 Steve Checkoway
|
d14323bd4a
|
Implement BACKTRACE_METHOD_X86_DARWIN. It isn't enabled since I can't test it though. There's a fair amount of duplicated code with the x86 linux version. This could be refactored but since I don't own a single x86 machine...
|
2007-02-02 13:58:17 +00:00 |
|
|
Steve Checkoway
|
631a595bce
|
Too noisy. This was only for debugging anyway. (Finding stuff in log.txt is a pain.)
|
2007-02-02 11:33:11 +00:00 |
|
|
Steve Checkoway
|
58715e319f
|
It seems that gcc decided that even though _mm_load_ps and _mm_loadu_ps are static inline functions calling a gcc builtin, it should emit a call to them because they were used as template parameters. Oh well, the code is short enough to just copy.
|
2007-01-28 23:12:06 +00:00 |
|
|
Steve Checkoway
|
1bb3190d62
|
Year.
|
2007-01-28 22:27:52 +00:00 |
|
|
Steve Checkoway
|
3813d5dbd2
|
Remove unneeded header and comment.
|
2007-01-28 22:27:25 +00:00 |
|
|
Steve Checkoway
|
f4d616e7a2
|
Update for floating point. No need to bother with the integer read because OS X doesn't use it.
|
2007-01-27 17:09:15 +00:00 |
|
|
Glenn Maynard
|
1b813dbd25
|
fix size check
|
2007-01-26 02:56:46 +00:00 |
|
|
Glenn Maynard
|
9f806c2373
|
cleanup, remove old comment
|
2007-01-26 02:10:11 +00:00 |
|
|
Glenn Maynard
|
ef525b19c5
|
fix crash
|
2007-01-24 11:49:53 +00:00 |
|
|
Glenn Maynard
|
7943380e6c
|
remove SignalHandler::ResetSignalHandlers
|
2007-01-24 11:41:31 +00:00 |
|
|
Glenn Maynard
|
0b26020d96
|
fix crashes in crash handler being handled by DoCleanShutdown
|
2007-01-24 11:38:29 +00:00 |
|
|
Glenn Maynard
|
7c96f57544
|
no, *nix builds in gcc
|
2007-01-24 09:44:23 +00:00 |
|
|
Glenn Maynard
|
675d991de4
|
remove HAVE_CRYPTOPP
|
2007-01-24 09:41:16 +00:00 |
|
|
Glenn Maynard
|
5ee4e793a7
|
don't do this in VC7
|
2007-01-24 08:18:38 +00:00 |
|
|
Glenn Maynard
|
08a4dffbe1
|
disable exceptions - much smaller binary
|
2007-01-24 07:51:52 +00:00 |
|
|
Steve Checkoway
|
2984456ba3
|
Simplify.
|
2007-01-18 12:46:02 +00:00 |
|
|
Steve Checkoway
|
cde616c5e8
|
Unused header.
|
2007-01-18 12:45:35 +00:00 |
|
|
Steve Checkoway
|
4f228d3a30
|
Use IsJoystick() and IsPump().
|
2007-01-18 12:35:13 +00:00 |
|
|
Steve Checkoway
|
ff32e1ff9d
|
Add support for the hat switch. This only works if the hid device uses 3 or 7 values plus a null position for the hat and furthermore, the minimum value needs to correspond to up with increasing values moving clockwise.
|
2007-01-18 12:30:35 +00:00 |
|
|
Steve Checkoway
|
44baf35a73
|
Actually check that PumpDevices have the correct vid/pid.
|
2007-01-18 07:18:41 +00:00 |
|
|
Chris Danford
|
87505ec245
|
change win32 lrintf to match other archs
|
2007-01-05 01:26:53 +00:00 |
|
|
AJ Kelly
|
c8d83fc337
|
revert; lroundf isn't needed with latest changes
|
2007-01-03 05:13:28 +00:00 |
|
|
Steve Checkoway
|
4517499f0f
|
lroundf -> lrintf.
|
2007-01-03 05:05:28 +00:00 |
|
|
AJ Kelly
|
92e5ea3a9f
|
added #define NEED_LROUNDF
|
2007-01-03 04:51:45 +00:00 |
|
|
Steve Checkoway
|
eb0b26cd82
|
Float to int is slow on ppc. It requires a store, a big stall, and then a load. It's even worse on a G5 if the compiler is dumb (and let's face it, I'm using gcc here...). Just use lroundf when we want an int. It will certainly be no slower.
|
2007-01-02 23:36:44 +00:00 |
|
|
Glenn Maynard
|
21de36d26a
|
fix warning
|
2006-12-19 01:42:26 +00:00 |
|
|
Glenn Maynard
|
a612f19a40
|
specify graphics drivers
(This isn't the best place for this, but RageDisplay drivers probably aren't
destined for arch.)
|
2006-12-17 08:10:22 +00:00 |
|
|
Glenn Maynard
|
7e327ef5b8
|
unused
|
2006-12-17 08:04:26 +00:00 |
|
|
Steve Checkoway
|
768c25ed1b
|
Comment.
|
2006-12-11 13:31:10 +00:00 |
|
|
Steve Checkoway
|
02282d4a14
|
Loosen sse read restrictions. Tested.
|
2006-12-11 13:26:12 +00:00 |
|
|
Steve Checkoway
|
0c539b83ce
|
Loosen altivec read restrictions. The src buffer still needs to be 16 byte aligned but the dest buffer only needs to be aligned for the type (2 bytes for int16_t and 4 for float).
|
2006-12-11 13:02:57 +00:00 |
|
|
Steve Checkoway
|
d5e67ae508
|
The data is not unaligned, it will be aligned at the boundary natural for its size or larger. (Typically at a sizeof(void*) boundary.) It is misaligned in the sense that it isn't at the 16 byte boundary we require.
|
2006-12-11 10:33:35 +00:00 |
|
|
Steve Checkoway
|
5e87581379
|
No need to assert, these are only used in once place and the preconditions are checked. Remove dependence on RageUtil (and global.h). Makes this easier to test.
|
2006-12-10 13:24:06 +00:00 |
|
|
Steve Checkoway
|
c850d0d86c
|
Remove bogus assert. As long as dest is on a 4 byte boundary and src is on a 2 byte boundary (which gcc will ensure unless you're casting into to those types), this works correct. (Retested.)
|
2006-12-10 11:08:17 +00:00 |
|
|
Steve Checkoway
|
1cf1734d5a
|
Fix window bug.
|
2006-12-10 10:20:32 +00:00 |
|
|
Vecais Dumais Laacis
|
1a2e57cb91
|
* using SM_main instead of main
* added some dummy functions in release mode (getenv, system, _popen, _pclose) which doesnt exist in libc.lib
|
2006-12-07 14:16:31 +00:00 |
|
|
Steve Checkoway
|
172033715f
|
Only one thing opens a connection, the low level window. Anything depending on it can check directly if it is open or not. A native X11 loading window would probably use it as well and we wouldn't want to close and reopen the Display. hmm.
|
2006-12-07 13:09:20 +00:00 |
|
|
Steve Checkoway
|
027b94d2c1
|
Use descriptive names.
|
2006-12-07 13:00:30 +00:00 |
|
|
Steve Checkoway
|
d1d9f1762e
|
Keep the event mask across window recreates. Close the display, I did not mean to remove that.
|
2006-12-07 12:54:24 +00:00 |
|
|
Steve Checkoway
|
a6016a9c03
|
Simplify. Only have a single window creation function.
|
2006-12-07 10:51:17 +00:00 |
|
|
Steve Checkoway
|
745fc68b9d
|
Simplify.
|
2006-12-07 08:41:54 +00:00 |
|
|
Steve Checkoway
|
19cee71210
|
Return values not used.
|
2006-12-07 08:31:40 +00:00 |
|
|
Steve Checkoway
|
0c21055858
|
Nothing returns false.
|
2006-11-28 17:57:16 +00:00 |
|
|
Glenn Maynard
|
1e21152174
|
fix compile
|
2006-11-02 03:47:02 +00:00 |
|
|
Steve Checkoway
|
e3e5cf77e5
|
Don't include Foreach.h in RageUtil.h which does not use it. Instead, include it in the individual files that actually use it.
|
2006-11-01 11:18:46 +00:00 |
|
|
Jason Felds
|
3c02509b9e
|
Well, if VC6 is removed, might as well remove the VC6 #ifs then.
|
2006-10-19 23:42:40 +00:00 |
|
|
Jason Felds
|
386bbc23e2
|
Revert: need to start clean again perhaps.
|
2006-10-19 04:38:13 +00:00 |
|
|
Jason Felds
|
332183a78d
|
Add pragma for warning C4652 under VC8. (Should this be pragmaed for all VC's?)
|
2006-10-19 04:16:09 +00:00 |
|
|
Glenn Maynard
|
001a57c2e1
|
exchange one nonportable API for another, since newer Linux kernels changed their mind
|
2006-10-18 19:48:20 +00:00 |
|
|
Steve Checkoway
|
3ef367e613
|
fix
|
2006-10-14 21:59:51 +00:00 |
|