[systemd-devel] journal segfault in (almost) latest git

[Dave Reisner]

On Wed, Mar 14, 2012 at 06:50:12PM +0100, Lennart Poettering wrote:
> On Wed, 14.03.12 18:08, Lennart Poettering (lennart at poettering.net) wrote:
> 
> > On Mon, 12.03.12 09:46, Dave Reisner (d at falconindy.com) wrote:
> > 
> > > I've captured a core dump from the journal with full debug symbols -- it
> > > looks like it imploded while trying to log a core dump from another
> > > crashing program. I haven't had a chance to poke around at the journal
> > > code or try to reproduce this, but I thought I'd set aside the core file [1]
> > > and the backtrace [2] for safe keeping...
> > > 
> > > [1] http://dev.archlinux.org/~dreisner/core.systemd-journald
> > > [2] http://dev.archlinux.org/~dreisner/journal-backtrace
> > 
> > Hmm, I am a bit puzzled about this. The backtrace shows some weirdness
> > with identifier/message being all mixed up in
> > dispatch_native_message().
> 
> Hmm, I really can't reproduce any such mixup here, and I really don't
> see how they ever could get mixed up. 
> 
> Is it possible that your debuginfo data was out of sync with the binary,
> or so?

Extremely unlikely. I'll see what I can do to reproduce.

> (Anyway, I did fix a small memory leak i found through this, so this
> actually has been really helpful so far ;-))

Great!

d