> From: Doug Evans <xdj...@gmail.com> > Date: Mon, 17 Feb 2014 02:39:25 -0500 > > Hi. > > Here's my modest contribution to the Guile anniversary potluck. > > The patch to selftest-support.exp could be done differently, > I've tried to keep it simple. The problem is that gdb with guile > will get SIGPWR from time to time when Guile's GC kicks in, > and we need this to not alter test behaviour. The patch just > tells the parent gdb to ignore SIGPWR, which is simple enough > without loss of coverage. A good question is what other signals > Guile GC might use.
SIGPWR? Doesn't exist on *BSD. Guess it uses some other random signal there? Didn't realize Guile used threads. I guess that's safe if the interpreter makes sure it never calls into GDB code concurrently. But it does make debugging GDB harder. And I hope this is not an excuse for other people to start using threads in other bits of GDB.
