+++ This bug was initially created as a clone of Bug #1371100 +++ https://treeherder.mozilla.org/logviewer.html#?job_id=105524186&repo=autoland [task 2017-06-08T16:26:31.082121Z] 16:26:31 INFO - TEST-PASS | dom/tests/browser/browser_noopener.js | Name should match for #test12 (private=false, container=true, alwaysNewWindow=true) - "uniquename3" == "uniquename3" - [task 2017-06-08T16:26:31.084279Z] 16:26:31 INFO - Buffered messages logged at 16:26:26 [task 2017-06-08T16:26:31.091505Z] 16:26:31 INFO - TEST-PASS | dom/tests/browser/browser_noopener.js | Opener should have been set for #test12 (private=false, container=true, alwaysNewWindow=true) - [object Window] == true - [task 2017-06-08T16:26:31.097381Z] 16:26:31 INFO - Leaving test bound newwindow_newproc [task 2017-06-08T16:26:31.099297Z] 16:26:31 INFO - Buffered messages finished [task 2017-06-08T16:26:31.101693Z] 16:26:31 INFO - TEST-UNEXPECTED-FAIL | dom/tests/browser/browser_noopener.js | This test exceeded the timeout threshold. It should be rewritten or split up. If that's not possible, use requestLongerTimeout(N), but only as a last resort. -

It looks like the last messages were logged just 5 seconds before the test timed out. I think I just need to bump up the test timeout even further, or remove some of the tests. Currently the test opens and closes ~166 windows (at least half of which are opened in brand new processes). This takes a really long time :S. Currently I am `requestLongerTimeout(15)`-ing in the test. I could bump it up even higher but that seems absurd. Any thoughts Olli?

Are there some tests which might not be needed? If not, requestLongerTimeout with 20 might not be too bad. I know it is high value.

(In reply to Olli Pettay [:smaug] from comment #2) > Are there some tests which might not be needed? > If not, requestLongerTimeout with 20 might not be too bad. I know it is high > value. I run the test suite twice, once with the pref for our of process noopener windows enabled, and one with it disabled. I could probably get away with not running the ones with it disabled which would potentially cut test time in half.

But do we have similar tests for the disabled case elsewhere?

(In reply to Olli Pettay [:smaug] from comment #4) > But do we have similar tests for the disabled case elsewhere? The disabled case is the current default, so it's tested by any other tests we have in the tree. I'm not sure if the specific things I'm testing are tested elsewhere. I originally tested all cases to make sure that I was keeping them consistent.

(In reply to Michael Layzell [:mystor] from comment #5) > (In reply to Olli Pettay [:smaug] from comment #4) > > But do we have similar tests for the disabled case elsewhere? > > The disabled case is the current default, so it's tested by any other tests > we have in the tree. I'm not sure if the specific things I'm testing are > tested elsewhere. I originally tested all cases to make sure that I was > keeping them consistent. Any thoughts as to what I should do here? I'm thinking that at least for now bumping up the requestLongerTimeout should help.

Here's a patch which requests an even longer timeout MozReview-Commit-ID: 7YhR1duEIJr

Pushed by michael@thelayzells.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/d412467b2891 Request an even longer timeout for browser_noopener.js, r=smaug