From b035ed05977fdef5ced4691028284b7f0ebaba19 Mon Sep 17 00:00:00 2001 From: rpj Date: Tue, 25 Jul 2000 11:27:23 +0000 Subject: 2000-07-25 Ross Johnson * dll.c (dllMain): Remove 2000-07-21 change - problem appears to be in pthread_create(). 2000-07-21 Ross Johnson * create.c (pthread_create): Set threadH to 0 (zero) everywhere. Some assignments were using NULL. Maybe it should be NULL everywhere - need to check. (I know they are nearly always the same thing - but not by definition.) * dll.c: Include resource leakage work-around. This is a partial FIXME which doesn't stop all leakage. The real problem needs to be found and fixed. - "David Baggett" * misc.c (pthread_self): Try to catch NULL thread handles at the point where they might be generated, even though they should always be valid at this point. * tsd.c (pthread_setspecific): return an error value if pthread_self() returns NULL. * sync.c (pthread_join): return an error value if pthread_self() returns NULL. * signal.c (pthread_sigmask): return an error value if pthread_self() returns NULL. --- ANNOUNCE | 8 -------- 1 file changed, 8 deletions(-) (limited to 'ANNOUNCE') diff --git a/ANNOUNCE b/ANNOUNCE index 3a46bab..2986e9f 100644 --- a/ANNOUNCE +++ b/ANNOUNCE @@ -242,14 +242,6 @@ points in applications and libraries: pthreadCancelableWait pthreadCancelableTimedWait -The library includes a non-portable function to force using the much -faster critical sections as the basis for POSIX mutexes when these are -not the default for a particular version of Windows (eg. W9x). - - pthread_mutexattr_setforcecs_np - -For example, if you don't ever use pthread_mutex_trylock() in your -application. Availability ------------ -- cgit v1.2.3