Bug#758619: more complete backtrace

[Eric Valette]

On 08/21/2014 10:26 AM, Simon McVittie wrote:

> In particular, the abort on incorrect locking is new; the original plan
> seems to have been for it to be runtime-optional.
> 
> I've sent this upstream to ask whether runtime-optional is feasible.


Well instead of correcting the symptom, I would prefer to correct the
bug itself. If locking mecahnism is hazardous, then it eman beahavior is
hazardous and that could lead to other bugs later.

Searching for the bug, I saw that, in the past, the very same undefined
behavior caused crash instead of and assert.

But I have no time (and probably not the gtk/glib basic knowledge) to
search for a solution.

-- eric