Tommi Mäkitalo <tommi@...> writes: > As I can see in your stack trace, this happens in tntnet. And tntnet > sets the signal handler for SIGPIPE. So it shouldn't happen anyway. > Setting the signal handler is global to the process. You do not need > to set it again in your components. > I wonder how the crash can happen. The kernel tells, that it stops > the process since it received a signal SIGPIPE although tntnet told > to ignore that signal by setting the handler to SIG_IGN. I have > currently no idea. Maybe someone else here has a idea? > Tommi
There is a SO_NOSIGPIPE option: http://stackoverflow.com/a/450130/257568 Perhaps it is available on PS3. There's MSG_NOSIGNAL: http://stackoverflow.com/a/1705705/257568 And here's something complex: http://stackoverflow.com/a/2347848/257568 First I would've checked if the handler is changed somewhere with struct sigaction have; if (sigaction (SIGPIPE, nullptr, &have) == 0) log_info ('SIGPIPE handler: ' << have.sa_handler); ------------------------------------------------------------------------------ Master HTML5, CSS3, ASP.NET, MVC, AJAX, Knockout.js, Web API and much more. Get web development skills now with LearnDevNow - 350+ hours of step-by-step video tutorials by Microsoft MVPs and experts. SALE $99.99 this month only -- learn more at: http://p.sf.net/sfu/learnmore_122812 _______________________________________________ Tntnet-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/tntnet-general
