Commit Graph

1522 Commits

Author SHA1 Message Date
luigi__ 3221fc2a88 Revert the last change because we lost the ability to limit framerate when using auto frameskip, which can be annoying.
I think we should rather add a "Limit framerate" option.
2009-01-31 13:13:54 +00:00
zeromus 36e1fee5fe disable throttling when frameskip is 0 in windows, so you can sort of fast forward, or test how fast the emu is running. also revert a couple hacks in oglrender which I do think cause a net glitch count increase 2009-01-31 05:20:06 +00:00
zeromus 19f9cd36fc fix: [ 2550645 ] Super Princess Peach: crash when starting a puzzle 2009-01-31 04:47:23 +00:00
gecko_reverse 98c68d8d89 added more keyboard buttons to the input settings 2009-01-31 04:36:48 +00:00
gecko_reverse 98528a75a7 fixed left and right key config (thanks to Arakun for finding the bugged code) 2009-01-31 04:20:53 +00:00
zeromus bf054c32ed fix vertex source texcoord generation (fixes environment mapping e.g. stars in sm64); this is a low risk fix and could be safely ported into release branch 2009-01-31 04:01:10 +00:00
zeromus 2578a0a291 fix a serious bug which crippled 3d engine loading from savestates. this should be ported to release branch. also add some missing dma state to savestate. 2009-01-31 03:40:20 +00:00
zeromus 09229a159e fix dual screen 3d bug: when capturing 2d+3d output, but the display mode is not producing 2d+3d data (such as in vram display mode), then we should peek directly at the 3d output as if we were capturing 3d output only 2009-01-31 02:58:45 +00:00
luigi__ e4a7588d4a Added casts from s16 to s32 in the PSG noise func.
Added the new channel_struct variables to savestates.
Fixed the wave write core (WMP doesn't complain anymore when playing a wave file recorded from DeSmuME)
2009-01-30 13:08:50 +00:00
gecko_reverse 8c2a951941 mac compile fix 2009-01-29 23:54:55 +00:00
zeromus e661331bd1 revert something crazymax did in dma which broke repeating dmas 2009-01-29 14:38:11 +00:00
luigi__ 18fa16d748 Added support for PSG white noise. 2009-01-29 14:34:28 +00:00
riccardom a87501a235 Put brackets around a possible empty macro. 2009-01-26 19:52:42 +00:00
riccardom 23cffd809c Make gfx3d_FlushFIFO() static. 2009-01-26 19:51:07 +00:00
zeromus 2df831cbb0 add some diagnostic messages 2009-01-26 18:43:34 +00:00
mtabachenko 4e9176a4c0 core:
- add busy flag in GXSTAT for BoxTest, PosTest & VecTest;
- fix pop matrix stack;
- GXFIFO emulation (with USE_GEOMETRY_FIFO_EMULATION definition in FIFO.h): I some corrected FIFO, but this have errors. On Homebrew roms (with small commands buffer) it`s work perfectly, but in commercial games with glitches. I can`t understand when it is needed flush GXFIFO to render. I always get the buffer overflow and loss of data. Where is a bug? I wrote small test for GXFIFO for comparing to the real DS (http://www.turboupload.com/files/get/wuf9IR3Oo1/fifo-3d.zip)
2009-01-26 13:18:17 +00:00
zeromus 18fed5b2ac speedup to x/y window handling code by precalculating as much as possible at beginning of scanline or when WIN0H* are set 2009-01-26 09:28:09 +00:00
zeromus fecd7ffd69 fix a bug that made some sprites get cut off in x=255; also some small speedups in GPU 2009-01-26 07:58:13 +00:00
zeromus 9689548ac8 change back to linear interpolation in SPU. it seems a bit extravagant now. perhaps we should make it configurable.. or, for luigi__'s consideration, I added a commented-out cos lookup table instead of doing the full call to cos() 2009-01-26 07:55:36 +00:00
zeromus 990c2a062e fix a bug in windows framerate throttling which made frameskip auto unable to skip frames. 2009-01-26 07:53:42 +00:00
zeromus 1d71a7d4b7 check me crazymax: it is a shame to have to memset the fifo so often. would it be sufficient just to reset the tail? 2009-01-26 07:02:50 +00:00
zeromus 0055130494 readme crazymax: matrix stack "size" is 31 for 32 entry stack (silly, i know) so your overrun check was wrong. 2009-01-25 22:11:50 +00:00
mtabachenko 63605b9967 core:
- added stack level into GXSTAT register;
2009-01-25 20:49:24 +00:00
zeromus b0fbd6a919 fix to sprite windows which I broke the other day. various speedups in win32 and in GPU 2009-01-25 20:16:19 +00:00
riccardom 26dc322b2c Remove extern stuff thas has been made static in GPU.cpp 2009-01-25 15:52:53 +00:00
zeromus 0c8e597533 finish some things i meant to do in the mosaic system and do some other gpu cleanups. 2009-01-25 07:26:01 +00:00
zeromus 1f2982fa5c fix recently introduced bug which made >256 wide BGs crash the map viewer debug tool.
remove redundant GPU blending code and add a lookup table.
2009-01-25 06:12:19 +00:00
riccardom 82b6c5c8ea Reenable gtk-glade compilation with offscreen mesa enabled for gtk frontend. 2009-01-24 13:43:46 +00:00
riccardom a3775a0138 gtk-glade: make a couple of functions static 2009-01-24 13:07:46 +00:00
riccardom 577a91896b Oops, i missed to commit the same change to header :) 2009-01-23 22:30:44 +00:00
riccardom c23054364d We can make quite a bit of data const. 2009-01-23 22:28:45 +00:00
riccardom 1f68f2de4a Make a few functions static and put parentheses around shift operators 2009-01-23 22:21:14 +00:00
riccardom 1b10babf27 Add proper includes to compile on linux (patch by gualteri) 2009-01-23 22:13:13 +00:00
yabause d2070ca6e5 Updated version number to something else, so users testing
svn won't get confused.
Also removed the check for a C compiler. It seems that
when searching for both a C and a C++ compiler, configure
don't care that there's no C++ if there's a C one.
2009-01-23 14:21:28 +00:00
mtabachenko dd167055b2 fix typos in addons and make functions is static (thanks xrmx) 2009-01-23 10:09:09 +00:00
mtabachenko f9d81fcd0d core:
- add experimental GFX FIFO emulation (It has a bit bugs and a revision is required. I will endeavour to correct it soon)
2009-01-22 21:50:51 +00:00
riccardom 992214a67a Fix a buffer overflow :) 2009-01-22 20:49:13 +00:00
riccardom 74a217400b noext is allocated by malloc in strdup so use free instead of delete []. 2009-01-22 20:40:18 +00:00
riccardom 3289c31d62 Make a couple of functions static. 2009-01-22 20:17:57 +00:00
riccardom bfc30a659e gcc does not like to skip variable initialization with gotos 2009-01-22 20:13:05 +00:00
zeromus 672b8e3c8f apply my planned mosaic optimizations to BG 2009-01-22 07:24:57 +00:00
zeromus 3e57278979 fix crash after loading savestate matrix stacks originating from revision 1513 2009-01-22 07:22:46 +00:00
zeromus a6b21e74e1 never mind. i was persuaded that i was wrong 2009-01-22 06:29:21 +00:00
zeromus b8b16c02df I think we should throw out touch input when it is out of bounds of the touchscreen, instead of clamping it. what do other people think? 2009-01-22 05:59:45 +00:00
luigi__ c6c4871020 Implemented a more efficient frameskip support, skipping frames at GPU level, which is loads faster.
It only works for Windows port atm, to make it work in the other ports :
call NDS_SkipFrame with param : true = skip the current frame, false = render the current frame.
2009-01-21 22:35:46 +00:00
luigi__ 47785d2015 Some minor optimiSome minor optimizations to the BG mosaic func.
Feel free to correct any problem I may have introduced!
2009-01-21 21:48:44 +00:00
mtabachenko d742dd7745 core:
- fix quick save states (bug in geometry)
2009-01-21 19:43:40 +00:00
luigi__ fbd32a12da Oops... the blending is forced only if the 3D layer has the highest prio, not if it's selected as 1st target... 2009-01-21 19:25:08 +00:00
riccardom 83fb5e2223 Make a couple of functions static and put () around & operands. 2009-01-21 18:59:47 +00:00
mtabachenko a5e3e92106 core:
- added VRAM mapping for DMA;
2009-01-21 14:06:16 +00:00