Wichert Akkerman wrote: > > Previously Jim Kingdon wrote: > > The answer I got was that the semantic differences from POSIX weren't > > really a problem, but they really, really wanted better debugging. > > gdb supports threads it seems. strace has basic thread support now > but it needs to do some black magic with signals since at the moment > any multithreaded app you trace will block on some signal magic. > Unfortunately I can't find any documentation at all on the magic needed > except the gdb source, which is cryptic enough to make it extremely > hard to figure out what is needed..
A little more data on gdb's thread support as experienced by this user: I've been using gdb on Red Hat 6.2 happily with pthreads programs. The only annoyances are: 1. you have to say "handle SIGUSR1 nostop" before starting the program. 2. you can't do postmortem debugging on core files (3. I also find lately that my pthreads programs will deadlock in gdb in my DPRINT routine, which uses a mutex; that may just be an application programming error on my part, though.) - Dan
