Re: [PATCH 0/6] Display GDB backtrace for internal errors

[Florian Weimer <fweimer@redhat.com>]

* Andrew Burgess:

> This series does two things:
>
>      1. Imports libbacktrace from gcc to present better backtraces of
>      GDB when we hit a fatal signal, and
>
>      2. Prints the backtrace when GDB hits an internal-error (e.g. an
>      assert).

Ideally, we would add this capability to bash, so that if a process dies
with certain signals, the shell prints a backtrace, possibly with help
from a core-catching daemon.  Backtrace generation from outside the
process is much more reliable.  It also avoids introducing further
opportunities for exploitation after memory corruption is detected.

In my opinion, it does not make much sense to add this functionality to
more and more GNU programs.

Thanks,
Florian