[sylpheed:36583] Re: Seg fault when reading directory with a large number of messages

Rich Coe rcoe at wi.rr.com
Thu Jun 15 06:16:09 JST 2017

On Mon, 12 Jun 2017 23:14:29 -0700
Dennis Carr <dennisthetiger at chez-vrolet.net> wrote:
> On Mon, 12 Jun 2017 06:28:29 -0500
> Rich Coe <rcoe at wi.rr.com> wrote:
> > A crash in malloc usually means that some code has overwritten it's
> > boundries and the structures that malloc uses to manage the heap have
> > been corrupted.
> Either the malloc arena is too small, the blob I'm loading is too big,
> or something went wild. 

I think you are definetly in the 'something went wild' category.

Looking at the gdb traceback, I'm guessing the stack grew so much that it
grew down into the top of the heap?  

That being the case, it looks like some call to libgail.so is recursively
doing something until it exhausts the stack.  libgail.so looks like a libgtk
interface to the 'gnome accessibility implementation library'.

If you load the libgtk debug symbols for you linux distro, gdb should print
out the symbolic names for the endless loop in the gdb output.
Also, check what version of libgtk you have installed.  The latest
version for version 2 is 2.24.31.

I've had a large folder of mail before, I think it was ~25 to ~30 (70?)
thousand.  I will try creating a ~50k folder and see if I see the same issue. 

Depending on your libgtk version, it might be a known issue, but we probably
need to know the api call from the stack trace to find it.

Rich Coe     rcoe at wi.rr.com

More information about the Sylpheed mailing list