[bt-devel] Updated Windows Build

[Gary Holmlund]

Eeli Kaikkonen wrote:
> On Fri, 23 Oct 2009, Gary Holmlund wrote:
>   
>> I got a crash that looks like it is related to the install thread. The
>> call stack does not have any BT functions involved,
>> but it looks like a QMutex is being used after it was deleted.
>>
>> Call Stack
>> QtCored4.dll!QMutex::lock()  Line 152 + 0x3 bytes    C++        // crash
>> here
>> QtCored4.dll!QThreadPrivate::finish(void * arg=0x01f9a2c8, bool
>> lockAnyway=true)  Line 320    C++
>> QtCored4.dll!QThreadPrivate::start(void * arg=0x01f9a2c8)  Line 309 +
>> 0xb bytes    C++
>> msvcr90d.dll!_callthreadstartex()  Line 348 + 0xf bytes    C
>> msvcr90d.dll!_threadstartex(void * ptd=0x01f83e50)  Line 331    C
>> kernel32.dll!_BaseThreadStart at 8()  + 0x37 bytes
>>
>> void QMutex::lock()
>> {
>>     Qt::HANDLE self;
>>
>>     if (d->recursive) {       // crashed on this line. It looks like the
>>                               // "this" pointer is not valid.
>>     
>
> I believe this may be a bug in our code. Could you attach the BT debug
> text?  Does it say "QThread: Destroyed while thread is still running"?
> It would also be important to know what exactly you were doing, i.e. how
> many modules installing, did you cancel any, did this happen after all
> was done... Sorry to ask these basic questions :) And also for the other
> crash you mentioned.
>
>   Yours,
> 	Eeli Kaikkonen (Mr.), Oulu, Finland
> 	e-mail: eekaikko at mailx.studentx.oulux.fix (with no x)
>   
The debug output does not seem to work in the windows version, so that 
is another issue to look into.

I was installing 1 module at a time and waiting for it to finish. The 
crash happened just after the progress dialog hit 100%. I did not use 
cancel.

The sword crash I mentioned was basically the same testing as above. I 
had a lengthy discussion with Matthew Talbert on irc last night. It 
seems clear that the issue is in sword file management, but we could not 
find the exact issue.

Gary