On Mon, Jul 27, 2015 at 10:11 AM, wm4 <nfx...@googlemail.com> wrote: > On Mon, 27 Jul 2015 08:30:38 -0400 > Ganesh Ajjanagadde <gajja...@mit.edu> wrote: > >> On Mon, Jul 27, 2015 at 7:35 AM, wm4 <nfx...@googlemail.com> wrote: >> > On Mon, 27 Jul 2015 10:39:16 +0200 >> > Nicolas George <geo...@nsup.org> wrote: >> > >> > >> >> > + signal(SIGSEGV, sigterm_handler); /* Segmentation fault (ANSI). >> >> > */ >> >> > + signal(SIGILL , sigterm_handler); /* Invalid instruction (ANSI). >> >> > */ >> >> > + signal(SIGFPE , sigterm_handler); /* Arithmetic error (ANSI). >> >> > */ >> >> >> >> NO!!! >> >> >> >> When a crash happens, we want it to happen right there, possibly leave a >> >> core. We do not want a signal handler to mess up the remains. >> > >> > +1 >> > >> >> > #ifdef SIGXCPU >> >> > signal(SIGXCPU, sigterm_handler); >> >> > #endif >> > >> > Why? >> >> Not sure; note this was not added by me. >> From Kerrisk's "Linux Programming Interface" book, >> SIGXCPU is raised when CPU limit is exceeded. >> It is a Linux thing, relevant when RLIMIT_CPU is set. > > You shouldn't just copy paste things you don't understand.
As I mentioned above, this was not part of my patch at all. Since you asked why, I looked it up out of curiousity and gave information that I could find. I felt this would be useful to others on this list as well, and cite a source for reference. Therefore, I do not see what the trouble is in posting it to the discussion, especially since I prefaced it with a "not sure". > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
