[Shotwell] killed by SIGTRAP

[oliver]

On Mon, Apr 09, 2012 at 09:59:49AM -0700, Adam Dingle wrote:
> On 04/09/2012 09:45 AM, oliver wrote:
> >Hello,
> >
> >
> >why is SIGTRAP not caught and irgnored?
> >
> >
> >I used ltrace to attach on an already running
> >shotwell, and it was killed by this attempt
> >and SIGTRAP was the reason.
> >
> >If SIGTRAP could be caught and ignored,
> >this problem would not stay here.
> >
> >Some days ago I thought, this problem only coccured with
> >old versions (0.7.2). But this time it occured in 0.11.6
> >also.
> >
> >Would be nice if this signal can be caught/ignored,
> >instead of that it kills shotwell.
> >
> >or are there any reasons why this signal should kill shotwell?
> 
> Oliver,
> 
> ltrace can't handle multi-threaded processes, and causes them to die
> with SIGTRAP.  This has nothing to do with Shotwell specifically.
> See, for example,
> 
> http://lists.alioth.debian.org/pipermail/ltrace-devel/2006-April/000036.html
> http://kazenotaiyo.blogspot.com/2010/02/tracing-system-and-library-calls-w.html
> https://bugzilla.redhat.com/show_bug.cgi?id=526007

Oh, thank you foir the links.
Din't knew that ltrace prolem.

> 
> Perhaps you'd have better luck with other tracing tools such as
> strace or latrace.  Cheers -

strace I already used and it has worked with shotwell,
but the output is rather Unix-API calls only, which not
necessarily is, what I was looking for.

latrace I didn't knew so far.

Thanks for the hint I will have a look at it.

Ciao,
  Oliver