This is the mail archive of the
frysk@sources.redhat.com
mailing list for the frysk project.
Re: patch for bug 3259
- From: Mike Cvet <mcvet at redhat dot com>
- To: Alexandre Oliva <aoliva at redhat dot com>
- Cc: Andrew Cagney <cagney at redhat dot com>, frysk at sources dot redhat dot com
- Date: Fri, 06 Oct 2006 16:47:42 -0400
- Subject: Re: patch for bug 3259
- References: <or7izhgq9s.fsf@free.oliva.athome.lsd.ic.unicamp.br> <4523BD82.3070302@redhat.com> <1160063678.2050.51.camel@towel.toronto.redhat.com> <or8xjt3ovv.fsf_-_@free.oliva.athome.lsd.ic.unicamp.br>
On Fri, 2006-10-06 at 04:05 -0300, Alexandre Oliva wrote:
> On Oct 5, 2006, Mike Cvet <mcvet@redhat.com> wrote:
>
> > http://sourceware.org/bugzilla/show_bug.cgi?id=3259
>
> > For whatever reason, the unwinding expects the innermost frame to be
> > looping at line 61, when in fact it should be 62. This doesn't happen in
> > the following backtrace testcase, or anywhere else I've seen... a
> > strange anomaly.
>
> I've just attached a patch to bugzilla that works around the compiler
> over-optimization of debug info that causes this symptom. Ok to
> install?
>
The following test method in the test class - testThreadedBacktrace -
uses funit-rt-threader.c as its test program. This program has three
threads, all of which cycle on a while(1), but the unwinding produces
the correct line numbers for them. What's going on here? If this is a
GCC bug shouldn't we file something against it?
- Mike