Hi L.D.,

thanks for looking at this. Would be interesting to see if the English
pages have the same effect like the German.

On Sun, 25 Mar 2001 15:46:38 -0500, L.D. Best wrote:

> Joerg,

> I think all of us have seen sites which are "memory eaters" ...
> About.com used to be really bad, IIRC. :<

> This site, however, shows some very interesting patterns with page
> loads/reloads ... Michael?

> The first time I got to the site I had 142 mem available; next page
> dropped it down to 141; next page to 136 or so  ... I pulled the ol'
> clear-the-memory stunt of shelling out to DOS [ALT-E] and returning.
> When I returned I was back to 142 available memory.

Right. Reloading Arachne will release the memory.

> Then I did something that showed results I've never seen before on a
> memory-eater site:  If I went back to the pages I'd already d/l'd there
> was no decrease in memory available.  Going back, going forward, if the
> page was already in cache it had no effect on available memory.

And it's really the very first time that a page will eat up the xSwap
mem.

> However, when I started downloading new pages once more, I got a
> dramatic drop in memory available == from 142 to 111 in one big bite.
> But after that the next half-dozen pages downloaded caused no further
> decrease in available memory.

> So, that's why I've cc'd this to bugs.  I don't know what's going on
> with the site, the page design, or Arachne. But this is the first time
> I've noticed that cached pages have a different effect on available
> memory than newly downloaded pages do.  This may be the "lab site" that
> will allow the xChaos folx to figure out where/why/how-to-fix the memory
> eater problem.

> l.d.
> ====
-- Arachne V1.70, NON-COMMERCIAL copy, http://arachne.cz/

Reply via email to