Date: Tue, 24 Feb 2004 17:26:13 -0500 (EST) From: Daniel Eischen <eischen@vigrid.com> To: Jeremy Messenger <mezz7@cox.net> Cc: Hye-Shik Chang <perky@i18n.org> Subject: Re: Need the help on debug the Python VS libpthread.. Message-ID: <Pine.GSO.4.10.10402241722080.3438-100000@pcnet5.pcnet.com> In-Reply-To: <opr3v3z30q8ckrg5@smtp.central.cox.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 24 Feb 2004, Jeremy Messenger wrote: > On Tue, 24 Feb 2004 10:57:32 -0800 (PST), Julian Elischer > <julian@elischer.org> wrote: > > > > > > > On Tue, 24 Feb 2004, Jeremy Messenger wrote: > > > >> Great, thank guys! I now can install net/straw with optimized on > >> libpthread. I am adding -thread in the CC, so the thread developers can > >> catch here too. :-) > > > > so what was fixed between teh quoted mail below and now? > > A workaround (not real fix?) as perky has committed[1]. There's something > wrong with waitpid if it runs as libpthread, since without the workaround > can run fine with libc_r. It is not necessarily a problem with libpthread. waitpid() can return EINTR when interrupted by a signal. If SIGCHLD isn't masked by the thread, then it is certainly plausible that it can get interrupted by it. You may not have seen it with libc_r because it handles signals differently as well as wrap waitpid() so it doesn't block the process. Libpthread doesn't wrap waitpid() except to make it a cancellation point. If its returning EINTR, I would suspect that it has received a signal. -- Dan Eischen
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.GSO.4.10.10402241722080.3438-100000>