Closed Bug 261843 Opened 20 years ago Closed 13 years ago

trunk builds causing Red Hat 7.x X server to leak memory

Categories

(Core Graveyard :: GFX: Gtk, defect)

x86
Linux
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED WORKSFORME

People

(Reporter: dbaron, Assigned: blizzard)

Details

Sometime around a month ago there was a regression on the trunk that causes the X servers on Red Hat 7.2 (and I think 7.1, and maybe others) to leak memory every time some of the Mozilla tinderbox tests are run. This is causing problems with our tinderbox automation since we need to restart some of the tinderbox's X servers every week or so (at least balsa and costarica, although I'm not sure that costarica is the same problem). This may affect other X servers as well. I'm not sure which tinderbox test is causing the problem, but probably somebody is going to need to figure out which one and then binary search builds.
(Note that the leaks are happening for both trunk builds on balsa -- a suite GTK1 build and a Firefox GTK1 build.)
Is there a quick way to detect the memory leak when one runs into it? I have a RH 7.2 system....
What do you mean by quick? A cron job I set up to record the amount of memory the X server was using showed it going up whenever the tinderbox tests ran on balsa -- but I haven't reduced it to which test or anything like that.
I meant some algorithmic way I could detect the problem when testing a build... I guess just start the build and then do stuff and then quit and see whether the X server memory usage went up, eh?
mlk keyword?
Does this problem still show up with the current Tinderboxes?
For what it's worth, the FC4 X server also seems to leak during Mozilla usage...
Product: Core → Core Graveyard
I would like to think this is long gone by now? I've been able to run Firefox for days on end without problems on my own X server, anyway.
Status: NEW → RESOLVED
Closed: 13 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.