The framebuffer size flows into Aurora as follows:
Into JUTVideo:
AURORA_WINDOW_RESIZED handler ->
mDoGph_gInf_c::setWindowSize ->
JUTVideo::setWindowSize
From JUTVideo -> dComIfG_play_c:
fpcCt_Handler ->
phase_4 ->
dComIfGp_setWindow
From dComIfGp_play_c -> Aurora:
mDoGph_Painter ->
drawDepth2 ->
GXSetTexCopySrc ->
GXCopyTex ->
aurora::gfx::resolve_pass (called by GXCopyTex)
The trouble is that 1) phase_4 is only called when creating a new scene/room,
and 2) even when it is called, mDoGph_Painter is executed before fpcCt_Handler
in the main loop so the new size isn't visible to aurora::gfx::resolve_pass
until one frame after it's set. This causes a crash due to the original size
being passed to CopyTextureToTexture despite the destination framebuffer now
being smaller.
As a workaround, we just cut out the whole middleman section and call
dComIfGp_setWindow immediately when we receive a resize event. This way the new
size is immediately visible to Aurora.
* launch.json cwd
* bodge to load gci for testing
* stub card stat
* gameplay bodges
* viewport, ub fixes
* add release with debug info cmake variant
* be fixes, sound stub
* viewport h
* d_msg_flow BE
* be fopAcM_createItemFromEnemyID
* update launch configuration to use iso
* more audio stubs
* Attempt to set viewport and get messages for brightness check
* skip opening scene again, fixed JMessage::TResourceContainer::TCResource::Do_destroy
* add guards for viewport changes
* moar endian swapping to get Link sitting in PROC_OPENING_SCENE and for dialogues
* BE d_msg_class i_data
* stub bgm start
* fix div by 0 error (for now)
* TEMP_BROKEN in d_menu_ring
* REQUIRES_GX_LINES
* properly stub renderingAmap::draw with REQUIRES_GX_LINES
* better stubbing outside of stubs
* fix event data getting swapped multiple times
* evil draw vp fix
* Stub log imgui
This redirects all spammy logs to an imgui window that is cleared per frame.
This fixes the serious performance dip of the logging, and makes the regular log readable.
* Oops move those optimization changes I accidentally committed behind a flag
DUSK_SELECTED_OPT
* gx_line macro in map
* fix audio stubbing
* switch to CARD API aurora impl
* remove kabufuda from link libs
* refactor imgui stuff and add input viewer
* merge stub log with refactor
* accidentally committed a metaforce header shh
* basic map loader
* ImGuiConsole: Add missing <thread> include
* you may now play as luigi (you may now load stages with bridges)
* bloom fix
* bloom leak fix
* cloud shadow fix
* add soft reset button to imgui menu
* if it broke dont not fix it
* i swear i committed this
* BE swap indMtx in JPAResource::setPTev
* wnark ct fix
* frsqrte implementation from kinoko
* Fix Clang compile error in JAISeq::prepare_getSeqData_
* Add endian conversions to dMsgFlow_c::getInitNodeIndex
This fixes a freeze when Fado tries to stop you from leaving the
starting area.
* Add RAII GXTexObj wrapper; fix almost all leaks
* Update aurora for indirect texturing
* Update aurora for CARD fix
* Fix Clang build
* More d_msg_flow endian fixes
Fixes softlock when trying to talk to Fado and possibly other NPCs.
* no frame limiter
* get pause menu working
* proper frame limiting
* particle pointer size fix
* improve map loader a bit
---------
Co-authored-by: Jasper St. Pierre <jstpierre@mecheye.net>
Co-authored-by: TakaRikka <takarikka@outlook.com>
Co-authored-by: CraftyBoss <talibabdulmaalik@gmail.com>
Co-authored-by: Luke Street <luke@street.dev>
Co-authored-by: Lurs <2795933+Lurs@users.noreply.github.com>
Co-authored-by: PJB3005 <pieterjan.briers+git@gmail.com>
Co-authored-by: tgsm <doodrabbit@hotmail.com>
Co-authored-by: Max Roncace <me@caseif.net>
Co-authored-by: Phillip Stephens <antidote.crk@gmail.com>
* Undo array allocation changes from #43
Doesn't work
* Expand dmeter heap sizes, give names
* Fix manual operator delete call in resource.cpp
* Disable map rendering for now
Aurora can't handle lines
* Re-enable assert heap on DVD thread
Should be fine?
* Some basic debug groups with the new Aurora API
* Allow Aurora backend to be set via CLI
* Give materials debug groups
* More debug groups
* JKRHeap separation: array edition
Pain
* fix kankyo UB
* Fix TEV Stage DL writing
BE issue
* BE Fixes in materials/shapes
* Move to Aurora GD impl
* JUTDataFileHeader
* j3d: load vertex and texture not through GD
* Endian swap vertex data (mostly)
* Just exit(0) when closing the game
Fix crashes :godo:
* fix fopAcM_ct_placement and remove memcpy
* J3D: track vertex arrays correctly, swap work
* fix visibility, turn off overridden new/delete when we call into aurora
* event: cut name be
* Default window improvements
Double size, allow OS to decide position
* survive TParse::parseHeader_next
* color swap fix
* swap endian/fixes oob function pointer
* Remove GXGetViewportv stub
Aurora has it now
* Set array pointers via GD again
Relies on https://github.com/encounter/aurora/pull/35
* Let Aurora decide graphics backend
* disable procbar drawing
* Fix TColor endianness conversion issues
Fixes the wrong color for the flashing logo
* cam param swap
* simplify vtx loading, mat/amb color fix
* endian swap stb/fvb data
* aurora stat changes
* fix storage buffer crash
dont unnecessarily reassign vertex buffers in a way that forces aurora to recache things
* set bgobj spec fix
* add camera debug, endian swap more stb data, d_a_bg_obj::spec_data_c swapped
* JKRExpHeap: fix bad allocator slowdown
* solid tex
* texture caching
* log level
* fix window aspect ratio, disable bloom, endian swap camera type data
camera is now actually playing opening cutscene!!!!
* add GXDestroyTexObj to a couple stack init'd GXTexObjs, remove some diag printfs
* endian swap joint weight envelope data
* move to dusk config
* verbose arg errors
* better stub logging (for now)
* less logging, more BE
* more stubbing, move logging stuff around
* move all logging to aurora logging
* fix STUB_LOG, __FUNCTION__ isn't a string literal, wasn't building as it
was.
* update aurora
* fix heap alignment mismatch and always head align for now
* prevent them from fucking up my shit
* forward jut warning to DuskLog
* remove report logging
* maybe and i must emphasize maybe fix JMessage parsing
* this was a dumb idea
* preserve negative alignment for JKRHeap
* use normal free on macos and linux
* ^
* fix JMAHermiteInterpolation c impl
* endian swap J2DScreen mColor (oops)
* swap more J3D anim data, remove weird pointer addr check in J3DMaterial getMaterialAnm
* typo fix
* Fix aligned_alloc() size issues on POSIX
aligned_alloc() requires its input size to be a multiple of alignment. This wasn't being upheld so there were allocation failures in init code that made the game fail to start outside Windows.
Also just cleaned up some of this code a bit and removed fallback cases that *shouldn't* get hit.
* _Exit instead of exit()
Seriously I don't want destructors to run. Let the OS reclaim that shit.
* Reapply "Isolate JKRHeap operator overloads" (#39)
This reverts commit 3623b27f37.
* Fix some oopsies
* Fix hardcoded pointer size in JUTCacheFont::allocArray
* More operator overload fixes
Add void template specialization for jkrDelete
Add new[] placement overload. Apparently.
* Fix delete macros on nullptr
TIL C++ allows that.
* fix delete[]
* fix new(std::nothrow) overload
* fix avoid ub
* swap remaining anim data needed for title logo
* get rid of op 7
* move aurora_end_frame to correct spot
* juttexture destroy tex
* j2d animation be
* shutdown crash
* link warp material fix
* mDoExt_3DlineMat1_c fix
* hacky keyboard controls
* endian swap J2DResReference, add kb_pad to files.cmake
* fix some missing endian swaps in J2D, remove addr alignment check
* Remove heap unsetting in aurora calls
Never worked properly and not the right solution even if it did
* Don't print in DC stubs
They're fine to never implement
* Fix alignment stuff again
* Compile GF from dolphin lib
Doesn't seem to break anything and shuts up some stub warnings
* j3dtexture tlut obj fix
* addTexMtxIndexInDL fix
* don't recreate null tex data every frame
* the actual fix i wanted to push
* its kinda fakematch city over here
* insert hte efb copy
* limited window size / viewport support
* IsDelete FIX
---------
Co-authored-by: madeline <qwertytrogi@gmail.com>
Co-authored-by: Jasper St. Pierre <jstpierre@mecheye.net>
Co-authored-by: Jeffrey Crowell <jeff@crowell.biz>
Co-authored-by: TakaRikka <takarikka@outlook.com>
Co-authored-by: CraftyBoss <talibabdulmaalik@gmail.com>
Co-authored-by: Lurs <2795933+Lurs@users.noreply.github.com>
* Wrap >4-char literals in a MULTI_CHAR macro
Modern compilers do not support CW's non-standard behavior with
>4 char literals. We can, however, use a constexpr function to
compute the u64 values directly. This leaves <=4 char literals
unchanged.
* Replace non-pointer usages of NULL with 0
* Define NULL to nullptr on C++11 and above
* Fix more -Wpointer-arith and -Woverflow warnings
* Replace u32/s32 with uintptr_t/intptr_t where appropriate
* JSUOutputStream: Overload all standard int types
* Fix remaining <string> -> <cstring> for GCC compilation (#3114 follow-up)
MWerks' <string> header transitively includes C string functions
(memcpy, strlen, strcmp, etc.), but GCC/Clang's <string> is the C++
std::string header. These files all use C string functions and should
include <cstring> instead.
* Use std::isnan instead of isnan for GCC compilation
GCC's <cmath> places isnan in the std namespace. Using the unqualified
isnan fails to compile with GCC/Clang.
* Fix cCcD_Src types: s32 -> u32 for bitmask fields
cCcD_SrcObjCommonBase::mSPrm, cCcD_SrcObjTg::mType, and
cCcD_SrcObjAt::mType are used as bitmasks (SetType/SetSPrm take u32,
MskType/MskSPrm use u32, values like 0xFFFFFFFF are common in
aggregate inits). Change from s32 to u32 to match usage.
Also fix AT_TYPE_WOLF_ATTACK/AT_TYPE_UNK to use unsigned literals,
and remove now-unnecessary (s32) casts on hex literals in collision
source data.
* Mark dummy() functions as static to avoid multiple definition errors
These decomp artifact functions have the same name and signature across
TUs, causing linker errors when building as a single binary.
Fixes#25
This isolates the JKRHeap operator new/delete overloads. Every single new/delete site in the code has been replaced with a macro.
Sadly for new[] and delete[] we have to keep global operators. The global new[] just allocates into malloc() however, and delete[] goes into free() if it's not in a JKRHeap. So that's fine.
PJB3005