Am Sonntag, 22. Januar 2006 17:03 schrieb Justin Pryzby:
> Hello David,
>
> On Sun, Jan 22, 2006 at 04:19:38PM +0100, David Schmitt wrote:
> > I have the same problems. Here some numbers:
>
> This doesn't answer precisely Eric's question.  He said:
> | Are you certain? If you open say 5 tabs, close them and then open 2
> | more, does memory usage increase markedly? Hard numbers please.
>
> Note that it involves closing existing tabs, and then reopening back
> to the same number.
>
> This is different from opening a tab (up to a new "record" number of
> tabs) and then closing it.
>
> FF is (probably) using a memory allocation scheme whereby memory
> doesn't get freed when it is no longer used, but retains allocation
> and reuses it when it is convenient.  I'm not sure I agree with this
> as a design decision, since memory allocation afaik isn't particularly
> expensive, and I agree that firefox seems to leak memory, in that it
> is *common* to end up with an FF process with a 200MB VM space.  But I
> don't know that your tests show this.

Indeed, running under valgrind showed only insignificant memory leaks (11kB) 
when loading the munin page. 

After further playing I also found, that FF indeed releases the pixmaps in the 
xserver when tabs are "replacing" the closed tabs with many graphics, i.e.: 3 
open tabs, close two, open two new empty, then FF seems to release most of 
the X resources (pixmaps) of the closed tabs. 



Regards, David
-- 
- hallo... wie gehts heute?
- *hust* gut *rotz* *keuch*
- gott sei dank kommunizieren wir über ein septisches medium ;)
 -- Matthias Leeb, Uni f. angewandte Kunst, 2005-02-15

Reply via email to