509313 - Huge amount of memory is used & does not stop climbing @ download.com

Status: RESOLVED DUPLICATE of bug 506428

(Core :: General, defect)

Description

[Noah (NoahSUMO/SolidSnake) [:Noah]]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2a1pre) Gecko/20090718 Firefox/3.0.11
Build Identifier: 

Just visit:
http://download.com.com/3000-2248-10201202.html?part=117075&subj=dlpage&tag=button

Don't need to do anything on the page, sit back, crack open Task Manager and in less than a minute, Firefox will be going over 280k will no sign of stopping. The browser becomes unresponsive and you have to force kill it.

I had JIT off on my trunk build, though I doubt it would help if it was on.

I only had one tab open, no addons, please don't try to troubleshoot my setup. Just try it yourself.

Reproducible: Always

Comment 1

[Milos Dinic [:Milos]]

Works fine for me on Linux x86_64 and Firefox 3.5.2
Anyone else?

Comment 2

[Jan Bambach [:Jan]]

Attachment: Screenshot (Website opened)

Screenshot (Windows Task Manager)

Comment 3

[Jan Bambach [:Jan]]

Attachment: Screenshot. All tabs are closed. FF is still using too much memory

All tabs are closed. FF is still using too much memory

Comment 4

[Milos Dinic [:Milos]]

WOW, you have 91 processes running? Try searching for spyware, malware or viruses.

Comment 5

[Milos Dinic [:Milos]]

Also, try running Firefox in safe mode. Check:

http://support.mozilla.com/kb/Safe+Mode

Comment 6

[Milos Dinic [:Milos]]

Marking this as WFM. If you have some other arguments to prove that bug, please comment here again.

Comment 7

[u348128]

Did you disable cookies? This might be a dupe of bug 506428.

cnet.com and download.com are related.
.

Comment 8

[u348128]

Yes, it reproduces when you disable cookies. I mark this as dupe.

Als, if you search on internet, it seems that IE8 has the same problem. Additional code is loaded, when cookies are disabled. This fires events and probably some kind of event escalation happens.

Comment 9

[u348128]

As workaround, I suggest you turn on cookies specifically for that site.

Comment 10

[Noah (NoahSUMO/SolidSnake) [:Noah]]

Hey Lucas, you're fast :P

I already figured that out but couldn't report that here in time. But I'll go ahead and put my story down anyway. :P

The troubleshooting:
**************************
I went back to the page a few more random times, and could not reproduce the issue. Even though I had reproduced it twice back to back in 2 separate Firefox sessions with no left over cache from the previous sessions because I was in permanent private browsing mode.

So I disabled & enabled Flash in my tests, since there was a flash ad present. In both tests, the page memory stayed @ 70mb.

Then I noticed a key issue I missed, my cookies (which I normally leave blocked all the time) were enabled because I needed to enable them so bugzilla could set its cookies when I was filing this bug.

Sure enough, I disabled cookies and the problem was 100% reproducible. I repro'ed it six times and had to force kill Firefox each time.

Do NOT visit the page until you disable cookies first, once the cookies are set, there are no problems.

Comment 11

[u348128]

I have been trying to reduce the testcase in the other bug, so I recognized it immediately.

The problem is, I can't reproduce it locally. Some file CookieXfer.html is loaded and is doing non-obvious things. It also uses 'fireEvent' method and I think that is non-standard Javascript. This also makes the bug less interesting for further analysis. With a 100% repro, one could catch a general bug, crashing other sites. But it seems that this very site specific.