1157468 - test_fetch_event.html leaks on e10s

Status: RESOLVED FIXED

(Core :: DOM: Workers, defect)

Description

[Andrew McCreight [:mccr8]]

This isn't visible right now because the content process leak threshold is fairly high (20kb) due to some GMP issues that have since been fixed.  If you can't fix this quickly, I'd like to disable this test for e10s.

Comment 1

[Josh Matthews [:jdm]]

Ah, I was operating under the assumption that those leaks were common to all e10s tests. I'll dig in.

Comment 2

[Andrew McCreight [:mccr8]]

Oh, to be clear, there are some leaks common to all e10s tests. ;)  The leaks should just be about 5kb rather than 14kb, which is what I see when I enable this test (but disable the test mentioned in bug 1157467).

Comment 3

[Josh Matthews [:jdm]]

Attachment: Avoid leaking the intercepting channel stream listener for redirected IPC channels

Comment 4

[Josh Matthews [:jdm]]

Since there's a cycle between HttpChannelChild and InterceptStreamListener, we missed a case where a redirect occurs and the original channel never receives OnStopRequest. This is where the cycle would usually be broken.

Comment 5

[Josh Matthews [:jdm]]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=17784a692550

Comment 6

[Josh Matthews [:jdm]]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=4cabed3fce17

Comment 7

[Josh Matthews [:jdm]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/fbc0045964c4

Comment 8

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/fbc0045964c4

Comment 9

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/fbc0045964c4