PySide Bugzilla Closed for New Bugs

PySide is now a Qt Add-on and uses the Qt Project's JIRA Bug Tracker instead of this Bugzilla instance. This Bugzilla is left for reference purposes.

Bug 799 - QTableWidget crash on clear()
: QTableWidget crash on clear()
Status: CLOSED ALREADYFIXED
Product: PySide
Classification: Unclassified
Component: PySide
: 1.0.0
: PC Linux
: P5 major
Assigned To: renato filho
:
:
:
  Show dependency treegraph
 
Reported: 2011-03-31 18:40 EEST by kegunderdev
Modified: 2011-04-29 18:07 EEST (History)
8 users (show)

See Also:


Attachments
Code and Data (26.10 KB, application/octet-stream)
2011-03-31 18:40 EEST, kegunderdev
Details

Note You need to log in before you can comment on or make changes to this bug.
Description kegunderdev 2011-03-31 18:40:45 EEST
Created attachment 296 [details]
Code and Data

Seen the crash in Windows and Linux.  Using 32 bit version of python 2.6 both
times.

Steps to produce the crash  (Note Numpy is required.) 
Run the NNsort.py script.
Adjust the Row count limit to 5 by using the lower spin box (row 347 is now
shown)

Click in the cell under column 0 in the table with the value 2667 (i.e. click
on cell id=2667)
Click on the Accept Selection button.
click on cell id=2448
Click on the Accept Selection button.
click on cell id=2645
Click on the Accept Selection button.
click on cell id=2676
Click on the Accept Selection button.


It should be noted that minor differences in input can cause the crash not to
occur (such as clicking in a cell from another column on the same row.)
Comment 1 Lauro Moura 2011-04-01 00:14:47 EEST
Couldn't manage to reproduce with PySide from git in a Ubuntu 64-bit debug
build.

Could you post the traceback and output of the crash?
Comment 2 kegunderdev 2011-04-01 21:46:39 EEST

Thanks for checking into it.  All I get as console output is:

*** glibc detected *** python: double free or corruption (fasttop): 0x088f03a0
***


I also can't consistantly reproduce the error since trying to get a debug build
made.  But, the error still occurs at times when I'm using the app.
Comment 3 kegunderdev 2011-04-05 22:32:05 EEST
Well, I never could get my app to run in debug mode using the openSUSE system. 
That distribution isn't really my cup of tea.

I built up a debian system using some spare cycles in the meantime.  I have no
issue running my app in gdb on debian.  Unfortunately, it never crashes on that
system (debug more or not.)

I did have to comment out one setMargin line in my mainmenu.py file on the
debian system (QVBoxLayout object had no attribute 'setMargin').  But other
than that, the app works perfectly.
Comment 4 Hugo Parente Lima 2011-04-06 16:28:12 EEST
(In reply to comment #3)
> Well, I never could get my app to run in debug mode using the openSUSE system. 
> That distribution isn't really my cup of tea.
> 
> I built up a debian system using some spare cycles in the meantime.  I have no
> issue running my app in gdb on debian.  Unfortunately, it never crashes on that
> system (debug more or not.)
> 
> I did have to comment out one setMargin line in my mainmenu.py file on the
> debian system (QVBoxLayout object had no attribute 'setMargin').  But other
> than that, the app works perfectly.

I have the setMargin issue on ArchLinux too.
Comment 5 Hugo Parente Lima 2011-04-06 16:29:25 EEST
Oops, setMargin is obsolete, so it's not binded, use setContentsMargins
instead.
Comment 6 Hugo Parente Lima 2011-04-06 16:34:23 EEST
I also could not reproduce this.
Comment 7 Hugo Parente Lima 2011-04-19 20:34:34 EEST
Could you provide a reproduceable test case? otherwise will be near impossible
to fix this bug... or I can tag it as ALREADYFIXED if you also can't reproduce
it anymore?
Comment 8 Hugo Parente Lima 2011-04-27 23:20:11 EEST
Marking as ALREADYFIXED as nobody were able to reproduce this, but for sure
sometime it existed.
Comment 9 renato filho 2011-04-29 18:07:59 EEST
PySide release 1.0.2