Closed Bug 1160774 Opened 10 years ago Closed 10 years ago

Intermittent e10s browser_toolbox_view_source_03.js | Test timed out

Categories

(DevTools :: General, defect)

defect
Not set
normal

Tracking

(e10s+, firefox39 unaffected, firefox40 wontfix, firefox41 fixed, firefox-esr31 unaffected, firefox-esr38 unaffected)

RESOLVED FIXED
Firefox 41
Tracking Status
e10s + ---
firefox39 --- unaffected
firefox40 --- wontfix
firefox41 --- fixed
firefox-esr31 --- unaffected
firefox-esr38 --- unaffected

People

(Reporter: philor, Assigned: sjakthol)

References

(Blocks 1 open bug)

Details

(Keywords: intermittent-failure)

Attachments

(1 file)

No description provided.
Blocks: e10s-tests
tracking-e10s: --- → +
The timeout occurs because source-utils.js makes an incorrect assumption that call to selectStyleSheet always causes 'stylesheet-selected' event to be emitted. If the style sheet that is to be selected is already selected, that event won't be emitted. In this test there's only one sheet in the document meaning the sheet was already selected during the call to loadTool. The failure is intermittent due to a pitfall in Style Editor initialization: the initial 'stylesheet-selected' event is not included in the initialization promise that loadTool waits to resolve. Often this event is emitted after loadTool resolves and the test works fine as the 'selected' promise is resolved by the initial event. However, if the initial event occurs before loadTool resolves, the 'selected' promise is never resolved as the event never triggers. Luckily we don't actually need that event as the promise returned by selectStyleSheet serves the same purpose.
Assignee: nobody → sjakthol
Status: NEW → ASSIGNED
Attachment #8609819 - Flags: review?(jsantell)
Attachment #8609819 - Flags: review?(jsantell) → review+
Looks good, and so does the try push; marked as checkin-needed; thanks, Sami!
Keywords: checkin-needed
Status: ASSIGNED → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
Target Milestone: --- → Firefox 41
Sami, is this something we should consider backporting?
Flags: needinfo?(sjakthol)
The patch in this bug doesn't make any difference to the user experience - it only affects this specific test. So if you're seeing this often on aurora and get tired of tagging the failures this should probably be uplifted. But the failure doesn't seem to be happening that often. Also this patch only applies to current aurora so Firefox 39 and esr38 should be fine.
Flags: needinfo?(sjakthol)
We don't run e10s tests on Aurora, so wontfix sounds like the way to go :)
Product: Firefox → DevTools
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: