[GSoC] Introduction

[George Kiagiadakis]

2010/3/4 Ilie Halip <lupuroshu at gmail.com>:
> On Wed, Mar 3, 2010 at 1:39 AM, George Kiagiadakis
> <kiagiadakis.george at gmail.com> wrote:
>>
>> Maybe you should also read our previous discussion on this
>> topic, which started a few days ago:
>>
>> http://mail.kde.org/pipermail/kde-windows/2010-February/004709.html
>> http://mail.kde.org/pipermail/kde-windows/2010-March/004712.html
>
>
> I took a look on those mails, and downloaded MinGW and your tests (I already
> have MSVC), and you were right: both gdb and windbg don't generate valid
> backtraces. Even writing a memory dump and analyzing it with windbg doesn't
> reveal much. I really don't understand why this happens. Can anyone please
> explain the problem here?

Well, I don't know what the problem is. That could be a part of your
investigation... If we could find the reason and a solution to it,
then probably another debugger would not be needed. However, I think
we should concentrate on the fact that the StackWalk api *does*
generate a correct stack trace and thus we need to take advantage of
that in kde. Let's just say the other debuggers are... well... buggy
:P

Regards,
George