1315042 - [e10s-multi] Re-enable tests after landing e10s-multi on nightly

Status: RESOLVED FIXED

(Testing :: General, defect, P2)

Description

[Gabor Krizsanits (INACTIVE)]

We had quite some difficulties with turning all the tests green with 2 content processes enabled, we had to temporarily turn off a few tests in Bug 1301340 as well in the end. Once 2 cp's are enabled on nightly, we want to find a way to turn these tests back on ASAP.

Comment 1

[Gabor Krizsanits (INACTIVE)]

Some of the tests seem to be fixed by the CPOW leak fix, I have not seen them failing on ash at least and has a green try push with them so I start with those.

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/1b619d3778de
Re-enable some tests. r=me

Comment 3

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/1b619d3778de

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/eaf48c4b002c
Re-enable some tests part2. r=me

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/c01643400ac9
Re-disable browser_referrer_middle_click.js because it fails often with e10s enabled. r=bustage-fix on a CLOSED TREE

Comment 6

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

Needinfo just to track that browser_referrer_middle_click.js got redisabled because of too many failures:

https://hg.mozilla.org/integration/mozilla-inbound/rev/c01643400ac9ac36c470c3e6840df098b3ad3f00

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/0c46a4a90613
redisable browser_referrer_middle_click_in_container.js for failures a=me

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/mozilla-central/rev/ece0e0fd16a3
redisable browser_referrer_middle_click_in_container.js for failures a=me

Comment 9

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

https://hg.mozilla.org/mozilla-central/rev/eaf48c4b002c
https://hg.mozilla.org/mozilla-central/rev/c01643400ac9
https://hg.mozilla.org/mozilla-central/rev/ece0e0fd16a3

Comment 10

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/ca140640db89
Re-disable browser_bug719271.js on Windows debug with e10s for frequent failures. r=bustage-fix

Comment 12

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/0c46a4a90613

Comment 13

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

https://hg.mozilla.org/mozilla-central/rev/ca140640db89

Comment 15

[:Gijs (he/him)]

Bug 1301340 forced browser_tabSwitchInPrintPreview.js to run in single-content-process-mode, but I don't see a bug to investigate that and/or remove that workaround, nor any comments in bug 1301340 as to why that was necessary. It's concerning to me that that change was necessary as the test relates to a security issue. Can you clarify if I'm just missing the bug for fixing that test / code properly, or if there isn't one (in which case, please can you file one) ?

Comment 16

[Gabor Krizsanits (INACTIVE)]

(In reply to :Gijs from comment #15)
> Bug 1301340 forced browser_tabSwitchInPrintPreview.js to run in
> single-content-process-mode, but I don't see a bug to investigate that
> and/or remove that workaround, nor any comments in bug 1301340 as to why
> that was necessary. It's concerning to me that that change was necessary as
> the test relates to a security issue. Can you clarify if I'm just missing
> the bug for fixing that test / code properly, or if there isn't one (in
> which case, please can you file one) ?

I haven't filed a bug for each of these tests, there is one bug where I plan to fix these issues (bug 1301015).

Most of the tests that used to fail with multi and timing related are working now, and there are a few tests where the reason for failure is some kind of anti pattern in the tests and in some cases assumptions about multiple tabs end up in the same content process. I went through most of these tests about a month ago and didn't find anything urgent to fix. That being said I'm going to get to it soon after/during the release as soon as I get time for it. The test you're worried about seem to be working fine locally, and if it failed in a scary way I would have investigated it back then at least, so I'm not too concerned, but feel free to try to reenable multi and push it to try, I expect it to work by now as many others on this list.

Comment 17

[BugBot [:suhaib / :marco/ :calixte]]

The leave-open keyword is there and there is no activity for 6 months.
:gbrown, maybe it's time to close this bug?

Comment 18

[Geoff Brown [:gbrown]]

I assume we should keep it open for eventual follow-up on disabled tests.

Comment 19

[BugBot [:suhaib / :marco/ :calixte]]

The leave-open keyword is there and there is no activity for 6 months.
:gbrown, maybe it's time to close this bug?

Comment 20

[Geoff Brown [:gbrown]]

Still comment 18?

Comment 21

[Mark Banner (:standard8)]

All the dependent bugs are now closed.