Bill Pringlemeir wrote: > I also remember I had a case without a core file. I think that a > break point instructions was hit.
Unless you run Gtk-Gnutella in a debugger, a break point instruction should cause a SIGTRAP and thus a core as well. For debugging purposes you should disable "fast assertion" support either by not compiling with -Dofficial or by passing -Ud_fast_assert to Configure. That's because this feature will usually mess with the stack trace. I think that's because you end up on the signal stack instead of the normal stack. > All of my builds are with GCC 3.4.3 and they have the IBM stack > protector patch. This patch actually generates quite a bit of extra > code from what I have seen... but I do have it enabled. Do you use -fstack-protector or -fstack-protector-all? -- Christian
pgpogeTpSWsvJ.pgp
Description: PGP signature
