The sem_wait(3) man page states that sem_wait and sem_trywait always return 0. That's not correct. The Opengroup man pages allow -1 as return value with errno set. This can happen on RHEL 3 as well, e. g. when running an application under GDB and pthread_create is called while another thread is waiting in a sem_wait call. In that case, sem_wait returns -1 and errno set to EINTR.
man page part of glibc-devel. fyi, the posix version (man 3p), which is part of man-pages, has the correct return value.
The LinuxThreads implementation always returns 0 and this is what the man page documents. The NPTL implementation is different. Just another reason why the man page should move to man-pages, it must be independent. Adrian, do you want to copy the pages, edit and distribute them? This need to be coordinated with Jakub stopping to distribute them.
Ugh, Andrian wasn't cc:ed. See comment #2.
comment 2: sure, i'll put them in. however, we'll need non-english versions as well to go in all the man pages.
Reassigning to man-pages. The files are in the glibc sources or, if you prefer, packaged in ~drepper/thread-man-pages.tar.bz2 on devserv. If you need help with making the changes requested in this bug, please let me. Also let Jakub know which man-pages RPM will contain the new pages so that the glibc RPM stops distributing them.
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on the solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHBA-2007-0408.html