s->func(s->context, (void*)head1, sizeof(head1)); + s->func(s->context, (void*)(std_dc_luminance_nrcodes+1), sizeof(std_dc_luminance_nrcodes)-1); + s->func(s->context, (void*)std_dc_luminance_values, sizeof(std_dc_luminance_values)); + stbiw__putc(s, 0x10); // HTYACinfo + s->func(s->context, (void*)(std_ac_luminance_nrcodes+1), sizeof(std_ac_luminance_nrcodes)-1); + s->func(s->context, (void*)std_ac_luminance_values, sizeof(std_ac_luminance_values)); + stbiw__putc(s, 1); // HTUDCinfo + s->func(s->context, (void*)(std_dc_chrominance_nrcodes+1), sizeof(std_dc_chrominance_nrcodes)-1); + s->func(s->context, (void*)std_dc_chrominance_values, sizeof(std_dc_chrominance_values)); + stbiw__putc(s, 0x11); // HTUACinfo + s->func(s->context, (void*)(std_ac_chrominance_nrcodes+1), sizeof(std_ac_chrominance_nrcodes)-1); + s->func(s->context, (void*)std_ac_chrominance_values, sizeof(std_ac_chrominance_values)); + s->func(s->context, (void*)head2, sizeof(head2)); + } + + // Encode 8x8 macroblocks + { + static const unsigned short fillBits[] = {0x7F, 7}; + const unsigned char *imageData = (const unsigned char *)data; + int DCY=0, DCU=0, DCV=0; + int bitBuf=0, bitCnt=0; + // comp == 2 is grey+alpha (alpha is ignored) + int ofsG = comp > 2 ? 1 : 0, ofsB = comp > 2 ? 2 : 0; + int x, y, pos; + for(y = 0; y < height; y += 8) { + for(x = 0; x < width; x += 8) { + float YDU[64], UDU[64], VDU[64]; + for(row = y, pos = 0; row < y+8; ++row) { + // row >= height => use last input row + int clamped_row = (row < height) ? row : height - 1; + int base_p = (stbi__flip_vertically_on_write ? (height-1-clamped_row) : clamped_row)*width*comp; + for(col = x; col < x+8; ++col, ++pos) { + float r, g, b; + // if col >= width => use pixel from last input column + int p = base_p + ((col < width) ? col : (width-1))*comp; + + r = imageData[p+0]; + g = imageData[p+ofsG]; + b = imageData[p+ofsB]; + YDU[pos]=+0.29900f*r+0.58700f*g+0.11400f*b-128; + UDU[pos]=-0.16874f*r-0.33126f*g+0.50000f*b; + VDU[pos]=+0.50000f*r-0.41869f*g-0.08131f*b; + } + } + + DCY = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, YDU, fdtbl_Y, DCY, YDC_HT, YAC_HT); + DCU = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, UDU, fdtbl_UV, DCU, UVDC_HT, UVAC_HT); + DCV = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, VDU, fdtbl_UV, DCV, UVDC_HT, UVAC_HT); + } + } + + // Do the bit alignment of the EOI marker + stbiw__jpg_writeBits(s, &bitBuf, &bitCnt, fillBits); + } + + // EOI + stbiw__putc(s, 0xFF); + stbiw__putc(s, 0xD9); + + return 1; +} + +STBIWDEF int stbi_write_jpg_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data, int quality) +{ + stbi__write_context s; + stbi__start_write_callbacks(&s, func, context); + return stbi_write_jpg_core(&s, x, y, comp, (void *) data, quality); +} + + +#ifndef STBI_WRITE_NO_STDIO +STBIWDEF int stbi_write_jpg(char const *filename, int x, int y, int comp, const void *data, int quality) +{ + stbi__write_context s; + if (stbi__start_write_file(&s,filename)) { + int r = stbi_write_jpg_core(&s, x, y, comp, data, quality); + stbi__end_write_file(&s); + return r; + } else + return 0; +} +#endif + +#endif // STB_IMAGE_WRITE_IMPLEMENTATION + +/* Revision history + 1.10 (2019-02-07) + support utf8 filenames in Windows; fix warnings and platform ifdefs + 1.09 (2018-02-11) + fix typo in zlib quality API, improve STB_I_W_STATIC in C++ + 1.08 (2018-01-29) + add stbi__flip_vertically_on_write, external zlib, zlib quality, choose PNG filter + 1.07 (2017-07-24) + doc fix + 1.06 (2017-07-23) + writing JPEG (using Jon Olick's code) + 1.05 ??? + 1.04 (2017-03-03) + monochrome BMP expansion + 1.03 ??? + 1.02 (2016-04-02) + avoid allocating large structures on the stack + 1.01 (2016-01-16) + STBIW_REALLOC_SIZED: support allocators with no realloc support + avoid race-condition in crc initialization + minor compile issues + 1.00 (2015-09-14) + installable file IO function + 0.99 (2015-09-13) + warning fixes; TGA rle support + 0.98 (2015-04-08) + added STBIW_MALLOC, STBIW_ASSERT etc + 0.97 (2015-01-18) + fixed HDR asserts, rewrote HDR rle logic + 0.96 (2015-01-17) + add HDR output + fix monochrome BMP + 0.95 (2014-08-17) + add monochrome TGA output + 0.94 (2014-05-31) + rename private functions to avoid conflicts with stb_image.h + 0.93 (2014-05-27) + warning fixes + 0.92 (2010-08-01) + casts to unsigned char to fix warnings + 0.91 (2010-07-17) + first public release + 0.90 first internal release +*/ + +/* +------------------------------------------------------------------------------ +This software is available under 2 licenses -- choose whichever you prefer. +------------------------------------------------------------------------------ +ALTERNATIVE A - MIT License +Copyright (c) 2017 Sean Barrett +Permission is hereby granted, free of charge, to any person obtaining a copy of +this software and associated documentation files (the "Software"), to deal in +the Software without restriction, including without limitation the rights to +use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies +of the Software, and to permit persons to whom the Software is furnished to do +so, subject to the following conditions: +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. +------------------------------------------------------------------------------ +ALTERNATIVE B - Public Domain (www.unlicense.org) +This is free and unencumbered software released into the public domain. +Anyone is free to copy, modify, publish, use, compile, sell, or distribute this +software, either in source code form or as a compiled binary, for any purpose, +commercial or non-commercial, and by any means. +In jurisdictions that recognize copyright laws, the author or authors of this +software dedicate any and all copyright interest in the software to the public +domain. We make this dedication for the benefit of the public at large and to +the detriment of our heirs and successors. We intend this dedication to be an +overt act of relinquishment in perpetuity of all present and future rights to +this software under copyright law. +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN +ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION +WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +------------------------------------------------------------------------------ +*/ \ No newline at end of file diff --git a/third-party/tiny_gltf/CMakeLists.txt b/third-party/tiny_gltf/CMakeLists.txt new file mode 100644 index 0000000000..639bd564c1 --- /dev/null +++ b/third-party/tiny_gltf/CMakeLists.txt @@ -0,0 +1,10 @@ +if (UNIX) + set(CMAKE_CXX_FLAGS "-O3") +else () + set(CMAKE_CXX_FLAGS "/EHsc") +endif (UNIX) + +include_directories(../ ../stb_image) +add_library(tiny_gltf tiny_gltf.cpp) +target_link_libraries(tiny_gltf stb_image) + diff --git a/third-party/tiny_gltf/tiny_gltf.cpp b/third-party/tiny_gltf/tiny_gltf.cpp new file mode 100644 index 0000000000..76b6542c00 --- /dev/null +++ b/third-party/tiny_gltf/tiny_gltf.cpp @@ -0,0 +1,4 @@ +#define TINYGLTF_IMPLEMENTATION + +#include "third-party/tiny_gltf/tiny_gltf.h" + diff --git a/third-party/tiny_gltf/tiny_gltf.h b/third-party/tiny_gltf/tiny_gltf.h new file mode 100644 index 0000000000..98cc4e42fc --- /dev/null +++ b/third-party/tiny_gltf/tiny_gltf.h @@ -0,0 +1,7748 @@ +// +// Header-only tiny glTF 2.0 loader and serializer. +// +// +// The MIT License (MIT) +// +// Copyright (c) 2015 - Present Syoyo Fujita, Aurélien Chatelain and many +// contributors. +// +// Permission is hereby granted, free of charge, to any person obtaining a copy +// of this software and associated documentation files (the "Software"), to deal +// in the Software without restriction, including without limitation the rights +// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +// copies of the Software, and to permit persons to whom the Software is +// furnished to do so, subject to the following conditions: +// +// The above copyright notice and this permission notice shall be included in +// all copies or substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN +// THE SOFTWARE. + +// Version: +// - v2.5.0 Add SetPreserveImageChannels() option to load image data as is. +// - v2.4.3 Fix null object output when when material has all default +// parameters. +// - v2.4.2 Decode percent-encoded URI. +// - v2.4.1 Fix some glTF object class does not have `extensions` and/or +// `extras` property. +// - v2.4.0 Experimental RapidJSON and C++14 support(Thanks to @jrkoone). +// - v2.3.1 Set default value of minFilter and magFilter in Sampler to -1. +// - v2.3.0 Modified Material representation according to glTF 2.0 schema +// (and introduced TextureInfo class) +// Change the behavior of `Value::IsNumber`. It return true either the +// value is int or real. +// - v2.2.0 Add loading 16bit PNG support. Add Sparse accessor support(Thanks +// to @Ybalrid) +// - v2.1.0 Add draco compression. +// - v2.0.1 Add comparsion feature(Thanks to @Selmar). +// - v2.0.0 glTF 2.0!. +// +// Tiny glTF loader is using following third party libraries: +// +// - jsonhpp: C++ JSON library. +// - base64: base64 decode/encode library. +// - stb_image: Image loading library. +// +#ifndef TINY_GLTF_H_ +#define TINY_GLTF_H_ + +#include +#include +#include // std::fabs +#include +#include +#include +#include +#include +#include +#include + +#ifndef TINYGLTF_USE_CPP14 +#include +#endif + +#ifdef __ANDROID__ +#ifdef TINYGLTF_ANDROID_LOAD_FROM_ASSETS +#include +#endif +#endif + +#ifdef __GNUC__ +#if (__GNUC__ < 4) || ((__GNUC__ == 4) && (__GNUC_MINOR__ <= 8)) +#define TINYGLTF_NOEXCEPT +#else +#define TINYGLTF_NOEXCEPT noexcept +#endif +#else +#define TINYGLTF_NOEXCEPT noexcept +#endif + +#define DEFAULT_METHODS(x) \ + ~x() = default; \ + x(const x &) = default; \ + x(x &&) TINYGLTF_NOEXCEPT = default; \ + x &operator=(const x &) = default; \ + x &operator=(x &&) TINYGLTF_NOEXCEPT = default; + +namespace tinygltf { + +#define TINYGLTF_MODE_POINTS (0) +#define TINYGLTF_MODE_LINE (1) +#define TINYGLTF_MODE_LINE_LOOP (2) +#define TINYGLTF_MODE_LINE_STRIP (3) +#define TINYGLTF_MODE_TRIANGLES (4) +#define TINYGLTF_MODE_TRIANGLE_STRIP (5) +#define TINYGLTF_MODE_TRIANGLE_FAN (6) + +#define TINYGLTF_COMPONENT_TYPE_BYTE (5120) +#define TINYGLTF_COMPONENT_TYPE_UNSIGNED_BYTE (5121) +#define TINYGLTF_COMPONENT_TYPE_SHORT (5122) +#define TINYGLTF_COMPONENT_TYPE_UNSIGNED_SHORT (5123) +#define TINYGLTF_COMPONENT_TYPE_INT (5124) +#define TINYGLTF_COMPONENT_TYPE_UNSIGNED_INT (5125) +#define TINYGLTF_COMPONENT_TYPE_FLOAT (5126) +#define TINYGLTF_COMPONENT_TYPE_DOUBLE (5130) // OpenGL double type. Note that some of glTF 2.0 validator does not support double type even the schema seems allow any value of integer: https://github.com/KhronosGroup/glTF/blob/b9884a2fd45130b4d673dd6c8a706ee21ee5c5f7/specification/2.0/schema/accessor.schema.json#L22 + +#define TINYGLTF_TEXTURE_FILTER_NEAREST (9728) +#define TINYGLTF_TEXTURE_FILTER_LINEAR (9729) +#define TINYGLTF_TEXTURE_FILTER_NEAREST_MIPMAP_NEAREST (9984) +#define TINYGLTF_TEXTURE_FILTER_LINEAR_MIPMAP_NEAREST (9985) +#define TINYGLTF_TEXTURE_FILTER_NEAREST_MIPMAP_LINEAR (9986) +#define TINYGLTF_TEXTURE_FILTER_LINEAR_MIPMAP_LINEAR (9987) + +#define TINYGLTF_TEXTURE_WRAP_REPEAT (10497) +#define TINYGLTF_TEXTURE_WRAP_CLAMP_TO_EDGE (33071) +#define TINYGLTF_TEXTURE_WRAP_MIRRORED_REPEAT (33648) + +// Redeclarations of the above for technique.parameters. +#define TINYGLTF_PARAMETER_TYPE_BYTE (5120) +#define TINYGLTF_PARAMETER_TYPE_UNSIGNED_BYTE (5121) +#define TINYGLTF_PARAMETER_TYPE_SHORT (5122) +#define TINYGLTF_PARAMETER_TYPE_UNSIGNED_SHORT (5123) +#define TINYGLTF_PARAMETER_TYPE_INT (5124) +#define TINYGLTF_PARAMETER_TYPE_UNSIGNED_INT (5125) +#define TINYGLTF_PARAMETER_TYPE_FLOAT (5126) + +#define TINYGLTF_PARAMETER_TYPE_FLOAT_VEC2 (35664) +#define TINYGLTF_PARAMETER_TYPE_FLOAT_VEC3 (35665) +#define TINYGLTF_PARAMETER_TYPE_FLOAT_VEC4 (35666) + +#define TINYGLTF_PARAMETER_TYPE_INT_VEC2 (35667) +#define TINYGLTF_PARAMETER_TYPE_INT_VEC3 (35668) +#define TINYGLTF_PARAMETER_TYPE_INT_VEC4 (35669) + +#define TINYGLTF_PARAMETER_TYPE_BOOL (35670) +#define TINYGLTF_PARAMETER_TYPE_BOOL_VEC2 (35671) +#define TINYGLTF_PARAMETER_TYPE_BOOL_VEC3 (35672) +#define TINYGLTF_PARAMETER_TYPE_BOOL_VEC4 (35673) + +#define TINYGLTF_PARAMETER_TYPE_FLOAT_MAT2 (35674) +#define TINYGLTF_PARAMETER_TYPE_FLOAT_MAT3 (35675) +#define TINYGLTF_PARAMETER_TYPE_FLOAT_MAT4 (35676) + +#define TINYGLTF_PARAMETER_TYPE_SAMPLER_2D (35678) + +// End parameter types + +#define TINYGLTF_TYPE_VEC2 (2) +#define TINYGLTF_TYPE_VEC3 (3) +#define TINYGLTF_TYPE_VEC4 (4) +#define TINYGLTF_TYPE_MAT2 (32 + 2) +#define TINYGLTF_TYPE_MAT3 (32 + 3) +#define TINYGLTF_TYPE_MAT4 (32 + 4) +#define TINYGLTF_TYPE_SCALAR (64 + 1) +#define TINYGLTF_TYPE_VECTOR (64 + 4) +#define TINYGLTF_TYPE_MATRIX (64 + 16) + +#define TINYGLTF_IMAGE_FORMAT_JPEG (0) +#define TINYGLTF_IMAGE_FORMAT_PNG (1) +#define TINYGLTF_IMAGE_FORMAT_BMP (2) +#define TINYGLTF_IMAGE_FORMAT_GIF (3) + +#define TINYGLTF_TEXTURE_FORMAT_ALPHA (6406) +#define TINYGLTF_TEXTURE_FORMAT_RGB (6407) +#define TINYGLTF_TEXTURE_FORMAT_RGBA (6408) +#define TINYGLTF_TEXTURE_FORMAT_LUMINANCE (6409) +#define TINYGLTF_TEXTURE_FORMAT_LUMINANCE_ALPHA (6410) + +#define TINYGLTF_TEXTURE_TARGET_TEXTURE2D (3553) +#define TINYGLTF_TEXTURE_TYPE_UNSIGNED_BYTE (5121) + +#define TINYGLTF_TARGET_ARRAY_BUFFER (34962) +#define TINYGLTF_TARGET_ELEMENT_ARRAY_BUFFER (34963) + +#define TINYGLTF_SHADER_TYPE_VERTEX_SHADER (35633) +#define TINYGLTF_SHADER_TYPE_FRAGMENT_SHADER (35632) + +#define TINYGLTF_DOUBLE_EPS (1.e-12) +#define TINYGLTF_DOUBLE_EQUAL(a, b) (std::fabs((b) - (a)) < TINYGLTF_DOUBLE_EPS) + +#ifdef __ANDROID__ +#ifdef TINYGLTF_ANDROID_LOAD_FROM_ASSETS +AAssetManager *asset_manager = nullptr; +#endif +#endif + +typedef enum { + NULL_TYPE, + REAL_TYPE, + INT_TYPE, + BOOL_TYPE, + STRING_TYPE, + ARRAY_TYPE, + BINARY_TYPE, + OBJECT_TYPE +} Type; + +static inline int32_t GetComponentSizeInBytes(uint32_t componentType) { + if (componentType == TINYGLTF_COMPONENT_TYPE_BYTE) { + return 1; + } else if (componentType == TINYGLTF_COMPONENT_TYPE_UNSIGNED_BYTE) { + return 1; + } else if (componentType == TINYGLTF_COMPONENT_TYPE_SHORT) { + return 2; + } else if (componentType == TINYGLTF_COMPONENT_TYPE_UNSIGNED_SHORT) { + return 2; + } else if (componentType == TINYGLTF_COMPONENT_TYPE_INT) { + return 4; + } else if (componentType == TINYGLTF_COMPONENT_TYPE_UNSIGNED_INT) { + return 4; + } else if (componentType == TINYGLTF_COMPONENT_TYPE_FLOAT) { + return 4; + } else if (componentType == TINYGLTF_COMPONENT_TYPE_DOUBLE) { + return 8; + } else { + // Unknown componenty type + return -1; + } +} + +static inline int32_t GetNumComponentsInType(uint32_t ty) { + if (ty == TINYGLTF_TYPE_SCALAR) { + return 1; + } else if (ty == TINYGLTF_TYPE_VEC2) { + return 2; + } else if (ty == TINYGLTF_TYPE_VEC3) { + return 3; + } else if (ty == TINYGLTF_TYPE_VEC4) { + return 4; + } else if (ty == TINYGLTF_TYPE_MAT2) { + return 4; + } else if (ty == TINYGLTF_TYPE_MAT3) { + return 9; + } else if (ty == TINYGLTF_TYPE_MAT4) { + return 16; + } else { + // Unknown componenty type + return -1; + } +} + +// TODO(syoyo): Move these functions to TinyGLTF class +bool IsDataURI(const std::string &in); +bool DecodeDataURI(std::vector *out, std::string &mime_type, + const std::string &in, size_t reqBytes, bool checkSize); + +#ifdef __clang__ +#pragma clang diagnostic push +// Suppress warning for : static Value null_value +#pragma clang diagnostic ignored "-Wexit-time-destructors" +#pragma clang diagnostic ignored "-Wpadded" +#endif + +// Simple class to represent JSON object +class Value { + public: + typedef std::vector Array; + typedef std::map Object; + + Value() + : type_(NULL_TYPE), + int_value_(0), + real_value_(0.0), + boolean_value_(false) {} + + explicit Value(bool b) : type_(BOOL_TYPE) { boolean_value_ = b; } + explicit Value(int i) : type_(INT_TYPE) { + int_value_ = i; + real_value_ = i; + } + explicit Value(double n) : type_(REAL_TYPE) { real_value_ = n; } + explicit Value(const std::string &s) : type_(STRING_TYPE) { + string_value_ = s; + } + explicit Value(std::string &&s) + : type_(STRING_TYPE), string_value_(std::move(s)) {} + explicit Value(const unsigned char *p, size_t n) : type_(BINARY_TYPE) { + binary_value_.resize(n); + memcpy(binary_value_.data(), p, n); + } + explicit Value(std::vector &&v) noexcept + : type_(BINARY_TYPE), + binary_value_(std::move(v)) {} + explicit Value(const Array &a) : type_(ARRAY_TYPE) { array_value_ = a; } + explicit Value(Array &&a) noexcept : type_(ARRAY_TYPE), + array_value_(std::move(a)) {} + + explicit Value(const Object &o) : type_(OBJECT_TYPE) { object_value_ = o; } + explicit Value(Object &&o) noexcept : type_(OBJECT_TYPE), + object_value_(std::move(o)) {} + + DEFAULT_METHODS(Value) + + char Type() const { return static_cast(type_); } + + bool IsBool() const { return (type_ == BOOL_TYPE); } + + bool IsInt() const { return (type_ == INT_TYPE); } + + bool IsNumber() const { return (type_ == REAL_TYPE) || (type_ == INT_TYPE); } + + bool IsReal() const { return (type_ == REAL_TYPE); } + + bool IsString() const { return (type_ == STRING_TYPE); } + + bool IsBinary() const { return (type_ == BINARY_TYPE); } + + bool IsArray() const { return (type_ == ARRAY_TYPE); } + + bool IsObject() const { return (type_ == OBJECT_TYPE); } + + // Use this function if you want to have number value as double. + double GetNumberAsDouble() const { + if (type_ == INT_TYPE) { + return double(int_value_); + } else { + return real_value_; + } + } + + // Use this function if you want to have number value as int. + // TODO(syoyo): Support int value larger than 32 bits + int GetNumberAsInt() const { + if (type_ == REAL_TYPE) { + return int(real_value_); + } else { + return int_value_; + } + } + + // Accessor + template + const T &Get() const; + template + T &Get(); + + // Lookup value from an array + const Value &Get(int idx) const { + static Value null_value; + assert(IsArray()); + assert(idx >= 0); + return (static_cast(idx) < array_value_.size()) + ? array_value_[static_cast(idx)] + : null_value; + } + + // Lookup value from a key-value pair + const Value &Get(const std::string &key) const { + static Value null_value; + assert(IsObject()); + Object::const_iterator it = object_value_.find(key); + return (it != object_value_.end()) ? it->second : null_value; + } + + size_t ArrayLen() const { + if (!IsArray()) return 0; + return array_value_.size(); + } + + // Valid only for object type. + bool Has(const std::string &key) const { + if (!IsObject()) return false; + Object::const_iterator it = object_value_.find(key); + return (it != object_value_.end()) ? true : false; + } + + // List keys + std::vector Keys() const { + std::vector keys; + if (!IsObject()) return keys; // empty + + for (Object::const_iterator it = object_value_.begin(); + it != object_value_.end(); ++it) { + keys.push_back(it->first); + } + + return keys; + } + + size_t Size() const { return (IsArray() ? ArrayLen() : Keys().size()); } + + bool operator==(const tinygltf::Value &other) const; + + protected: + int type_ = NULL_TYPE; + + int int_value_ = 0; + double real_value_ = 0.0; + std::string string_value_; + std::vector binary_value_; + Array array_value_; + Object object_value_; + bool boolean_value_ = false; +}; + +#ifdef __clang__ +#pragma clang diagnostic pop +#endif + +#define TINYGLTF_VALUE_GET(ctype, var) \ + template <> \ + inline const ctype &Value::Get() const { \ + return var; \ + } \ + template <> \ + inline ctype &Value::Get() { \ + return var; \ + } +TINYGLTF_VALUE_GET(bool, boolean_value_) +TINYGLTF_VALUE_GET(double, real_value_) +TINYGLTF_VALUE_GET(int, int_value_) +TINYGLTF_VALUE_GET(std::string, string_value_) +TINYGLTF_VALUE_GET(std::vector, binary_value_) +TINYGLTF_VALUE_GET(Value::Array, array_value_) +TINYGLTF_VALUE_GET(Value::Object, object_value_) +#undef TINYGLTF_VALUE_GET + +#ifdef __clang__ +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wc++98-compat" +#pragma clang diagnostic ignored "-Wpadded" +#endif + +/// Agregate object for representing a color +using ColorValue = std::array; + +// === legacy interface ==== +// TODO(syoyo): Deprecate `Parameter` class. +struct Parameter { + bool bool_value = false; + bool has_number_value = false; + std::string string_value; + std::vector number_array; + std::map json_double_value; + double number_value = 0.0; + + // context sensitive methods. depending the type of the Parameter you are + // accessing, these are either valid or not + // If this parameter represent a texture map in a material, will return the + // texture index + + /// Return the index of a texture if this Parameter is a texture map. + /// Returned value is only valid if the parameter represent a texture from a + /// material + int TextureIndex() const { + const auto it = json_double_value.find("index"); + if (it != std::end(json_double_value)) { + return int(it->second); + } + return -1; + } + + /// Return the index of a texture coordinate set if this Parameter is a + /// texture map. Returned value is only valid if the parameter represent a + /// texture from a material + int TextureTexCoord() const { + const auto it = json_double_value.find("texCoord"); + if (it != std::end(json_double_value)) { + return int(it->second); + } + // As per the spec, if texCoord is ommited, this parameter is 0 + return 0; + } + + /// Return the scale of a texture if this Parameter is a normal texture map. + /// Returned value is only valid if the parameter represent a normal texture + /// from a material + double TextureScale() const { + const auto it = json_double_value.find("scale"); + if (it != std::end(json_double_value)) { + return it->second; + } + // As per the spec, if scale is ommited, this paramter is 1 + return 1; + } + + /// Return the strength of a texture if this Parameter is a an occlusion map. + /// Returned value is only valid if the parameter represent an occlusion map + /// from a material + double TextureStrength() const { + const auto it = json_double_value.find("strength"); + if (it != std::end(json_double_value)) { + return it->second; + } + // As per the spec, if strenghth is ommited, this parameter is 1 + return 1; + } + + /// Material factor, like the roughness or metalness of a material + /// Returned value is only valid if the parameter represent a texture from a + /// material + double Factor() const { return number_value; } + + /// Return the color of a material + /// Returned value is only valid if the parameter represent a texture from a + /// material + ColorValue ColorFactor() const { + return { + {// this agregate intialize the std::array object, and uses C++11 RVO. + number_array[0], number_array[1], number_array[2], + (number_array.size() > 3 ? number_array[3] : 1.0)}}; + } + + Parameter() = default; + DEFAULT_METHODS(Parameter) + bool operator==(const Parameter &) const; +}; + +#ifdef __clang__ +#pragma clang diagnostic pop +#endif + +#ifdef __clang__ +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wpadded" +#endif + +typedef std::map ParameterMap; +typedef std::map ExtensionMap; + +struct AnimationChannel { + int sampler; // required + int target_node; // required (index of the node to target) + std::string target_path; // required in ["translation", "rotation", "scale", + // "weights"] + Value extras; + ExtensionMap extensions; + ExtensionMap target_extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + std::string target_extensions_json_string; + + AnimationChannel() : sampler(-1), target_node(-1) {} + DEFAULT_METHODS(AnimationChannel) + bool operator==(const AnimationChannel &) const; +}; + +struct AnimationSampler { + int input; // required + int output; // required + std::string interpolation; // "LINEAR", "STEP","CUBICSPLINE" or user defined + // string. default "LINEAR" + Value extras; + ExtensionMap extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + AnimationSampler() : input(-1), output(-1), interpolation("LINEAR") {} + DEFAULT_METHODS(AnimationSampler) + bool operator==(const AnimationSampler &) const; +}; + +struct Animation { + std::string name; + std::vector channels; + std::vector samplers; + Value extras; + ExtensionMap extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + Animation() = default; + DEFAULT_METHODS(Animation) + bool operator==(const Animation &) const; +}; + +struct Skin { + std::string name; + int inverseBindMatrices; // required here but not in the spec + int skeleton; // The index of the node used as a skeleton root + std::vector joints; // Indices of skeleton nodes + + Value extras; + ExtensionMap extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + Skin() { + inverseBindMatrices = -1; + skeleton = -1; + } + DEFAULT_METHODS(Skin) + bool operator==(const Skin &) const; +}; + +struct Sampler { + std::string name; + // glTF 2.0 spec does not define default value for `minFilter` and + // `magFilter`. Set -1 in TinyGLTF(issue #186) + int minFilter = + -1; // optional. -1 = no filter defined. ["NEAREST", "LINEAR", + // "NEAREST_MIPMAP_NEAREST", "LINEAR_MIPMAP_NEAREST", + // "NEAREST_MIPMAP_LINEAR", "LINEAR_MIPMAP_LINEAR"] + int magFilter = + -1; // optional. -1 = no filter defined. ["NEAREST", "LINEAR"] + int wrapS = + TINYGLTF_TEXTURE_WRAP_REPEAT; // ["CLAMP_TO_EDGE", "MIRRORED_REPEAT", + // "REPEAT"], default "REPEAT" + int wrapT = + TINYGLTF_TEXTURE_WRAP_REPEAT; // ["CLAMP_TO_EDGE", "MIRRORED_REPEAT", + // "REPEAT"], default "REPEAT" + //int wrapR = TINYGLTF_TEXTURE_WRAP_REPEAT; // TinyGLTF extension. currently not used. + + Value extras; + ExtensionMap extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + Sampler() + : minFilter(-1), + magFilter(-1), + wrapS(TINYGLTF_TEXTURE_WRAP_REPEAT), + wrapT(TINYGLTF_TEXTURE_WRAP_REPEAT) {} + DEFAULT_METHODS(Sampler) + bool operator==(const Sampler &) const; +}; + +struct Image { + std::string name; + int width; + int height; + int component; + int bits; // bit depth per channel. 8(byte), 16 or 32. + int pixel_type; // pixel type(TINYGLTF_COMPONENT_TYPE_***). usually + // UBYTE(bits = 8) or USHORT(bits = 16) + std::vector image; + int bufferView; // (required if no uri) + std::string mimeType; // (required if no uri) ["image/jpeg", "image/png", + // "image/bmp", "image/gif"] + std::string uri; // (required if no mimeType) uri is not decoded(e.g. + // whitespace may be represented as %20) + Value extras; + ExtensionMap extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + // When this flag is true, data is stored to `image` in as-is format(e.g. jpeg + // compressed for "image/jpeg" mime) This feature is good if you use custom + // image loader function. (e.g. delayed decoding of images for faster glTF + // parsing) Default parser for Image does not provide as-is loading feature at + // the moment. (You can manipulate this by providing your own LoadImageData + // function) + bool as_is; + + Image() : as_is(false) { + bufferView = -1; + width = -1; + height = -1; + component = -1; + bits = -1; + pixel_type = -1; + } + DEFAULT_METHODS(Image) + + bool operator==(const Image &) const; +}; + +struct Texture { + std::string name; + + int sampler; + int source; + Value extras; + ExtensionMap extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + Texture() : sampler(-1), source(-1) {} + DEFAULT_METHODS(Texture) + + bool operator==(const Texture &) const; +}; + +struct TextureInfo { + int index = -1; // required. + int texCoord; // The set index of texture's TEXCOORD attribute used for + // texture coordinate mapping. + + Value extras; + ExtensionMap extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + TextureInfo() : index(-1), texCoord(0) {} + DEFAULT_METHODS(TextureInfo) + bool operator==(const TextureInfo &) const; +}; + +struct NormalTextureInfo { + int index = -1; // required + int texCoord; // The set index of texture's TEXCOORD attribute used for + // texture coordinate mapping. + double scale; // scaledNormal = normalize(( + // * 2.0 - 1.0) * vec3(, , 1.0)) + + Value extras; + ExtensionMap extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + NormalTextureInfo() : index(-1), texCoord(0), scale(1.0) {} + DEFAULT_METHODS(NormalTextureInfo) + bool operator==(const NormalTextureInfo &) const; +}; + +struct OcclusionTextureInfo { + int index = -1; // required + int texCoord; // The set index of texture's TEXCOORD attribute used for + // texture coordinate mapping. + double strength; // occludedColor = lerp(color, color * , ) + + Value extras; + ExtensionMap extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + OcclusionTextureInfo() : index(-1), texCoord(0), strength(1.0) {} + DEFAULT_METHODS(OcclusionTextureInfo) + bool operator==(const OcclusionTextureInfo &) const; +}; + +// pbrMetallicRoughness class defined in glTF 2.0 spec. +struct PbrMetallicRoughness { + std::vector baseColorFactor; // len = 4. default [1,1,1,1] + TextureInfo baseColorTexture; + double metallicFactor; // default 1 + double roughnessFactor; // default 1 + TextureInfo metallicRoughnessTexture; + + Value extras; + ExtensionMap extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + PbrMetallicRoughness() + : baseColorFactor(std::vector{1.0, 1.0, 1.0, 1.0}), + metallicFactor(1.0), + roughnessFactor(1.0) {} + DEFAULT_METHODS(PbrMetallicRoughness) + bool operator==(const PbrMetallicRoughness &) const; +}; + +// Each extension should be stored in a ParameterMap. +// members not in the values could be included in the ParameterMap +// to keep a single material model +struct Material { + std::string name; + + std::vector emissiveFactor; // length 3. default [0, 0, 0] + std::string alphaMode; // default "OPAQUE" + double alphaCutoff; // default 0.5 + bool doubleSided; // default false; + + PbrMetallicRoughness pbrMetallicRoughness; + + NormalTextureInfo normalTexture; + OcclusionTextureInfo occlusionTexture; + TextureInfo emissiveTexture; + + // For backward compatibility + // TODO(syoyo): Remove `values` and `additionalValues` in the next release. + ParameterMap values; + ParameterMap additionalValues; + + ExtensionMap extensions; + Value extras; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + Material() : alphaMode("OPAQUE"), alphaCutoff(0.5), doubleSided(false) {} + DEFAULT_METHODS(Material) + + bool operator==(const Material &) const; +}; + +struct BufferView { + std::string name; + int buffer{-1}; // Required + size_t byteOffset{0}; // minimum 0, default 0 + size_t byteLength{0}; // required, minimum 1. 0 = invalid + size_t byteStride{0}; // minimum 4, maximum 252 (multiple of 4), default 0 = + // understood to be tightly packed + int target{0}; // ["ARRAY_BUFFER", "ELEMENT_ARRAY_BUFFER"] for vertex indices + // or atttribs. Could be 0 for other data + Value extras; + ExtensionMap extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + bool dracoDecoded{false}; // Flag indicating this has been draco decoded + + BufferView() + : buffer(-1), + byteOffset(0), + byteLength(0), + byteStride(0), + target(0), + dracoDecoded(false) {} + DEFAULT_METHODS(BufferView) + bool operator==(const BufferView &) const; +}; + +struct Accessor { + int bufferView; // optional in spec but required here since sparse accessor + // are not supported + std::string name; + size_t byteOffset; + bool normalized; // optional. + int componentType; // (required) One of TINYGLTF_COMPONENT_TYPE_*** + size_t count; // required + int type; // (required) One of TINYGLTF_TYPE_*** .. + Value extras; + ExtensionMap extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + std::vector + minValues; // optional. integer value is promoted to double + std::vector + maxValues; // optional. integer value is promoted to double + + struct { + int count; + bool isSparse; + struct { + int byteOffset; + int bufferView; + int componentType; // a TINYGLTF_COMPONENT_TYPE_ value + } indices; + struct { + int bufferView; + int byteOffset; + } values; + } sparse; + + /// + /// Utility function to compute byteStride for a given bufferView object. + /// Returns -1 upon invalid glTF value or parameter configuration. + /// + int ByteStride(const BufferView &bufferViewObject) const { + if (bufferViewObject.byteStride == 0) { + // Assume data is tightly packed. + int componentSizeInBytes = + GetComponentSizeInBytes(static_cast(componentType)); + if (componentSizeInBytes <= 0) { + return -1; + } + + int numComponents = GetNumComponentsInType(static_cast(type)); + if (numComponents <= 0) { + return -1; + } + + return componentSizeInBytes * numComponents; + } else { + // Check if byteStride is a mulple of the size of the accessor's component + // type. + int componentSizeInBytes = + GetComponentSizeInBytes(static_cast(componentType)); + if (componentSizeInBytes <= 0) { + return -1; + } + + if ((bufferViewObject.byteStride % uint32_t(componentSizeInBytes)) != 0) { + return -1; + } + return static_cast(bufferViewObject.byteStride); + } + + // unreachable return 0; + } + + Accessor() + : bufferView(-1), + byteOffset(0), + normalized(false), + componentType(-1), + count(0), + type(-1) { + sparse.isSparse = false; + } + DEFAULT_METHODS(Accessor) + bool operator==(const tinygltf::Accessor &) const; +}; + +struct PerspectiveCamera { + double aspectRatio; // min > 0 + double yfov; // required. min > 0 + double zfar; // min > 0 + double znear; // required. min > 0 + + PerspectiveCamera() + : aspectRatio(0.0), + yfov(0.0), + zfar(0.0) // 0 = use infinite projecton matrix + , + znear(0.0) {} + DEFAULT_METHODS(PerspectiveCamera) + bool operator==(const PerspectiveCamera &) const; + + ExtensionMap extensions; + Value extras; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; +}; + +struct OrthographicCamera { + double xmag; // required. must not be zero. + double ymag; // required. must not be zero. + double zfar; // required. `zfar` must be greater than `znear`. + double znear; // required + + OrthographicCamera() : xmag(0.0), ymag(0.0), zfar(0.0), znear(0.0) {} + DEFAULT_METHODS(OrthographicCamera) + bool operator==(const OrthographicCamera &) const; + + ExtensionMap extensions; + Value extras; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; +}; + +struct Camera { + std::string type; // required. "perspective" or "orthographic" + std::string name; + + PerspectiveCamera perspective; + OrthographicCamera orthographic; + + Camera() {} + DEFAULT_METHODS(Camera) + bool operator==(const Camera &) const; + + ExtensionMap extensions; + Value extras; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; +}; + +struct Primitive { + std::map attributes; // (required) A dictionary object of + // integer, where each integer + // is the index of the accessor + // containing an attribute. + int material; // The index of the material to apply to this primitive + // when rendering. + int indices; // The index of the accessor that contains the indices. + int mode; // one of TINYGLTF_MODE_*** + std::vector > targets; // array of morph targets, + // where each target is a dict with attribues in ["POSITION, "NORMAL", + // "TANGENT"] pointing + // to their corresponding accessors + ExtensionMap extensions; + Value extras; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + Primitive() { + material = -1; + indices = -1; + mode = -1; + } + DEFAULT_METHODS(Primitive) + bool operator==(const Primitive &) const; +}; + +struct Mesh { + std::string name; + std::vector primitives; + std::vector weights; // weights to be applied to the Morph Targets + ExtensionMap extensions; + Value extras; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + Mesh() = default; + DEFAULT_METHODS(Mesh) + bool operator==(const Mesh &) const; +}; + +class Node { + public: + Node() : camera(-1), skin(-1), mesh(-1) {} + + DEFAULT_METHODS(Node) + + bool operator==(const Node &) const; + + int camera; // the index of the camera referenced by this node + + std::string name; + int skin; + int mesh; + std::vector children; + std::vector rotation; // length must be 0 or 4 + std::vector scale; // length must be 0 or 3 + std::vector translation; // length must be 0 or 3 + std::vector matrix; // length must be 0 or 16 + std::vector weights; // The weights of the instantiated Morph Target + + ExtensionMap extensions; + Value extras; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; +}; + +struct Buffer { + std::string name; + std::vector data; + std::string + uri; // considered as required here but not in the spec (need to clarify) + // uri is not decoded(e.g. whitespace may be represented as %20) + Value extras; + ExtensionMap extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + Buffer() = default; + DEFAULT_METHODS(Buffer) + bool operator==(const Buffer &) const; +}; + +struct Asset { + std::string version = "2.0"; // required + std::string generator; + std::string minVersion; + std::string copyright; + ExtensionMap extensions; + Value extras; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + Asset() = default; + DEFAULT_METHODS(Asset) + bool operator==(const Asset &) const; +}; + +struct Scene { + std::string name; + std::vector nodes; + + ExtensionMap extensions; + Value extras; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; + + Scene() = default; + DEFAULT_METHODS(Scene) + bool operator==(const Scene &) const; +}; + +struct SpotLight { + double innerConeAngle; + double outerConeAngle; + + SpotLight() : innerConeAngle(0.0), outerConeAngle(0.7853981634) {} + DEFAULT_METHODS(SpotLight) + bool operator==(const SpotLight &) const; + + ExtensionMap extensions; + Value extras; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; +}; + +struct Light { + std::string name; + std::vector color; + double intensity{1.0}; + std::string type; + double range{0.0}; // 0.0 = inifinite + SpotLight spot; + + Light() : intensity(1.0), range(0.0) {} + DEFAULT_METHODS(Light) + + bool operator==(const Light &) const; + + ExtensionMap extensions; + Value extras; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; +}; + +class Model { + public: + Model() = default; + DEFAULT_METHODS(Model) + + bool operator==(const Model &) const; + + std::vector accessors; + std::vector animations; + std::vector buffers; + std::vector bufferViews; + std::vector materials; + std::vector meshes; + std::vector nodes; + std::vector textures; + std::vector images; + std::vector skins; + std::vector samplers; + std::vector cameras; + std::vector scenes; + std::vector lights; + + int defaultScene = -1; + std::vector extensionsUsed; + std::vector extensionsRequired; + + Asset asset; + + Value extras; + ExtensionMap extensions; + + // Filled when SetStoreOriginalJSONForExtrasAndExtensions is enabled. + std::string extras_json_string; + std::string extensions_json_string; +}; + +enum SectionCheck { + NO_REQUIRE = 0x00, + REQUIRE_VERSION = 0x01, + REQUIRE_SCENE = 0x02, + REQUIRE_SCENES = 0x04, + REQUIRE_NODES = 0x08, + REQUIRE_ACCESSORS = 0x10, + REQUIRE_BUFFERS = 0x20, + REQUIRE_BUFFER_VIEWS = 0x40, + REQUIRE_ALL = 0x7f +}; + +/// +/// LoadImageDataFunction type. Signature for custom image loading callbacks. +/// +typedef bool (*LoadImageDataFunction)(Image *, const int, std::string *, + std::string *, int, int, + const unsigned char *, int, + void *user_pointer); + +/// +/// WriteImageDataFunction type. Signature for custom image writing callbacks. +/// +typedef bool (*WriteImageDataFunction)(const std::string *, const std::string *, + Image *, bool, void *); + +#ifndef TINYGLTF_NO_STB_IMAGE +// Declaration of default image loader callback +bool LoadImageData(Image *image, const int image_idx, std::string *err, + std::string *warn, int req_width, int req_height, + const unsigned char *bytes, int size, void *); +#endif + +#ifndef TINYGLTF_NO_STB_IMAGE_WRITE +// Declaration of default image writer callback +bool WriteImageData(const std::string *basepath, const std::string *filename, + Image *image, bool embedImages, void *); +#endif + +/// +/// FilExistsFunction type. Signature for custom filesystem callbacks. +/// +typedef bool (*FileExistsFunction)(const std::string &abs_filename, void *); + +/// +/// ExpandFilePathFunction type. Signature for custom filesystem callbacks. +/// +typedef std::string (*ExpandFilePathFunction)(const std::string &, void *); + +/// +/// ReadWholeFileFunction type. Signature for custom filesystem callbacks. +/// +typedef bool (*ReadWholeFileFunction)(std::vector *, + std::string *, const std::string &, + void *); + +/// +/// WriteWholeFileFunction type. Signature for custom filesystem callbacks. +/// +typedef bool (*WriteWholeFileFunction)(std::string *, const std::string &, + const std::vector &, + void *); + +/// +/// A structure containing all required filesystem callbacks and a pointer to +/// their user data. +/// +struct FsCallbacks { + FileExistsFunction FileExists; + ExpandFilePathFunction ExpandFilePath; + ReadWholeFileFunction ReadWholeFile; + WriteWholeFileFunction WriteWholeFile; + + void *user_data; // An argument that is passed to all fs callbacks +}; + +#ifndef TINYGLTF_NO_FS +// Declaration of default filesystem callbacks + +bool FileExists(const std::string &abs_filename, void *); + +/// +/// Expand file path(e.g. `~` to home directory on posix, `%APPDATA%` to +/// `C:\\Users\\tinygltf\\AppData`) +/// +/// @param[in] filepath File path string. Assume UTF-8 +/// @param[in] userdata User data. Set to `nullptr` if you don't need it. +/// +std::string ExpandFilePath(const std::string &filepath, void *userdata); + +bool ReadWholeFile(std::vector *out, std::string *err, + const std::string &filepath, void *); + +bool WriteWholeFile(std::string *err, const std::string &filepath, + const std::vector &contents, void *); +#endif + +/// +/// glTF Parser/Serialier context. +/// +class TinyGLTF { + public: +#ifdef __clang__ + #pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wc++98-compat" +#endif + + TinyGLTF() : bin_data_(nullptr), bin_size_(0), is_binary_(false) {} + +#ifdef __clang__ +#pragma clang diagnostic pop +#endif + + ~TinyGLTF() {} + + /// + /// Loads glTF ASCII asset from a file. + /// Set warning message to `warn` for example it fails to load asserts. + /// Returns false and set error string to `err` if there's an error. + /// + bool LoadASCIIFromFile(Model *model, std::string *err, std::string *warn, + const std::string &filename, + unsigned int check_sections = REQUIRE_VERSION); + + /// + /// Loads glTF ASCII asset from string(memory). + /// `length` = strlen(str); + /// Set warning message to `warn` for example it fails to load asserts. + /// Returns false and set error string to `err` if there's an error. + /// + bool LoadASCIIFromString(Model *model, std::string *err, std::string *warn, + const char *str, const unsigned int length, + const std::string &base_dir, + unsigned int check_sections = REQUIRE_VERSION); + + /// + /// Loads glTF binary asset from a file. + /// Set warning message to `warn` for example it fails to load asserts. + /// Returns false and set error string to `err` if there's an error. + /// + bool LoadBinaryFromFile(Model *model, std::string *err, std::string *warn, + const std::string &filename, + unsigned int check_sections = REQUIRE_VERSION); + + /// + /// Loads glTF binary asset from memory. + /// `length` = strlen(str); + /// Set warning message to `warn` for example it fails to load asserts. + /// Returns false and set error string to `err` if there's an error. + /// + bool LoadBinaryFromMemory(Model *model, std::string *err, std::string *warn, + const unsigned char *bytes, + const unsigned int length, + const std::string &base_dir = "", + unsigned int check_sections = REQUIRE_VERSION); + + /// + /// Write glTF to stream, buffers and images will be embeded + /// + bool WriteGltfSceneToStream(Model *model, std::ostream &stream, + bool prettyPrint, bool writeBinary); + + /// + /// Write glTF to file. + /// + bool WriteGltfSceneToFile(Model *model, const std::string &filename, + bool embedImages, bool embedBuffers, + bool prettyPrint, bool writeBinary); + + /// + /// Set callback to use for loading image data + /// + void SetImageLoader(LoadImageDataFunction LoadImageData, void *user_data); + + /// + /// Unset(remove) callback of loading image data + /// + void RemoveImageLoader(); + + /// + /// Set callback to use for writing image data + /// + void SetImageWriter(WriteImageDataFunction WriteImageData, void *user_data); + + /// + /// Set callbacks to use for filesystem (fs) access and their user data + /// + void SetFsCallbacks(FsCallbacks callbacks); + + /// + /// Set serializing default values(default = false). + /// When true, default values are force serialized to .glTF. + /// This may be helpfull if you want to serialize a full description of glTF + /// data. + /// + /// TODO(LTE): Supply parsing option as function arguments to + /// `LoadASCIIFromFile()` and others, not by a class method + /// + void SetSerializeDefaultValues(const bool enabled) { + serialize_default_values_ = enabled; + } + + bool GetSerializeDefaultValues() const { return serialize_default_values_; } + + /// + /// Store original JSON string for `extras` and `extensions`. + /// This feature will be useful when the user want to reconstruct custom data + /// structure from JSON string. + /// + void SetStoreOriginalJSONForExtrasAndExtensions(const bool enabled) { + store_original_json_for_extras_and_extensions_ = enabled; + } + + bool GetStoreOriginalJSONForExtrasAndExtensions() const { + return store_original_json_for_extras_and_extensions_; + } + + /// + /// Specify whether preserve image channales when loading images or not. + /// (Not effective when the user suppy their own LoadImageData callbacks) + /// + void SetPreserveImageChannels(bool onoff) { + preserve_image_channels_ = onoff; + } + + bool GetPreserveImageChannels() const { return preserve_image_channels_; } + + private: + /// + /// Loads glTF asset from string(memory). + /// `length` = strlen(str); + /// Set warning message to `warn` for example it fails to load asserts + /// Returns false and set error string to `err` if there's an error. + /// + bool LoadFromString(Model *model, std::string *err, std::string *warn, + const char *str, const unsigned int length, + const std::string &base_dir, unsigned int check_sections); + + const unsigned char *bin_data_ = nullptr; + size_t bin_size_ = 0; + bool is_binary_ = false; + + bool serialize_default_values_ = false; ///< Serialize default values? + + bool store_original_json_for_extras_and_extensions_ = false; + + bool preserve_image_channels_ = false; /// Default false(expand channels to + /// RGBA) for backward compatibility. + + FsCallbacks fs = { +#ifndef TINYGLTF_NO_FS + &tinygltf::FileExists, &tinygltf::ExpandFilePath, + &tinygltf::ReadWholeFile, &tinygltf::WriteWholeFile, + + nullptr // Fs callback user data +#else + nullptr, nullptr, nullptr, nullptr, + + nullptr // Fs callback user data +#endif + }; + + LoadImageDataFunction LoadImageData = +#ifndef TINYGLTF_NO_STB_IMAGE + &tinygltf::LoadImageData; +#else + nullptr; +#endif + void *load_image_user_data_{nullptr}; + bool user_image_loader_{false}; + + WriteImageDataFunction WriteImageData = +#ifndef TINYGLTF_NO_STB_IMAGE_WRITE + &tinygltf::WriteImageData; +#else + nullptr; +#endif + void *write_image_user_data_{nullptr}; +}; + +#ifdef __clang__ +#pragma clang diagnostic pop // -Wpadded +#endif + +} // namespace tinygltf + +#endif // TINY_GLTF_H_ + +#if defined(TINYGLTF_IMPLEMENTATION) || defined(__INTELLISENSE__) +#include +//#include +#ifndef TINYGLTF_NO_FS +#include +#include +#endif +#include + +#ifdef __clang__ +// Disable some warnings for external files. +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wfloat-equal" +#pragma clang diagnostic ignored "-Wexit-time-destructors" +#pragma clang diagnostic ignored "-Wconversion" +#pragma clang diagnostic ignored "-Wold-style-cast" +#pragma clang diagnostic ignored "-Wglobal-constructors" +#if __has_warning("-Wreserved-id-macro") +#pragma clang diagnostic ignored "-Wreserved-id-macro" +#endif +#pragma clang diagnostic ignored "-Wdisabled-macro-expansion" +#pragma clang diagnostic ignored "-Wpadded" +#pragma clang diagnostic ignored "-Wc++98-compat" +#pragma clang diagnostic ignored "-Wc++98-compat-pedantic" +#pragma clang diagnostic ignored "-Wdocumentation-unknown-command" +#pragma clang diagnostic ignored "-Wswitch-enum" +#pragma clang diagnostic ignored "-Wimplicit-fallthrough" +#pragma clang diagnostic ignored "-Wweak-vtables" +#pragma clang diagnostic ignored "-Wcovered-switch-default" +#if __has_warning("-Wdouble-promotion") +#pragma clang diagnostic ignored "-Wdouble-promotion" +#endif +#if __has_warning("-Wcomma") +#pragma clang diagnostic ignored "-Wcomma" +#endif +#if __has_warning("-Wzero-as-null-pointer-constant") +#pragma clang diagnostic ignored "-Wzero-as-null-pointer-constant" +#endif +#if __has_warning("-Wcast-qual") +#pragma clang diagnostic ignored "-Wcast-qual" +#endif +#if __has_warning("-Wmissing-variable-declarations") +#pragma clang diagnostic ignored "-Wmissing-variable-declarations" +#endif +#if __has_warning("-Wmissing-prototypes") +#pragma clang diagnostic ignored "-Wmissing-prototypes" +#endif +#if __has_warning("-Wcast-align") +#pragma clang diagnostic ignored "-Wcast-align" +#endif +#if __has_warning("-Wnewline-eof") +#pragma clang diagnostic ignored "-Wnewline-eof" +#endif +#if __has_warning("-Wunused-parameter") +#pragma clang diagnostic ignored "-Wunused-parameter" +#endif +#if __has_warning("-Wmismatched-tags") +#pragma clang diagnostic ignored "-Wmismatched-tags" +#endif +#if __has_warning("-Wextra-semi-stmt") +#pragma clang diagnostic ignored "-Wextra-semi-stmt" +#endif +#endif + +// Disable GCC warnigs +#ifdef __GNUC__ +#pragma GCC diagnostic push +#pragma GCC diagnostic ignored "-Wtype-limits" +#endif // __GNUC__ + +#ifndef TINYGLTF_NO_INCLUDE_JSON +#ifndef TINYGLTF_USE_RAPIDJSON +#include "json.hpp" +#else +#ifndef TINYGLTF_NO_INCLUDE_RAPIDJSON +#include "document.h" +#include "prettywriter.h" +#include "rapidjson.h" +#include "stringbuffer.h" +#include "writer.h" +#endif +#endif +#endif + +#ifdef TINYGLTF_ENABLE_DRACO +#include "draco/compression/decode.h" +#include "draco/core/decoder_buffer.h" +#endif + +#ifndef TINYGLTF_NO_STB_IMAGE +#ifndef TINYGLTF_NO_INCLUDE_STB_IMAGE +#include "stb_image.h" +#endif +#endif + +#ifndef TINYGLTF_NO_STB_IMAGE_WRITE +#ifndef TINYGLTF_NO_INCLUDE_STB_IMAGE_WRITE +#include "stb_image_write.h" +#endif +#endif + +#ifdef __clang__ +#pragma clang diagnostic pop +#endif + +#ifdef __GNUC__ +#pragma GCC diagnostic pop +#endif + +#ifdef _WIN32 + +// issue 143. +// Define NOMINMAX to avoid min/max defines, +// but undef it after included windows.h +#ifndef NOMINMAX +#define TINYGLTF_INTERNAL_NOMINMAX +#define NOMINMAX +#endif + +#ifndef WIN32_LEAN_AND_MEAN +#define WIN32_LEAN_AND_MEAN +#define TINYGLTF_INTERNAL_WIN32_LEAN_AND_MEAN +#endif +#include // include API for expanding a file path + +#ifdef TINYGLTF_INTERNAL_WIN32_LEAN_AND_MEAN +#undef WIN32_LEAN_AND_MEAN +#endif + +#if defined(TINYGLTF_INTERNAL_NOMINMAX) +#undef NOMINMAX +#endif + +#if defined(__GLIBCXX__) // mingw + +#include // _O_RDONLY + +#include // fstream (all sorts of IO stuff) + stdio_filebuf (=streambuf) + +#endif + +#elif !defined(__ANDROID__) && !defined(__OpenBSD__) +#include +#endif + +#if defined(__sparcv9) || defined(__powerpc__) +// Big endian +#else +#if (__BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) || MINIZ_X86_OR_X64_CPU +#define TINYGLTF_LITTLE_ENDIAN 1 +#endif +#endif + +namespace { +#ifdef TINYGLTF_USE_RAPIDJSON + +#ifdef TINYGLTF_USE_RAPIDJSON_CRTALLOCATOR +// This uses the RapidJSON CRTAllocator. It is thread safe and multiple +// documents may be active at once. +using json = + rapidjson::GenericValue, rapidjson::CrtAllocator>; +using json_const_iterator = json::ConstMemberIterator; +using json_const_array_iterator = json const *; +using JsonDocument = + rapidjson::GenericDocument, rapidjson::CrtAllocator>; +rapidjson::CrtAllocator s_CrtAllocator; // stateless and thread safe +rapidjson::CrtAllocator &GetAllocator() { return s_CrtAllocator; } +#else +// This uses the default RapidJSON MemoryPoolAllocator. It is very fast, but +// not thread safe. Only a single JsonDocument may be active at any one time, +// meaning only a single gltf load/save can be active any one time. +using json = rapidjson::Value; +using json_const_iterator = json::ConstMemberIterator; +using json_const_array_iterator = json const *; +rapidjson::Document *s_pActiveDocument = nullptr; +rapidjson::Document::AllocatorType &GetAllocator() { + assert(s_pActiveDocument); // Root json node must be JsonDocument type + return s_pActiveDocument->GetAllocator(); +} + +#ifdef __clang__ +#pragma clang diagnostic push +// Suppress JsonDocument(JsonDocument &&rhs) noexcept +#pragma clang diagnostic ignored "-Wunused-member-function" +#endif + +struct JsonDocument : public rapidjson::Document { + JsonDocument() { + assert(s_pActiveDocument == + nullptr); // When using default allocator, only one document can be + // active at a time, if you need multiple active at once, + // define TINYGLTF_USE_RAPIDJSON_CRTALLOCATOR + s_pActiveDocument = this; + } + JsonDocument(const JsonDocument &) = delete; + JsonDocument(JsonDocument &&rhs) noexcept + : rapidjson::Document(std::move(rhs)) { + s_pActiveDocument = this; + rhs.isNil = true; + } + ~JsonDocument() { + if (!isNil) { + s_pActiveDocument = nullptr; + } + } + + private: + bool isNil = false; +}; + +#ifdef __clang__ +#pragma clang diagnostic pop +#endif + +#endif // TINYGLTF_USE_RAPIDJSON_CRTALLOCATOR + +#else +using nlohmann::json; +using json_const_iterator = json::const_iterator; +using json_const_array_iterator = json_const_iterator; +using JsonDocument = json; +#endif + +void JsonParse(JsonDocument &doc, const char *str, size_t length, + bool throwExc = false) { +#ifdef TINYGLTF_USE_RAPIDJSON + (void)throwExc; + doc.Parse(str, length); +#else + doc = json::parse(str, str + length, nullptr, throwExc); +#endif +} +} // namespace + +#ifdef __APPLE__ +#include "TargetConditionals.h" +#endif + +#ifdef __clang__ +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wc++98-compat" +#endif + +namespace tinygltf { + +/// +/// Internal LoadImageDataOption struct. +/// This struct is passed through `user_pointer` in LoadImageData. +/// The struct is not passed when the user supply their own LoadImageData +/// callbacks. +/// +struct LoadImageDataOption { + // true: preserve image channels(e.g. load as RGB image if the image has RGB + // channels) default `false`(channels are expanded to RGBA for backward + // compatiblity). + bool preserve_channels{false}; +}; + +// Equals function for Value, for recursivity +static bool Equals(const tinygltf::Value &one, const tinygltf::Value &other) { + if (one.Type() != other.Type()) return false; + + switch (one.Type()) { + case NULL_TYPE: + return true; + case BOOL_TYPE: + return one.Get() == other.Get(); + case REAL_TYPE: + return TINYGLTF_DOUBLE_EQUAL(one.Get(), other.Get()); + case INT_TYPE: + return one.Get() == other.Get(); + case OBJECT_TYPE: { + auto oneObj = one.Get(); + auto otherObj = other.Get(); + if (oneObj.size() != otherObj.size()) return false; + for (auto &it : oneObj) { + auto otherIt = otherObj.find(it.first); + if (otherIt == otherObj.end()) return false; + + if (!Equals(it.second, otherIt->second)) return false; + } + return true; + } + case ARRAY_TYPE: { + if (one.Size() != other.Size()) return false; + for (int i = 0; i < int(one.Size()); ++i) + if (!Equals(one.Get(i), other.Get(i))) return false; + return true; + } + case STRING_TYPE: + return one.Get() == other.Get(); + case BINARY_TYPE: + return one.Get >() == + other.Get >(); + default: { + // unhandled type + return false; + } + } +} + +// Equals function for std::vector using TINYGLTF_DOUBLE_EPSILON +static bool Equals(const std::vector &one, + const std::vector &other) { + if (one.size() != other.size()) return false; + for (int i = 0; i < int(one.size()); ++i) { + if (!TINYGLTF_DOUBLE_EQUAL(one[size_t(i)], other[size_t(i)])) return false; + } + return true; +} + +bool Accessor::operator==(const Accessor &other) const { + return this->bufferView == other.bufferView && + this->byteOffset == other.byteOffset && + this->componentType == other.componentType && + this->count == other.count && this->extensions == other.extensions && + this->extras == other.extras && + Equals(this->maxValues, other.maxValues) && + Equals(this->minValues, other.minValues) && this->name == other.name && + this->normalized == other.normalized && this->type == other.type; +} +bool Animation::operator==(const Animation &other) const { + return this->channels == other.channels && + this->extensions == other.extensions && this->extras == other.extras && + this->name == other.name && this->samplers == other.samplers; +} +bool AnimationChannel::operator==(const AnimationChannel &other) const { + return this->extensions == other.extensions && this->extras == other.extras && + this->target_node == other.target_node && + this->target_path == other.target_path && + this->sampler == other.sampler; +} +bool AnimationSampler::operator==(const AnimationSampler &other) const { + return this->extras == other.extras && this->extensions == other.extensions && + this->input == other.input && + this->interpolation == other.interpolation && + this->output == other.output; +} +bool Asset::operator==(const Asset &other) const { + return this->copyright == other.copyright && + this->extensions == other.extensions && this->extras == other.extras && + this->generator == other.generator && + this->minVersion == other.minVersion && this->version == other.version; +} +bool Buffer::operator==(const Buffer &other) const { + return this->data == other.data && this->extensions == other.extensions && + this->extras == other.extras && this->name == other.name && + this->uri == other.uri; +} +bool BufferView::operator==(const BufferView &other) const { + return this->buffer == other.buffer && this->byteLength == other.byteLength && + this->byteOffset == other.byteOffset && + this->byteStride == other.byteStride && this->name == other.name && + this->target == other.target && this->extensions == other.extensions && + this->extras == other.extras && + this->dracoDecoded == other.dracoDecoded; +} +bool Camera::operator==(const Camera &other) const { + return this->name == other.name && this->extensions == other.extensions && + this->extras == other.extras && + this->orthographic == other.orthographic && + this->perspective == other.perspective && this->type == other.type; +} +bool Image::operator==(const Image &other) const { + return this->bufferView == other.bufferView && + this->component == other.component && + this->extensions == other.extensions && this->extras == other.extras && + this->height == other.height && this->image == other.image && + this->mimeType == other.mimeType && this->name == other.name && + this->uri == other.uri && this->width == other.width; +} +bool Light::operator==(const Light &other) const { + return Equals(this->color, other.color) && this->name == other.name && + this->type == other.type; +} +bool Material::operator==(const Material &other) const { + return (this->pbrMetallicRoughness == other.pbrMetallicRoughness) && + (this->normalTexture == other.normalTexture) && + (this->occlusionTexture == other.occlusionTexture) && + (this->emissiveTexture == other.emissiveTexture) && + Equals(this->emissiveFactor, other.emissiveFactor) && + (this->alphaMode == other.alphaMode) && + TINYGLTF_DOUBLE_EQUAL(this->alphaCutoff, other.alphaCutoff) && + (this->doubleSided == other.doubleSided) && + (this->extensions == other.extensions) && + (this->extras == other.extras) && (this->values == other.values) && + (this->additionalValues == other.additionalValues) && + (this->name == other.name); +} +bool Mesh::operator==(const Mesh &other) const { + return this->extensions == other.extensions && this->extras == other.extras && + this->name == other.name && Equals(this->weights, other.weights) && + this->primitives == other.primitives; +} +bool Model::operator==(const Model &other) const { + return this->accessors == other.accessors && + this->animations == other.animations && this->asset == other.asset && + this->buffers == other.buffers && + this->bufferViews == other.bufferViews && + this->cameras == other.cameras && + this->defaultScene == other.defaultScene && + this->extensions == other.extensions && + this->extensionsRequired == other.extensionsRequired && + this->extensionsUsed == other.extensionsUsed && + this->extras == other.extras && this->images == other.images && + this->lights == other.lights && this->materials == other.materials && + this->meshes == other.meshes && this->nodes == other.nodes && + this->samplers == other.samplers && this->scenes == other.scenes && + this->skins == other.skins && this->textures == other.textures; +} +bool Node::operator==(const Node &other) const { + return this->camera == other.camera && this->children == other.children && + this->extensions == other.extensions && this->extras == other.extras && + Equals(this->matrix, other.matrix) && this->mesh == other.mesh && + this->name == other.name && Equals(this->rotation, other.rotation) && + Equals(this->scale, other.scale) && this->skin == other.skin && + Equals(this->translation, other.translation) && + Equals(this->weights, other.weights); +} +bool SpotLight::operator==(const SpotLight &other) const { + return this->extensions == other.extensions && this->extras == other.extras && + TINYGLTF_DOUBLE_EQUAL(this->innerConeAngle, other.innerConeAngle) && + TINYGLTF_DOUBLE_EQUAL(this->outerConeAngle, other.outerConeAngle); +} +bool OrthographicCamera::operator==(const OrthographicCamera &other) const { + return this->extensions == other.extensions && this->extras == other.extras && + TINYGLTF_DOUBLE_EQUAL(this->xmag, other.xmag) && + TINYGLTF_DOUBLE_EQUAL(this->ymag, other.ymag) && + TINYGLTF_DOUBLE_EQUAL(this->zfar, other.zfar) && + TINYGLTF_DOUBLE_EQUAL(this->znear, other.znear); +} +bool Parameter::operator==(const Parameter &other) const { + if (this->bool_value != other.bool_value || + this->has_number_value != other.has_number_value) + return false; + + if (!TINYGLTF_DOUBLE_EQUAL(this->number_value, other.number_value)) + return false; + + if (this->json_double_value.size() != other.json_double_value.size()) + return false; + for (auto &it : this->json_double_value) { + auto otherIt = other.json_double_value.find(it.first); + if (otherIt == other.json_double_value.end()) return false; + + if (!TINYGLTF_DOUBLE_EQUAL(it.second, otherIt->second)) return false; + } + + if (!Equals(this->number_array, other.number_array)) return false; + + if (this->string_value != other.string_value) return false; + + return true; +} +bool PerspectiveCamera::operator==(const PerspectiveCamera &other) const { + return TINYGLTF_DOUBLE_EQUAL(this->aspectRatio, other.aspectRatio) && + this->extensions == other.extensions && this->extras == other.extras && + TINYGLTF_DOUBLE_EQUAL(this->yfov, other.yfov) && + TINYGLTF_DOUBLE_EQUAL(this->zfar, other.zfar) && + TINYGLTF_DOUBLE_EQUAL(this->znear, other.znear); +} +bool Primitive::operator==(const Primitive &other) const { + return this->attributes == other.attributes && this->extras == other.extras && + this->indices == other.indices && this->material == other.material && + this->mode == other.mode && this->targets == other.targets; +} +bool Sampler::operator==(const Sampler &other) const { + return this->extensions == other.extensions && this->extras == other.extras && + this->magFilter == other.magFilter && + this->minFilter == other.minFilter && this->name == other.name && + this->wrapS == other.wrapS && + this->wrapT == other.wrapT; + + //this->wrapR == other.wrapR +} +bool Scene::operator==(const Scene &other) const { + return this->extensions == other.extensions && this->extras == other.extras && + this->name == other.name && this->nodes == other.nodes; +} +bool Skin::operator==(const Skin &other) const { + return this->extensions == other.extensions && this->extras == other.extras && + this->inverseBindMatrices == other.inverseBindMatrices && + this->joints == other.joints && this->name == other.name && + this->skeleton == other.skeleton; +} +bool Texture::operator==(const Texture &other) const { + return this->extensions == other.extensions && this->extras == other.extras && + this->name == other.name && this->sampler == other.sampler && + this->source == other.source; +} +bool TextureInfo::operator==(const TextureInfo &other) const { + return this->extensions == other.extensions && this->extras == other.extras && + this->index == other.index && this->texCoord == other.texCoord; +} +bool NormalTextureInfo::operator==(const NormalTextureInfo &other) const { + return this->extensions == other.extensions && this->extras == other.extras && + this->index == other.index && this->texCoord == other.texCoord && + TINYGLTF_DOUBLE_EQUAL(this->scale, other.scale); +} +bool OcclusionTextureInfo::operator==(const OcclusionTextureInfo &other) const { + return this->extensions == other.extensions && this->extras == other.extras && + this->index == other.index && this->texCoord == other.texCoord && + TINYGLTF_DOUBLE_EQUAL(this->strength, other.strength); +} +bool PbrMetallicRoughness::operator==(const PbrMetallicRoughness &other) const { + return this->extensions == other.extensions && this->extras == other.extras && + (this->baseColorTexture == other.baseColorTexture) && + (this->metallicRoughnessTexture == other.metallicRoughnessTexture) && + Equals(this->baseColorFactor, other.baseColorFactor) && + TINYGLTF_DOUBLE_EQUAL(this->metallicFactor, other.metallicFactor) && + TINYGLTF_DOUBLE_EQUAL(this->roughnessFactor, other.roughnessFactor); +} +bool Value::operator==(const Value &other) const { + return Equals(*this, other); +} + +static void swap4(unsigned int *val) { +#ifdef TINYGLTF_LITTLE_ENDIAN + (void)val; +#else + unsigned int tmp = *val; + unsigned char *dst = reinterpret_cast(val); + unsigned char *src = reinterpret_cast(&tmp); + + dst[0] = src[3]; + dst[1] = src[2]; + dst[2] = src[1]; + dst[3] = src[0]; +#endif +} + +static std::string JoinPath(const std::string &path0, + const std::string &path1) { + if (path0.empty()) { + return path1; + } else { + // check '/' + char lastChar = *path0.rbegin(); + if (lastChar != '/') { + return path0 + std::string("/") + path1; + } else { + return path0 + path1; + } + } +} + +static std::string FindFile(const std::vector &paths, + const std::string &filepath, FsCallbacks *fs) { + if (fs == nullptr || fs->ExpandFilePath == nullptr || + fs->FileExists == nullptr) { + // Error, fs callback[s] missing + return std::string(); + } + + for (size_t i = 0; i < paths.size(); i++) { + std::string absPath = + fs->ExpandFilePath(JoinPath(paths[i], filepath), fs->user_data); + if (fs->FileExists(absPath, fs->user_data)) { + return absPath; + } + } + + return std::string(); +} + +static std::string GetFilePathExtension(const std::string &FileName) { + if (FileName.find_last_of(".") != std::string::npos) + return FileName.substr(FileName.find_last_of(".") + 1); + return ""; +} + +static std::string GetBaseDir(const std::string &filepath) { + if (filepath.find_last_of("/\\") != std::string::npos) + return filepath.substr(0, filepath.find_last_of("/\\")); + return ""; +} + +static std::string GetBaseFilename(const std::string &filepath) { + auto idx = filepath.find_last_of("/\\"); + if (idx != std::string::npos) + return filepath.substr(idx + 1); + return filepath; +} + +std::string base64_encode(unsigned char const *, unsigned int len); +std::string base64_decode(std::string const &s); + +/* + base64.cpp and base64.h + + Copyright (C) 2004-2008 René Nyffenegger + + This source code is provided 'as-is', without any express or implied + warranty. In no event will the author be held liable for any damages + arising from the use of this software. + + Permission is granted to anyone to use this software for any purpose, + including commercial applications, and to alter it and redistribute it + freely, subject to the following restrictions: + + 1. The origin of this source code must not be misrepresented; you must not + claim that you wrote the original source code. If you use this source code + in a product, an acknowledgment in the product documentation would be + appreciated but is not required. + + 2. Altered source versions must be plainly marked as such, and must not be + misrepresented as being the original source code. + + 3. This notice may not be removed or altered from any source distribution. + + René Nyffenegger rene.nyffenegger@adp-gmbh.ch + +*/ + +#ifdef __clang__ +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wsign-conversion" +#pragma clang diagnostic ignored "-Wconversion" +#endif + +static inline bool is_base64(unsigned char c) { + return (isalnum(c) || (c == '+') || (c == '/')); +} + +std::string base64_encode(unsigned char const *bytes_to_encode, + unsigned int in_len) { + std::string ret; + int i = 0; + int j = 0; + unsigned char char_array_3[3]; + unsigned char char_array_4[4]; + + const char *base64_chars = + "ABCDEFGHIJKLMNOPQRSTUVWXYZ" + "abcdefghijklmnopqrstuvwxyz" + "0123456789+/"; + + while (in_len--) { + char_array_3[i++] = *(bytes_to_encode++); + if (i == 3) { + char_array_4[0] = (char_array_3[0] & 0xfc) >> 2; + char_array_4[1] = + ((char_array_3[0] & 0x03) << 4) + ((char_array_3[1] & 0xf0) >> 4); + char_array_4[2] = + ((char_array_3[1] & 0x0f) << 2) + ((char_array_3[2] & 0xc0) >> 6); + char_array_4[3] = char_array_3[2] & 0x3f; + + for (i = 0; (i < 4); i++) ret += base64_chars[char_array_4[i]]; + i = 0; + } + } + + if (i) { + for (j = i; j < 3; j++) char_array_3[j] = '\0'; + + char_array_4[0] = (char_array_3[0] & 0xfc) >> 2; + char_array_4[1] = + ((char_array_3[0] & 0x03) << 4) + ((char_array_3[1] & 0xf0) >> 4); + char_array_4[2] = + ((char_array_3[1] & 0x0f) << 2) + ((char_array_3[2] & 0xc0) >> 6); + + for (j = 0; (j < i + 1); j++) ret += base64_chars[char_array_4[j]]; + + while ((i++ < 3)) ret += '='; + } + + return ret; +} + +std::string base64_decode(std::string const &encoded_string) { + int in_len = static_cast(encoded_string.size()); + int i = 0; + int j = 0; + int in_ = 0; + unsigned char char_array_4[4], char_array_3[3]; + std::string ret; + + const std::string base64_chars = + "ABCDEFGHIJKLMNOPQRSTUVWXYZ" + "abcdefghijklmnopqrstuvwxyz" + "0123456789+/"; + + while (in_len-- && (encoded_string[in_] != '=') && + is_base64(encoded_string[in_])) { + char_array_4[i++] = encoded_string[in_]; + in_++; + if (i == 4) { + for (i = 0; i < 4; i++) + char_array_4[i] = + static_cast(base64_chars.find(char_array_4[i])); + + char_array_3[0] = + (char_array_4[0] << 2) + ((char_array_4[1] & 0x30) >> 4); + char_array_3[1] = + ((char_array_4[1] & 0xf) << 4) + ((char_array_4[2] & 0x3c) >> 2); + char_array_3[2] = ((char_array_4[2] & 0x3) << 6) + char_array_4[3]; + + for (i = 0; (i < 3); i++) ret += char_array_3[i]; + i = 0; + } + } + + if (i) { + for (j = i; j < 4; j++) char_array_4[j] = 0; + + for (j = 0; j < 4; j++) + char_array_4[j] = + static_cast(base64_chars.find(char_array_4[j])); + + char_array_3[0] = (char_array_4[0] << 2) + ((char_array_4[1] & 0x30) >> 4); + char_array_3[1] = + ((char_array_4[1] & 0xf) << 4) + ((char_array_4[2] & 0x3c) >> 2); + char_array_3[2] = ((char_array_4[2] & 0x3) << 6) + char_array_4[3]; + + for (j = 0; (j < i - 1); j++) ret += char_array_3[j]; + } + + return ret; +} +#ifdef __clang__ +#pragma clang diagnostic pop +#endif + +// https://github.com/syoyo/tinygltf/issues/228 +// TODO(syoyo): Use uriparser https://uriparser.github.io/ for stricter Uri +// decoding? +// +// Uri Decoding from DLIB +// http://dlib.net/dlib/server/server_http.cpp.html +// --- dlib begin ------------------------------------------------------------ +// Copyright (C) 2003 Davis E. King (davis@dlib.net) +// License: Boost Software License +// Boost Software License - Version 1.0 - August 17th, 2003 + +// Permission is hereby granted, free of charge, to any person or organization +// obtaining a copy of the software and accompanying documentation covered by +// this license (the "Software") to use, reproduce, display, distribute, +// execute, and transmit the Software, and to prepare derivative works of the +// Software, and to permit third-parties to whom the Software is furnished to +// do so, all subject to the following: +// The copyright notices in the Software and this entire statement, including +// the above license grant, this restriction and the following disclaimer, +// must be included in all copies of the Software, in whole or in part, and +// all derivative works of the Software, unless such copies or derivative +// works are solely in the form of machine-executable object code generated by +// a source language processor. +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +// FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT. IN NO EVENT +// SHALL THE COPYRIGHT HOLDERS OR ANYONE DISTRIBUTING THE SOFTWARE BE LIABLE +// FOR ANY DAMAGES OR OTHER LIABILITY, WHETHER IN CONTRACT, TORT OR OTHERWISE, +// ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER +// DEALINGS IN THE SOFTWARE. +// +namespace dlib { + +inline unsigned char from_hex(unsigned char ch) { + if (ch <= '9' && ch >= '0') + ch -= '0'; + else if (ch <= 'f' && ch >= 'a') + ch -= 'a' - 10; + else if (ch <= 'F' && ch >= 'A') + ch -= 'A' - 10; + else + ch = 0; + return ch; +} + +static const std::string urldecode(const std::string &str) { + using namespace std; + string result; + string::size_type i; + for (i = 0; i < str.size(); ++i) { + if (str[i] == '+') { + result += ' '; + } else if (str[i] == '%' && str.size() > i + 2) { + const unsigned char ch1 = + from_hex(static_cast(str[i + 1])); + const unsigned char ch2 = + from_hex(static_cast(str[i + 2])); + const unsigned char ch = static_cast((ch1 << 4) | ch2); + result += static_cast(ch); + i += 2; + } else { + result += str[i]; + } + } + return result; +} + +} // namespace dlib +// --- dlib end -------------------------------------------------------------- + +static bool LoadExternalFile(std::vector *out, std::string *err, + std::string *warn, const std::string &filename, + const std::string &basedir, bool required, + size_t reqBytes, bool checkSize, FsCallbacks *fs) { + if (fs == nullptr || fs->FileExists == nullptr || + fs->ExpandFilePath == nullptr || fs->ReadWholeFile == nullptr) { + // This is a developer error, assert() ? + if (err) { + (*err) += "FS callback[s] not set\n"; + } + return false; + } + + std::string *failMsgOut = required ? err : warn; + + out->clear(); + + std::vector paths; + paths.push_back(basedir); + paths.push_back("."); + + std::string filepath = FindFile(paths, filename, fs); + if (filepath.empty() || filename.empty()) { + if (failMsgOut) { + (*failMsgOut) += "File not found : " + filename + "\n"; + } + return false; + } + + std::vector buf; + std::string fileReadErr; + bool fileRead = + fs->ReadWholeFile(&buf, &fileReadErr, filepath, fs->user_data); + if (!fileRead) { + if (failMsgOut) { + (*failMsgOut) += + "File read error : " + filepath + " : " + fileReadErr + "\n"; + } + return false; + } + + size_t sz = buf.size(); + if (sz == 0) { + if (failMsgOut) { + (*failMsgOut) += "File is empty : " + filepath + "\n"; + } + return false; + } + + if (checkSize) { + if (reqBytes == sz) { + out->swap(buf); + return true; + } else { + std::stringstream ss; + ss << "File size mismatch : " << filepath << ", requestedBytes " + << reqBytes << ", but got " << sz << std::endl; + if (failMsgOut) { + (*failMsgOut) += ss.str(); + } + return false; + } + } + + out->swap(buf); + return true; +} + +void TinyGLTF::SetImageLoader(LoadImageDataFunction func, void *user_data) { + LoadImageData = func; + load_image_user_data_ = user_data; + user_image_loader_ = true; +} + +void TinyGLTF::RemoveImageLoader() { + LoadImageData = +#ifndef TINYGLTF_NO_STB_IMAGE + &tinygltf::LoadImageData; +#else + nullptr; +#endif + + load_image_user_data_ = nullptr; + user_image_loader_ = false; +} + +#ifndef TINYGLTF_NO_STB_IMAGE +bool LoadImageData(Image *image, const int image_idx, std::string *err, + std::string *warn, int req_width, int req_height, + const unsigned char *bytes, int size, void *user_data) { + (void)warn; + + LoadImageDataOption option; + if (user_data) { + option = *reinterpret_cast(user_data); + } + + int w = 0, h = 0, comp = 0, req_comp = 0; + + unsigned char *data = nullptr; + + // preserve_channels true: Use channels stored in the image file. + // false: force 32-bit textures for common Vulkan compatibility. It appears + // that some GPU drivers do not support 24-bit images for Vulkan + req_comp = option.preserve_channels ? 0 : 4; + int bits = 8; + int pixel_type = TINYGLTF_COMPONENT_TYPE_UNSIGNED_BYTE; + + // It is possible that the image we want to load is a 16bit per channel image + // We are going to attempt to load it as 16bit per channel, and if it worked, + // set the image data accodingly. We are casting the returned pointer into + // unsigned char, because we are representing "bytes". But we are updating + // the Image metadata to signal that this image uses 2 bytes (16bits) per + // channel: + if (stbi_is_16_bit_from_memory(bytes, size)) { + data = reinterpret_cast( + stbi_load_16_from_memory(bytes, size, &w, &h, &comp, req_comp)); + if (data) { + bits = 16; + pixel_type = TINYGLTF_COMPONENT_TYPE_UNSIGNED_SHORT; + } + } + + // at this point, if data is still NULL, it means that the image wasn't + // 16bit per channel, we are going to load it as a normal 8bit per channel + // mage as we used to do: + // if image cannot be decoded, ignore parsing and keep it by its path + // don't break in this case + // FIXME we should only enter this function if the image is embedded. If + // image->uri references + // an image file, it should be left as it is. Image loading should not be + // mandatory (to support other formats) + if (!data) data = stbi_load_from_memory(bytes, size, &w, &h, &comp, req_comp); + if (!data) { + // NOTE: you can use `warn` instead of `err` + if (err) { + (*err) += + "Unknown image format. STB cannot decode image data for image[" + + std::to_string(image_idx) + "] name = \"" + image->name + "\".\n"; + } + return false; + } + + if ((w < 1) || (h < 1)) { + stbi_image_free(data); + if (err) { + (*err) += "Invalid image data for image[" + std::to_string(image_idx) + + "] name = \"" + image->name + "\"\n"; + } + return false; + } + + if (req_width > 0) { + if (req_width != w) { + stbi_image_free(data); + if (err) { + (*err) += "Image width mismatch for image[" + + std::to_string(image_idx) + "] name = \"" + image->name + + "\"\n"; + } + return false; + } + } + + if (req_height > 0) { + if (req_height != h) { + stbi_image_free(data); + if (err) { + (*err) += "Image height mismatch. for image[" + + std::to_string(image_idx) + "] name = \"" + image->name + + "\"\n"; + } + return false; + } + } + + if (req_comp != 0) { + // loaded data has `req_comp` channels(components) + comp = req_comp; + } + + image->width = w; + image->height = h; + image->component = comp; + image->bits = bits; + image->pixel_type = pixel_type; + image->image.resize(static_cast(w * h * comp) * size_t(bits / 8)); + std::copy(data, data + w * h * comp * (bits / 8), image->image.begin()); + stbi_image_free(data); + + return true; +} +#endif + +void TinyGLTF::SetImageWriter(WriteImageDataFunction func, void *user_data) { + WriteImageData = func; + write_image_user_data_ = user_data; +} + +#ifndef TINYGLTF_NO_STB_IMAGE_WRITE +static void WriteToMemory_stbi(void *context, void *data, int size) { + std::vector *buffer = + reinterpret_cast *>(context); + + unsigned char *pData = reinterpret_cast(data); + + buffer->insert(buffer->end(), pData, pData + size); +} + +bool WriteImageData(const std::string *basepath, const std::string *filename, + Image *image, bool embedImages, void *fsPtr) { + const std::string ext = GetFilePathExtension(*filename); + + // Write image to temporary buffer + std::string header; + std::vector data; + + if (ext == "png") { + if ((image->bits != 8) || + (image->pixel_type != TINYGLTF_COMPONENT_TYPE_UNSIGNED_BYTE)) { + // Unsupported pixel format + return false; + } + + if (!stbi_write_png_to_func(WriteToMemory_stbi, &data, image->width, + image->height, image->component, + &image->image[0], 0)) { + return false; + } + header = "data:image/png;base64,"; + } else if (ext == "jpg") { + if (!stbi_write_jpg_to_func(WriteToMemory_stbi, &data, image->width, + image->height, image->component, + &image->image[0], 100)) { + return false; + } + header = "data:image/jpeg;base64,"; + } else if (ext == "bmp") { + if (!stbi_write_bmp_to_func(WriteToMemory_stbi, &data, image->width, + image->height, image->component, + &image->image[0])) { + return false; + } + header = "data:image/bmp;base64,"; + } else if (!embedImages) { + // Error: can't output requested format to file + return false; + } + + if (embedImages) { + // Embed base64-encoded image into URI + if (data.size()) { + image->uri = + header + + base64_encode(&data[0], static_cast(data.size())); + } else { + // Throw error? + } + } else { + // Write image to disc + FsCallbacks *fs = reinterpret_cast(fsPtr); + if ((fs != nullptr) && (fs->WriteWholeFile != nullptr)) { + const std::string imagefilepath = JoinPath(*basepath, *filename); + std::string writeError; + if (!fs->WriteWholeFile(&writeError, imagefilepath, data, + fs->user_data)) { + // Could not write image file to disc; Throw error ? + return false; + } + } else { + // Throw error? + } + image->uri = *filename; + } + + return true; +} +#endif + +void TinyGLTF::SetFsCallbacks(FsCallbacks callbacks) { fs = callbacks; } + +#ifdef _WIN32 +static inline std::wstring UTF8ToWchar(const std::string &str) { + int wstr_size = + MultiByteToWideChar(CP_UTF8, 0, str.data(), (int)str.size(), nullptr, 0); + std::wstring wstr(wstr_size, 0); + MultiByteToWideChar(CP_UTF8, 0, str.data(), (int)str.size(), &wstr[0], + (int)wstr.size()); + return wstr; +} + +static inline std::string WcharToUTF8(const std::wstring &wstr) { + int str_size = WideCharToMultiByte(CP_UTF8, 0, wstr.data(), (int)wstr.size(), + nullptr, 0, NULL, NULL); + std::string str(str_size, 0); + WideCharToMultiByte(CP_UTF8, 0, wstr.data(), (int)wstr.size(), &str[0], + (int)str.size(), NULL, NULL); + return str; +} +#endif + +#ifndef TINYGLTF_NO_FS +// Default implementations of filesystem functions + +bool FileExists(const std::string &abs_filename, void *) { + bool ret; +#ifdef TINYGLTF_ANDROID_LOAD_FROM_ASSETS + if (asset_manager) { + AAsset *asset = AAssetManager_open(asset_manager, abs_filename.c_str(), + AASSET_MODE_STREAMING); + if (!asset) { + return false; + } + AAsset_close(asset); + ret = true; + } else { + return false; + } +#else +#ifdef _WIN32 +#if defined(_MSC_VER) || defined(__GLIBCXX__) + FILE *fp = nullptr; + errno_t err = _wfopen_s(&fp, UTF8ToWchar(abs_filename).c_str(), L"rb"); + if (err != 0) { + return false; + } +#else + FILE *fp = nullptr; + errno_t err = fopen_s(&fp, abs_filename.c_str(), "rb"); + if (err != 0) { + return false; + } +#endif + +#else + FILE *fp = fopen(abs_filename.c_str(), "rb"); +#endif + if (fp) { + ret = true; + fclose(fp); + } else { + ret = false; + } +#endif + + return ret; +} + +std::string ExpandFilePath(const std::string &filepath, void *) { +#ifdef _WIN32 + // Assume input `filepath` is encoded in UTF-8 + std::wstring wfilepath = UTF8ToWchar(filepath); + DWORD wlen = ExpandEnvironmentStringsW(wfilepath.c_str(), nullptr, 0); + wchar_t *wstr = new wchar_t[wlen]; + ExpandEnvironmentStringsW(wfilepath.c_str(), wstr, wlen); + + std::wstring ws(wstr); + delete[] wstr; + return WcharToUTF8(ws); + +#else + +#if defined(TARGET_OS_IPHONE) || defined(TARGET_IPHONE_SIMULATOR) || \ + defined(__ANDROID__) || defined(__EMSCRIPTEN__) || defined(__OpenBSD__) + // no expansion + std::string s = filepath; +#else + std::string s; + wordexp_t p; + + if (filepath.empty()) { + return ""; + } + + // Quote the string to keep any spaces in filepath intact. + std::string quoted_path = "\"" + filepath + "\""; + // char** w; + int ret = wordexp(quoted_path.c_str(), &p, 0); + if (ret) { + // err + s = filepath; + return s; + } + + // Use first element only. + if (p.we_wordv) { + s = std::string(p.we_wordv[0]); + wordfree(&p); + } else { + s = filepath; + } + +#endif + + return s; +#endif +} + +bool ReadWholeFile(std::vector *out, std::string *err, + const std::string &filepath, void *) { +#ifdef TINYGLTF_ANDROID_LOAD_FROM_ASSETS + if (asset_manager) { + AAsset *asset = AAssetManager_open(asset_manager, filepath.c_str(), + AASSET_MODE_STREAMING); + if (!asset) { + if (err) { + (*err) += "File open error : " + filepath + "\n"; + } + return false; + } + size_t size = AAsset_getLength(asset); + if (size == 0) { + if (err) { + (*err) += "Invalid file size : " + filepath + + " (does the path point to a directory?)"; + } + return false; + } + out->resize(size); + AAsset_read(asset, reinterpret_cast(&out->at(0)), size); + AAsset_close(asset); + return true; + } else { + if (err) { + (*err) += "No asset manager specified : " + filepath + "\n"; + } + return false; + } +#else +#ifdef _WIN32 +#if defined(__GLIBCXX__) // mingw + int file_descriptor = + _wopen(UTF8ToWchar(filepath).c_str(), _O_RDONLY | _O_BINARY); + __gnu_cxx::stdio_filebuf wfile_buf(file_descriptor, std::ios_base::in); + std::istream f(&wfile_buf); +#elif defined(_MSC_VER) || defined(_LIBCPP_VERSION) + // For libcxx, assume _LIBCPP_HAS_OPEN_WITH_WCHAR is defined to accept + // `wchar_t *` + std::ifstream f(UTF8ToWchar(filepath).c_str(), std::ifstream::binary); +#else + // Unknown compiler/runtime + std::ifstream f(filepath.c_str(), std::ifstream::binary); +#endif +#else + std::ifstream f(filepath.c_str(), std::ifstream::binary); +#endif + if (!f) { + if (err) { + (*err) += "File open error : " + filepath + "\n"; + } + return false; + } + + f.seekg(0, f.end); + size_t sz = static_cast(f.tellg()); + f.seekg(0, f.beg); + + if (int64_t(sz) < 0) { + if (err) { + (*err) += "Invalid file size : " + filepath + + " (does the path point to a directory?)"; + } + return false; + } else if (sz == 0) { + if (err) { + (*err) += "File is empty : " + filepath + "\n"; + } + return false; + } + + out->resize(sz); + f.read(reinterpret_cast(&out->at(0)), + static_cast(sz)); + + return true; +#endif +} + +bool WriteWholeFile(std::string *err, const std::string &filepath, + const std::vector &contents, void *) { +#ifdef _WIN32 +#if defined(__GLIBCXX__) // mingw + int file_descriptor = _wopen(UTF8ToWchar(filepath).c_str(), + _O_CREAT | _O_WRONLY | _O_TRUNC | _O_BINARY); + __gnu_cxx::stdio_filebuf wfile_buf( + file_descriptor, std::ios_base::out | std::ios_base::binary); + std::ostream f(&wfile_buf); +#elif defined(_MSC_VER) + std::ofstream f(UTF8ToWchar(filepath).c_str(), std::ofstream::binary); +#else // clang? + std::ofstream f(filepath.c_str(), std::ofstream::binary); +#endif +#else + std::ofstream f(filepath.c_str(), std::ofstream::binary); +#endif + if (!f) { + if (err) { + (*err) += "File open error for writing : " + filepath + "\n"; + } + return false; + } + + f.write(reinterpret_cast(&contents.at(0)), + static_cast(contents.size())); + if (!f) { + if (err) { + (*err) += "File write error: " + filepath + "\n"; + } + return false; + } + + return true; +} + +#endif // TINYGLTF_NO_FS + +static std::string MimeToExt(const std::string &mimeType) { + if (mimeType == "image/jpeg") { + return "jpg"; + } else if (mimeType == "image/png") { + return "png"; + } else if (mimeType == "image/bmp") { + return "bmp"; + } else if (mimeType == "image/gif") { + return "gif"; + } + + return ""; +} + +static void UpdateImageObject(Image &image, std::string &baseDir, int index, + bool embedImages, + WriteImageDataFunction *WriteImageData = nullptr, + void *user_data = nullptr) { + std::string filename; + std::string ext; + // If image has uri, use it it as a filename + if (image.uri.size()) { + filename = GetBaseFilename(image.uri); + ext = GetFilePathExtension(filename); + } else if (image.bufferView != -1) { + // If there's no URI and the data exists in a buffer, + // don't change properties or write images + } else if (image.name.size()) { + ext = MimeToExt(image.mimeType); + // Otherwise use name as filename + filename = image.name + "." + ext; + } else { + ext = MimeToExt(image.mimeType); + // Fallback to index of image as filename + filename = std::to_string(index) + "." + ext; + } + + // If callback is set, modify image data object + if (*WriteImageData != nullptr && !filename.empty()) { + std::string uri; + (*WriteImageData)(&baseDir, &filename, &image, embedImages, user_data); + } +} + +bool IsDataURI(const std::string &in) { + std::string header = "data:application/octet-stream;base64,"; + if (in.find(header) == 0) { + return true; + } + + header = "data:image/jpeg;base64,"; + if (in.find(header) == 0) { + return true; + } + + header = "data:image/png;base64,"; + if (in.find(header) == 0) { + return true; + } + + header = "data:image/bmp;base64,"; + if (in.find(header) == 0) { + return true; + } + + header = "data:image/gif;base64,"; + if (in.find(header) == 0) { + return true; + } + + header = "data:text/plain;base64,"; + if (in.find(header) == 0) { + return true; + } + + header = "data:application/gltf-buffer;base64,"; + if (in.find(header) == 0) { + return true; + } + + return false; +} + +bool DecodeDataURI(std::vector *out, std::string &mime_type, + const std::string &in, size_t reqBytes, bool checkSize) { + std::string header = "data:application/octet-stream;base64,"; + std::string data; + if (in.find(header) == 0) { + data = base64_decode(in.substr(header.size())); // cut mime string. + } + + if (data.empty()) { + header = "data:image/jpeg;base64,"; + if (in.find(header) == 0) { + mime_type = "image/jpeg"; + data = base64_decode(in.substr(header.size())); // cut mime string. + } + } + + if (data.empty()) { + header = "data:image/png;base64,"; + if (in.find(header) == 0) { + mime_type = "image/png"; + data = base64_decode(in.substr(header.size())); // cut mime string. + } + } + + if (data.empty()) { + header = "data:image/bmp;base64,"; + if (in.find(header) == 0) { + mime_type = "image/bmp"; + data = base64_decode(in.substr(header.size())); // cut mime string. + } + } + + if (data.empty()) { + header = "data:image/gif;base64,"; + if (in.find(header) == 0) { + mime_type = "image/gif"; + data = base64_decode(in.substr(header.size())); // cut mime string. + } + } + + if (data.empty()) { + header = "data:text/plain;base64,"; + if (in.find(header) == 0) { + mime_type = "text/plain"; + data = base64_decode(in.substr(header.size())); + } + } + + if (data.empty()) { + header = "data:application/gltf-buffer;base64,"; + if (in.find(header) == 0) { + data = base64_decode(in.substr(header.size())); + } + } + + // TODO(syoyo): Allow empty buffer? #229 + if (data.empty()) { + return false; + } + + if (checkSize) { + if (data.size() != reqBytes) { + return false; + } + out->resize(reqBytes); + } else { + out->resize(data.size()); + } + std::copy(data.begin(), data.end(), out->begin()); + return true; +} + +namespace { +bool GetInt(const json &o, int &val) { +#ifdef TINYGLTF_USE_RAPIDJSON + if (!o.IsDouble()) { + if (o.IsInt()) { + val = o.GetInt(); + return true; + } else if (o.IsUint()) { + val = static_cast(o.GetUint()); + return true; + } else if (o.IsInt64()) { + val = static_cast(o.GetInt64()); + return true; + } else if (o.IsUint64()) { + val = static_cast(o.GetUint64()); + return true; + } + } + + return false; +#else + auto type = o.type(); + + if ((type == json::value_t::number_integer) || + (type == json::value_t::number_unsigned)) { + val = static_cast(o.get()); + return true; + } + + return false; +#endif +} + +#ifdef TINYGLTF_USE_RAPIDJSON +bool GetDouble(const json &o, double &val) { + if (o.IsDouble()) { + val = o.GetDouble(); + return true; + } + + return false; +} +#endif + +bool GetNumber(const json &o, double &val) { +#ifdef TINYGLTF_USE_RAPIDJSON + if (o.IsNumber()) { + val = o.GetDouble(); + return true; + } + + return false; +#else + if (o.is_number()) { + val = o.get(); + return true; + } + + return false; +#endif +} + +bool GetString(const json &o, std::string &val) { +#ifdef TINYGLTF_USE_RAPIDJSON + if (o.IsString()) { + val = o.GetString(); + return true; + } + + return false; +#else + if (o.type() == json::value_t::string) { + val = o.get(); + return true; + } + + return false; +#endif +} + +bool IsArray(const json &o) { +#ifdef TINYGLTF_USE_RAPIDJSON + return o.IsArray(); +#else + return o.is_array(); +#endif +} + +json_const_array_iterator ArrayBegin(const json &o) { +#ifdef TINYGLTF_USE_RAPIDJSON + return o.Begin(); +#else + return o.begin(); +#endif +} + +json_const_array_iterator ArrayEnd(const json &o) { +#ifdef TINYGLTF_USE_RAPIDJSON + return o.End(); +#else + return o.end(); +#endif +} + +bool IsObject(const json &o) { +#ifdef TINYGLTF_USE_RAPIDJSON + return o.IsObject(); +#else + return o.is_object(); +#endif +} + +json_const_iterator ObjectBegin(const json &o) { +#ifdef TINYGLTF_USE_RAPIDJSON + return o.MemberBegin(); +#else + return o.begin(); +#endif +} + +json_const_iterator ObjectEnd(const json &o) { +#ifdef TINYGLTF_USE_RAPIDJSON + return o.MemberEnd(); +#else + return o.end(); +#endif +} + +// Making this a const char* results in a pointer to a temporary when +// TINYGLTF_USE_RAPIDJSON is off. +std::string GetKey(json_const_iterator &it) { +#ifdef TINYGLTF_USE_RAPIDJSON + return it->name.GetString(); +#else + return it.key().c_str(); +#endif +} + +bool FindMember(const json &o, const char *member, json_const_iterator &it) { +#ifdef TINYGLTF_USE_RAPIDJSON + if (!o.IsObject()) { + return false; + } + it = o.FindMember(member); + return it != o.MemberEnd(); +#else + it = o.find(member); + return it != o.end(); +#endif +} + +const json &GetValue(json_const_iterator &it) { +#ifdef TINYGLTF_USE_RAPIDJSON + return it->value; +#else + return it.value(); +#endif +} + +std::string JsonToString(const json &o, int spacing = -1) { +#ifdef TINYGLTF_USE_RAPIDJSON + using namespace rapidjson; + StringBuffer buffer; + if (spacing == -1) { + Writer writer(buffer); + o.Accept(writer); + } else { + PrettyWriter writer(buffer); + writer.SetIndent(' ', uint32_t(spacing)); + o.Accept(writer); + } + return buffer.GetString(); +#else + return o.dump(spacing); +#endif +} + +} // namespace + +static bool ParseJsonAsValue(Value *ret, const json &o) { + Value val{}; +#ifdef TINYGLTF_USE_RAPIDJSON + using rapidjson::Type; + switch (o.GetType()) { + case Type::kObjectType: { + Value::Object value_object; + for (auto it = o.MemberBegin(); it != o.MemberEnd(); ++it) { + Value entry; + ParseJsonAsValue(&entry, it->value); + if (entry.Type() != NULL_TYPE) + value_object.emplace(GetKey(it), std::move(entry)); + } + if (value_object.size() > 0) val = Value(std::move(value_object)); + } break; + case Type::kArrayType: { + Value::Array value_array; + value_array.reserve(o.Size()); + for (auto it = o.Begin(); it != o.End(); ++it) { + Value entry; + ParseJsonAsValue(&entry, *it); + if (entry.Type() != NULL_TYPE) + value_array.emplace_back(std::move(entry)); + } + if (value_array.size() > 0) val = Value(std::move(value_array)); + } break; + case Type::kStringType: + val = Value(std::string(o.GetString())); + break; + case Type::kFalseType: + case Type::kTrueType: + val = Value(o.GetBool()); + break; + case Type::kNumberType: + if (!o.IsDouble()) { + int i = 0; + GetInt(o, i); + val = Value(i); + } else { + double d = 0.0; + GetDouble(o, d); + val = Value(d); + } + break; + case Type::kNullType: + break; + // all types are covered, so no `case default` + } +#else + switch (o.type()) { + case json::value_t::object: { + Value::Object value_object; + for (auto it = o.begin(); it != o.end(); it++) { + Value entry; + ParseJsonAsValue(&entry, it.value()); + if (entry.Type() != NULL_TYPE) + value_object.emplace(it.key(), std::move(entry)); + } + if (value_object.size() > 0) val = Value(std::move(value_object)); + } break; + case json::value_t::array: { + Value::Array value_array; + value_array.reserve(o.size()); + for (auto it = o.begin(); it != o.end(); it++) { + Value entry; + ParseJsonAsValue(&entry, it.value()); + if (entry.Type() != NULL_TYPE) + value_array.emplace_back(std::move(entry)); + } + if (value_array.size() > 0) val = Value(std::move(value_array)); + } break; + case json::value_t::string: + val = Value(o.get()); + break; + case json::value_t::boolean: + val = Value(o.get()); + break; + case json::value_t::number_integer: + case json::value_t::number_unsigned: + val = Value(static_cast(o.get())); + break; + case json::value_t::number_float: + val = Value(o.get()); + break; + case json::value_t::null: + case json::value_t::discarded: + case json::value_t::binary: + // default: + break; + } +#endif + if (ret) *ret = std::move(val); + + return val.Type() != NULL_TYPE; +} + +static bool ParseExtrasProperty(Value *ret, const json &o) { + json_const_iterator it; + if (!FindMember(o, "extras", it)) { + return false; + } + + return ParseJsonAsValue(ret, GetValue(it)); +} + +static bool ParseBooleanProperty(bool *ret, std::string *err, const json &o, + const std::string &property, + const bool required, + const std::string &parent_node = "") { + json_const_iterator it; + if (!FindMember(o, property.c_str(), it)) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is missing"; + if (!parent_node.empty()) { + (*err) += " in " + parent_node; + } + (*err) += ".\n"; + } + } + return false; + } + + auto &value = GetValue(it); + + bool isBoolean; + bool boolValue = false; +#ifdef TINYGLTF_USE_RAPIDJSON + isBoolean = value.IsBool(); + if (isBoolean) { + boolValue = value.GetBool(); + } +#else + isBoolean = value.is_boolean(); + if (isBoolean) { + boolValue = value.get(); + } +#endif + if (!isBoolean) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is not a bool type.\n"; + } + } + return false; + } + + if (ret) { + (*ret) = boolValue; + } + + return true; +} + +static bool ParseIntegerProperty(int *ret, std::string *err, const json &o, + const std::string &property, + const bool required, + const std::string &parent_node = "") { + json_const_iterator it; + if (!FindMember(o, property.c_str(), it)) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is missing"; + if (!parent_node.empty()) { + (*err) += " in " + parent_node; + } + (*err) += ".\n"; + } + } + return false; + } + + int intValue; + bool isInt = GetInt(GetValue(it), intValue); + if (!isInt) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is not an integer type.\n"; + } + } + return false; + } + + if (ret) { + (*ret) = intValue; + } + + return true; +} + +static bool ParseUnsignedProperty(size_t *ret, std::string *err, const json &o, + const std::string &property, + const bool required, + const std::string &parent_node = "") { + json_const_iterator it; + if (!FindMember(o, property.c_str(), it)) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is missing"; + if (!parent_node.empty()) { + (*err) += " in " + parent_node; + } + (*err) += ".\n"; + } + } + return false; + } + + auto &value = GetValue(it); + + size_t uValue = 0; + bool isUValue; +#ifdef TINYGLTF_USE_RAPIDJSON + isUValue = false; + if (value.IsUint()) { + uValue = value.GetUint(); + isUValue = true; + } else if (value.IsUint64()) { + uValue = value.GetUint64(); + isUValue = true; + } +#else + isUValue = value.is_number_unsigned(); + if (isUValue) { + uValue = value.get(); + } +#endif + if (!isUValue) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is not a positive integer.\n"; + } + } + return false; + } + + if (ret) { + (*ret) = uValue; + } + + return true; +} + +static bool ParseNumberProperty(double *ret, std::string *err, const json &o, + const std::string &property, + const bool required, + const std::string &parent_node = "") { + json_const_iterator it; + + if (!FindMember(o, property.c_str(), it)) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is missing"; + if (!parent_node.empty()) { + (*err) += " in " + parent_node; + } + (*err) += ".\n"; + } + } + return false; + } + + double numberValue; + bool isNumber = GetNumber(GetValue(it), numberValue); + + if (!isNumber) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is not a number type.\n"; + } + } + return false; + } + + if (ret) { + (*ret) = numberValue; + } + + return true; +} + +static bool ParseNumberArrayProperty(std::vector *ret, std::string *err, + const json &o, const std::string &property, + bool required, + const std::string &parent_node = "") { + json_const_iterator it; + if (!FindMember(o, property.c_str(), it)) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is missing"; + if (!parent_node.empty()) { + (*err) += " in " + parent_node; + } + (*err) += ".\n"; + } + } + return false; + } + + if (!IsArray(GetValue(it))) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is not an array"; + if (!parent_node.empty()) { + (*err) += " in " + parent_node; + } + (*err) += ".\n"; + } + } + return false; + } + + ret->clear(); + auto end = ArrayEnd(GetValue(it)); + for (auto i = ArrayBegin(GetValue(it)); i != end; ++i) { + double numberValue; + const bool isNumber = GetNumber(*i, numberValue); + if (!isNumber) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is not a number.\n"; + if (!parent_node.empty()) { + (*err) += " in " + parent_node; + } + (*err) += ".\n"; + } + } + return false; + } + ret->push_back(numberValue); + } + + return true; +} + +static bool ParseIntegerArrayProperty(std::vector *ret, std::string *err, + const json &o, + const std::string &property, + bool required, + const std::string &parent_node = "") { + json_const_iterator it; + if (!FindMember(o, property.c_str(), it)) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is missing"; + if (!parent_node.empty()) { + (*err) += " in " + parent_node; + } + (*err) += ".\n"; + } + } + return false; + } + + if (!IsArray(GetValue(it))) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is not an array"; + if (!parent_node.empty()) { + (*err) += " in " + parent_node; + } + (*err) += ".\n"; + } + } + return false; + } + + ret->clear(); + auto end = ArrayEnd(GetValue(it)); + for (auto i = ArrayBegin(GetValue(it)); i != end; ++i) { + int numberValue; + bool isNumber = GetInt(*i, numberValue); + if (!isNumber) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is not an integer type.\n"; + if (!parent_node.empty()) { + (*err) += " in " + parent_node; + } + (*err) += ".\n"; + } + } + return false; + } + ret->push_back(numberValue); + } + + return true; +} + +static bool ParseStringProperty( + std::string *ret, std::string *err, const json &o, + const std::string &property, bool required, + const std::string &parent_node = std::string()) { + json_const_iterator it; + if (!FindMember(o, property.c_str(), it)) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is missing"; + if (parent_node.empty()) { + (*err) += ".\n"; + } else { + (*err) += " in `" + parent_node + "'.\n"; + } + } + } + return false; + } + + std::string strValue; + if (!GetString(GetValue(it), strValue)) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is not a string type.\n"; + } + } + return false; + } + + if (ret) { + (*ret) = std::move(strValue); + } + + return true; +} + +static bool ParseStringIntegerProperty(std::map *ret, + std::string *err, const json &o, + const std::string &property, + bool required, + const std::string &parent = "") { + json_const_iterator it; + if (!FindMember(o, property.c_str(), it)) { + if (required) { + if (err) { + if (!parent.empty()) { + (*err) += + "'" + property + "' property is missing in " + parent + ".\n"; + } else { + (*err) += "'" + property + "' property is missing.\n"; + } + } + } + return false; + } + + const json &dict = GetValue(it); + + // Make sure we are dealing with an object / dictionary. + if (!IsObject(dict)) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is not an object.\n"; + } + } + return false; + } + + ret->clear(); + + json_const_iterator dictIt(ObjectBegin(dict)); + json_const_iterator dictItEnd(ObjectEnd(dict)); + + for (; dictIt != dictItEnd; ++dictIt) { + int intVal; + if (!GetInt(GetValue(dictIt), intVal)) { + if (required) { + if (err) { + (*err) += "'" + property + "' value is not an integer type.\n"; + } + } + return false; + } + + // Insert into the list. + (*ret)[GetKey(dictIt)] = intVal; + } + return true; +} + +static bool ParseJSONProperty(std::map *ret, + std::string *err, const json &o, + const std::string &property, bool required) { + json_const_iterator it; + if (!FindMember(o, property.c_str(), it)) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is missing. \n'"; + } + } + return false; + } + + const json &obj = GetValue(it); + + if (!IsObject(obj)) { + if (required) { + if (err) { + (*err) += "'" + property + "' property is not a JSON object.\n"; + } + } + return false; + } + + ret->clear(); + + json_const_iterator it2(ObjectBegin(obj)); + json_const_iterator itEnd(ObjectEnd(obj)); + for (; it2 != itEnd; ++it2) { + double numVal; + if (GetNumber(GetValue(it2), numVal)) + ret->emplace(std::string(GetKey(it2)), numVal); + } + + return true; +} + +static bool ParseParameterProperty(Parameter *param, std::string *err, + const json &o, const std::string &prop, + bool required) { + // A parameter value can either be a string or an array of either a boolean or + // a number. Booleans of any kind aren't supported here. Granted, it + // complicates the Parameter structure and breaks it semantically in the sense + // that the client probably works off the assumption that if the string is + // empty the vector is used, etc. Would a tagged union work? + if (ParseStringProperty(¶m->string_value, err, o, prop, false)) { + // Found string property. + return true; + } else if (ParseNumberArrayProperty(¶m->number_array, err, o, prop, + false)) { + // Found a number array. + return true; + } else if (ParseNumberProperty(¶m->number_value, err, o, prop, false)) { + return param->has_number_value = true; + } else if (ParseJSONProperty(¶m->json_double_value, err, o, prop, + false)) { + return true; + } else if (ParseBooleanProperty(¶m->bool_value, err, o, prop, false)) { + return true; + } else { + if (required) { + if (err) { + (*err) += "parameter must be a string or number / number array.\n"; + } + } + return false; + } +} + +static bool ParseExtensionsProperty(ExtensionMap *ret, std::string *err, + const json &o) { + (void)err; + + json_const_iterator it; + if (!FindMember(o, "extensions", it)) { + return false; + } + + auto &obj = GetValue(it); + if (!IsObject(obj)) { + return false; + } + ExtensionMap extensions; + json_const_iterator extIt = ObjectBegin(obj); // it.value().begin(); + json_const_iterator extEnd = ObjectEnd(obj); + for (; extIt != extEnd; ++extIt) { + auto &itObj = GetValue(extIt); + if (!IsObject(itObj)) continue; + std::string key(GetKey(extIt)); + if (!ParseJsonAsValue(&extensions[key], itObj)) { + if (!key.empty()) { + // create empty object so that an extension object is still of type + // object + extensions[key] = Value{Value::Object{}}; + } + } + } + if (ret) { + (*ret) = std::move(extensions); + } + return true; +} + +static bool ParseAsset(Asset *asset, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + ParseStringProperty(&asset->version, err, o, "version", true, "Asset"); + ParseStringProperty(&asset->generator, err, o, "generator", false, "Asset"); + ParseStringProperty(&asset->minVersion, err, o, "minVersion", false, "Asset"); + ParseStringProperty(&asset->copyright, err, o, "copyright", false, "Asset"); + + ParseExtensionsProperty(&asset->extensions, err, o); + + // Unity exporter version is added as extra here + ParseExtrasProperty(&(asset->extras), o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + asset->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + asset->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + return true; +} + +static bool ParseImage(Image *image, const int image_idx, std::string *err, + std::string *warn, const json &o, + bool store_original_json_for_extras_and_extensions, + const std::string &basedir, FsCallbacks *fs, + LoadImageDataFunction *LoadImageData = nullptr, + void *load_image_user_data = nullptr) { + // A glTF image must either reference a bufferView or an image uri + + // schema says oneOf [`bufferView`, `uri`] + // TODO(syoyo): Check the type of each parameters. + json_const_iterator it; + bool hasBufferView = FindMember(o, "bufferView", it); + bool hasURI = FindMember(o, "uri", it); + + ParseStringProperty(&image->name, err, o, "name", false); + + if (hasBufferView && hasURI) { + // Should not both defined. + if (err) { + (*err) += + "Only one of `bufferView` or `uri` should be defined, but both are " + "defined for image[" + + std::to_string(image_idx) + "] name = \"" + image->name + "\"\n"; + } + return false; + } + + if (!hasBufferView && !hasURI) { + if (err) { + (*err) += "Neither required `bufferView` nor `uri` defined for image[" + + std::to_string(image_idx) + "] name = \"" + image->name + + "\"\n"; + } + return false; + } + + ParseExtensionsProperty(&image->extensions, err, o); + ParseExtrasProperty(&image->extras, o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator eit; + if (FindMember(o, "extensions", eit)) { + image->extensions_json_string = JsonToString(GetValue(eit)); + } + } + { + json_const_iterator eit; + if (FindMember(o, "extras", eit)) { + image->extras_json_string = JsonToString(GetValue(eit)); + } + } + } + + if (hasBufferView) { + int bufferView = -1; + if (!ParseIntegerProperty(&bufferView, err, o, "bufferView", true)) { + if (err) { + (*err) += "Failed to parse `bufferView` for image[" + + std::to_string(image_idx) + "] name = \"" + image->name + + "\"\n"; + } + return false; + } + + std::string mime_type; + ParseStringProperty(&mime_type, err, o, "mimeType", false); + + int width = 0; + ParseIntegerProperty(&width, err, o, "width", false); + + int height = 0; + ParseIntegerProperty(&height, err, o, "height", false); + + // Just only save some information here. Loading actual image data from + // bufferView is done after this `ParseImage` function. + image->bufferView = bufferView; + image->mimeType = mime_type; + image->width = width; + image->height = height; + + return true; + } + + // Parse URI & Load image data. + + std::string uri; + std::string tmp_err; + if (!ParseStringProperty(&uri, &tmp_err, o, "uri", true)) { + if (err) { + (*err) += "Failed to parse `uri` for image[" + std::to_string(image_idx) + + "] name = \"" + image->name + "\".\n"; + } + return false; + } + + std::vector img; + + if (IsDataURI(uri)) { + if (!DecodeDataURI(&img, image->mimeType, uri, 0, false)) { + if (err) { + (*err) += "Failed to decode 'uri' for image[" + + std::to_string(image_idx) + "] name = [" + image->name + + "]\n"; + } + return false; + } + } else { + // Assume external file + // Keep texture path (for textures that cannot be decoded) + image->uri = uri; +#ifdef TINYGLTF_NO_EXTERNAL_IMAGE + return true; +#endif + std::string decoded_uri = dlib::urldecode(uri); + if (!LoadExternalFile(&img, err, warn, decoded_uri, basedir, + /* required */ false, /* required bytes */ 0, + /* checksize */ false, fs)) { + if (warn) { + (*warn) += "Failed to load external 'uri' for image[" + + std::to_string(image_idx) + "] name = [" + image->name + + "]\n"; + } + // If the image cannot be loaded, keep uri as image->uri. + return true; + } + + if (img.empty()) { + if (warn) { + (*warn) += "Image data is empty for image[" + + std::to_string(image_idx) + "] name = [" + image->name + + "] \n"; + } + return false; + } + } + + if (*LoadImageData == nullptr) { + if (err) { + (*err) += "No LoadImageData callback specified.\n"; + } + return false; + } + return (*LoadImageData)(image, image_idx, err, warn, 0, 0, &img.at(0), + static_cast(img.size()), load_image_user_data); +} + +static bool ParseTexture(Texture *texture, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions, + const std::string &basedir) { + (void)basedir; + int sampler = -1; + int source = -1; + ParseIntegerProperty(&sampler, err, o, "sampler", false); + + ParseIntegerProperty(&source, err, o, "source", false); + + texture->sampler = sampler; + texture->source = source; + + ParseExtensionsProperty(&texture->extensions, err, o); + ParseExtrasProperty(&texture->extras, o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + texture->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + texture->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + ParseStringProperty(&texture->name, err, o, "name", false); + + return true; +} + +static bool ParseTextureInfo( + TextureInfo *texinfo, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + if (texinfo == nullptr) { + return false; + } + + if (!ParseIntegerProperty(&texinfo->index, err, o, "index", + /* required */ true, "TextureInfo")) { + return false; + } + + ParseIntegerProperty(&texinfo->texCoord, err, o, "texCoord", false); + + ParseExtensionsProperty(&texinfo->extensions, err, o); + ParseExtrasProperty(&texinfo->extras, o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + texinfo->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + texinfo->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + return true; +} + +static bool ParseNormalTextureInfo( + NormalTextureInfo *texinfo, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + if (texinfo == nullptr) { + return false; + } + + if (!ParseIntegerProperty(&texinfo->index, err, o, "index", + /* required */ true, "NormalTextureInfo")) { + return false; + } + + ParseIntegerProperty(&texinfo->texCoord, err, o, "texCoord", false); + ParseNumberProperty(&texinfo->scale, err, o, "scale", false); + + ParseExtensionsProperty(&texinfo->extensions, err, o); + ParseExtrasProperty(&texinfo->extras, o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + texinfo->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + texinfo->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + return true; +} + +static bool ParseOcclusionTextureInfo( + OcclusionTextureInfo *texinfo, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + if (texinfo == nullptr) { + return false; + } + + if (!ParseIntegerProperty(&texinfo->index, err, o, "index", + /* required */ true, "NormalTextureInfo")) { + return false; + } + + ParseIntegerProperty(&texinfo->texCoord, err, o, "texCoord", false); + ParseNumberProperty(&texinfo->strength, err, o, "strength", false); + + ParseExtensionsProperty(&texinfo->extensions, err, o); + ParseExtrasProperty(&texinfo->extras, o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + texinfo->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + texinfo->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + return true; +} + +static bool ParseBuffer(Buffer *buffer, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions, + FsCallbacks *fs, const std::string &basedir, + bool is_binary = false, + const unsigned char *bin_data = nullptr, + size_t bin_size = 0) { + size_t byteLength; + if (!ParseUnsignedProperty(&byteLength, err, o, "byteLength", true, + "Buffer")) { + return false; + } + + // In glTF 2.0, uri is not mandatory anymore + buffer->uri.clear(); + ParseStringProperty(&buffer->uri, err, o, "uri", false, "Buffer"); + + // having an empty uri for a non embedded image should not be valid + if (!is_binary && buffer->uri.empty()) { + if (err) { + (*err) += "'uri' is missing from non binary glTF file buffer.\n"; + } + } + + json_const_iterator type; + if (FindMember(o, "type", type)) { + std::string typeStr; + if (GetString(GetValue(type), typeStr)) { + if (typeStr.compare("arraybuffer") == 0) { + // buffer.type = "arraybuffer"; + } + } + } + + if (is_binary) { + // Still binary glTF accepts external dataURI. + if (!buffer->uri.empty()) { + // First try embedded data URI. + if (IsDataURI(buffer->uri)) { + std::string mime_type; + if (!DecodeDataURI(&buffer->data, mime_type, buffer->uri, byteLength, + true)) { + if (err) { + (*err) += + "Failed to decode 'uri' : " + buffer->uri + " in Buffer\n"; + } + return false; + } + } else { + // External .bin file. + std::string decoded_uri = dlib::urldecode(buffer->uri); + if (!LoadExternalFile(&buffer->data, err, /* warn */ nullptr, + decoded_uri, basedir, /* required */ true, + byteLength, /* checkSize */ true, fs)) { + return false; + } + } + } else { + // load data from (embedded) binary data + + if ((bin_size == 0) || (bin_data == nullptr)) { + if (err) { + (*err) += "Invalid binary data in `Buffer'.\n"; + } + return false; + } + + if (byteLength > bin_size) { + if (err) { + std::stringstream ss; + ss << "Invalid `byteLength'. Must be equal or less than binary size: " + "`byteLength' = " + << byteLength << ", binary size = " << bin_size << std::endl; + (*err) += ss.str(); + } + return false; + } + + // Read buffer data + buffer->data.resize(static_cast(byteLength)); + memcpy(&(buffer->data.at(0)), bin_data, static_cast(byteLength)); + } + + } else { + if (IsDataURI(buffer->uri)) { + std::string mime_type; + if (!DecodeDataURI(&buffer->data, mime_type, buffer->uri, byteLength, + true)) { + if (err) { + (*err) += "Failed to decode 'uri' : " + buffer->uri + " in Buffer\n"; + } + return false; + } + } else { + // Assume external .bin file. + std::string decoded_uri = dlib::urldecode(buffer->uri); + if (!LoadExternalFile(&buffer->data, err, /* warn */ nullptr, decoded_uri, + basedir, /* required */ true, byteLength, + /* checkSize */ true, fs)) { + return false; + } + } + } + + ParseStringProperty(&buffer->name, err, o, "name", false); + + ParseExtensionsProperty(&buffer->extensions, err, o); + ParseExtrasProperty(&buffer->extras, o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + buffer->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + buffer->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + return true; +} + +static bool ParseBufferView( + BufferView *bufferView, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + int buffer = -1; + if (!ParseIntegerProperty(&buffer, err, o, "buffer", true, "BufferView")) { + return false; + } + + size_t byteOffset = 0; + ParseUnsignedProperty(&byteOffset, err, o, "byteOffset", false); + + size_t byteLength = 1; + if (!ParseUnsignedProperty(&byteLength, err, o, "byteLength", true, + "BufferView")) { + return false; + } + + size_t byteStride = 0; + if (!ParseUnsignedProperty(&byteStride, err, o, "byteStride", false)) { + // Spec says: When byteStride of referenced bufferView is not defined, it + // means that accessor elements are tightly packed, i.e., effective stride + // equals the size of the element. + // We cannot determine the actual byteStride until Accessor are parsed, thus + // set 0(= tightly packed) here(as done in OpenGL's VertexAttribPoiner) + byteStride = 0; + } + + if ((byteStride > 252) || ((byteStride % 4) != 0)) { + if (err) { + std::stringstream ss; + ss << "Invalid `byteStride' value. `byteStride' must be the multiple of " + "4 : " + << byteStride << std::endl; + + (*err) += ss.str(); + } + return false; + } + + int target = 0; + ParseIntegerProperty(&target, err, o, "target", false); + if ((target == TINYGLTF_TARGET_ARRAY_BUFFER) || + (target == TINYGLTF_TARGET_ELEMENT_ARRAY_BUFFER)) { + // OK + } else { + target = 0; + } + bufferView->target = target; + + ParseStringProperty(&bufferView->name, err, o, "name", false); + + ParseExtensionsProperty(&bufferView->extensions, err, o); + ParseExtrasProperty(&bufferView->extras, o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + bufferView->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + bufferView->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + bufferView->buffer = buffer; + bufferView->byteOffset = byteOffset; + bufferView->byteLength = byteLength; + bufferView->byteStride = byteStride; + return true; +} + +static bool ParseSparseAccessor(Accessor *accessor, std::string *err, + const json &o) { + accessor->sparse.isSparse = true; + + int count = 0; + if (!ParseIntegerProperty(&count, err, o, "count", true, "SparseAccessor")) { + return false; + } + + json_const_iterator indices_iterator; + json_const_iterator values_iterator; + if (!FindMember(o, "indices", indices_iterator)) { + (*err) = "the sparse object of this accessor doesn't have indices"; + return false; + } + + if (!FindMember(o, "values", values_iterator)) { + (*err) = "the sparse object ob ths accessor doesn't have values"; + return false; + } + + const json &indices_obj = GetValue(indices_iterator); + const json &values_obj = GetValue(values_iterator); + + int indices_buffer_view = 0, indices_byte_offset = 0, component_type = 0; + if (!ParseIntegerProperty(&indices_buffer_view, err, indices_obj, "bufferView", + true, "SparseAccessor")) { + return false; + } + ParseIntegerProperty(&indices_byte_offset, err, indices_obj, "byteOffset", + false); + if (!ParseIntegerProperty(&component_type, err, indices_obj, "componentType", + true, "SparseAccessor")) { + return false; + } + + int values_buffer_view = 0, values_byte_offset = 0; + if (!ParseIntegerProperty(&values_buffer_view, err, values_obj, "bufferView", + true, "SparseAccessor")) { + return false; + } + ParseIntegerProperty(&values_byte_offset, err, values_obj, "byteOffset", + false); + + accessor->sparse.count = count; + accessor->sparse.indices.bufferView = indices_buffer_view; + accessor->sparse.indices.byteOffset = indices_byte_offset; + accessor->sparse.indices.componentType = component_type; + accessor->sparse.values.bufferView = values_buffer_view; + accessor->sparse.values.byteOffset = values_byte_offset; + + return true; +} + +static bool ParseAccessor(Accessor *accessor, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + int bufferView = -1; + ParseIntegerProperty(&bufferView, err, o, "bufferView", false, "Accessor"); + + size_t byteOffset = 0; + ParseUnsignedProperty(&byteOffset, err, o, "byteOffset", false, "Accessor"); + + bool normalized = false; + ParseBooleanProperty(&normalized, err, o, "normalized", false, "Accessor"); + + size_t componentType = 0; + if (!ParseUnsignedProperty(&componentType, err, o, "componentType", true, + "Accessor")) { + return false; + } + + size_t count = 0; + if (!ParseUnsignedProperty(&count, err, o, "count", true, "Accessor")) { + return false; + } + + std::string type; + if (!ParseStringProperty(&type, err, o, "type", true, "Accessor")) { + return false; + } + + if (type.compare("SCALAR") == 0) { + accessor->type = TINYGLTF_TYPE_SCALAR; + } else if (type.compare("VEC2") == 0) { + accessor->type = TINYGLTF_TYPE_VEC2; + } else if (type.compare("VEC3") == 0) { + accessor->type = TINYGLTF_TYPE_VEC3; + } else if (type.compare("VEC4") == 0) { + accessor->type = TINYGLTF_TYPE_VEC4; + } else if (type.compare("MAT2") == 0) { + accessor->type = TINYGLTF_TYPE_MAT2; + } else if (type.compare("MAT3") == 0) { + accessor->type = TINYGLTF_TYPE_MAT3; + } else if (type.compare("MAT4") == 0) { + accessor->type = TINYGLTF_TYPE_MAT4; + } else { + std::stringstream ss; + ss << "Unsupported `type` for accessor object. Got \"" << type << "\"\n"; + if (err) { + (*err) += ss.str(); + } + return false; + } + + ParseStringProperty(&accessor->name, err, o, "name", false); + + accessor->minValues.clear(); + accessor->maxValues.clear(); + ParseNumberArrayProperty(&accessor->minValues, err, o, "min", false, + "Accessor"); + + ParseNumberArrayProperty(&accessor->maxValues, err, o, "max", false, + "Accessor"); + + accessor->count = count; + accessor->bufferView = bufferView; + accessor->byteOffset = byteOffset; + accessor->normalized = normalized; + { + if (componentType >= TINYGLTF_COMPONENT_TYPE_BYTE && + componentType <= TINYGLTF_COMPONENT_TYPE_DOUBLE) { + // OK + accessor->componentType = int(componentType); + } else { + std::stringstream ss; + ss << "Invalid `componentType` in accessor. Got " << componentType + << "\n"; + if (err) { + (*err) += ss.str(); + } + return false; + } + } + + ParseExtensionsProperty(&(accessor->extensions), err, o); + ParseExtrasProperty(&(accessor->extras), o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + accessor->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + accessor->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + // check if accessor has a "sparse" object: + json_const_iterator iterator; + if (FindMember(o, "sparse", iterator)) { + // here this accessor has a "sparse" subobject + return ParseSparseAccessor(accessor, err, GetValue(iterator)); + } + + return true; +} + +#ifdef TINYGLTF_ENABLE_DRACO + +static void DecodeIndexBuffer(draco::Mesh *mesh, size_t componentSize, + std::vector &outBuffer) { + if (componentSize == 4) { + assert(sizeof(mesh->face(draco::FaceIndex(0))[0]) == componentSize); + memcpy(outBuffer.data(), &mesh->face(draco::FaceIndex(0))[0], + outBuffer.size()); + } else { + size_t faceStride = componentSize * 3; + for (draco::FaceIndex f(0); f < mesh->num_faces(); ++f) { + const draco::Mesh::Face &face = mesh->face(f); + if (componentSize == 2) { + uint16_t indices[3] = {(uint16_t)face[0].value(), + (uint16_t)face[1].value(), + (uint16_t)face[2].value()}; + memcpy(outBuffer.data() + f.value() * faceStride, &indices[0], + faceStride); + } else { + uint8_t indices[3] = {(uint8_t)face[0].value(), + (uint8_t)face[1].value(), + (uint8_t)face[2].value()}; + memcpy(outBuffer.data() + f.value() * faceStride, &indices[0], + faceStride); + } + } + } +} + +template +static bool GetAttributeForAllPoints(draco::Mesh *mesh, + const draco::PointAttribute *pAttribute, + std::vector &outBuffer) { + size_t byteOffset = 0; + T values[4] = {0, 0, 0, 0}; + for (draco::PointIndex i(0); i < mesh->num_points(); ++i) { + const draco::AttributeValueIndex val_index = pAttribute->mapped_index(i); + if (!pAttribute->ConvertValue(val_index, pAttribute->num_components(), + values)) + return false; + + memcpy(outBuffer.data() + byteOffset, &values[0], + sizeof(T) * pAttribute->num_components()); + byteOffset += sizeof(T) * pAttribute->num_components(); + } + + return true; +} + +static bool GetAttributeForAllPoints(uint32_t componentType, draco::Mesh *mesh, + const draco::PointAttribute *pAttribute, + std::vector &outBuffer) { + bool decodeResult = false; + switch (componentType) { + case TINYGLTF_COMPONENT_TYPE_UNSIGNED_BYTE: + decodeResult = + GetAttributeForAllPoints(mesh, pAttribute, outBuffer); + break; + case TINYGLTF_COMPONENT_TYPE_BYTE: + decodeResult = + GetAttributeForAllPoints(mesh, pAttribute, outBuffer); + break; + case TINYGLTF_COMPONENT_TYPE_UNSIGNED_SHORT: + decodeResult = + GetAttributeForAllPoints(mesh, pAttribute, outBuffer); + break; + case TINYGLTF_COMPONENT_TYPE_SHORT: + decodeResult = + GetAttributeForAllPoints(mesh, pAttribute, outBuffer); + break; + case TINYGLTF_COMPONENT_TYPE_INT: + decodeResult = + GetAttributeForAllPoints(mesh, pAttribute, outBuffer); + break; + case TINYGLTF_COMPONENT_TYPE_UNSIGNED_INT: + decodeResult = + GetAttributeForAllPoints(mesh, pAttribute, outBuffer); + break; + case TINYGLTF_COMPONENT_TYPE_FLOAT: + decodeResult = + GetAttributeForAllPoints(mesh, pAttribute, outBuffer); + break; + case TINYGLTF_COMPONENT_TYPE_DOUBLE: + decodeResult = + GetAttributeForAllPoints(mesh, pAttribute, outBuffer); + break; + default: + return false; + } + + return decodeResult; +} + +static bool ParseDracoExtension(Primitive *primitive, Model *model, + std::string *err, + const Value &dracoExtensionValue) { + (void)err; + auto bufferViewValue = dracoExtensionValue.Get("bufferView"); + if (!bufferViewValue.IsInt()) return false; + auto attributesValue = dracoExtensionValue.Get("attributes"); + if (!attributesValue.IsObject()) return false; + + auto attributesObject = attributesValue.Get(); + int bufferView = bufferViewValue.Get(); + + BufferView &view = model->bufferViews[bufferView]; + Buffer &buffer = model->buffers[view.buffer]; + // BufferView has already been decoded + if (view.dracoDecoded) return true; + view.dracoDecoded = true; + + const char *bufferViewData = + reinterpret_cast(buffer.data.data() + view.byteOffset); + size_t bufferViewSize = view.byteLength; + + // decode draco + draco::DecoderBuffer decoderBuffer; + decoderBuffer.Init(bufferViewData, bufferViewSize); + draco::Decoder decoder; + auto decodeResult = decoder.DecodeMeshFromBuffer(&decoderBuffer); + if (!decodeResult.ok()) { + return false; + } + const std::unique_ptr &mesh = decodeResult.value(); + + // create new bufferView for indices + if (primitive->indices >= 0) { + int32_t componentSize = GetComponentSizeInBytes( + model->accessors[primitive->indices].componentType); + Buffer decodedIndexBuffer; + decodedIndexBuffer.data.resize(mesh->num_faces() * 3 * componentSize); + + DecodeIndexBuffer(mesh.get(), componentSize, decodedIndexBuffer.data); + + model->buffers.emplace_back(std::move(decodedIndexBuffer)); + + BufferView decodedIndexBufferView; + decodedIndexBufferView.buffer = int(model->buffers.size() - 1); + decodedIndexBufferView.byteLength = + int(mesh->num_faces() * 3 * componentSize); + decodedIndexBufferView.byteOffset = 0; + decodedIndexBufferView.byteStride = 0; + decodedIndexBufferView.target = TINYGLTF_TARGET_ARRAY_BUFFER; + model->bufferViews.emplace_back(std::move(decodedIndexBufferView)); + + model->accessors[primitive->indices].bufferView = + int(model->bufferViews.size() - 1); + model->accessors[primitive->indices].count = int(mesh->num_faces() * 3); + } + + for (const auto &attribute : attributesObject) { + if (!attribute.second.IsInt()) return false; + auto primitiveAttribute = primitive->attributes.find(attribute.first); + if (primitiveAttribute == primitive->attributes.end()) return false; + + int dracoAttributeIndex = attribute.second.Get(); + const auto pAttribute = mesh->GetAttributeByUniqueId(dracoAttributeIndex); + const auto componentType = + model->accessors[primitiveAttribute->second].componentType; + + // Create a new buffer for this decoded buffer + Buffer decodedBuffer; + size_t bufferSize = mesh->num_points() * pAttribute->num_components() * + GetComponentSizeInBytes(componentType); + decodedBuffer.data.resize(bufferSize); + + if (!GetAttributeForAllPoints(componentType, mesh.get(), pAttribute, + decodedBuffer.data)) + return false; + + model->buffers.emplace_back(std::move(decodedBuffer)); + + BufferView decodedBufferView; + decodedBufferView.buffer = int(model->buffers.size() - 1); + decodedBufferView.byteLength = bufferSize; + decodedBufferView.byteOffset = pAttribute->byte_offset(); + decodedBufferView.byteStride = pAttribute->byte_stride(); + decodedBufferView.target = primitive->indices >= 0 + ? TINYGLTF_TARGET_ELEMENT_ARRAY_BUFFER + : TINYGLTF_TARGET_ARRAY_BUFFER; + model->bufferViews.emplace_back(std::move(decodedBufferView)); + + model->accessors[primitiveAttribute->second].bufferView = + int(model->bufferViews.size() - 1); + model->accessors[primitiveAttribute->second].count = + int(mesh->num_points()); + } + + return true; +} +#endif + +static bool ParsePrimitive(Primitive *primitive, Model *model, std::string *err, + const json &o, + bool store_original_json_for_extras_and_extensions) { + int material = -1; + ParseIntegerProperty(&material, err, o, "material", false); + primitive->material = material; + + int mode = TINYGLTF_MODE_TRIANGLES; + ParseIntegerProperty(&mode, err, o, "mode", false); + primitive->mode = mode; // Why only triangled were supported ? + + int indices = -1; + ParseIntegerProperty(&indices, err, o, "indices", false); + primitive->indices = indices; + if (!ParseStringIntegerProperty(&primitive->attributes, err, o, "attributes", + true, "Primitive")) { + return false; + } + + // Look for morph targets + json_const_iterator targetsObject; + if (FindMember(o, "targets", targetsObject) && + IsArray(GetValue(targetsObject))) { + auto targetsObjectEnd = ArrayEnd(GetValue(targetsObject)); + for (json_const_array_iterator i = ArrayBegin(GetValue(targetsObject)); + i != targetsObjectEnd; ++i) { + std::map targetAttribues; + + const json &dict = *i; + if (IsObject(dict)) { + json_const_iterator dictIt(ObjectBegin(dict)); + json_const_iterator dictItEnd(ObjectEnd(dict)); + + for (; dictIt != dictItEnd; ++dictIt) { + int iVal; + if (GetInt(GetValue(dictIt), iVal)) + targetAttribues[GetKey(dictIt)] = iVal; + } + primitive->targets.emplace_back(std::move(targetAttribues)); + } + } + } + + ParseExtrasProperty(&(primitive->extras), o); + ParseExtensionsProperty(&primitive->extensions, err, o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + primitive->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + primitive->extras_json_string = JsonToString(GetValue(it)); + } + } + } + +#ifdef TINYGLTF_ENABLE_DRACO + auto dracoExtension = + primitive->extensions.find("KHR_draco_mesh_compression"); + if (dracoExtension != primitive->extensions.end()) { + ParseDracoExtension(primitive, model, err, dracoExtension->second); + } +#else + (void)model; +#endif + + return true; +} + +static bool ParseMesh(Mesh *mesh, Model *model, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + ParseStringProperty(&mesh->name, err, o, "name", false); + + mesh->primitives.clear(); + json_const_iterator primObject; + if (FindMember(o, "primitives", primObject) && + IsArray(GetValue(primObject))) { + json_const_array_iterator primEnd = ArrayEnd(GetValue(primObject)); + for (json_const_array_iterator i = ArrayBegin(GetValue(primObject)); + i != primEnd; ++i) { + Primitive primitive; + if (ParsePrimitive(&primitive, model, err, *i, + store_original_json_for_extras_and_extensions)) { + // Only add the primitive if the parsing succeeds. + mesh->primitives.emplace_back(std::move(primitive)); + } + } + } + + // Should probably check if has targets and if dimensions fit + ParseNumberArrayProperty(&mesh->weights, err, o, "weights", false); + + ParseExtensionsProperty(&mesh->extensions, err, o); + ParseExtrasProperty(&(mesh->extras), o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + mesh->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + mesh->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + return true; +} + +static bool ParseNode(Node *node, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + ParseStringProperty(&node->name, err, o, "name", false); + + int skin = -1; + ParseIntegerProperty(&skin, err, o, "skin", false); + node->skin = skin; + + // Matrix and T/R/S are exclusive + if (!ParseNumberArrayProperty(&node->matrix, err, o, "matrix", false)) { + ParseNumberArrayProperty(&node->rotation, err, o, "rotation", false); + ParseNumberArrayProperty(&node->scale, err, o, "scale", false); + ParseNumberArrayProperty(&node->translation, err, o, "translation", false); + } + + int camera = -1; + ParseIntegerProperty(&camera, err, o, "camera", false); + node->camera = camera; + + int mesh = -1; + ParseIntegerProperty(&mesh, err, o, "mesh", false); + node->mesh = mesh; + + node->children.clear(); + ParseIntegerArrayProperty(&node->children, err, o, "children", false); + + ParseNumberArrayProperty(&node->weights, err, o, "weights", false); + + ParseExtensionsProperty(&node->extensions, err, o); + ParseExtrasProperty(&(node->extras), o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + node->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + node->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + return true; +} + +static bool ParsePbrMetallicRoughness( + PbrMetallicRoughness *pbr, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + if (pbr == nullptr) { + return false; + } + + std::vector baseColorFactor; + if (ParseNumberArrayProperty(&baseColorFactor, err, o, "baseColorFactor", + /* required */ false)) { + if (baseColorFactor.size() != 4) { + if (err) { + (*err) += + "Array length of `baseColorFactor` parameter in " + "pbrMetallicRoughness must be 4, but got " + + std::to_string(baseColorFactor.size()) + "\n"; + } + return false; + } + pbr->baseColorFactor = baseColorFactor; + } + + { + json_const_iterator it; + if (FindMember(o, "baseColorTexture", it)) { + ParseTextureInfo(&pbr->baseColorTexture, err, GetValue(it), + store_original_json_for_extras_and_extensions); + } + } + + { + json_const_iterator it; + if (FindMember(o, "metallicRoughnessTexture", it)) { + ParseTextureInfo(&pbr->metallicRoughnessTexture, err, GetValue(it), + store_original_json_for_extras_and_extensions); + } + } + + ParseNumberProperty(&pbr->metallicFactor, err, o, "metallicFactor", false); + ParseNumberProperty(&pbr->roughnessFactor, err, o, "roughnessFactor", false); + + ParseExtensionsProperty(&pbr->extensions, err, o); + ParseExtrasProperty(&pbr->extras, o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + pbr->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + pbr->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + return true; +} + +static bool ParseMaterial(Material *material, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + ParseStringProperty(&material->name, err, o, "name", /* required */ false); + + if (ParseNumberArrayProperty(&material->emissiveFactor, err, o, + "emissiveFactor", + /* required */ false)) { + if (material->emissiveFactor.size() != 3) { + if (err) { + (*err) += + "Array length of `emissiveFactor` parameter in " + "material must be 3, but got " + + std::to_string(material->emissiveFactor.size()) + "\n"; + } + return false; + } + } else { + // fill with default values + material->emissiveFactor = {0.0, 0.0, 0.0}; + } + + ParseStringProperty(&material->alphaMode, err, o, "alphaMode", + /* required */ false); + ParseNumberProperty(&material->alphaCutoff, err, o, "alphaCutoff", + /* required */ false); + ParseBooleanProperty(&material->doubleSided, err, o, "doubleSided", + /* required */ false); + + { + json_const_iterator it; + if (FindMember(o, "pbrMetallicRoughness", it)) { + ParsePbrMetallicRoughness(&material->pbrMetallicRoughness, err, + GetValue(it), + store_original_json_for_extras_and_extensions); + } + } + + { + json_const_iterator it; + if (FindMember(o, "normalTexture", it)) { + ParseNormalTextureInfo(&material->normalTexture, err, GetValue(it), + store_original_json_for_extras_and_extensions); + } + } + + { + json_const_iterator it; + if (FindMember(o, "occlusionTexture", it)) { + ParseOcclusionTextureInfo(&material->occlusionTexture, err, GetValue(it), + store_original_json_for_extras_and_extensions); + } + } + + { + json_const_iterator it; + if (FindMember(o, "emissiveTexture", it)) { + ParseTextureInfo(&material->emissiveTexture, err, GetValue(it), + store_original_json_for_extras_and_extensions); + } + } + + // Old code path. For backward compatibility, we still store material values + // as Parameter. This will create duplicated information for + // example(pbrMetallicRoughness), but should be neglible in terms of memory + // consumption. + // TODO(syoyo): Remove in the next major release. + material->values.clear(); + material->additionalValues.clear(); + + json_const_iterator it(ObjectBegin(o)); + json_const_iterator itEnd(ObjectEnd(o)); + + for (; it != itEnd; ++it) { + std::string key(GetKey(it)); + if (key == "pbrMetallicRoughness") { + if (IsObject(GetValue(it))) { + const json &values_object = GetValue(it); + + json_const_iterator itVal(ObjectBegin(values_object)); + json_const_iterator itValEnd(ObjectEnd(values_object)); + + for (; itVal != itValEnd; ++itVal) { + Parameter param; + if (ParseParameterProperty(¶m, err, values_object, GetKey(itVal), + false)) { + material->values.emplace(GetKey(itVal), std::move(param)); + } + } + } + } else if (key == "extensions" || key == "extras") { + // done later, skip, otherwise poorly parsed contents will be saved in the + // parametermap and serialized again later + } else { + Parameter param; + if (ParseParameterProperty(¶m, err, o, key, false)) { + // names of materials have already been parsed. Putting it in this map + // doesn't correctly reflext the glTF specification + if (key != "name") + material->additionalValues.emplace(std::move(key), std::move(param)); + } + } + } + + material->extensions.clear(); + ParseExtensionsProperty(&material->extensions, err, o); + ParseExtrasProperty(&(material->extras), o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator eit; + if (FindMember(o, "extensions", eit)) { + material->extensions_json_string = JsonToString(GetValue(eit)); + } + } + { + json_const_iterator eit; + if (FindMember(o, "extras", eit)) { + material->extras_json_string = JsonToString(GetValue(eit)); + } + } + } + + return true; +} + +static bool ParseAnimationChannel( + AnimationChannel *channel, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + int samplerIndex = -1; + int targetIndex = -1; + if (!ParseIntegerProperty(&samplerIndex, err, o, "sampler", true, + "AnimationChannel")) { + if (err) { + (*err) += "`sampler` field is missing in animation channels\n"; + } + return false; + } + + json_const_iterator targetIt; + if (FindMember(o, "target", targetIt) && IsObject(GetValue(targetIt))) { + const json &target_object = GetValue(targetIt); + + if (!ParseIntegerProperty(&targetIndex, err, target_object, "node", true)) { + if (err) { + (*err) += "`node` field is missing in animation.channels.target\n"; + } + return false; + } + + if (!ParseStringProperty(&channel->target_path, err, target_object, "path", + true)) { + if (err) { + (*err) += "`path` field is missing in animation.channels.target\n"; + } + return false; + } + ParseExtensionsProperty(&channel->target_extensions, err, target_object); + if (store_original_json_for_extras_and_extensions) { + json_const_iterator it; + if (FindMember(target_object, "extensions", it)) { + channel->target_extensions_json_string = JsonToString(GetValue(it)); + } + } + } + + channel->sampler = samplerIndex; + channel->target_node = targetIndex; + + ParseExtensionsProperty(&channel->extensions, err, o); + ParseExtrasProperty(&(channel->extras), o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + channel->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + channel->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + return true; +} + +static bool ParseAnimation(Animation *animation, std::string *err, + const json &o, + bool store_original_json_for_extras_and_extensions) { + { + json_const_iterator channelsIt; + if (FindMember(o, "channels", channelsIt) && + IsArray(GetValue(channelsIt))) { + json_const_array_iterator channelEnd = ArrayEnd(GetValue(channelsIt)); + for (json_const_array_iterator i = ArrayBegin(GetValue(channelsIt)); + i != channelEnd; ++i) { + AnimationChannel channel; + if (ParseAnimationChannel( + &channel, err, *i, + store_original_json_for_extras_and_extensions)) { + // Only add the channel if the parsing succeeds. + animation->channels.emplace_back(std::move(channel)); + } + } + } + } + + { + json_const_iterator samplerIt; + if (FindMember(o, "samplers", samplerIt) && IsArray(GetValue(samplerIt))) { + const json &sampler_array = GetValue(samplerIt); + + json_const_array_iterator it = ArrayBegin(sampler_array); + json_const_array_iterator itEnd = ArrayEnd(sampler_array); + + for (; it != itEnd; ++it) { + const json &s = *it; + + AnimationSampler sampler; + int inputIndex = -1; + int outputIndex = -1; + if (!ParseIntegerProperty(&inputIndex, err, s, "input", true)) { + if (err) { + (*err) += "`input` field is missing in animation.sampler\n"; + } + return false; + } + ParseStringProperty(&sampler.interpolation, err, s, "interpolation", + false); + if (!ParseIntegerProperty(&outputIndex, err, s, "output", true)) { + if (err) { + (*err) += "`output` field is missing in animation.sampler\n"; + } + return false; + } + sampler.input = inputIndex; + sampler.output = outputIndex; + ParseExtensionsProperty(&(sampler.extensions), err, o); + ParseExtrasProperty(&(sampler.extras), s); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator eit; + if (FindMember(o, "extensions", eit)) { + sampler.extensions_json_string = JsonToString(GetValue(eit)); + } + } + { + json_const_iterator eit; + if (FindMember(o, "extras", eit)) { + sampler.extras_json_string = JsonToString(GetValue(eit)); + } + } + } + + animation->samplers.emplace_back(std::move(sampler)); + } + } + } + + ParseStringProperty(&animation->name, err, o, "name", false); + + ParseExtensionsProperty(&animation->extensions, err, o); + ParseExtrasProperty(&(animation->extras), o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + animation->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + animation->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + return true; +} + +static bool ParseSampler(Sampler *sampler, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + ParseStringProperty(&sampler->name, err, o, "name", false); + + int minFilter = -1; + int magFilter = -1; + int wrapS = TINYGLTF_TEXTURE_WRAP_REPEAT; + int wrapT = TINYGLTF_TEXTURE_WRAP_REPEAT; + //int wrapR = TINYGLTF_TEXTURE_WRAP_REPEAT; + ParseIntegerProperty(&minFilter, err, o, "minFilter", false); + ParseIntegerProperty(&magFilter, err, o, "magFilter", false); + ParseIntegerProperty(&wrapS, err, o, "wrapS", false); + ParseIntegerProperty(&wrapT, err, o, "wrapT", false); + //ParseIntegerProperty(&wrapR, err, o, "wrapR", false); // tinygltf extension + + // TODO(syoyo): Check the value is alloed one. + // (e.g. we allow 9728(NEAREST), but don't allow 9727) + + sampler->minFilter = minFilter; + sampler->magFilter = magFilter; + sampler->wrapS = wrapS; + sampler->wrapT = wrapT; + //sampler->wrapR = wrapR; + + ParseExtensionsProperty(&(sampler->extensions), err, o); + ParseExtrasProperty(&(sampler->extras), o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + sampler->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + sampler->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + return true; +} + +static bool ParseSkin(Skin *skin, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + ParseStringProperty(&skin->name, err, o, "name", false, "Skin"); + + std::vector joints; + if (!ParseIntegerArrayProperty(&joints, err, o, "joints", false, "Skin")) { + return false; + } + skin->joints = std::move(joints); + + int skeleton = -1; + ParseIntegerProperty(&skeleton, err, o, "skeleton", false, "Skin"); + skin->skeleton = skeleton; + + int invBind = -1; + ParseIntegerProperty(&invBind, err, o, "inverseBindMatrices", true, "Skin"); + skin->inverseBindMatrices = invBind; + + ParseExtensionsProperty(&(skin->extensions), err, o); + ParseExtrasProperty(&(skin->extras), o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + skin->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + skin->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + return true; +} + +static bool ParsePerspectiveCamera( + PerspectiveCamera *camera, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + double yfov = 0.0; + if (!ParseNumberProperty(&yfov, err, o, "yfov", true, "OrthographicCamera")) { + return false; + } + + double znear = 0.0; + if (!ParseNumberProperty(&znear, err, o, "znear", true, + "PerspectiveCamera")) { + return false; + } + + double aspectRatio = 0.0; // = invalid + ParseNumberProperty(&aspectRatio, err, o, "aspectRatio", false, + "PerspectiveCamera"); + + double zfar = 0.0; // = invalid + ParseNumberProperty(&zfar, err, o, "zfar", false, "PerspectiveCamera"); + + camera->aspectRatio = aspectRatio; + camera->zfar = zfar; + camera->yfov = yfov; + camera->znear = znear; + + ParseExtensionsProperty(&camera->extensions, err, o); + ParseExtrasProperty(&(camera->extras), o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + camera->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + camera->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + // TODO(syoyo): Validate parameter values. + + return true; +} + +static bool ParseSpotLight(SpotLight *light, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + ParseNumberProperty(&light->innerConeAngle, err, o, "innerConeAngle", false); + ParseNumberProperty(&light->outerConeAngle, err, o, "outerConeAngle", false); + + ParseExtensionsProperty(&light->extensions, err, o); + ParseExtrasProperty(&light->extras, o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + light->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + light->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + // TODO(syoyo): Validate parameter values. + + return true; +} + +static bool ParseOrthographicCamera( + OrthographicCamera *camera, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + double xmag = 0.0; + if (!ParseNumberProperty(&xmag, err, o, "xmag", true, "OrthographicCamera")) { + return false; + } + + double ymag = 0.0; + if (!ParseNumberProperty(&ymag, err, o, "ymag", true, "OrthographicCamera")) { + return false; + } + + double zfar = 0.0; + if (!ParseNumberProperty(&zfar, err, o, "zfar", true, "OrthographicCamera")) { + return false; + } + + double znear = 0.0; + if (!ParseNumberProperty(&znear, err, o, "znear", true, + "OrthographicCamera")) { + return false; + } + + ParseExtensionsProperty(&camera->extensions, err, o); + ParseExtrasProperty(&(camera->extras), o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + camera->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + camera->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + camera->xmag = xmag; + camera->ymag = ymag; + camera->zfar = zfar; + camera->znear = znear; + + // TODO(syoyo): Validate parameter values. + + return true; +} + +static bool ParseCamera(Camera *camera, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + if (!ParseStringProperty(&camera->type, err, o, "type", true, "Camera")) { + return false; + } + + if (camera->type.compare("orthographic") == 0) { + json_const_iterator orthoIt; + if (!FindMember(o, "orthographic", orthoIt)) { + if (err) { + std::stringstream ss; + ss << "Orhographic camera description not found." << std::endl; + (*err) += ss.str(); + } + return false; + } + + const json &v = GetValue(orthoIt); + if (!IsObject(v)) { + if (err) { + std::stringstream ss; + ss << "\"orthographic\" is not a JSON object." << std::endl; + (*err) += ss.str(); + } + return false; + } + + if (!ParseOrthographicCamera( + &camera->orthographic, err, v, + store_original_json_for_extras_and_extensions)) { + return false; + } + } else if (camera->type.compare("perspective") == 0) { + json_const_iterator perspIt; + if (!FindMember(o, "perspective", perspIt)) { + if (err) { + std::stringstream ss; + ss << "Perspective camera description not found." << std::endl; + (*err) += ss.str(); + } + return false; + } + + const json &v = GetValue(perspIt); + if (!IsObject(v)) { + if (err) { + std::stringstream ss; + ss << "\"perspective\" is not a JSON object." << std::endl; + (*err) += ss.str(); + } + return false; + } + + if (!ParsePerspectiveCamera( + &camera->perspective, err, v, + store_original_json_for_extras_and_extensions)) { + return false; + } + } else { + if (err) { + std::stringstream ss; + ss << "Invalid camera type: \"" << camera->type + << "\". Must be \"perspective\" or \"orthographic\"" << std::endl; + (*err) += ss.str(); + } + return false; + } + + ParseStringProperty(&camera->name, err, o, "name", false); + + ParseExtensionsProperty(&camera->extensions, err, o); + ParseExtrasProperty(&(camera->extras), o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + camera->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + camera->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + return true; +} + +static bool ParseLight(Light *light, std::string *err, const json &o, + bool store_original_json_for_extras_and_extensions) { + if (!ParseStringProperty(&light->type, err, o, "type", true)) { + return false; + } + + if (light->type == "spot") { + json_const_iterator spotIt; + if (!FindMember(o, "spot", spotIt)) { + if (err) { + std::stringstream ss; + ss << "Spot light description not found." << std::endl; + (*err) += ss.str(); + } + return false; + } + + const json &v = GetValue(spotIt); + if (!IsObject(v)) { + if (err) { + std::stringstream ss; + ss << "\"spot\" is not a JSON object." << std::endl; + (*err) += ss.str(); + } + return false; + } + + if (!ParseSpotLight(&light->spot, err, v, + store_original_json_for_extras_and_extensions)) { + return false; + } + } + + ParseStringProperty(&light->name, err, o, "name", false); + ParseNumberArrayProperty(&light->color, err, o, "color", false); + ParseNumberProperty(&light->range, err, o, "range", false); + ParseNumberProperty(&light->intensity, err, o, "intensity", false); + ParseExtensionsProperty(&light->extensions, err, o); + ParseExtrasProperty(&(light->extras), o); + + if (store_original_json_for_extras_and_extensions) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + light->extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + light->extras_json_string = JsonToString(GetValue(it)); + } + } + } + + return true; +} + +bool TinyGLTF::LoadFromString(Model *model, std::string *err, std::string *warn, + const char *json_str, + unsigned int json_str_length, + const std::string &base_dir, + unsigned int check_sections) { + if (json_str_length < 4) { + if (err) { + (*err) = "JSON string too short.\n"; + } + return false; + } + + JsonDocument v; + +#if (defined(__cpp_exceptions) || defined(__EXCEPTIONS) || \ + defined(_CPPUNWIND)) && \ + !defined(TINYGLTF_NOEXCEPTION) + try { + JsonParse(v, json_str, json_str_length, true); + + } catch (const std::exception &e) { + if (err) { + (*err) = e.what(); + } + return false; + } +#else + { + JsonParse(v, json_str, json_str_length); + + if (!IsObject(v)) { + // Assume parsing was failed. + if (err) { + (*err) = "Failed to parse JSON object\n"; + } + return false; + } + } +#endif + + if (!IsObject(v)) { + // root is not an object. + if (err) { + (*err) = "Root element is not a JSON object\n"; + } + return false; + } + + { + bool version_found = false; + json_const_iterator it; + if (FindMember(v, "asset", it) && IsObject(GetValue(it))) { + auto &itObj = GetValue(it); + json_const_iterator version_it; + std::string versionStr; + if (FindMember(itObj, "version", version_it) && + GetString(GetValue(version_it), versionStr)) { + version_found = true; + } + } + if (version_found) { + // OK + } else if (check_sections & REQUIRE_VERSION) { + if (err) { + (*err) += "\"asset\" object not found in .gltf or not an object type\n"; + } + return false; + } + } + + // scene is not mandatory. + // FIXME Maybe a better way to handle it than removing the code + + auto IsArrayMemberPresent = [](const json &_v, const char *name) -> bool { + json_const_iterator it; + return FindMember(_v, name, it) && IsArray(GetValue(it)); + }; + + { + if ((check_sections & REQUIRE_SCENES) && + !IsArrayMemberPresent(v, "scenes")) { + if (err) { + (*err) += "\"scenes\" object not found in .gltf or not an array type\n"; + } + return false; + } + } + + { + if ((check_sections & REQUIRE_NODES) && !IsArrayMemberPresent(v, "nodes")) { + if (err) { + (*err) += "\"nodes\" object not found in .gltf\n"; + } + return false; + } + } + + { + if ((check_sections & REQUIRE_ACCESSORS) && + !IsArrayMemberPresent(v, "accessors")) { + if (err) { + (*err) += "\"accessors\" object not found in .gltf\n"; + } + return false; + } + } + + { + if ((check_sections & REQUIRE_BUFFERS) && + !IsArrayMemberPresent(v, "buffers")) { + if (err) { + (*err) += "\"buffers\" object not found in .gltf\n"; + } + return false; + } + } + + { + if ((check_sections & REQUIRE_BUFFER_VIEWS) && + !IsArrayMemberPresent(v, "bufferViews")) { + if (err) { + (*err) += "\"bufferViews\" object not found in .gltf\n"; + } + return false; + } + } + + model->buffers.clear(); + model->bufferViews.clear(); + model->accessors.clear(); + model->meshes.clear(); + model->cameras.clear(); + model->nodes.clear(); + model->extensionsUsed.clear(); + model->extensionsRequired.clear(); + model->extensions.clear(); + model->defaultScene = -1; + + // 1. Parse Asset + { + json_const_iterator it; + if (FindMember(v, "asset", it) && IsObject(GetValue(it))) { + const json &root = GetValue(it); + + ParseAsset(&model->asset, err, root, + store_original_json_for_extras_and_extensions_); + } + } + +#ifdef TINYGLTF_USE_CPP14 + auto ForEachInArray = [](const json &_v, const char *member, + const auto &cb) -> bool +#else + // The std::function<> implementation can be less efficient because it will + // allocate heap when the size of the captured lambda is above 16 bytes with + // clang and gcc, but it does not require C++14. + auto ForEachInArray = [](const json &_v, const char *member, + const std::function &cb) -> bool +#endif + { + json_const_iterator itm; + if (FindMember(_v, member, itm) && IsArray(GetValue(itm))) { + const json &root = GetValue(itm); + auto it = ArrayBegin(root); + auto end = ArrayEnd(root); + for (; it != end; ++it) { + if (!cb(*it)) return false; + } + } + return true; + }; + + // 2. Parse extensionUsed + { + ForEachInArray(v, "extensionsUsed", [&](const json &o) { + std::string str; + GetString(o, str); + model->extensionsUsed.emplace_back(std::move(str)); + return true; + }); + } + + { + ForEachInArray(v, "extensionsRequired", [&](const json &o) { + std::string str; + GetString(o, str); + model->extensionsRequired.emplace_back(std::move(str)); + return true; + }); + } + + // 3. Parse Buffer + { + bool success = ForEachInArray(v, "buffers", [&](const json &o) { + if (!IsObject(o)) { + if (err) { + (*err) += "`buffers' does not contain an JSON object."; + } + return false; + } + Buffer buffer; + if (!ParseBuffer(&buffer, err, o, + store_original_json_for_extras_and_extensions_, &fs, + base_dir, is_binary_, bin_data_, bin_size_)) { + return false; + } + + model->buffers.emplace_back(std::move(buffer)); + return true; + }); + + if (!success) { + return false; + } + } + // 4. Parse BufferView + { + bool success = ForEachInArray(v, "bufferViews", [&](const json &o) { + if (!IsObject(o)) { + if (err) { + (*err) += "`bufferViews' does not contain an JSON object."; + } + return false; + } + BufferView bufferView; + if (!ParseBufferView(&bufferView, err, o, + store_original_json_for_extras_and_extensions_)) { + return false; + } + + model->bufferViews.emplace_back(std::move(bufferView)); + return true; + }); + + if (!success) { + return false; + } + } + + // 5. Parse Accessor + { + bool success = ForEachInArray(v, "accessors", [&](const json &o) { + if (!IsObject(o)) { + if (err) { + (*err) += "`accessors' does not contain an JSON object."; + } + return false; + } + Accessor accessor; + if (!ParseAccessor(&accessor, err, o, + store_original_json_for_extras_and_extensions_)) { + return false; + } + + model->accessors.emplace_back(std::move(accessor)); + return true; + }); + + if (!success) { + return false; + } + } + + // 6. Parse Mesh + { + bool success = ForEachInArray(v, "meshes", [&](const json &o) { + if (!IsObject(o)) { + if (err) { + (*err) += "`meshes' does not contain an JSON object."; + } + return false; + } + Mesh mesh; + if (!ParseMesh(&mesh, model, err, o, + store_original_json_for_extras_and_extensions_)) { + return false; + } + + model->meshes.emplace_back(std::move(mesh)); + return true; + }); + + if (!success) { + return false; + } + } + + // Assign missing bufferView target types + // - Look for missing Mesh indices + // - Look for missing Mesh attributes + for (auto &mesh : model->meshes) { + for (auto &primitive : mesh.primitives) { + if (primitive.indices > + -1) // has indices from parsing step, must be Element Array Buffer + { + if (size_t(primitive.indices) >= model->accessors.size()) { + if (err) { + (*err) += "primitive indices accessor out of bounds"; + } + return false; + } + + auto bufferView = + model->accessors[size_t(primitive.indices)].bufferView; + if (bufferView < 0 || size_t(bufferView) >= model->bufferViews.size()) { + if (err) { + (*err) += "accessor[" + std::to_string(primitive.indices) + + "] invalid bufferView"; + } + return false; + } + + model->bufferViews[size_t(bufferView)].target = + TINYGLTF_TARGET_ELEMENT_ARRAY_BUFFER; + // we could optionally check if acessors' bufferView type is Scalar, as + // it should be + } + + for (auto &attribute : primitive.attributes) { + model + ->bufferViews[size_t( + model->accessors[size_t(attribute.second)].bufferView)] + .target = TINYGLTF_TARGET_ARRAY_BUFFER; + } + + for (auto &target : primitive.targets) { + for (auto &attribute : target) { + auto bufferView = + model->accessors[size_t(attribute.second)].bufferView; + // bufferView could be null(-1) for sparse morph target + if (bufferView >= 0) { + model->bufferViews[size_t(bufferView)].target = + TINYGLTF_TARGET_ARRAY_BUFFER; + } + } + } + } + } + + // 7. Parse Node + { + bool success = ForEachInArray(v, "nodes", [&](const json &o) { + if (!IsObject(o)) { + if (err) { + (*err) += "`nodes' does not contain an JSON object."; + } + return false; + } + Node node; + if (!ParseNode(&node, err, o, + store_original_json_for_extras_and_extensions_)) { + return false; + } + + model->nodes.emplace_back(std::move(node)); + return true; + }); + + if (!success) { + return false; + } + } + + // 8. Parse scenes. + { + bool success = ForEachInArray(v, "scenes", [&](const json &o) { + if (!IsObject(o)) { + if (err) { + (*err) += "`scenes' does not contain an JSON object."; + } + return false; + } + std::vector nodes; + ParseIntegerArrayProperty(&nodes, err, o, "nodes", false); + + Scene scene; + scene.nodes = std::move(nodes); + + ParseStringProperty(&scene.name, err, o, "name", false); + + ParseExtensionsProperty(&scene.extensions, err, o); + ParseExtrasProperty(&scene.extras, o); + + if (store_original_json_for_extras_and_extensions_) { + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + scene.extensions_json_string = JsonToString(GetValue(it)); + } + } + { + json_const_iterator it; + if (FindMember(o, "extras", it)) { + scene.extras_json_string = JsonToString(GetValue(it)); + } + } + } + + model->scenes.emplace_back(std::move(scene)); + return true; + }); + + if (!success) { + return false; + } + } + + // 9. Parse default scenes. + { + json_const_iterator rootIt; + int iVal; + if (FindMember(v, "scene", rootIt) && GetInt(GetValue(rootIt), iVal)) { + model->defaultScene = iVal; + } + } + + // 10. Parse Material + { + bool success = ForEachInArray(v, "materials", [&](const json &o) { + if (!IsObject(o)) { + if (err) { + (*err) += "`materials' does not contain an JSON object."; + } + return false; + } + Material material; + ParseStringProperty(&material.name, err, o, "name", false); + + if (!ParseMaterial(&material, err, o, + store_original_json_for_extras_and_extensions_)) { + return false; + } + + model->materials.emplace_back(std::move(material)); + return true; + }); + + if (!success) { + return false; + } + } + + // 11. Parse Image + void *load_image_user_data{nullptr}; + + LoadImageDataOption load_image_option; + + if (user_image_loader_) { + // Use user supplied pointer + load_image_user_data = load_image_user_data_; + } else { + load_image_option.preserve_channels = preserve_image_channels_; + load_image_user_data = reinterpret_cast(&load_image_option); + } + + { + int idx = 0; + bool success = ForEachInArray(v, "images", [&](const json &o) { + if (!IsObject(o)) { + if (err) { + (*err) += "image[" + std::to_string(idx) + "] is not a JSON object."; + } + return false; + } + Image image; + if (!ParseImage(&image, idx, err, warn, o, + store_original_json_for_extras_and_extensions_, base_dir, + &fs, &this->LoadImageData, load_image_user_data)) { + return false; + } + + if (image.bufferView != -1) { + // Load image from the buffer view. + if (size_t(image.bufferView) >= model->bufferViews.size()) { + if (err) { + std::stringstream ss; + ss << "image[" << idx << "] bufferView \"" << image.bufferView + << "\" not found in the scene." << std::endl; + (*err) += ss.str(); + } + return false; + } + + const BufferView &bufferView = + model->bufferViews[size_t(image.bufferView)]; + if (size_t(bufferView.buffer) >= model->buffers.size()) { + if (err) { + std::stringstream ss; + ss << "image[" << idx << "] buffer \"" << bufferView.buffer + << "\" not found in the scene." << std::endl; + (*err) += ss.str(); + } + return false; + } + const Buffer &buffer = model->buffers[size_t(bufferView.buffer)]; + + if (*LoadImageData == nullptr) { + if (err) { + (*err) += "No LoadImageData callback specified.\n"; + } + return false; + } + bool ret = LoadImageData( + &image, idx, err, warn, image.width, image.height, + &buffer.data[bufferView.byteOffset], + static_cast(bufferView.byteLength), load_image_user_data); + if (!ret) { + return false; + } + } + + model->images.emplace_back(std::move(image)); + ++idx; + return true; + }); + + if (!success) { + return false; + } + } + + // 12. Parse Texture + { + bool success = ForEachInArray(v, "textures", [&](const json &o) { + if (!IsObject(o)) { + if (err) { + (*err) += "`textures' does not contain an JSON object."; + } + return false; + } + Texture texture; + if (!ParseTexture(&texture, err, o, + store_original_json_for_extras_and_extensions_, + base_dir)) { + return false; + } + + model->textures.emplace_back(std::move(texture)); + return true; + }); + + if (!success) { + return false; + } + } + + // 13. Parse Animation + { + bool success = ForEachInArray(v, "animations", [&](const json &o) { + if (!IsObject(o)) { + if (err) { + (*err) += "`animations' does not contain an JSON object."; + } + return false; + } + Animation animation; + if (!ParseAnimation(&animation, err, o, + store_original_json_for_extras_and_extensions_)) { + return false; + } + + model->animations.emplace_back(std::move(animation)); + return true; + }); + + if (!success) { + return false; + } + } + + // 14. Parse Skin + { + bool success = ForEachInArray(v, "skins", [&](const json &o) { + if (!IsObject(o)) { + if (err) { + (*err) += "`skins' does not contain an JSON object."; + } + return false; + } + Skin skin; + if (!ParseSkin(&skin, err, o, + store_original_json_for_extras_and_extensions_)) { + return false; + } + + model->skins.emplace_back(std::move(skin)); + return true; + }); + + if (!success) { + return false; + } + } + + // 15. Parse Sampler + { + bool success = ForEachInArray(v, "samplers", [&](const json &o) { + if (!IsObject(o)) { + if (err) { + (*err) += "`samplers' does not contain an JSON object."; + } + return false; + } + Sampler sampler; + if (!ParseSampler(&sampler, err, o, + store_original_json_for_extras_and_extensions_)) { + return false; + } + + model->samplers.emplace_back(std::move(sampler)); + return true; + }); + + if (!success) { + return false; + } + } + + // 16. Parse Camera + { + bool success = ForEachInArray(v, "cameras", [&](const json &o) { + if (!IsObject(o)) { + if (err) { + (*err) += "`cameras' does not contain an JSON object."; + } + return false; + } + Camera camera; + if (!ParseCamera(&camera, err, o, + store_original_json_for_extras_and_extensions_)) { + return false; + } + + model->cameras.emplace_back(std::move(camera)); + return true; + }); + + if (!success) { + return false; + } + } + + // 17. Parse Extensions + ParseExtensionsProperty(&model->extensions, err, v); + + // 18. Specific extension implementations + { + json_const_iterator rootIt; + if (FindMember(v, "extensions", rootIt) && IsObject(GetValue(rootIt))) { + const json &root = GetValue(rootIt); + + json_const_iterator it(ObjectBegin(root)); + json_const_iterator itEnd(ObjectEnd(root)); + for (; it != itEnd; ++it) { + // parse KHR_lights_punctual extension + std::string key(GetKey(it)); + if ((key == "KHR_lights_punctual") && IsObject(GetValue(it))) { + const json &object = GetValue(it); + json_const_iterator itLight; + if (FindMember(object, "lights", itLight)) { + const json &lights = GetValue(itLight); + if (!IsArray(lights)) { + continue; + } + + auto arrayIt(ArrayBegin(lights)); + auto arrayItEnd(ArrayEnd(lights)); + for (; arrayIt != arrayItEnd; ++arrayIt) { + Light light; + if (!ParseLight(&light, err, *arrayIt, + store_original_json_for_extras_and_extensions_)) { + return false; + } + model->lights.emplace_back(std::move(light)); + } + } + } + } + } + } + + // 19. Parse Extras + ParseExtrasProperty(&model->extras, v); + + if (store_original_json_for_extras_and_extensions_) { + model->extras_json_string = JsonToString(v["extras"]); + model->extensions_json_string = JsonToString(v["extensions"]); + } + + return true; +} + +bool TinyGLTF::LoadASCIIFromString(Model *model, std::string *err, + std::string *warn, const char *str, + unsigned int length, + const std::string &base_dir, + unsigned int check_sections) { + is_binary_ = false; + bin_data_ = nullptr; + bin_size_ = 0; + + return LoadFromString(model, err, warn, str, length, base_dir, + check_sections); +} + +bool TinyGLTF::LoadASCIIFromFile(Model *model, std::string *err, + std::string *warn, const std::string &filename, + unsigned int check_sections) { + std::stringstream ss; + + if (fs.ReadWholeFile == nullptr) { + // Programmer error, assert() ? + ss << "Failed to read file: " << filename + << ": one or more FS callback not set" << std::endl; + if (err) { + (*err) = ss.str(); + } + return false; + } + + std::vector data; + std::string fileerr; + bool fileread = fs.ReadWholeFile(&data, &fileerr, filename, fs.user_data); + if (!fileread) { + ss << "Failed to read file: " << filename << ": " << fileerr << std::endl; + if (err) { + (*err) = ss.str(); + } + return false; + } + + size_t sz = data.size(); + if (sz == 0) { + if (err) { + (*err) = "Empty file."; + } + return false; + } + + std::string basedir = GetBaseDir(filename); + + bool ret = LoadASCIIFromString( + model, err, warn, reinterpret_cast(&data.at(0)), + static_cast(data.size()), basedir, check_sections); + + return ret; +} + +bool TinyGLTF::LoadBinaryFromMemory(Model *model, std::string *err, + std::string *warn, + const unsigned char *bytes, + unsigned int size, + const std::string &base_dir, + unsigned int check_sections) { + if (size < 20) { + if (err) { + (*err) = "Too short data size for glTF Binary."; + } + return false; + } + + if (bytes[0] == 'g' && bytes[1] == 'l' && bytes[2] == 'T' && + bytes[3] == 'F') { + // ok + } else { + if (err) { + (*err) = "Invalid magic."; + } + return false; + } + + unsigned int version; // 4 bytes + unsigned int length; // 4 bytes + unsigned int model_length; // 4 bytes + unsigned int model_format; // 4 bytes; + + // @todo { Endian swap for big endian machine. } + memcpy(&version, bytes + 4, 4); + swap4(&version); + memcpy(&length, bytes + 8, 4); + swap4(&length); + memcpy(&model_length, bytes + 12, 4); + swap4(&model_length); + memcpy(&model_format, bytes + 16, 4); + swap4(&model_format); + + // In case the Bin buffer is not present, the size is exactly 20 + size of + // JSON contents, + // so use "greater than" operator. + if ((20 + model_length > size) || (model_length < 1) || (length > size) || + (20 + model_length > length) || + (model_format != 0x4E4F534A)) { // 0x4E4F534A = JSON format. + if (err) { + (*err) = "Invalid glTF binary."; + } + return false; + } + + // Extract JSON string. + std::string jsonString(reinterpret_cast(&bytes[20]), + model_length); + + is_binary_ = true; + bin_data_ = bytes + 20 + model_length + + 8; // 4 bytes (buffer_length) + 4 bytes(buffer_format) + bin_size_ = + length - (20 + model_length); // extract header + JSON scene data. + + bool ret = LoadFromString(model, err, warn, + reinterpret_cast(&bytes[20]), + model_length, base_dir, check_sections); + if (!ret) { + return ret; + } + + return true; +} + +bool TinyGLTF::LoadBinaryFromFile(Model *model, std::string *err, + std::string *warn, + const std::string &filename, + unsigned int check_sections) { + std::stringstream ss; + + if (fs.ReadWholeFile == nullptr) { + // Programmer error, assert() ? + ss << "Failed to read file: " << filename + << ": one or more FS callback not set" << std::endl; + if (err) { + (*err) = ss.str(); + } + return false; + } + + std::vector data; + std::string fileerr; + bool fileread = fs.ReadWholeFile(&data, &fileerr, filename, fs.user_data); + if (!fileread) { + ss << "Failed to read file: " << filename << ": " << fileerr << std::endl; + if (err) { + (*err) = ss.str(); + } + return false; + } + + std::string basedir = GetBaseDir(filename); + + bool ret = LoadBinaryFromMemory(model, err, warn, &data.at(0), + static_cast(data.size()), + basedir, check_sections); + + return ret; +} + +/////////////////////// +// GLTF Serialization +/////////////////////// +namespace { +json JsonFromString(const char *s) { +#ifdef TINYGLTF_USE_RAPIDJSON + return json(s, GetAllocator()); +#else + return json(s); +#endif +} + +void JsonAssign(json &dest, const json &src) { +#ifdef TINYGLTF_USE_RAPIDJSON + dest.CopyFrom(src, GetAllocator()); +#else + dest = src; +#endif +} + +void JsonAddMember(json &o, const char *key, json &&value) { +#ifdef TINYGLTF_USE_RAPIDJSON + if (!o.IsObject()) { + o.SetObject(); + } + o.AddMember(json(key, GetAllocator()), std::move(value), GetAllocator()); +#else + o[key] = std::move(value); +#endif +} + +void JsonPushBack(json &o, json &&value) { +#ifdef TINYGLTF_USE_RAPIDJSON + o.PushBack(std::move(value), GetAllocator()); +#else + o.push_back(std::move(value)); +#endif +} + +bool JsonIsNull(const json &o) { +#ifdef TINYGLTF_USE_RAPIDJSON + return o.IsNull(); +#else + return o.is_null(); +#endif +} + +void JsonSetObject(json &o) { +#ifdef TINYGLTF_USE_RAPIDJSON + o.SetObject(); +#else + o = o.object({}); +#endif +} + +void JsonReserveArray(json &o, size_t s) { +#ifdef TINYGLTF_USE_RAPIDJSON + o.SetArray(); + o.Reserve(static_cast(s), GetAllocator()); +#endif + (void)(o); + (void)(s); +} +} // namespace + +// typedef std::pair json_object_pair; + +template +static void SerializeNumberProperty(const std::string &key, T number, + json &obj) { + // obj.insert( + // json_object_pair(key, json(static_cast(number)))); + // obj[key] = static_cast(number); + JsonAddMember(obj, key.c_str(), json(number)); +} + +#ifdef TINYGLTF_USE_RAPIDJSON +template <> +void SerializeNumberProperty(const std::string &key, size_t number, json &obj) { + JsonAddMember(obj, key.c_str(), json(static_cast(number))); +} +#endif + +template +static void SerializeNumberArrayProperty(const std::string &key, + const std::vector &value, + json &obj) { + if (value.empty()) return; + + json ary; + JsonReserveArray(ary, value.size()); + for (const auto &s : value) { + JsonPushBack(ary, json(s)); + } + JsonAddMember(obj, key.c_str(), std::move(ary)); +} + +static void SerializeStringProperty(const std::string &key, + const std::string &value, json &obj) { + JsonAddMember(obj, key.c_str(), JsonFromString(value.c_str())); +} + +static void SerializeStringArrayProperty(const std::string &key, + const std::vector &value, + json &obj) { + json ary; + JsonReserveArray(ary, value.size()); + for (auto &s : value) { + JsonPushBack(ary, JsonFromString(s.c_str())); + } + JsonAddMember(obj, key.c_str(), std::move(ary)); +} + +static bool ValueToJson(const Value &value, json *ret) { + json obj; +#ifdef TINYGLTF_USE_RAPIDJSON + switch (value.Type()) { + case REAL_TYPE: + obj.SetDouble(value.Get()); + break; + case INT_TYPE: + obj.SetInt(value.Get()); + break; + case BOOL_TYPE: + obj.SetBool(value.Get()); + break; + case STRING_TYPE: + obj.SetString(value.Get().c_str(), GetAllocator()); + break; + case ARRAY_TYPE: { + obj.SetArray(); + obj.Reserve(static_cast(value.ArrayLen()), + GetAllocator()); + for (unsigned int i = 0; i < value.ArrayLen(); ++i) { + Value elementValue = value.Get(int(i)); + json elementJson; + if (ValueToJson(value.Get(int(i)), &elementJson)) + obj.PushBack(std::move(elementJson), GetAllocator()); + } + break; + } + case BINARY_TYPE: + // TODO + // obj = json(value.Get>()); + return false; + break; + case OBJECT_TYPE: { + obj.SetObject(); + Value::Object objMap = value.Get(); + for (auto &it : objMap) { + json elementJson; + if (ValueToJson(it.second, &elementJson)) { + obj.AddMember(json(it.first.c_str(), GetAllocator()), + std::move(elementJson), GetAllocator()); + } + } + break; + } + case NULL_TYPE: + default: + return false; + } +#else + switch (value.Type()) { + case REAL_TYPE: + obj = json(value.Get()); + break; + case INT_TYPE: + obj = json(value.Get()); + break; + case BOOL_TYPE: + obj = json(value.Get()); + break; + case STRING_TYPE: + obj = json(value.Get()); + break; + case ARRAY_TYPE: { + for (unsigned int i = 0; i < value.ArrayLen(); ++i) { + Value elementValue = value.Get(int(i)); + json elementJson; + if (ValueToJson(value.Get(int(i)), &elementJson)) + obj.push_back(elementJson); + } + break; + } + case BINARY_TYPE: + // TODO + // obj = json(value.Get>()); + return false; + break; + case OBJECT_TYPE: { + Value::Object objMap = value.Get(); + for (auto &it : objMap) { + json elementJson; + if (ValueToJson(it.second, &elementJson)) obj[it.first] = elementJson; + } + break; + } + case NULL_TYPE: + default: + return false; + } +#endif + if (ret) *ret = std::move(obj); + return true; +} + +static void SerializeValue(const std::string &key, const Value &value, + json &obj) { + json ret; + if (ValueToJson(value, &ret)) { + JsonAddMember(obj, key.c_str(), std::move(ret)); + } +} + +static void SerializeGltfBufferData(const std::vector &data, + json &o) { + std::string header = "data:application/octet-stream;base64,"; + if (data.size() > 0) { + std::string encodedData = + base64_encode(&data[0], static_cast(data.size())); + SerializeStringProperty("uri", header + encodedData, o); + } else { + // Issue #229 + // size 0 is allowd. Just emit mime header. + SerializeStringProperty("uri", header, o); + } +} + +static bool SerializeGltfBufferData(const std::vector &data, + const std::string &binFilename) { +#ifdef _WIN32 +#if defined(__GLIBCXX__) // mingw + int file_descriptor = _wopen(UTF8ToWchar(binFilename).c_str(), + _O_CREAT | _O_WRONLY | _O_TRUNC | _O_BINARY); + __gnu_cxx::stdio_filebuf wfile_buf( + file_descriptor, std::ios_base::out | std::ios_base::binary); + std::ostream output(&wfile_buf); + if (!wfile_buf.is_open()) return false; +#elif defined(_MSC_VER) + std::ofstream output(UTF8ToWchar(binFilename).c_str(), std::ofstream::binary); + if (!output.is_open()) return false; +#else + std::ofstream output(binFilename.c_str(), std::ofstream::binary); + if (!output.is_open()) return false; +#endif +#else + std::ofstream output(binFilename.c_str(), std::ofstream::binary); + if (!output.is_open()) return false; +#endif + if (data.size() > 0) { + output.write(reinterpret_cast(&data[0]), + std::streamsize(data.size())); + } else { + // Issue #229 + // size 0 will be still valid buffer data. + // write empty file. + } + return true; +} + +#if 0 // FIXME(syoyo): not used. will be removed in the future release. +static void SerializeParameterMap(ParameterMap ¶m, json &o) { + for (ParameterMap::iterator paramIt = param.begin(); paramIt != param.end(); + ++paramIt) { + if (paramIt->second.number_array.size()) { + SerializeNumberArrayProperty(paramIt->first, + paramIt->second.number_array, o); + } else if (paramIt->second.json_double_value.size()) { + json json_double_value; + for (std::map::iterator it = + paramIt->second.json_double_value.begin(); + it != paramIt->second.json_double_value.end(); ++it) { + if (it->first == "index") { + json_double_value[it->first] = paramIt->second.TextureIndex(); + } else { + json_double_value[it->first] = it->second; + } + } + + o[paramIt->first] = json_double_value; + } else if (!paramIt->second.string_value.empty()) { + SerializeStringProperty(paramIt->first, paramIt->second.string_value, o); + } else if (paramIt->second.has_number_value) { + o[paramIt->first] = paramIt->second.number_value; + } else { + o[paramIt->first] = paramIt->second.bool_value; + } + } +} +#endif + +static void SerializeExtensionMap(const ExtensionMap &extensions, json &o) { + if (!extensions.size()) return; + + json extMap; + for (ExtensionMap::const_iterator extIt = extensions.begin(); + extIt != extensions.end(); ++extIt) { + // Allow an empty object for extension(#97) + json ret; + bool isNull = true; + if (ValueToJson(extIt->second, &ret)) { + isNull = JsonIsNull(ret); + JsonAddMember(extMap, extIt->first.c_str(), std::move(ret)); + } + if (isNull) { + if (!(extIt->first.empty())) { // name should not be empty, but for sure + // create empty object so that an extension name is still included in + // json. + json empty; + JsonSetObject(empty); + JsonAddMember(extMap, extIt->first.c_str(), std::move(empty)); + } + } + } + JsonAddMember(o, "extensions", std::move(extMap)); +} + +static void SerializeGltfAccessor(Accessor &accessor, json &o) { + if (accessor.bufferView >= 0) + SerializeNumberProperty("bufferView", accessor.bufferView, o); + + if (accessor.byteOffset != 0) + SerializeNumberProperty("byteOffset", int(accessor.byteOffset), o); + + SerializeNumberProperty("componentType", accessor.componentType, o); + SerializeNumberProperty("count", accessor.count, o); + + if ((accessor.componentType == TINYGLTF_COMPONENT_TYPE_FLOAT) || + (accessor.componentType == TINYGLTF_COMPONENT_TYPE_DOUBLE)) { + SerializeNumberArrayProperty("min", accessor.minValues, o); + SerializeNumberArrayProperty("max", accessor.maxValues, o); + } else { + // Issue #301. Serialize as integer. + // Assume int value is within [-2**31-1, 2**31-1] + { + std::vector values; + std::transform(accessor.minValues.begin(), accessor.minValues.end(), + std::back_inserter(values), + [](double v) { return static_cast(v); }); + + SerializeNumberArrayProperty("min", values, o); + } + + { + std::vector values; + std::transform(accessor.maxValues.begin(), accessor.maxValues.end(), + std::back_inserter(values), + [](double v) { return static_cast(v); }); + + SerializeNumberArrayProperty("max", values, o); + } + } + + if (accessor.normalized) + SerializeValue("normalized", Value(accessor.normalized), o); + std::string type; + switch (accessor.type) { + case TINYGLTF_TYPE_SCALAR: + type = "SCALAR"; + break; + case TINYGLTF_TYPE_VEC2: + type = "VEC2"; + break; + case TINYGLTF_TYPE_VEC3: + type = "VEC3"; + break; + case TINYGLTF_TYPE_VEC4: + type = "VEC4"; + break; + case TINYGLTF_TYPE_MAT2: + type = "MAT2"; + break; + case TINYGLTF_TYPE_MAT3: + type = "MAT3"; + break; + case TINYGLTF_TYPE_MAT4: + type = "MAT4"; + break; + } + + SerializeStringProperty("type", type, o); + if (!accessor.name.empty()) SerializeStringProperty("name", accessor.name, o); + + if (accessor.extras.Type() != NULL_TYPE) { + SerializeValue("extras", accessor.extras, o); + } +} + +static void SerializeGltfAnimationChannel(AnimationChannel &channel, json &o) { + SerializeNumberProperty("sampler", channel.sampler, o); + { + json target; + SerializeNumberProperty("node", channel.target_node, target); + SerializeStringProperty("path", channel.target_path, target); + + SerializeExtensionMap(channel.target_extensions, target); + + JsonAddMember(o, "target", std::move(target)); + } + + if (channel.extras.Type() != NULL_TYPE) { + SerializeValue("extras", channel.extras, o); + } + + SerializeExtensionMap(channel.extensions, o); +} + +static void SerializeGltfAnimationSampler(AnimationSampler &sampler, json &o) { + SerializeNumberProperty("input", sampler.input, o); + SerializeNumberProperty("output", sampler.output, o); + SerializeStringProperty("interpolation", sampler.interpolation, o); + + if (sampler.extras.Type() != NULL_TYPE) { + SerializeValue("extras", sampler.extras, o); + } +} + +static void SerializeGltfAnimation(Animation &animation, json &o) { + if (!animation.name.empty()) + SerializeStringProperty("name", animation.name, o); + + { + json channels; + JsonReserveArray(channels, animation.channels.size()); + for (unsigned int i = 0; i < animation.channels.size(); ++i) { + json channel; + AnimationChannel gltfChannel = animation.channels[i]; + SerializeGltfAnimationChannel(gltfChannel, channel); + JsonPushBack(channels, std::move(channel)); + } + + JsonAddMember(o, "channels", std::move(channels)); + } + + { + json samplers; + JsonReserveArray(samplers, animation.samplers.size()); + for (unsigned int i = 0; i < animation.samplers.size(); ++i) { + json sampler; + AnimationSampler gltfSampler = animation.samplers[i]; + SerializeGltfAnimationSampler(gltfSampler, sampler); + JsonPushBack(samplers, std::move(sampler)); + } + JsonAddMember(o, "samplers", std::move(samplers)); + } + + if (animation.extras.Type() != NULL_TYPE) { + SerializeValue("extras", animation.extras, o); + } + + SerializeExtensionMap(animation.extensions, o); +} + +static void SerializeGltfAsset(Asset &asset, json &o) { + if (!asset.generator.empty()) { + SerializeStringProperty("generator", asset.generator, o); + } + + if (!asset.copyright.empty()) { + SerializeStringProperty("copyright", asset.copyright, o); + } + + if (asset.version.empty()) { + // Just in case + // `version` must be defined + asset.version = "2.0"; + } + + // TODO(syoyo): Do we need to check if `version` is greater or equal to 2.0? + SerializeStringProperty("version", asset.version, o); + + if (asset.extras.Keys().size()) { + SerializeValue("extras", asset.extras, o); + } + + SerializeExtensionMap(asset.extensions, o); +} + +static void SerializeGltfBufferBin(Buffer &buffer, json &o, + std::vector &binBuffer) { + SerializeNumberProperty("byteLength", buffer.data.size(), o); + binBuffer = buffer.data; + + if (buffer.name.size()) SerializeStringProperty("name", buffer.name, o); + + if (buffer.extras.Type() != NULL_TYPE) { + SerializeValue("extras", buffer.extras, o); + } +} + +static void SerializeGltfBuffer(Buffer &buffer, json &o) { + SerializeNumberProperty("byteLength", buffer.data.size(), o); + SerializeGltfBufferData(buffer.data, o); + + if (buffer.name.size()) SerializeStringProperty("name", buffer.name, o); + + if (buffer.extras.Type() != NULL_TYPE) { + SerializeValue("extras", buffer.extras, o); + } +} + +static bool SerializeGltfBuffer(Buffer &buffer, json &o, + const std::string &binFilename, + const std::string &binBaseFilename) { + if (!SerializeGltfBufferData(buffer.data, binFilename)) return false; + SerializeNumberProperty("byteLength", buffer.data.size(), o); + SerializeStringProperty("uri", binBaseFilename, o); + + if (buffer.name.size()) SerializeStringProperty("name", buffer.name, o); + + if (buffer.extras.Type() != NULL_TYPE) { + SerializeValue("extras", buffer.extras, o); + } + return true; +} + +static void SerializeGltfBufferView(BufferView &bufferView, json &o) { + SerializeNumberProperty("buffer", bufferView.buffer, o); + SerializeNumberProperty("byteLength", bufferView.byteLength, o); + + // byteStride is optional, minimum allowed is 4 + if (bufferView.byteStride >= 4) { + SerializeNumberProperty("byteStride", bufferView.byteStride, o); + } + // byteOffset is optional, default is 0 + if (bufferView.byteOffset > 0) { + SerializeNumberProperty("byteOffset", bufferView.byteOffset, o); + } + // Target is optional, check if it contains a valid value + if (bufferView.target == TINYGLTF_TARGET_ARRAY_BUFFER || + bufferView.target == TINYGLTF_TARGET_ELEMENT_ARRAY_BUFFER) { + SerializeNumberProperty("target", bufferView.target, o); + } + if (bufferView.name.size()) { + SerializeStringProperty("name", bufferView.name, o); + } + + if (bufferView.extras.Type() != NULL_TYPE) { + SerializeValue("extras", bufferView.extras, o); + } +} + +static void SerializeGltfImage(Image &image, json &o) { + // if uri empty, the mimeType and bufferview should be set + if (image.uri.empty()) { + SerializeStringProperty("mimeType", image.mimeType, o); + SerializeNumberProperty("bufferView", image.bufferView, o); + } else { + // TODO(syoyo): dlib::urilencode? + SerializeStringProperty("uri", image.uri, o); + } + + if (image.name.size()) { + SerializeStringProperty("name", image.name, o); + } + + if (image.extras.Type() != NULL_TYPE) { + SerializeValue("extras", image.extras, o); + } + + SerializeExtensionMap(image.extensions, o); +} + +static void SerializeGltfTextureInfo(TextureInfo &texinfo, json &o) { + SerializeNumberProperty("index", texinfo.index, o); + + if (texinfo.texCoord != 0) { + SerializeNumberProperty("texCoord", texinfo.texCoord, o); + } + + if (texinfo.extras.Type() != NULL_TYPE) { + SerializeValue("extras", texinfo.extras, o); + } + + SerializeExtensionMap(texinfo.extensions, o); +} + +static void SerializeGltfNormalTextureInfo(NormalTextureInfo &texinfo, + json &o) { + SerializeNumberProperty("index", texinfo.index, o); + + if (texinfo.texCoord != 0) { + SerializeNumberProperty("texCoord", texinfo.texCoord, o); + } + + if (!TINYGLTF_DOUBLE_EQUAL(texinfo.scale, 1.0)) { + SerializeNumberProperty("scale", texinfo.scale, o); + } + + if (texinfo.extras.Type() != NULL_TYPE) { + SerializeValue("extras", texinfo.extras, o); + } + + SerializeExtensionMap(texinfo.extensions, o); +} + +static void SerializeGltfOcclusionTextureInfo(OcclusionTextureInfo &texinfo, + json &o) { + SerializeNumberProperty("index", texinfo.index, o); + + if (texinfo.texCoord != 0) { + SerializeNumberProperty("texCoord", texinfo.texCoord, o); + } + + if (!TINYGLTF_DOUBLE_EQUAL(texinfo.strength, 1.0)) { + SerializeNumberProperty("strength", texinfo.strength, o); + } + + if (texinfo.extras.Type() != NULL_TYPE) { + SerializeValue("extras", texinfo.extras, o); + } + + SerializeExtensionMap(texinfo.extensions, o); +} + +static void SerializeGltfPbrMetallicRoughness(PbrMetallicRoughness &pbr, + json &o) { + std::vector default_baseColorFactor = {1.0, 1.0, 1.0, 1.0}; + if (!Equals(pbr.baseColorFactor, default_baseColorFactor)) { + SerializeNumberArrayProperty("baseColorFactor", pbr.baseColorFactor, + o); + } + + if (!TINYGLTF_DOUBLE_EQUAL(pbr.metallicFactor, 1.0)) { + SerializeNumberProperty("metallicFactor", pbr.metallicFactor, o); + } + + if (!TINYGLTF_DOUBLE_EQUAL(pbr.roughnessFactor, 1.0)) { + SerializeNumberProperty("roughnessFactor", pbr.roughnessFactor, o); + } + + if (pbr.baseColorTexture.index > -1) { + json texinfo; + SerializeGltfTextureInfo(pbr.baseColorTexture, texinfo); + JsonAddMember(o, "baseColorTexture", std::move(texinfo)); + } + + if (pbr.metallicRoughnessTexture.index > -1) { + json texinfo; + SerializeGltfTextureInfo(pbr.metallicRoughnessTexture, texinfo); + JsonAddMember(o, "metallicRoughnessTexture", std::move(texinfo)); + } + + SerializeExtensionMap(pbr.extensions, o); + + if (pbr.extras.Type() != NULL_TYPE) { + SerializeValue("extras", pbr.extras, o); + } +} + +static void SerializeGltfMaterial(Material &material, json &o) { + if (material.name.size()) { + SerializeStringProperty("name", material.name, o); + } + + // QUESTION(syoyo): Write material parameters regardless of its default value? + + if (!TINYGLTF_DOUBLE_EQUAL(material.alphaCutoff, 0.5)) { + SerializeNumberProperty("alphaCutoff", material.alphaCutoff, o); + } + + if (material.alphaMode.compare("OPAQUE") != 0) { + SerializeStringProperty("alphaMode", material.alphaMode, o); + } + + if (material.doubleSided != false) + JsonAddMember(o, "doubleSided", json(material.doubleSided)); + + if (material.normalTexture.index > -1) { + json texinfo; + SerializeGltfNormalTextureInfo(material.normalTexture, texinfo); + JsonAddMember(o, "normalTexture", std::move(texinfo)); + } + + if (material.occlusionTexture.index > -1) { + json texinfo; + SerializeGltfOcclusionTextureInfo(material.occlusionTexture, texinfo); + JsonAddMember(o, "occlusionTexture", std::move(texinfo)); + } + + if (material.emissiveTexture.index > -1) { + json texinfo; + SerializeGltfTextureInfo(material.emissiveTexture, texinfo); + JsonAddMember(o, "emissiveTexture", std::move(texinfo)); + } + + std::vector default_emissiveFactor = {0.0, 0.0, 0.0}; + if (!Equals(material.emissiveFactor, default_emissiveFactor)) { + SerializeNumberArrayProperty("emissiveFactor", + material.emissiveFactor, o); + } + + { + json pbrMetallicRoughness; + SerializeGltfPbrMetallicRoughness(material.pbrMetallicRoughness, + pbrMetallicRoughness); + // Issue 204 + // Do not serialize `pbrMetallicRoughness` if pbrMetallicRoughness has all + // default values(json is null). Otherwise it will serialize to + // `pbrMetallicRoughness : null`, which cannot be read by other glTF + // importers(and validators). + // + if (!JsonIsNull(pbrMetallicRoughness)) { + JsonAddMember(o, "pbrMetallicRoughness", std::move(pbrMetallicRoughness)); + } + } + +#if 0 // legacy way. just for the record. + if (material.values.size()) { + json pbrMetallicRoughness; + SerializeParameterMap(material.values, pbrMetallicRoughness); + JsonAddMember(o, "pbrMetallicRoughness", std::move(pbrMetallicRoughness)); + } + + SerializeParameterMap(material.additionalValues, o); +#else + +#endif + + SerializeExtensionMap(material.extensions, o); + + if (material.extras.Type() != NULL_TYPE) { + SerializeValue("extras", material.extras, o); + } +} + +static void SerializeGltfMesh(Mesh &mesh, json &o) { + json primitives; + JsonReserveArray(primitives, mesh.primitives.size()); + for (unsigned int i = 0; i < mesh.primitives.size(); ++i) { + json primitive; + const Primitive &gltfPrimitive = mesh.primitives[i]; // don't make a copy + { + json attributes; + for (auto attrIt = gltfPrimitive.attributes.begin(); + attrIt != gltfPrimitive.attributes.end(); ++attrIt) { + SerializeNumberProperty(attrIt->first, attrIt->second, attributes); + } + + JsonAddMember(primitive, "attributes", std::move(attributes)); + } + + // Indicies is optional + if (gltfPrimitive.indices > -1) { + SerializeNumberProperty("indices", gltfPrimitive.indices, primitive); + } + // Material is optional + if (gltfPrimitive.material > -1) { + SerializeNumberProperty("material", gltfPrimitive.material, + primitive); + } + SerializeNumberProperty("mode", gltfPrimitive.mode, primitive); + + // Morph targets + if (gltfPrimitive.targets.size()) { + json targets; + JsonReserveArray(targets, gltfPrimitive.targets.size()); + for (unsigned int k = 0; k < gltfPrimitive.targets.size(); ++k) { + json targetAttributes; + std::map targetData = gltfPrimitive.targets[k]; + for (std::map::iterator attrIt = targetData.begin(); + attrIt != targetData.end(); ++attrIt) { + SerializeNumberProperty(attrIt->first, attrIt->second, + targetAttributes); + } + JsonPushBack(targets, std::move(targetAttributes)); + } + JsonAddMember(primitive, "targets", std::move(targets)); + } + + SerializeExtensionMap(gltfPrimitive.extensions, primitive); + + if (gltfPrimitive.extras.Type() != NULL_TYPE) { + SerializeValue("extras", gltfPrimitive.extras, primitive); + } + + JsonPushBack(primitives, std::move(primitive)); + } + + JsonAddMember(o, "primitives", std::move(primitives)); + + if (mesh.weights.size()) { + SerializeNumberArrayProperty("weights", mesh.weights, o); + } + + if (mesh.name.size()) { + SerializeStringProperty("name", mesh.name, o); + } + + SerializeExtensionMap(mesh.extensions, o); + if (mesh.extras.Type() != NULL_TYPE) { + SerializeValue("extras", mesh.extras, o); + } +} + +static void SerializeSpotLight(SpotLight &spot, json &o) { + SerializeNumberProperty("innerConeAngle", spot.innerConeAngle, o); + SerializeNumberProperty("outerConeAngle", spot.outerConeAngle, o); + SerializeExtensionMap(spot.extensions, o); + if (spot.extras.Type() != NULL_TYPE) { + SerializeValue("extras", spot.extras, o); + } +} + +static void SerializeGltfLight(Light &light, json &o) { + if (!light.name.empty()) SerializeStringProperty("name", light.name, o); + SerializeNumberProperty("intensity", light.intensity, o); + if (light.range > 0.0) { + SerializeNumberProperty("range", light.range, o); + } + SerializeNumberArrayProperty("color", light.color, o); + SerializeStringProperty("type", light.type, o); + if (light.type == "spot") { + json spot; + SerializeSpotLight(light.spot, spot); + JsonAddMember(o, "spot", std::move(spot)); + } + SerializeExtensionMap(light.extensions, o); + if (light.extras.Type() != NULL_TYPE) { + SerializeValue("extras", light.extras, o); + } +} + +static void SerializeGltfNode(Node &node, json &o) { + if (node.translation.size() > 0) { + SerializeNumberArrayProperty("translation", node.translation, o); + } + if (node.rotation.size() > 0) { + SerializeNumberArrayProperty("rotation", node.rotation, o); + } + if (node.scale.size() > 0) { + SerializeNumberArrayProperty("scale", node.scale, o); + } + if (node.matrix.size() > 0) { + SerializeNumberArrayProperty("matrix", node.matrix, o); + } + if (node.mesh != -1) { + SerializeNumberProperty("mesh", node.mesh, o); + } + + if (node.skin != -1) { + SerializeNumberProperty("skin", node.skin, o); + } + + if (node.camera != -1) { + SerializeNumberProperty("camera", node.camera, o); + } + + if (node.weights.size() > 0) { + SerializeNumberArrayProperty("weights", node.weights, o); + } + + if (node.extras.Type() != NULL_TYPE) { + SerializeValue("extras", node.extras, o); + } + + SerializeExtensionMap(node.extensions, o); + if (!node.name.empty()) SerializeStringProperty("name", node.name, o); + SerializeNumberArrayProperty("children", node.children, o); +} + +static void SerializeGltfSampler(Sampler &sampler, json &o) { + if (sampler.magFilter != -1) { + SerializeNumberProperty("magFilter", sampler.magFilter, o); + } + if (sampler.minFilter != -1) { + SerializeNumberProperty("minFilter", sampler.minFilter, o); + } + //SerializeNumberProperty("wrapR", sampler.wrapR, o); + SerializeNumberProperty("wrapS", sampler.wrapS, o); + SerializeNumberProperty("wrapT", sampler.wrapT, o); + + if (sampler.extras.Type() != NULL_TYPE) { + SerializeValue("extras", sampler.extras, o); + } +} + +static void SerializeGltfOrthographicCamera(const OrthographicCamera &camera, + json &o) { + SerializeNumberProperty("zfar", camera.zfar, o); + SerializeNumberProperty("znear", camera.znear, o); + SerializeNumberProperty("xmag", camera.xmag, o); + SerializeNumberProperty("ymag", camera.ymag, o); + + if (camera.extras.Type() != NULL_TYPE) { + SerializeValue("extras", camera.extras, o); + } +} + +static void SerializeGltfPerspectiveCamera(const PerspectiveCamera &camera, + json &o) { + SerializeNumberProperty("zfar", camera.zfar, o); + SerializeNumberProperty("znear", camera.znear, o); + if (camera.aspectRatio > 0) { + SerializeNumberProperty("aspectRatio", camera.aspectRatio, o); + } + + if (camera.yfov > 0) { + SerializeNumberProperty("yfov", camera.yfov, o); + } + + if (camera.extras.Type() != NULL_TYPE) { + SerializeValue("extras", camera.extras, o); + } +} + +static void SerializeGltfCamera(const Camera &camera, json &o) { + SerializeStringProperty("type", camera.type, o); + if (!camera.name.empty()) { + SerializeStringProperty("name", camera.name, o); + } + + if (camera.type.compare("orthographic") == 0) { + json orthographic; + SerializeGltfOrthographicCamera(camera.orthographic, orthographic); + JsonAddMember(o, "orthographic", std::move(orthographic)); + } else if (camera.type.compare("perspective") == 0) { + json perspective; + SerializeGltfPerspectiveCamera(camera.perspective, perspective); + JsonAddMember(o, "perspective", std::move(perspective)); + } else { + // ??? + } + + if (camera.extras.Type() != NULL_TYPE) { + SerializeValue("extras", camera.extras, o); + } + SerializeExtensionMap(camera.extensions, o); +} + +static void SerializeGltfScene(Scene &scene, json &o) { + SerializeNumberArrayProperty("nodes", scene.nodes, o); + + if (scene.name.size()) { + SerializeStringProperty("name", scene.name, o); + } + if (scene.extras.Type() != NULL_TYPE) { + SerializeValue("extras", scene.extras, o); + } + SerializeExtensionMap(scene.extensions, o); +} + +static void SerializeGltfSkin(Skin &skin, json &o) { + // required + SerializeNumberArrayProperty("joints", skin.joints, o); + + if (skin.inverseBindMatrices >= 0) { + SerializeNumberProperty("inverseBindMatrices", skin.inverseBindMatrices, o); + } + + if (skin.skeleton >= 0) { + SerializeNumberProperty("skeleton", skin.skeleton, o); + } + + if (skin.name.size()) { + SerializeStringProperty("name", skin.name, o); + } +} + +static void SerializeGltfTexture(Texture &texture, json &o) { + if (texture.sampler > -1) { + SerializeNumberProperty("sampler", texture.sampler, o); + } + if (texture.source > -1) { + SerializeNumberProperty("source", texture.source, o); + } + if (texture.name.size()) { + SerializeStringProperty("name", texture.name, o); + } + if (texture.extras.Type() != NULL_TYPE) { + SerializeValue("extras", texture.extras, o); + } + SerializeExtensionMap(texture.extensions, o); +} + +/// +/// Serialize all properties except buffers and images. +/// +static void SerializeGltfModel(Model *model, json &o) { + // ACCESSORS + if (model->accessors.size()) { + json accessors; + JsonReserveArray(accessors, model->accessors.size()); + for (unsigned int i = 0; i < model->accessors.size(); ++i) { + json accessor; + SerializeGltfAccessor(model->accessors[i], accessor); + JsonPushBack(accessors, std::move(accessor)); + } + JsonAddMember(o, "accessors", std::move(accessors)); + } + + // ANIMATIONS + if (model->animations.size()) { + json animations; + JsonReserveArray(animations, model->animations.size()); + for (unsigned int i = 0; i < model->animations.size(); ++i) { + if (model->animations[i].channels.size()) { + json animation; + SerializeGltfAnimation(model->animations[i], animation); + JsonPushBack(animations, std::move(animation)); + } + } + + JsonAddMember(o, "animations", std::move(animations)); + } + + // ASSET + json asset; + SerializeGltfAsset(model->asset, asset); + JsonAddMember(o, "asset", std::move(asset)); + + // BUFFERVIEWS + if (model->bufferViews.size()) { + json bufferViews; + JsonReserveArray(bufferViews, model->bufferViews.size()); + for (unsigned int i = 0; i < model->bufferViews.size(); ++i) { + json bufferView; + SerializeGltfBufferView(model->bufferViews[i], bufferView); + JsonPushBack(bufferViews, std::move(bufferView)); + } + JsonAddMember(o, "bufferViews", std::move(bufferViews)); + } + + // Extensions required + if (model->extensionsRequired.size()) { + SerializeStringArrayProperty("extensionsRequired", + model->extensionsRequired, o); + } + + // MATERIALS + if (model->materials.size()) { + json materials; + JsonReserveArray(materials, model->materials.size()); + for (unsigned int i = 0; i < model->materials.size(); ++i) { + json material; + SerializeGltfMaterial(model->materials[i], material); + + if (JsonIsNull(material)) { + // Issue 294. + // `material` does not have any required parameters + // so the result may be null(unmodified) when all material parameters + // have default value. + // + // null is not allowed thus we create an empty JSON object. + JsonSetObject(material); + } + JsonPushBack(materials, std::move(material)); + } + JsonAddMember(o, "materials", std::move(materials)); + } + + // MESHES + if (model->meshes.size()) { + json meshes; + JsonReserveArray(meshes, model->meshes.size()); + for (unsigned int i = 0; i < model->meshes.size(); ++i) { + json mesh; + SerializeGltfMesh(model->meshes[i], mesh); + JsonPushBack(meshes, std::move(mesh)); + } + JsonAddMember(o, "meshes", std::move(meshes)); + } + + // NODES + if (model->nodes.size()) { + json nodes; + JsonReserveArray(nodes, model->nodes.size()); + for (unsigned int i = 0; i < model->nodes.size(); ++i) { + json node; + SerializeGltfNode(model->nodes[i], node); + JsonPushBack(nodes, std::move(node)); + } + JsonAddMember(o, "nodes", std::move(nodes)); + } + + // SCENE + if (model->defaultScene > -1) { + SerializeNumberProperty("scene", model->defaultScene, o); + } + + // SCENES + if (model->scenes.size()) { + json scenes; + JsonReserveArray(scenes, model->scenes.size()); + for (unsigned int i = 0; i < model->scenes.size(); ++i) { + json currentScene; + SerializeGltfScene(model->scenes[i], currentScene); + JsonPushBack(scenes, std::move(currentScene)); + } + JsonAddMember(o, "scenes", std::move(scenes)); + } + + // SKINS + if (model->skins.size()) { + json skins; + JsonReserveArray(skins, model->skins.size()); + for (unsigned int i = 0; i < model->skins.size(); ++i) { + json skin; + SerializeGltfSkin(model->skins[i], skin); + JsonPushBack(skins, std::move(skin)); + } + JsonAddMember(o, "skins", std::move(skins)); + } + + // TEXTURES + if (model->textures.size()) { + json textures; + JsonReserveArray(textures, model->textures.size()); + for (unsigned int i = 0; i < model->textures.size(); ++i) { + json texture; + SerializeGltfTexture(model->textures[i], texture); + JsonPushBack(textures, std::move(texture)); + } + JsonAddMember(o, "textures", std::move(textures)); + } + + // SAMPLERS + if (model->samplers.size()) { + json samplers; + JsonReserveArray(samplers, model->samplers.size()); + for (unsigned int i = 0; i < model->samplers.size(); ++i) { + json sampler; + SerializeGltfSampler(model->samplers[i], sampler); + JsonPushBack(samplers, std::move(sampler)); + } + JsonAddMember(o, "samplers", std::move(samplers)); + } + + // CAMERAS + if (model->cameras.size()) { + json cameras; + JsonReserveArray(cameras, model->cameras.size()); + for (unsigned int i = 0; i < model->cameras.size(); ++i) { + json camera; + SerializeGltfCamera(model->cameras[i], camera); + JsonPushBack(cameras, std::move(camera)); + } + JsonAddMember(o, "cameras", std::move(cameras)); + } + + // EXTENSIONS + SerializeExtensionMap(model->extensions, o); + + auto extensionsUsed = model->extensionsUsed; + + // LIGHTS as KHR_lights_punctual + if (model->lights.size()) { + json lights; + JsonReserveArray(lights, model->lights.size()); + for (unsigned int i = 0; i < model->lights.size(); ++i) { + json light; + SerializeGltfLight(model->lights[i], light); + JsonPushBack(lights, std::move(light)); + } + json khr_lights_cmn; + JsonAddMember(khr_lights_cmn, "lights", std::move(lights)); + json ext_j; + + { + json_const_iterator it; + if (FindMember(o, "extensions", it)) { + JsonAssign(ext_j, GetValue(it)); + } + } + + JsonAddMember(ext_j, "KHR_lights_punctual", std::move(khr_lights_cmn)); + + JsonAddMember(o, "extensions", std::move(ext_j)); + + // Also add "KHR_lights_punctual" to `extensionsUsed` + { + auto has_khr_lights_punctual = + std::find_if(extensionsUsed.begin(), extensionsUsed.end(), + [](const std::string &s) { + return (s.compare("KHR_lights_punctual") == 0); + }); + + if (has_khr_lights_punctual == extensionsUsed.end()) { + extensionsUsed.push_back("KHR_lights_punctual"); + } + } + } + + // Extensions used + if (extensionsUsed.size()) { + SerializeStringArrayProperty("extensionsUsed", extensionsUsed, o); + } + + // EXTRAS + if (model->extras.Type() != NULL_TYPE) { + SerializeValue("extras", model->extras, o); + } +} + +static bool WriteGltfStream(std::ostream &stream, const std::string &content) { + stream << content << std::endl; + return true; +} + +static bool WriteGltfFile(const std::string &output, + const std::string &content) { +#ifdef _WIN32 +#if defined(_MSC_VER) + std::ofstream gltfFile(UTF8ToWchar(output).c_str()); +#elif defined(__GLIBCXX__) + int file_descriptor = _wopen(UTF8ToWchar(output).c_str(), + _O_CREAT | _O_WRONLY | _O_TRUNC | _O_BINARY); + __gnu_cxx::stdio_filebuf wfile_buf( + file_descriptor, std::ios_base::out | std::ios_base::binary); + std::ostream gltfFile(&wfile_buf); + if (!wfile_buf.is_open()) return false; +#else + std::ofstream gltfFile(output.c_str()); + if (!gltfFile.is_open()) return false; +#endif +#else + std::ofstream gltfFile(output.c_str()); + if (!gltfFile.is_open()) return false; +#endif + return WriteGltfStream(gltfFile, content); +} + +static void WriteBinaryGltfStream(std::ostream &stream, + const std::string &content, + const std::vector &binBuffer) { + const std::string header = "glTF"; + const int version = 2; + + const uint32_t content_size = uint32_t(content.size()); + const uint32_t binBuffer_size = uint32_t(binBuffer.size()); + // determine number of padding bytes required to ensure 4 byte alignment + const uint32_t content_padding_size = content_size % 4 == 0 ? 0 : 4 - content_size % 4; + const uint32_t bin_padding_size = binBuffer_size % 4 == 0 ? 0 : 4 - binBuffer_size % 4; + + // 12 bytes for header, JSON content length, 8 bytes for JSON chunk info. + // Chunk data must be located at 4-byte boundary, which may require padding + const uint32_t length = + 12 + 8 + content_size + content_padding_size + + (binBuffer_size ? (8 + binBuffer_size + bin_padding_size) : 0); + + stream.write(header.c_str(), std::streamsize(header.size())); + stream.write(reinterpret_cast(&version), sizeof(version)); + stream.write(reinterpret_cast(&length), sizeof(length)); + + // JSON chunk info, then JSON data + const uint32_t model_length = uint32_t(content.size()) + content_padding_size; + const uint32_t model_format = 0x4E4F534A; + stream.write(reinterpret_cast(&model_length), + sizeof(model_length)); + stream.write(reinterpret_cast(&model_format), + sizeof(model_format)); + stream.write(content.c_str(), std::streamsize(content.size())); + + // Chunk must be multiplies of 4, so pad with spaces + if (content_padding_size > 0) { + const std::string padding = std::string(size_t(content_padding_size), ' '); + stream.write(padding.c_str(), std::streamsize(padding.size())); + } + if (binBuffer.size() > 0) { + // BIN chunk info, then BIN data + const uint32_t bin_length = uint32_t(binBuffer.size()) + bin_padding_size; + const uint32_t bin_format = 0x004e4942; + stream.write(reinterpret_cast(&bin_length), + sizeof(bin_length)); + stream.write(reinterpret_cast(&bin_format), + sizeof(bin_format)); + stream.write(reinterpret_cast(binBuffer.data()), + std::streamsize(binBuffer.size())); + // Chunksize must be multiplies of 4, so pad with zeroes + if (bin_padding_size > 0) { + const std::vector padding = + std::vector(size_t(bin_padding_size), 0); + stream.write(reinterpret_cast(padding.data()), + std::streamsize(padding.size())); + } + } +} + +static void WriteBinaryGltfFile(const std::string &output, + const std::string &content, + const std::vector &binBuffer) { +#ifdef _WIN32 +#if defined(_MSC_VER) + std::ofstream gltfFile(UTF8ToWchar(output).c_str(), std::ios::binary); +#elif defined(__GLIBCXX__) + int file_descriptor = _wopen(UTF8ToWchar(output).c_str(), + _O_CREAT | _O_WRONLY | _O_TRUNC | _O_BINARY); + __gnu_cxx::stdio_filebuf wfile_buf( + file_descriptor, std::ios_base::out | std::ios_base::binary); + std::ostream gltfFile(&wfile_buf); +#else + std::ofstream gltfFile(output.c_str(), std::ios::binary); +#endif +#else + std::ofstream gltfFile(output.c_str(), std::ios::binary); +#endif + WriteBinaryGltfStream(gltfFile, content, binBuffer); +} + +bool TinyGLTF::WriteGltfSceneToStream(Model *model, std::ostream &stream, + bool prettyPrint = true, + bool writeBinary = false) { + JsonDocument output; + + /// Serialize all properties except buffers and images. + SerializeGltfModel(model, output); + + // BUFFERS + std::vector binBuffer; + if (model->buffers.size()) { + json buffers; + JsonReserveArray(buffers, model->buffers.size()); + for (unsigned int i = 0; i < model->buffers.size(); ++i) { + json buffer; + if (writeBinary && i == 0 && model->buffers[i].uri.empty()) { + SerializeGltfBufferBin(model->buffers[i], buffer, binBuffer); + } else { + SerializeGltfBuffer(model->buffers[i], buffer); + } + JsonPushBack(buffers, std::move(buffer)); + } + JsonAddMember(output, "buffers", std::move(buffers)); + } + + // IMAGES + if (model->images.size()) { + json images; + JsonReserveArray(images, model->images.size()); + for (unsigned int i = 0; i < model->images.size(); ++i) { + json image; + + std::string dummystring = ""; + // UpdateImageObject need baseDir but only uses it if embeddedImages is + // enabled, since we won't write separate images when writing to a stream + // we + UpdateImageObject(model->images[i], dummystring, int(i), false, + &this->WriteImageData, this->write_image_user_data_); + SerializeGltfImage(model->images[i], image); + JsonPushBack(images, std::move(image)); + } + JsonAddMember(output, "images", std::move(images)); + } + + if (writeBinary) { + WriteBinaryGltfStream(stream, JsonToString(output), binBuffer); + } else { + WriteGltfStream(stream, JsonToString(output, prettyPrint ? 2 : -1)); + } + + return true; +} + +bool TinyGLTF::WriteGltfSceneToFile(Model *model, const std::string &filename, + bool embedImages = false, + bool embedBuffers = false, + bool prettyPrint = true, + bool writeBinary = false) { + JsonDocument output; + std::string defaultBinFilename = GetBaseFilename(filename); + std::string defaultBinFileExt = ".bin"; + std::string::size_type pos = + defaultBinFilename.rfind('.', defaultBinFilename.length()); + + if (pos != std::string::npos) { + defaultBinFilename = defaultBinFilename.substr(0, pos); + } + std::string baseDir = GetBaseDir(filename); + if (baseDir.empty()) { + baseDir = "./"; + } + /// Serialize all properties except buffers and images. + SerializeGltfModel(model, output); + + // BUFFERS + std::vector usedUris; + std::vector binBuffer; + if (model->buffers.size()) { + json buffers; + JsonReserveArray(buffers, model->buffers.size()); + for (unsigned int i = 0; i < model->buffers.size(); ++i) { + json buffer; + if (writeBinary && i == 0 && model->buffers[i].uri.empty()) { + SerializeGltfBufferBin(model->buffers[i], buffer, binBuffer); + } else if (embedBuffers) { + SerializeGltfBuffer(model->buffers[i], buffer); + } else { + std::string binSavePath; + std::string binUri; + if (!model->buffers[i].uri.empty() && + !IsDataURI(model->buffers[i].uri)) { + binUri = model->buffers[i].uri; + } else { + binUri = defaultBinFilename + defaultBinFileExt; + bool inUse = true; + int numUsed = 0; + while (inUse) { + inUse = false; + for (const std::string &usedName : usedUris) { + if (binUri.compare(usedName) != 0) continue; + inUse = true; + binUri = defaultBinFilename + std::to_string(numUsed++) + + defaultBinFileExt; + break; + } + } + } + usedUris.push_back(binUri); + binSavePath = JoinPath(baseDir, binUri); + if (!SerializeGltfBuffer(model->buffers[i], buffer, binSavePath, + binUri)) { + return false; + } + } + JsonPushBack(buffers, std::move(buffer)); + } + JsonAddMember(output, "buffers", std::move(buffers)); + } + + // IMAGES + if (model->images.size()) { + json images; + JsonReserveArray(images, model->images.size()); + for (unsigned int i = 0; i < model->images.size(); ++i) { + json image; + + UpdateImageObject(model->images[i], baseDir, int(i), embedImages, + &this->WriteImageData, this->write_image_user_data_); + SerializeGltfImage(model->images[i], image); + JsonPushBack(images, std::move(image)); + } + JsonAddMember(output, "images", std::move(images)); + } + + if (writeBinary) { + WriteBinaryGltfFile(filename, JsonToString(output), binBuffer); + } else { + WriteGltfFile(filename, JsonToString(output, (prettyPrint ? 2 : -1))); + } + + return true; +} + +} // namespace tinygltf + +#ifdef __clang__ +#pragma clang diagnostic pop +#endif + +#endif // TINYGLTF_IMPLEMENTATION \ No newline at end of file