Glenn Maynard
b97d527d02
add EventImpl::WaitTimeoutSupported
2006-12-22 10:39:55 +00:00
Glenn Maynard
4418f202f4
update comment
2006-12-13 22:30:35 +00:00
Steve Checkoway
2250322d41
trylock
2006-12-01 18:17:16 +00:00
Glenn Maynard
b77024f3bc
include
2006-09-18 22:45:42 +00:00
Steve Checkoway
61739160fe
Handle the new clock api "correctly." At least differentiate CLOCK_MONOTONIC and CLOCK_REALTIME. Mac OS X uses the former for timers and is forced to use the latter for the pthread functions.
2006-07-18 04:39:30 +00:00
Steve Checkoway
bea4177222
Remove unneeded assert.
2006-07-18 03:57:31 +00:00
Steve Checkoway
4f8049377b
The default ctor gets the time. Avoid that since we're setting it to the value of the timeval.
2006-07-18 03:53:21 +00:00
Steve Checkoway
e349c58ba9
Fix compile for systems that use pthreads but don't define UNIX and don't have clockid_t defined. If necessary, this can be redone in a more general way using configure.ac.
2006-07-13 06:31:10 +00:00
Glenn Maynard
6e245f46ae
use pthread_condattr_setclock to avoid race conditions with the clock
2006-07-12 01:47:10 +00:00
Glenn Maynard
be5ff68d58
Fix race condition: thread->ThreadId is not necessarily assigned before
...
StartThread is run.
2006-04-05 06:45:03 +00:00
Glenn Maynard
fcab277ec9
keep this around as a reminder; this doesn't matter for normal clocks, but it's not good for CLOCK_MONOTONIC
2006-03-23 21:45:53 +00:00
Chris Danford
96d75f8cee
fix timeout value passed to pthread_cond_timedwait
2006-03-23 17:36:51 +00:00
Steve Checkoway
bb1b56aed4
BSD
2006-03-20 06:52:06 +00:00
Glenn Maynard
01629e7f53
spacing
2006-02-14 11:35:00 +00:00
Steve Checkoway
3cd26abb99
Just in case these are ever implemented, fix. sem_* return -1 for all error conditions. Only sem_destroy sets errno to EBUSY anway.
...
Also, ASSERT so we get a backtrace.
2006-02-05 13:24:44 +00:00
Steve Checkoway
79218ae6ad
Bah. Revert. The functions sem_init() and sem_destroy() are implemented as stubs only. They return -1 with errno=ENOSYS.
2006-02-05 13:21:58 +00:00
Steve Checkoway
48bf80b5f6
OS X has semaphores for every version we support (10.2.8 and on).
2006-02-05 13:00:09 +00:00
Steve Checkoway
833d06d1f3
pthread_join() returns an error code, it does not set errno.
2005-12-17 13:17:35 +00:00
Glenn Maynard
d03f1ed31f
cleanup
2005-12-13 01:10:48 +00:00
Steve Checkoway
aa0f90feae
__MACOSX__ --> MACOSX, keep __MACOSX__ for ogg.
2005-10-24 10:37:56 +00:00
Glenn Maynard
1783fe71ef
simplify
2005-07-19 03:15:24 +00:00
Glenn Maynard
a79f1381f3
fix casting problem?
2005-06-22 20:58:41 +00:00
Steve Checkoway
99387583f9
DARWIN -> __MACOSX__
2005-06-21 08:56:36 +00:00
Ben Anderson
d137be895e
Thanks Glenn, your blatant 'it worked for me' syndrome really bit me in the ass.
2005-04-05 23:51:38 +00:00
Glenn Maynard
af231d1a63
cleanup
2005-04-05 05:07:13 +00:00
Ted Percival
e1ab8c5b77
fix timing errors in thread wait functions on unix. patch by makovick fixes bug #1154212
2005-03-24 19:46:50 +00:00
Glenn Maynard
23d907e3d4
remove ill-formatted comment (I feel like I've asked him to space comments
...
readably, like the rest of the code, half a dozen times).
The function is probably just returning ERROR_CALL_NOT_IMPLEMENTED.
If it's returning some other error, find out what it is and add it to the conditional. Don't
just break out without understanding the logic of the rest of the function.
2005-02-12 06:21:29 +00:00
Charles Lohr
0051938177
We really shouldn't crash on systems this command is not fully supported on. Windows 9x will always respond with a 1.
2005-02-12 05:32:35 +00:00
Glenn Maynard
d38fedfcaa
fix thread handle leak (these don't show up as threads anywhere, but
...
sysinternals handle.exe shows them)
2005-02-06 10:25:13 +00:00
Glenn Maynard
46f414e7e5
Fix race condition.
...
Work around undocumented SignalObjectAndWait behavior; the docs
refer to "the object", but the function takes two objects. (Cut and paste
documentation ...)
2005-01-29 20:25:52 +00:00
Glenn Maynard
8bf00eb4e8
implement event wait timeouts in win32
2005-01-27 01:54:33 +00:00
Glenn Maynard
903aaadcc5
After signalling an event, always wait for the signalled thread to touch
...
m_WaitersDone. This fixes problems with
event.Lock();
event.Signal();
while(x) event.Wait();
picking up the signal that it just sent (which is wrong), especially in the
debugger.
2005-01-27 01:19:24 +00:00
Glenn Maynard
1b49d06b5d
oops
2005-01-26 21:41:40 +00:00
Glenn Maynard
08ddf9af77
fix up EventImpl_Pthreads::Wait
2005-01-26 21:13:50 +00:00
Glenn Maynard
63604a6536
implement RageEvent::Wait timeouts
2005-01-26 20:59:41 +00:00
Glenn Maynard
0330fa9e87
untested pthreads event implementation
2004-11-07 22:56:03 +00:00
Glenn Maynard
4de15ff3a7
untested thread events implementation for win32
2004-11-07 22:43:39 +00:00
Chris Danford
79beff5e26
revert timeout length
2004-10-09 16:56:29 +00:00
Chris Danford
47e2e2d48d
increase mutex timeout
2004-10-09 09:00:24 +00:00
Glenn Maynard
00b2d18a8e
show thread names in the VC debugger
2004-09-17 02:36:04 +00:00
Glenn Maynard
fefe7bad49
fix thread handles on win9x
2004-09-14 04:39:52 +00:00
Glenn Maynard
34f22e588c
fix up SemaImpl_Pthreads::Wait
2004-09-08 21:58:47 +00:00
Glenn Maynard
3e7f55ae30
no, don't use pthreads extensions if we don't at least have timedlock (not going to bother making it two tests unless it's actually needed)
2004-09-08 08:56:19 +00:00
Steve Checkoway
bf3ce581af
OS X has pthread_cond_timedwait but not pthread_mutex_timedlock.
2004-09-08 08:36:43 +00:00
Glenn Maynard
2625da14f2
cleanup
...
support semaphore timeouts
2004-09-08 07:56:32 +00:00
Glenn Maynard
81f2f6460a
hopefully fix init order in vc6
2004-09-08 06:33:07 +00:00
Steve Checkoway
bdd4988829
Simplify.
2004-09-07 12:49:49 +00:00
Steve Checkoway
8adbeda32a
Use DarwinThreadHelpers, remove mach stuff.
2004-09-07 12:38:10 +00:00
Glenn Maynard
9cf85e4b84
remove comment
2004-09-07 04:24:26 +00:00
Glenn Maynard
7f805dcc87
simplify
2004-09-07 03:26:23 +00:00
Glenn Maynard