Re : Interesting UI bug

["E. ALLAUD" <e allaud wanadoo fr>]

On 2003.04.29 12:02, E. ALLAUD wrote:
> This bug bites me sometimes, and it seems like a gtk+ bug. I don't 
> know how to reproduce it exaclty, but it happens when I check my POP3 
> mboxes and browse my mails at the same time (I suspect that it is 
> triggered when interacting with the gtk tree view and balsa is 
> updating it by adding the incoming mails). The tree is still 
> functional but when you click on a mail it does not get displayed (ie 
> the balsamessage callbakc is not called or blocked), and the cursor 
> flickers (it seems to be refreshed a lot). Here is the bt obtained by 
> hitting ctrl-c :
> 
> (gdb) bt
> #0  0x410cfa0b in writev () from /lib/i686/libc.so.6
> #1  0x4117e710 in XUnlockDisplay () from /usr/X11R6/lib/libX11.so.6
> #2  0x4117f32f in _X11TransWritev () from /usr/X11R6/lib/libX11.so.6
> #3  0x4115fdc7 in _XSend () from /usr/X11R6/lib/libX11.so.6
> #4  0x4115524a in _XReverse_Bytes () from /usr/X11R6/lib/libX11.so.6
> #5  0x4115598b in XPutImage () from /usr/X11R6/lib/libX11.so.6
> 
> This seems like a loop in X events. It is sufficient to click on 
> another mailbox and back to the one the bug occured in to have all 
> things functional again.
> Weird, isn't it?
> Bye
> Mau

More on this bug : it can be triggered only by deleting mails in a 
mailbox (I did it on local mboxes). So this means that it is only the 
delete path which is buggy or which triggers a gtk bug. This narrows 
the thing a bit. Moreover I saw that gtk_tree method are also called 
(the backtrace I obtain by pressing ctrl-c changes each time I do it, 
because there is an infinite loop, this is confirmed by top which says 
that balsa eats 60% of cpu and X the remaining 40%).

Bye
Manu

PGP signature