* 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.
Needed because we're using true OS threading, so we can't do the normal JSystem heap changing on thread change.
Requires encapsulating it (removing the public static field)
uintptr_t == u32 == unsigned int on 32-bit MSVC since u32 was changed from unsigned long to unsigned int for TARGET_PC. Guard the uintptr_t overload to x64 only where uintptr_t != u32.
They store the data address in the same struct that's directly loaded from archive files. This means it's impossible to expand to 64-bit directly, and instead we need to allocate the data pointers in a separate array.
There are a few places where JASGlobalInstance have their sInstance storage implemented via template specialization. The problem is that these have no initializer, which means that they are not proper definitions (only declarations) in standards-compliant C++. MSVC and (evidently) MWCC accept this, but modern Clang and GCC do not and won't emit symbols.
I've added a macro that properly initializes these outside MWCC.
Also, JASGlobalInstance<JAUSectionHeap> was only being declared in each actor file that used it, which sounds incorrect? Not sure about this but I added it to m_Do_main too, again behind MWERKS check.
Major changes:
- Implement Big-Endian to Little-Endian byte-swapping for all RARC archive
types (JKRCompArchive, JKRMemArchive, JKRDvdArchive, JKRAramArchive)
- Implement DVD file I/O via DvdEmu (DVDOpen, DVDFastOpen, DVDReadPrio,
DVDReadAsyncPrio, DVDConvertPathToEntrynum)
- Fix YAZ0 decompression endianness in JKRDvdRipper, JKRDecomp, JKRAram,
and JKRDvdAramRipper (use JKRDecompExpandSize instead of direct header read)
- Emulate ARAM with 16MB malloc buffer and synchronous memcpy in ARQPostRequest
instead of hardware DMA transfers that hang on PC
- Add real OS threading implementation (OSContext, OSThread, OSMutex) using
native Windows threads with side-table pattern for GC struct compatibility
- Fix font endianness for JUTResFont and JUTCacheFont
- Redirect GXVert.h to Aurora's PC implementation to prevent FIFO writes to
the GameCube hardware address 0xCC008000
- Add Z-buffer texture format support (GX_TF_Z24X8, GX_TF_Z16, GX_TF_Z8)
in Aurora's texture converter
Luke Street