[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-core-devel
Subject:    Re: kde2 crashes over time?
From:       Don Sanders <sanders () kde ! org>
Date:       2000-10-09 16:44:27
[Download RAW message or body]

On Monday 09 October 2000 18:22, Don Sanders wrote:
> On Monday 09 October 2000 18:03, Mario Weilguni wrote:
> > Am Mon, 09 Okt 2000 schrieb George Staikos:
> > > Has anyone tried running KDE2 over a long period of time without
> > > logging out? (like 3-4 days at least)    I have noticed consistently
> > > that running it for this long causes it to crash out of the X server. 
> > > I'm running it in Xnest on XF864.0 (also seen on XF863.x)
> >
> > I've had a running 1.94 for 2 weeks or so until I'd to take down my
> > system for a kernel upgrade. I worked on it 8 hours a day, 5 days per
> > week without any problems beside kmail taking too much memory every 2
> > days or so.
>
> That should definitely NOT happen. Can you reproduce the problem? I run
> KMail for many days without any increase in memory used, (after the first
> few hours).
>
> If you receive or send a large attachment the amount of memory used will
> increase (and even though that memory will be freed later top doesn't
> report that). Also there is a problem with file->save as excessively
> pooling memory when many messages are saved, but I doubt that's your
> problem.
>
> I went over all of KMail very carefully to make sure the memory use
> characteristics of all functions were good and tested this pretty
> thoroughly.
>
> Hmm, do you perhaps have a fast ethernet connection and machine and receive
> thousands of messages at once?

Damn I never got around to fixing mysterious crashes caused by the following 
code in kmacctexppop.cpp

    if ((*cur)->parent()) {
      int count = (*cur)->parent()->count();
      if ((*cur)->parent()->operator[](count - 1) == *cur)
	(*cur)->parent()->unGetMsg(count - 1);
    }

Maybe you can uncomment that code and see if it helps. I can't think of any 
other factor that could cause memory use to grow over time.

BFN,
Don.

[prev in list] [next in list] [prev in thread] [next in thread]