High CPU usage (Seamonkey 2)

[Dave Yeo-3]

For the last few months I've noticed that after having Seamonkey open
for a few days if I scroll, change tabs or even sometimes just move the
mouse over the open tab that CPU usage spikes. I get the clock pointer,
Seamonkey becomes unusable, switching to other applications is very
slow. Then things go back to normal. These spikes start out lasting a
second or two and gradually get worse. Right now Seamonkey has been
running for 5 days and at times it is freezing for 5 to 10 seconds. A
lot of the time if I stick to one tab the problem disappears.
System is a Athlon XP 2000 with 512 MBs memory. Currently swap is at 330
MBs with 70 MBs main memory free. This is not a swap problem though it
feels like it is as the disk light does not come on.
Does anyone else see similar?
Dave
ps Killing the session and letting session restore restore fixes the
problem.
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2

[Ilya Zakharevich]

[A complimentary Cc of this posting was sent to
Dave Yeo
<[hidden email]>], who wrote in article <[hidden email]>:
> For the last few months I've noticed that after having Seamonkey open
> for a few days if I scroll, change tabs or even sometimes just move the
> mouse over the open tab that CPU usage spikes. I get the clock pointer,
> Seamonkey becomes unusable, switching to other applications is very
> slow. Then things go back to normal. These spikes start out lasting a
> second or two and gradually get worse.

Same here with FF 2.* - after my daughter visits some of her favorite
sites, there is a certain "contamination" of UI (very similar to what
you describe).  I suspect "JS poisoning": some callbacks for UI
installed by web pages might be (?) not deleted when a page is left;
when hundreds of callbacks accumulate, any UI action would slow down
the operation.

Of course, this is just a conjecture - but it is the only conjecture I
know which would explain these slowdowns...  (BTW, I have no swap file
growth from the initial 24M.)

Hope this helps,
Ilya
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2

[rafe-2]

On May 17, 7:12 pm, Dave Yeo <[hidden email]> wrote:

> Does anyone else see similar?
> Dave
> ps Killing the session and letting session restore restore fixes the
> problem.

yes, I have the same sorts of problems -- very often when I return to
the computer, SeaMonkey 1.1.13 has somehow hung and I have to press
Ctrl-Esc numerous times to get the hanging app-dismiss screen to wake
up Doodle's Screen Saver and turn the monitor on.

it certainly has the feel of a memory leak -- after 2-3 days
frequently everything else running grinds down to a virtual halt & I
can hear lots of swapping.  if closing and restarting SM doesn't fix
it, I am obliged to reboot.

-rafe t.

_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2

[Peter Weilbacher]

On 05/18/09 11:28, Ilya Zakharevich wrote:

> Same here with FF 2.* - after my daughter visits some of her favorite
> sites, there is a certain "contamination" of UI (very similar to what
> you describe).  I suspect "JS poisoning": some callbacks for UI
> installed by web pages might be (?) not deleted when a page is left;
> when hundreds of callbacks accumulate, any UI action would slow down
> the operation.
>
> Of course, this is just a conjecture - but it is the only conjecture I
> know which would explain these slowdowns...  (BTW, I have no swap file
> growth from the initial 24M.)

If this is really JS-related then you and Dave are almost definitely
seeing different problems, because the JS implementation has changed so
much between your FF 2.x and his SM 2.0pre.
--
Please                      Enhanced OS/2 builds of Mozilla applications
reply in                                    http://pmw-warpzilla.sf.net/
newsgroup
     Steve's Warpzilla Tips: http://www.os2bbs.com/os2news/Warpzilla.html
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2

[Peter Weilbacher]

On 05/18/09 18:52, rafe wrote:

> yes, I have the same sorts of problems -- very often when I return to
> the computer, SeaMonkey 1.1.13 has somehow hung and I have to press
> Ctrl-Esc numerous times to get the hanging app-dismiss screen to wake
> up Doodle's Screen Saver and turn the monitor on.
>
> it certainly has the feel of a memory leak -- after 2-3 days
> frequently everything else running grinds down to a virtual halt&  I
> can hear lots of swapping.  if closing and restarting SM doesn't fix
> it, I am obliged to reboot.

There are no memory leaks, at least none of this size in Mozilla. Debug
builds have built-in checks that would show leaks immediately. (That's
how I know of one that leaks a few bytes but I don't know how to find
that.) As have said every time this "leak" discussion starts again, this
is most likely due to a non-optimal implementation of the C-library heap.

Especially, if a SM restart doesn't clear the problem then it is
unlikely to be Mozilla problem...

But why keep a system running for days anyway? If it's a server, don't
run browsers on it. If it's a desktop/workstation then switch it off
when the work is done. What a waste of energy. I cannot leave mine
running for more than a day anyway. So discuss all you want, I won't be
able to reproduce, let alone fix it. ;-)
--
Please                      Enhanced OS/2 builds of Mozilla applications
reply in                                    http://pmw-warpzilla.sf.net/
newsgroup
     Steve's Warpzilla Tips: http://www.os2bbs.com/os2news/Warpzilla.html
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2

[Ilya Zakharevich]

[A complimentary Cc of this posting was NOT [per weedlist] sent to
Peter Weilbacher
<[hidden email]>], who wrote in article <[hidden email]>:
> > Same here with FF 2.* - after my daughter visits some of her favorite
> > sites, there is a certain "contamination" of UI (very similar to what
> > you describe).  I suspect "JS poisoning": some callbacks for UI
> > installed by web pages might be (?) not deleted when a page is left;
> > when hundreds of callbacks accumulate, any UI action would slow down
> > the operation.

> > Of course, this is just a conjecture - but it is the only conjecture I
> > know which would explain these slowdowns...  (BTW, I have no swap file
> > growth from the initial 24M.)

Silly me: of course I should have written "JS or FLASH -related" -
there is no way to distinguish for which API the callbacks are
accumulating...

> If this is really JS-related then you and Dave are almost definitely
> seeing different problems, because the JS implementation has changed so
> much between your FF 2.x and his SM 2.0pre.

This does not guarantie that the callback-freeing code changed...

Yours,
Ilya
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2

[Ray Davison]

Dave Leo wrote:
> For the last few months I've noticed that after having Seamonkey open
> for a few days if I scroll, change tabs or even sometimes just move the
> mouse over the open tab that CPU usage spikes. I get the clock pointer,
> Seamonkey becomes unusable, switching to other applications is very
> slow. Then things go back to normal.

I have gotten such behaver on OS/2 SM for a long time.  And I have a
dedicated on-line machine that only runs when I am using it.  Sometimes
it seems to be caused by trying to do things too fast without letting
one activity settle before initiating the next.  Like maybe things get
jammed in a que.  Sometimes I can close SM and fix it.  Sometimes it is
so jammed I cannot close and must CAD.

This is a fairly current machine, and I don't recall seeing it on Win
SM.  Same thing with some other issues.  More and more I have to boot
Win to use SM.

Ray
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2

[David T. Johnson]

I've seen this as well.  The system is obviously running out of memory,
either through a SM memory leak or a memory leak in a process that SM
relys on.  Apparently small, though.


Dave Yeo wrote:

> For the last few months I've noticed that after having Seamonkey open
> for a few days if I scroll, change tabs or even sometimes just move the
> mouse over the open tab that CPU usage spikes. I get the clock pointer,
> Seamonkey becomes unusable, switching to other applications is very
> slow. Then things go back to normal. These spikes start out lasting a
> second or two and gradually get worse. Right now Seamonkey has been
> running for 5 days and at times it is freezing for 5 to 10 seconds. A
> lot of the time if I stick to one tab the problem disappears.
> System is a Athlon XP 2000 with 512 MBs memory. Currently swap is at 330
> MBs with 70 MBs main memory free. This is not a swap problem though it
> feels like it is as the disk light does not come on.
> Does anyone else see similar?
> Dave
> ps Killing the session and letting session restore restore fixes the
> problem.

--
Posted with OS/2 Warp 4.52
and Sea Monkey 1.5a
_______________________________________________
dev-ports-os2 mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-ports-os2