1355998 - Fix & re-enable browser_inlinesettings_browser.js on Windows

Status: RESOLVED FIXED

(Toolkit :: Add-ons Manager, defect, P2)

Description

[Mark Banner (:standard8)]

In landing screenshots (bug 1346825), we found an issue with browser_inlinesettings_browser.js was pretty much (not quite) perma-failing on Windows.

Typical log (from https://treeherder.mozilla.org/logviewer.html#?job_id=91025656&repo=try&lineNumber=6366):

12:02:19     INFO - Entering test bound test_inline_browser_addon
12:02:19     INFO - TEST-PASS | toolkit/mozapps/extensions/test/browser/test-window/browser_inlinesettings_browser.js | view controller is tracking the displayed view correctly - 
12:02:19     INFO - TEST-PASS | toolkit/mozapps/extensions/test/browser/test-window/browser_inlinesettings_browser.js | Options should be inline browser type - 
12:02:19     INFO - TEST-PASS | toolkit/mozapps/extensions/test/browser/test-window/browser_inlinesettings_browser.js | Element should not be null, when checking visibility - 
12:02:19     INFO - Buffered messages logged at 12:02:11
12:02:19     INFO - TEST-PASS | toolkit/mozapps/extensions/test/browser/test-window/browser_inlinesettings_browser.js | Preferences button should be visible - 


12:02:19     INFO - TEST-UNEXPECTED-FAIL | toolkit/mozapps/extensions/test/browser/test-window/browser_inlinesettings_browser.js | A promise chain failed to handle a rejection:  - at chrome://global/content/bindings/browser.xml:151 - Component returned failure code: 0x80040111 (NS_ERROR_NOT_AVAILABLE) [nsIWebNavigation.loadURIWithOptions]
12:02:19     INFO - Stack trace:
12:02:19     INFO - JS frame :: chrome://global/content/bindings/browser.xml :: loadURIWithFlags/< :: line 151
12:02:19     INFO - JS frame :: chrome://global/content/bindings/browser.xml :: _wrapURIChangeCall :: line 43
12:02:19     INFO - JS frame :: chrome://global/content/bindings/browser.xml :: loadURIWithFlags :: line 150
12:02:19     INFO - JS frame :: chrome://global/content/bindings/browser.xml :: loadURI/< :: line 114
12:02:19     INFO - JS frame :: chrome://global/content/bindings/browser.xml :: _wrapURIChangeCall :: line 43
12:02:19     INFO - JS frame :: chrome://global/content/bindings/browser.xml :: loadURI :: line 113
12:02:19     INFO - JS frame :: chrome://mozapps/content/extensions/extensions.js :: gDetailView.createOptionsBrowser</< :: line 3678
12:02:19     INFO - JS frame :: resource://gre/modules/Promise.jsm -> resource://gre/modules/Promise-backend.js :: Promise :: line 390
12:02:19     INFO - JS frame :: chrome://mozapps/content/extensions/extensions.js :: gDetailView.createOptionsBrowser< :: line 3651
12:02:19     INFO - JS frame :: resource://gre/modules/Task.jsm :: TaskImpl_run :: line 319
12:02:19     INFO - JS frame :: resource://gre/modules/Promise.jsm -> resource://gre/modules/Promise-backend.js :: process :: line 922
12:02:19     INFO - JS frame :: resource://gre/modules/Promise.jsm -> resource://gre/modules/Promise-backend.js :: walkerLoop :: line 806
12:02:55     INFO - Longer timeout required, waiting longer...  Remaining timeouts: 1
12:03:40     INFO - Not taking screenshot here: see the one that was previously logged
12:03:40     INFO - TEST-UNEXPECTED-FAIL | toolkit/mozapps/extensions/test/browser/test-window/browser_inlinesettings_browser.js | Test timed out - 


Logging showed that the URL passed to loadURIWithOptions was "moz-extension://70b66e9e-6247-406c-9b2f-755ea93bd72c/options.html".

On IRC, Rob said that it looked like the add-on created for the test was id aa69531f-2c04-4290-8efc-584533735b2a.

Comment 1

[Robert Helmer [:rhelmer]]

(In reply to Mark Banner (:standard8) from comment #0)
> Logging showed that the URL passed to loadURIWithOptions was
> "moz-extension://70b66e9e-6247-406c-9b2f-755ea93bd72c/options.html".
> 
> On IRC, Rob said that it looked like the add-on created for the test was id
> aa69531f-2c04-4290-8efc-584533735b2a.

FTR I was reading the log wrong - that was the moz-extension:// UUID so isn't helpful.

Comment 2

[Robert Helmer [:rhelmer]]

Attachment: Bug 1355998 - re-enable browser_inlinesettings_browser.js on Windows. Screenshots delaying webextension startup fixes this intermittent.

Review commit: https://reviewboard.mozilla.org/r/134412/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/134412/

Comment 3

[Robert Helmer [:rhelmer]]

I can't reproduce this after several tries on tryserver - I suspect delaying start of the webextension in bug 1346825 helped. What do you think Mark?

Comment 4

[Mark Banner (:standard8)]

Comment on attachment 8862511 [details]
Bug 1355998 - re-enable browser_inlinesettings_browser.js on Windows. Screenshots delaying webextension startup fixes this intermittent.

https://reviewboard.mozilla.org/r/134412/#review138468

I did a try build with Screenshots enabled within tests and we still hit this failure:

https://treeherder.mozilla.org/#/jobs?repo=try&revision=e0c05d0a7cd0da5363cb049dcc7d7224e9f32997&selectedJob=95864785

https://treeherder.mozilla.org/logviewer.html#?job_id=95866141&repo=try&lineNumber=16818

So I think we should try fixing those before we re-enable this.

Comment 5

[Robert Helmer [:rhelmer]]

Ugh, yes I should have done my try run with screenshots enabled.

Comment 6

[Robert Helmer [:rhelmer]]

Comment on attachment 8862511 [details]
Bug 1355998 - re-enable browser_inlinesettings_browser.js on Windows. Screenshots delaying webextension startup fixes this intermittent.

https://reviewboard.mozilla.org/r/134412/#review138468

Rebasing on a relatively new mozilla-central doesn't seem to show this anymore: https://treeherder.mozilla.org/#/jobs?repo=try&revision=ce979729295872bfaee08e601e00419bd1224406

Comment 7

[Mark Banner (:standard8)]

Comment on attachment 8862511 [details]
Bug 1355998 - re-enable browser_inlinesettings_browser.js on Windows. Screenshots delaying webextension startup fixes this intermittent.

https://reviewboard.mozilla.org/r/134412/#review139316

Per discussions & further try runs, the issue seems to have gone away, so lets re-enable this and see how it goes.

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/661b2bd5e032
re-enable browser_inlinesettings_browser.js on Windows. Screenshots delaying webextension startup fixes this intermittent. r=standard8

Comment 9

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/661b2bd5e032

Comment 10

[Mark Banner (:standard8)]

Unfortunately this still seems to be happening when screenshots is enabled for tests:

https://treeherder.mozilla.org/#/jobs?repo=try&revision=ac58b8b3c33146661ffdb796e80cd98bcd19fd03&selectedJob=96867938

Comment 11

[Robert Helmer [:rhelmer]]

Hm. I wonder why my try run from comment 6 didn't show it - I guess it's possible this only happens for non-debug builds :/

Once I can repro this somewhere (try or locally) I should be able to get it fixed up, thanks for the needinfo Mark.

Comment 12

[Robert Helmer [:rhelmer]]

Hm. Well I only see this on windows 7, and only for opt builds, but I can repro it now:

https://treeherder.mozilla.org/#/jobs?repo=try&revision=9aced73760659ce7ca1d815352511b9103f7dd9e&selectedJob=97418432

Comment 13

[Robert Helmer [:rhelmer]]

OK. So this is only happening when about:addons is loaded into a window (as opposed to a browser tab), as far as I can tell:

https://treeherder.mozilla.org/#/jobs?repo=try&revision=b1dca8c87da41ffaf40445f2d1116e92c6064d41&selectedJob=97589408

I recently announced intent to remove this from our tests, because there are a number of bugs that we don't want to fix that only show up in this mode and Firefox hasn't used it for years. I'll go ahead and do that in bug 1349723.

Comment 14

[Mark Banner (:standard8)]

Rob, should we try this again now that bug 1349723 has landed?

Comment 15

[Robert Helmer [:rhelmer]]

(In reply to Mark Banner (:standard8) from comment #14)
> Rob, should we try this again now that bug 1349723 has landed?

Yes we should - I will try again today.

Comment 16

[Robert Helmer [:rhelmer]]

Actually looks like we already landed this and it stuck:
https://hg.mozilla.org/mozilla-central/rev/661b2bd5e032

Please reopen if I'm mistaken.