On Thu, Nov 02, 2017 at 02:01:20PM +0100, Eugene Syromiatnikov wrote: > On Thu, Nov 02, 2017 at 06:38:39AM +0300, Dmitry V. Levin wrote: > > On Tue, Oct 17, 2017 at 02:29:31PM -0400, Stan Cox wrote: > > [...] > > > + /* Make sure SIGCHLD has the default action so that waitpid > > > + definitely works without losing track of children. The user > > > + should not have given us a bogus state to inherit, but he might > > > + have. Arguably we should detect SIG_IGN here and pass it on > > > + to children, but probably noone really needs that. */ > > > + signal(SIGCHLD, SIG_DFL); > > > > Why do you ever need this change? > > This change is needed for handling the child gdbserver process; I assume > that ought to be handled in a way similar to the way -o| child process > is handled now. And that leads to the fact that it ought to be refactored > into something that can handle both, as there are some parts of the code > that do the same things twice; moreover, personally I'd prefer Josh's > way of spawning child process that employs posix_spawn rather than > that current implementation which incorrectly uses vfork.
OK, but this signal(SIGCHLD, SIG_DFL) is not needed regardless of the method of handling child processes. -- ldv
signature.asc
Description: PGP signature
------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________ Strace-devel mailing list Strace-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/strace-devel
