summaryrefslogtreecommitdiff
path: root/semaphore.c
AgeCommit message (Collapse)Author
1999-01-14Fri Jan 15 15:41:28 1999 Ross Johnson <rpj@swan.canberra.edu.au>rpj
* pthread.h: Add new 'abstime' arg to pthreadCancelableWait() prototype. * condvar.c (cond_timedwait): New generalised function called by both pthread_cond_wait() and pthread_cond_timedwait(). This is essentially pthread_cond_wait() renamed and modified to add the 'abstime' arg and call the new _pthread_sem_timedwait() instead of sem_wait(). (pthread_cond_wait): Now just calls the internal static function cond_timedwait() with an INFINITE wait. (pthread_cond_timedwait): Now implemented. Calls the internal static function cond_timedwait(). * implement.h (_pthread_sem_timedwait): New internal function prototype. * misc.c (pthreadCancelableWait): Added new 'abstime' argument to allow shorter than INFINITE wait. * semaphore.c (_pthread_sem_timedwait): New function for internal use. This is essentially sem_wait() modified to add the 'abstime' arg and call the modified (see above) pthreadCancelableWait().
1999-01-12Wed Jan 13 09:34:52 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au>rpj
* build.bat: Delete old binaries before compiling/linking. Tue Jan 12 09:58:38 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au> * dll.c: The Microsoft compiler pragmas probably are more appropriately protected by _MSC_VER than by _WIN32. - Tor Lillqvist <tml@iki.fi>. * condvar.c (pthread_cond_timedwait): Fix function description comments. * pthread.h: Define ETIMEDOUT. This should be returned by pthread_cond_timedwait which is not implemented yet as of snapshot-1999-01-04-1305. It was implemented in the older version. The Microsoft compiler pragmas probably are more appropriately protected by _MSC_VER than by _WIN32. - Tor Lillqvist <tml@iki.fi>. * pthread.def: pthread_mutex_destroy was missing from the def file - Tor Lillqvist <tml@iki.fi>. * condvar.c (pthread_cond_broadcast): Ensure we only wait on threads if there were any waiting on the condition. I think pthread_cond_broadcast should do the WaitForSingleObject only if cv->waiters > 0? Otherwise it seems to hang, at least in the testg thread program from glib. - Tor Lillqvist <tml@iki.fi>. * semaphore.c (sem_post): Correct typo in comment. Mon Jan 11 20:33:19 1999 Ross Johnson <rpj@ixobrychus.canberra.edu.au> * pthread.h: Re-arrange conditional compile of pthread_cleanup-* macros. * cleanup.c (_pthread_push_cleanup): Provide conditional compile of cleanup->prev.
1999-01-03Merge John Bossom's code into the main trunk. See ChangeLog for details.snapshot-1999-01-04-1305rpj
This will be tagged as snapshot-1999-01-04-1305