288167 - memory leak in javascript animation

Status: RESOLVED WORKSFORME

(Firefox :: General, defect)

Description

[Hermann Stamm-Wilbrandt]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; de-DE; rv:1.7.6) Gecko/20050226 Firefox/1.0.1
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; de-DE; rv:1.7.6) Gecko/20050226 Firefox/1.0.1

This page results in a huge memory leak only for Firefox -- Mozilla, Opera and
IE do not have a memory leak here.
The most condensed version of the leak may be found here (leaks more slowly than
above):
http://www.stamm-wilbrandt.de/memory_leak/single_file_simple.html
The leak is caused by the statement 'animimg.obj.src = "../images/anim3.gif";',
although it always references the same image!
If you remove this statement, there is no memory leak!



Reproducible: Always

Steps to Reproduce:
1. Start Firefox
2. Open "http://www.stamm-wilbrandt.de/memory_leak/"
3. watch Mem Usage in Task Manager for Firefox

Actual Results:  
Firefox will eat up all memory

Expected Results:  
work with fixed amount of memory

Comment 1

[Giles Robertson]

Works for me.

Displays (both URLs) with constant memory usage.

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-GB; rv:1.7.6) Gecko/20050321
Firefox/1.0.2

Comment 2

[José Jeria]

WFM: Windows XP, Firefox 1.0.1 and latest trunk build. Memory usage stays the
same. Can you retest in safe mode?

Comment 3

[Henrik Lynggaard Hansen]

I am confirming this defect. I have tested on my linux box using 
Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.7.6) Gecko/20050324 Firefox/1.0
(Ubuntu package 1.0.2) (a.k.a. firefox 1.0.2 without the restricted artwork).

In ten minutes by rss memory grew from roughly 32 mb to 112 mb. Firefox was
doing nothing else and it was a clean start. Screenshots of system monitor comming.

also changing os to all as I saw it on linux

Comment 4

[Henrik Lynggaard Hansen]

Attachment: sys mon on clean start

Comment 5

[Henrik Lynggaard Hansen]

Attachment: sys mon after 5 minutes

This should be plenty of time for memory usage to reach a stable level.

Comment 6

[Henrik Lynggaard Hansen]

Attachment: sys mon after 10 minutes

Memory usage just kept increasing.

Comment 7

[Henrik Lynggaard Hansen]

suggest blocking 1.0.3 since this is one hefty memory leak, more than 5 mb a minute.

Comment 8

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

We're not taking changes on 1.0.x other than security fixes and regressions
along the 1.0.x series (i.e., regressions from security fixes).

This is also pretty clearly in the wrong component and needs an appropriate one.

Comment 9

[Hermann Stamm-Wilbrandt]

(In reply to comment #8)
>... This is also pretty clearly in the wrong component and needs an 
appropriate one.
The appropriate component would be "layout engine" (Gecko), but this is not 
selectable under "Components" above.

Comment 10

[José Jeria]

Since this bug works fine on trunk, it should be marked as WORKSFORME. Only
security fixes will land on future 1.0.x releases.

Comment 11

[Giacomo Magnini]

Why this is not marked as WONTFIX?

Comment 12

[José Jeria]

This wfm on trunk --> wfm

Plese reopen if you can reproduce with a nightly trunk build.

Comment 13

[Hermann Stamm-Wilbrandt]

I installed the nightly build from 12.4.2005 and can reproduce the bug on the 
link:
http://www.stamm-wilbrandt.de/memory_leak/

Memory gets eaten up at a rate of 100KB per second!

(In reply to comment #12)
> This wfm on trunk --> wfm
> Plese reopen if you can reproduce with a nightly trunk build.

Comment 14

[Pavel Cvrcek [:JasnaPaka]]

(In reply to comment #13)
> I installed the nightly build from 12.4.2005 and can reproduce the bug on the 
> link:
> http://www.stamm-wilbrandt.de/memory_leak/
> 
> Memory gets eaten up at a rate of 100KB per second!

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8b2) Gecko/20050505
Firefox/1.0+

I think that there isn't problem with page above. Memory doesn't get eaten up. 

Comment 15

[Wayne Mery (:wsmwk)]

both URL WFM windows trunk Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a1) Gecko/20061116 Minefield/3.0a1

and 20060510 - although CPU is through the roof on this build.

FF 2.0 
* http://www.stamm-wilbrandt.de/memory_leak/ is solid as a rock on windows.  
* single_file_simple.html settles down after a minute or two to no memory increase.

If you can reproduce a high magnitude of loss please reopen the bug and it will be helpful if you can narrow the testcase to something more specific - eg try non-animated gif or alter some aspect of the script.