← Back to team overview

registry team mailing list archive

Re: [Bug 593290] Re: Terminator crashes, when trying to split in too many windows

 

Hi Stephen,

I can't live anymore without Terminator, so of course I want to be more
detailed! ;)

Attached is my configuration which I use on all my computers/laptops and
the output of Terminator with the added print statement.

The freeze indeed uses 100% CPU, no exceptions are thrown or whatever,
it just locks up until I kill it.

It's indeed getting a small area when it crashes but to my feeling still
more than one line left, I have a 30 inch monitor and there seems to be
some space left.

Since I can reproduce the issue in less than 2 seconds it's no problem.
I hope this information is helpful. If not tell me what to do more, I
know quite some Python (not really into PyGTK) but if you want to add me
more debug statements, no problem.

Eventually I could give you temporarily access to one of my boxes so you
could try it by yourself.

Thanks for your time!

Greetings,

Poppe

> Date: Mon, 13 Sep 2010 17:20:02 +0000
> From: 593290@xxxxxxxxxxxxxxxxxx
> To: dwaalspoor98@xxxxxxxxxxx
> Subject: [Bug 593290] Re: Terminator crashes,	when trying to split in too many windows
> 
> Can I ask those of you experiencing this to be a little more descriptive
> of what you mean by "crash".
> 
> Do you mean it exits with segfault? Exits with stack trace? Or locks up,
> and soaks up lots of CPU, i.e 100% of a core?
> 
> I'm seeing the 100% core lockup. I've so far got it down to something where two vte's seem to get into a "fight" trying to get space. It's like:
> 1) A says "I need P",
> 2) the splitter tells B "you now have Q",
> 3) B says "no way! I need at least R",
> 4) splitter tells A "Sorry, can you cope with S",
> 5) A says "Nope, I need P",
> 6) Goto 2
> 
> The key thing that seems to kick it off is dividing a space where the
> two vte's get <1 line of height. The strange thing is that it's not 100%
> reproducible in trunk. I have been working on some optimisations, and
> one of them makes this happen very easily. It seems like a way to set
> minimum height/widths, and either deny a split, or to force splitters
> higher up the hierarchy to give us more space. (Yes, I have also seen
> this trigger on vertical splitters too, but we don't typically split as
> small horizontally as we do vertically.)
> 
> You can add "print self, repr(allocation)" to the top of the on_vte_size_allocate method in terminatorlib/terminal.py and then try to trigger the issue. What you should see on the stdout is lots of cycling. i.e.
> <Terminal object at 0x2716d70 (terminatorlib+terminal+Terminal at 0x2413700)> gtk.gdk.Rectangle(0, 27, 642, 23)
> <Terminal object at 0x2716d70 (terminatorlib+terminal+Terminal at 0x2413700)> gtk.gdk.Rectangle(0, 63, 642, 1)
> <Terminal object at 0x2702be0 (terminatorlib+terminal+Terminal at 0x2413160)> gtk.gdk.Rectangle(0, 17, 642, 23)
> <Terminal object at 0x2702be0 (terminatorlib+terminal+Terminal at 0x2413160)> gtk.gdk.Rectangle(0, 17, 642, 1)
> <Terminal object at 0x2716d70 (terminatorlib+terminal+Terminal at 0x2413700)> gtk.gdk.Rectangle(0, 27, 642, 23)
> <Terminal object at 0x2716d70 (terminatorlib+terminal+Terminal at 0x2413700)> gtk.gdk.Rectangle(0, 63, 642, 1)
> <Terminal object at 0x2702be0 (terminatorlib+terminal+Terminal at 0x2413160)> gtk.gdk.Rectangle(0, 17, 642, 23)
> <Terminal object at 0x2702be0 (terminatorlib+terminal+Terminal at 0x2413160)> gtk.gdk.Rectangle(0, 17, 642, 1)
> 
> -- 
> Terminator crashes, when trying to split in too many windows
> https://bugs.launchpad.net/bugs/593290
> You received this bug notification because you are a direct subscriber
> of the bug.
 		 	   		  

** Attachment added: "crashlog.tgz"
   https://bugs.launchpad.net/bugs/593290/+attachment/1579891/+files/crashlog.tgz

** Attachment added: "config"
   https://bugs.launchpad.net/bugs/593290/+attachment/1579892/+files/config

-- 
Terminator crashes, when trying to split in too many windows
https://bugs.launchpad.net/bugs/593290
You received this bug notification because you are a member of Registry
Administrators, which is the registrant for Fedora.



References