PR_OpenSemaphore is not implemented on Mac OS X, which causes the NSPR tests nameshm1, sema, semaerr, and semaping to fail. In Mac OS X 10.4, both System V and POSIX semaphores are available. So we can implement PR_OpenSemaphore now. This patch uses System V semaphores, but POSIX semaphores also work. I chose System V semaphores because we use System V semaphores on FreeBSD, NetBSD, and OpenBSD (three systems similar to Darwin).

I checked in the patch on the NSPR trunk (NSPR 4.7.4). Checking in pr/src/pthreads/ptsynch.c; /cvsroot/mozilla/nsprpub/pr/src/pthreads/ptsynch.c,v <-- ptsynch.c new revision: 3.31; previous revision: 3.30 done Checking in pr/include/md/_darwin.h; /cvsroot/mozilla/nsprpub/pr/include/md/_darwin.h,v <-- _darwin.h new revision: 3.22; previous revision: 3.21 done

Comment on attachment 353123 [details] [diff] [review] Proposed patch Tested in 32/64 bit. Note the remaining nspr tests that fail in 32/64 bit are cltsrv, layer, nblayer, and provider.

Thanks for reviewing and testing the patch. Does nameshm1 complete on your Mac? On my Mac (10.5), nameshm1 hangs. The cltsrv, layer, nblayer, and provider test failures are caused by bug 469745.

nameshm1 succeeded in my testing. Yesterday I upgraded my mac to 10.6, I am not sure what the results were for 10.5. I am now on-site at work and do not have access to a mac.

(In reply to comment #4) > nameshm1 succeeded in my testing. Yesterday I upgraded my mac to 10.6, I am not > sure what the results were for 10.5. I am now on-site at work and do not have > access to a mac. sorry I meant I updated to 10.5.6, then tested on 10.5.6. I was using 10.5.5 and not sure if nameshm1 failed or suceeded.