bugGNU Octave - Bugs: bug #54507, Segfault in QT5 when resizing...

 
 

bug #54507: Segfault in QT5 when resizing window during initialization

Submitted by:  None
Submitted on:  Mon 13 Aug 2018 06:40:56 PM UTC  
 
Category:  GUI Severity:  3 - Normal
Priority:  5 - Normal Item Group:  Segfault, Bus Error, etc.
Status:  Need Info Assigned to:  None
Originator Name:  Steve Kargl Originator Email:  -email is unavailable-
Open/Closed:  Open Release:  4.4.1
Operating System:  BSD

Add a New Comment(Rich Markup)
   

You are not logged in

Please log in, so followups can be emailed to you.

 

Fri 17 Aug 2018 12:04:11 AM UTC, comment #4:

There could be something in Octave. For a half year or more I've been experiencing a startup crash only after compilation and build when working on a bug, and not every time after compilation. Instead maybe every tenth or fifteen time. I'm using Mint linux.

It's the sort of thing in which, I'm guessing, the OS has to do a bit more work to locate and load the executable for some reason. (Maybe find the program and put it in some hash table for quick future access? Again, just guessing.)

Every time I see this crash, I try thinking of a way I could devise a way to find it, maybe make it occur regularly. But then I just move on because the next launch will work fine.

I keep thinking it might have something to do with loading GUI settings from the previous session, that the long load delay causes the settings to clash with whatever else is happening at start up. But again, that is only a general feeling.

So, the comment here about moving the window makes me wonder. Window managers often do some fancy stuff. For example, it wouldn't surprise me if launching an application in GTK or whatever redraws a window from tiny to small to create some neat-o effect. So, maybe if there is a delay in loading and the settings restores the window size later than usual, maybe there is a thread clash of what the Window Manager is doing and what Octave GUI window resize is doing.

If you can make this problem more common than once every one or two days (which is what I see), it might lead to a route of investigation. Perhaps Torsten could give simple instructions for a line to comment out so that the Octave GUI does not restore settings at start up, then you build and try to repeat the start-up crashes on FreeBSD; if the crashes go away, that could be a lead.

Dan Sebald <sebald>
Tue 14 Aug 2018 08:24:50 PM UTC, comment #3:

Posted to the mailing list:

The crash doesn't occur on each execution. I can get the
segfault in 1 out of every 5 or so executions. It seems to
be related to how fast I can grab the window to resize/move
it. I haven't noticed whether the command prompt is present
when I do the grabbing, but it is present as octave dumps
core.

Someone else has reported the bug on FreeBSD, and my FreeBSD bug
report is here:

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230072

I'm not convinced it an octave problem, but I don't know
C++ so cannot delve into Qt5 with much success. It alsoi
could be a FreeBSD pthread problem.

If you need me to recompile a support library with debug
symbols, just ask.

John W. Eaton <jwe>
Project Administrator
Tue 14 Aug 2018 07:00:47 PM UTC, comment #2:

But seriously, I can't reproduce the problem because for me, once the window appears on the screen, Octave's initialization seems to be complete.

Could you post the result of executing

in your gdb session to see what the other threads are doing at the time of the crash?

Also, on your system, how long is the delay between the window first appearing on the screen and the Octave prompt showing up in the command window? Or the last bit of initialization of the other windows if the command prompt is not the last thing to appear?

John W. Eaton <jwe>
Project Administrator
Tue 14 Aug 2018 06:51:58 PM UTC, comment #1:

It seems you may have found a Henny Youngman problem.

John W. Eaton <jwe>
Project Administrator
Mon 13 Aug 2018 06:40:56 PM UTC, original submission:

On FreeBSD-12, amd64, system where I have installed octave 4.4.0 and 4.4.1 from FreeBSD ports collection, I see a seqfault
when starting 'octave --gui' from the command line. The segfault is reproducible for me by trying to resize the windows as soon as it appears on screen. AFAICT, this appears to be a race issue in the initialization of octave

Anonymous

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach Files:
   
   
Comment:
   

No files currently attached

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -email is unavailable- added by sebald (Posted a comment)
  • -email is unavailable- added by jwe (Posted a comment)
  • -email is unavailable- added by None (Submitted the item)
  •  

    Do you think this task is very important?
    If so, you can add your encouragement to it.
    This task has 0 encouragements so far.

    Only project members can vote.

     

     

     

    Follows 1 latest change.

    Date Changed by Updated Field Previous Value => Replaced by
    2018-08-14 jwe StatusNone => Need Info

    Back to the top


    Powered by Savane 3.3