fix: [ 2550645 ] Super Princess Peach: crash when starting a puzzle

This commit is contained in:
zeromus 2009-01-31 04:47:23 +00:00
parent 98c68d8d89
commit 19f9cd36fc
2 changed files with 21 additions and 8 deletions

View File

@ -1,18 +1,25 @@
0.9+ 0.9 -> 0/10
General/Core:
bug: SF [ 2550645 ] Super Princess Peach: crash when starting a puzzle [zeromus]
bug: win32: fix spu wav file writer (regression from 0.8) [zeromus]
bug: More variables added to savestate. Savestates probably invalidated.
Graphics: Graphics:
bug: Fix some errors in rendering 512 tall BG [zeromus,CrazyMax] (still not fully resolved) bug: Fix some errors in rendering 512 tall BG [zeromus,CrazyMax] (still not fully resolved)
bug: 3d compositing integrated into GPU engine (more precision and speed in cases where no compositing needed) [luigi__] bug: 3d compositing integrated into GPU engine (more precision and speed in cases where no compositing needed) [luigi__]
bug: Add optional fragment shading pipeline (more precision) [luigi__] bug: Add optional fragment shading pipeline (more precision) [luigi__]
bug: fix texture mapping across texture banks [zeromus] bug: fix texture mapping across texture and palette banks [zeromus]
bug: fix texture handling for texels with MSB set (fixes some texture corruption) [zeromus] bug: fix texture handling for texels with MSB set (fixes some texture corruption) [zeromus]
bug: fix thumb ROR [zeromus] bug: fix arm/thumb ROR [zeromus]
bug: win32: fix spu wav file writer (regression from 0.8) [zeromus] bug: fix texture coordinate generation from normal (fixes environment mapping) [zeromus]
enh: dont memcmp for texcache validity when the texture banks arent unmapping [zeromus] bug: fix display capturing from 2d+3d source when display mode is display-vram [zeromus]
bug: more correct mosaic effects [zeromus]
enh: dont memcmp for texcache validity when the texture banks arent unmapping (speedup) [zeromus]
Mac OS X Port: Mac OS X Port:
bug: left key setting no longer gets confused with the right key [jeff] bug: left key setting no longer gets confused with the right key [jeff]
enh: more keys can be mapped enh: more keys can be mapped
0.8 -> 0.9 0.8 -> 0.9
There have been so many changes that this list can hardly be considered complete. There have been so many changes that this list can hardly be considered complete.
For this release, we have formatted the list into a high level overview of big changes and bad bugs relative to the previous release. Hopefully what you're interested in. For this release, we have formatted the list into a high level overview of big changes and bad bugs relative to the previous release. Hopefully what you're interested in.

View File

@ -536,6 +536,12 @@ u8 *MMU_RenderMapToLCD(u32 vram_addr)
static FORCEINLINE u32 MMU_LCDmap(u32 addr) static FORCEINLINE u32 MMU_LCDmap(u32 addr)
{ {
//handle LCD memory mirroring
if ((addr < 0x07000000) && (addr>=0x068A4000))
return 0x06800000 +
//(addr%0xA4000); //yuck!! is this even how it mirrors? but we have to keep from overrunning the buffer somehow
(addr&0x80000); //just as likely to be right (I have no clue how it should work) but faster.
if ((addr < 0x6000000)) return addr; if ((addr < 0x6000000)) return addr;
if ((addr > 0x661FFFF)) return addr; // Engine BOBJ max 128KB if ((addr > 0x661FFFF)) return addr; // Engine BOBJ max 128KB
@ -921,7 +927,7 @@ void FASTCALL MMU_doDMA(u32 num)
/* word count */ /* word count */
taille = (MMU.DMACrt[PROCNUM][num]&0xFFFF); taille = (MMU.DMACrt[PROCNUM][num]&0x1FFFF);
// If we are in "Main memory display" mode just copy an entire // If we are in "Main memory display" mode just copy an entire
// screen (256x192 pixels). // screen (256x192 pixels).