Bug report: Crash on adding dive

Dirk Hohndel dirk at hohndel.org
Mon Sep 10 08:04:47 PDT 2012


On Sep 10, 2012, at 7:33 AM, Pierre-Yves Chibon wrote:

> On Mon, 2012-09-10 at 07:27 -0700, Dirk Hohndel wrote:
>> On Sep 10, 2012, at 7:24 AM, Pierre-Yves Chibon wrote:
>> 
>>> On Mon, 2012-09-10 at 11:33 +0300, Lubomir I. Ivanov wrote:
>>>> i cannot reproduce with:
>>>> 2.20.1-1 linux
>>>> 2.24.0-1 win32
>>> GLib (gthread-posix.c): Unexpected error from C library during
>>> 'pthread_setspecific': Invalid argument.  Aborting.
>>> Aborted (core dumped)
>>> 
>>> gtk2-2.24.11-1.fc17.x86_64
>>> gtk3-3.4.4-1.fc17.x86_64
>>> pango-1.30.0-1.fc17.x86_64
>>> cairo-1.10.2-7.fc17.x86_64
>> 
>> This is interesting.  You are running Fedora (just like Linus and me). And neither Linus nor I have been able to reproduce the crash.
>> 
>> Can you give step by step instructions how to create the crash?
> 
> $ git remote -v
> origin	git://git.hohndel.org/subsurface.git (fetch)
> origin	git://git.hohndel.org/subsurface.git (push)
> $ git branch
> * master
> ...
> $ git log -1
> commit f51066fe40cb8d0bf9b6130adc1c85af3934ffe7
> ...
> $ make clean && make && ./subsurface dives/test1.xml 
> ...
> <subsurface opens>
> Ctrl+o or File -> open, search and select test2.xml
> <crash>

Oh good (sort of). This is not the pthreads bug but a bug in the program logic.
My code tries to merge a dive into a trip but the tree node for the trip is not where we think it is.
Since I know that I tested that code and it worked fine I need to find the test cases and figure out why it worked there and fails here. The logic breakdown is obvious to me (now), but I'm kind of puzzled how we got here…

/D



More information about the subsurface mailing list