[Wine] Debugging a STATUS_ACCESS_VIOLATION

[James McKenzie]

Burzmali wrote:
> James McKenzie wrote:
>   
>> Burzmali wrote:
>>
>>     
>>> Here's a link to the error and the prior 200 lines run with +d3d http://pastebin.com/Qqkc3gGu
>>>
>>> Here's 100 priors the with +relay
>>> http://pastebin.com/JJsDBNcs
>>>
>>> After the first seh, the system loops throwing an seh, waiting for a resolution, and then throwing the same seh again until something timeouts or breaks the deadlock or whatever.  The issue always seems to raise when the application handles "context 0x32f6620"
>>>
>>>
>>>
>>>       
>> Did someone ask you to post a log with +relay?  If not, please repost 
>> the log file without it.  +relay shows inter-dll communication and 
>> generates LARGE log files (which pastebin.com will not accept) and makes 
>> it hard to determine what triggered the first seh error.
>>
>> James McKenzie
>>     
>
> I've been working on this for a few days and the DIB (or lack thereof) is to blame.  Basically, the seh I am getting is Wine flipping a flag to mark that X needs to sync.  In the application, the developers choose to do something pixel by pixel, we sends the flipping locks on and off and syncing to achieve in 3 minutes what Windows does in less than a second.
>
> I'll update the bug that relates to this (http://bugs.winehq.org/show_bug.cgi?id=22968) and maybe poke through the code some more, but it isn't hopeful.
>   
If it is DIB related, look at bug 421 (yes, we have bugs that old) and 
see if the code supplied by Max will work.  It is from before the Wine 
1.2 release (he specifies where you have to set Wine's git to in the 
instructions).

Yes, the DIB engine is a long and very difficult item to incorporate 
into Wine.

James McKenzie