Update of bug #23233 (project gnustep):
Status: None => Confirmed
_______________________________________________________
Follow-up Comment #2:
I did some analysis and it turns out that this is really caused by the fix I
made to NSWindow. Now the memory gets freed correctly, but not in the right
order. As the graphics context and the info dictionary in it are autoreleased
objects, they only get freed when the corresponding autorelease pool is
released. As by this time the NSWindow is already deallocated (although it
still has a reference) and this results in a segmentation fault.
Now what can we do about it? The only solution I can think of is to split up
the dealloc in NSWindow and if the windowNumber is set call _terminateBackend
and return without actually deallocating the NSWindow, which will get done
when the autorelease pool if cleared then the NSWindow is freeed.
_______________________________________________________
Reply to this item at:
<http://savannah.gnu.org/bugs/?23233>
_______________________________________________
Nachricht geschickt von/durch Savannah
http://savannah.gnu.org/
_______________________________________________
Bug-gnustep mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/bug-gnustep
