Mednafen 0.9.32.1-WIP [message #3252] |
Sat, 23 November 2013 13:58 |
|
Development on 0.9.33-WIP is taking too long, and some of the changes are going to be contentious, so this release should be considered a backport of some of the simple changes and bug fixes from the in-development 0.9.33-WIP.
Highlights include:
- Atari Lynx II stereo support(thanks to "sage").
- More stable PSX DualShock rumble on Linux(fewer kernel panics, yay).
- PSX DualShock rumble is now working in Final Fantasy 8, and possibly other games where it was previously broken.
Toggle Spoiler
-- 0.9.32.1-WIP: --
November 23, 2013:
PCE_FAST: Fixed a potential crash after a garbage/corrupted ROM image is loaded.
PCE, PCE_FAST: Don't generate a command error if a track number too high is specified via CD commands 0xD8 and/or 0xD9; instead,
treat it as if the leadout track was specified. I'm not certain if this is correct behavior as per the actual system, but
that'll have to wait until I'm able to run some tests. Fixes a startup lockup in "Star Mobile".
Lynx: Added emulation of Lynx 2 stereo audio registers, based on patches submitted by "sage".
NES: Reworked some real/floating-point setting's minimum and maximum value strings to try to work around a reported
issue with the settings parsing code.
PCE: During HES playback initialization, the second emulated VDC's registers are now initialized properly so that the
emulation event system isn't overtaxed(which drove up CPU usage considerably during HES playback).
PSX: Improved DualShock emulation accuracy per tests on a DualShock; rumble now works in "Final Fantasy 8", and possibly
other games where rumble was previously non-working.
Reworked how joystick/gamepad rumble is handled on Linux, to perform better and hopefully not cause kernel panics
anymore(or at least not so often as the old code did).
PSX: Removed an unnecessary and borked memset() statement from mdec.cpp.
Added a very simple double-precision floating-point range test to tests.cpp.
Compile with "-fno-aggressive-loop-optimizations", if the option is detected to be supported by the compiler. I don't
particularly trust that new optimization in gcc 4.8.x.
mednafen-0.9.32.1-wip.tar.bz2
SHA1: aa57533321b79ae63f5a64e2bbbdb31a397902d3
mednafen-0.9.32.1-wip-win32.zip
SHA1: 441ea51bcb118a2286908a814c9bc3f920efa7ba
[Updated on: Fri, 14 November 2014 22:00]
|
|
Re: Mednafen 0.9.32.1-WIP [message #3255 is a reply to message #3252 ] |
Thu, 28 November 2013 09:06 |
|
Administrator wrote on Sat, 23 November 2013 14:58 | Development on 0.9.33-WIP is taking too long, and some of the changes are going to be contentious [...]
|
Please, can you elaborate a bit about this?
|
|
Re: Mednafen 0.9.32.1-WIP [message #3256 is a reply to message #3252 ] |
Thu, 28 November 2013 18:00 |
|
0.9.33-WIP will have increased CPU usage for PC Engine and PC-FX emulation due to sound resampling and filtering changes, and a fair amount of old code throughout Mednafen will be modified, which runs the risk of introducing regressions.
[Updated on: Thu, 28 November 2013 18:01]
|
|
Re: Mednafen 0.9.32.1-WIP [message #3285 is a reply to message #3252 ] |
Fri, 20 December 2013 15:20 |
|
Thanks for your consideration !
|
|
Re: Mednafen 0.9.32.1-WIP [message #3302 is a reply to message #3252 ] |
Sun, 26 January 2014 13:25 |
|
Hi,
I have one problem with the current version (although I am not really sure which version brought this error, last one I have (which compiled fine back than is 0.9.28):
depbase=`echo vdc.o | sed 's|[^/]*$|.deps/&|;s|\.o$||'`;\
g++ -DLOCALEDIR=\"/usr/local/share/locale\" -DHAVE_CONFIG_H -I../../include -I../../intl -I../../src/hw_misc -I../../src/hw_sound -fsigned-char -Wall -Winline -Wshadow -Wempty-body -Wignored-qualifiers -fno-fast-math -fno-unsafe-math-optimizations -fno-aggressive-loop-optimizations -fomit-frame-pointer -finline-limit=6000 --param large-function-growth=800 --param inline-unit-growth=175 --param max-inline-insns-single=10000 -fno-strict-overflow -g -O2 -MT vdc.o -MD -MP -MF $depbase.Tpo -c -o vdc.o vdc.cpp &&\
mv -f $depbase.Tpo $depbase.Po
vdc.cpp: In function ‘void PCE_Fast::VDC_RunFrame(MDFN_Surface*, MDFN_Rect*, MDFN_Rect*, int)’:
vdc.cpp:1652:10: warning: variable ‘dummy_ne’ set but not used [-Wunused-but-set-variable]
int32 dummy_ne;
^
vdc.cpp: In function ‘void PCE_Fast::MixBGSPR_x86(uint32, const uint8*, const uint16*, uint32*)’:
vdc.cpp:1079:1: error: bp cannot be used in asm here
}
^
Makefile:398: recipe for target 'vdc.o' failed
make[2]: *** [vdc.o] Error 1
make[2]: Leaving directory '/opt/mednafen/src/pce_fast'
Makefile:911: recipe for target 'all-recursive' failed
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory '/opt/mednafen/src'
Makefile:435: recipe for target 'all-recursive' failed
make: *** [all-recursive] Error 1
This is on gentoo, gcc 4.8.2
[Updated on: Sun, 26 January 2014 13:28]
|
|
Re: Mednafen 0.9.32.1-WIP [message #3306 is a reply to message #3252 ] |
Wed, 29 January 2014 18:21 |
|
Hi,
Just dropped by to report that the title and intermission screens on Wonder Momo for the PC Engine display incorrectly (parts of the screens look cropped).
Has for the game itself, I've managed to make it past the first stage with no problems whatsoever. Controls seemed a bit slow/laggy, but that may be the game's fault not the emulator's.
|
|
Re: Mednafen 0.9.32.1-WIP [message #3330 is a reply to message #3302 ] |
Thu, 27 February 2014 17:28 |
|
ktrhn wrote on Sun, 26 January 2014 14:25 | Hi,
I have one problem with the current version (although I am not really sure which version brought this error, last one I have (which compiled fine back than is 0.9.28):
This is on gentoo, gcc 4.8.2
|
I'll look into it, but try using gcc 4.7.x in the meantime(assuming this is still an issue a month later ).
[Updated on: Thu, 27 February 2014 17:28]
|
|