493547 - [SeaMonkey, cn-sea-qm-win2k3-01] crashtest: new 188 bytes leak, including nsPluginHostImpl, on 2009/05/17 after big new tests landing

Status: RESOLVED FIXED

(SeaMonkey :: Build Config, defect)

Description

[Serge Gautherie (:sgautherie)]

First build is:
{
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1242616339.1242620567.15305.gz
WINNT 5.2 comm-central unit test on 2009/05/17 20:12:19

TEST-UNEXPECTED-FAIL | runtests-leaks | leaked 188 bytes during test execution
TEST-UNEXPECTED-FAIL | runtests-leaks | leaked 1 instance of nsLocalFile with size 88 bytes
TEST-UNEXPECTED-FAIL | runtests-leaks | leaked 1 instance of nsPluginHostImpl with size 88 bytes
TEST-UNEXPECTED-FAIL | runtests-leaks | leaked 1 instance of nsStringBuffer with size 8 bytes
TEST-UNEXPECTED-FAIL | runtests-leaks | leaked 1 instance of nsVoidArray with size 4 bytes
}

Regression timeframe:
http://hg.mozilla.org/releases/mozilla-1.9.1/pushloghtml?startdate=2009-05-17+16%3A17%3A21&enddate=2009-05-17+20%3A09%3A20

To be noted: Windows SeaMonkey-Ports is still green.!.

Comment 1

[Serge Gautherie (:sgautherie)]

[Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9.2a1pre) Gecko/20090518 SeaMonkey/2.0b1pre] (experimental/_m-c_, home, optim default) (W2Ksp4)
(http://hg.mozilla.org/mozilla-central/rev/a7c0b3588242
 +http://hg.mozilla.org/comm-central/rev/6786ebf24275 + bug 493008 patches)

Fwiw, I don't reproduce it locally.

It looks like to be specific to this machine:
related to some (bad / badly handled) plugin?

Comment 2

[Robert Kaiser]

Yes, this looks somewhat strange, esp. as the boxes in SeaMonkey-Ports, which are testing the same code, are passing crashtests without failures.

Comment 3

[Serge Gautherie (:sgautherie)]

KaiRo, I would suggest, ftb, to code (in buildbot factory and config) and use a crashtest leak threshold...

Comment 4

[Serge Gautherie (:sgautherie)]

KaiRo, could you compare <about:plugins> on SeaMonkey and SeaMonkey-Ports boxes?

Comment 5

[Robert Kaiser]

about:plugins is identical on builds from both buildbot configs.

Comment 6

[Serge Gautherie (:sgautherie)]

Though I'd like to figure this out,
I guess you don't have much time to spend on this issue,
so the best course of action will be to just WontFix this after you have moved the new boxes to production...

Comment 7

[Serge Gautherie (:sgautherie)]

R.WontFix, now that the box(es) has been recycled.

Comment 8

[Serge Gautherie (:sgautherie)]

It looks like cn-sea-qm-win2k3-01 has actually been reused and is still leaking.

What would be the difference from cb-seamonkey-win32-01 and cb-seamonkey-win32-02 which don't leak?

Comment 9

[Robert Kaiser]

(In reply to comment #8)
> It looks like cn-sea-qm-win2k3-01 has actually been reused and is still
> leaking.

Sure it was "reused", the plan was to join the old VMs to the new pools all along, or else we wouldn't have/get the machine power to run multiple trees. It's interesting that it's still leaking, though.

> What would be the difference from cb-seamonkey-win32-01 and
> cb-seamonkey-win32-02 which don't leak?

I have no clue, I don't actually know of any difference.

Comment 10

[Serge Gautherie (:sgautherie)]

Attachment: (Av1) Add a threshold

Work around the leak until we get a better solution.

Comment 11

[Robert Kaiser]

Not sure if I like it, and I really want to know what's up with that box. In addition, we want to upgrade Java anyway, which will change some things regarding plugins, which this is related to, so I'd rather leave this as it is until I come around to care about this Java upgrade. And for this week, I don't care about anything but Beta 1 anyhow. ;-)

Comment 12

[Serge Gautherie (:sgautherie)]

Bug 503807 fixed this. (Unless bug 471647 would now be "masking" it.)

Comment 13

[Robert Kaiser]

Nice, we now have the same 484 leak than on the other boxes - at least that inconsistency on that one VM seems to be gone :)