Closed Bug 1751733 Opened 2 years ago Closed 2 years ago

High frequency Linux/Windows browser/base/content/test/performance/browser_tabopen.js | unexpected changed rect: ({x1:213, x2:230, y1:4, y2:39, w:18, h:36}), window width: 1280 -

Categories

(Firefox :: General, defect, P5)

defect

Tracking

()

RESOLVED FIXED
98 Branch
Tracking Status
firefox-esr91 --- unaffected
firefox96 --- unaffected
firefox97 --- unaffected
firefox98 --- fixed

People

(Reporter: intermittent-bug-filer, Assigned: emilio)

References

(Regression)

Details

(Keywords: intermittent-failure, regression, Whiteboard: [stockwell disable-recommended])

Attachments

(6 files, 1 obsolete file)

Filed by: smolnar [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer?job_id=365236037&repo=autoland
Full log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/J1fbjOsJTwaztdgAayNBCg/runs/0/artifacts/public/logs/live_backing.log


INFO - TEST-PASS | browser/base/content/test/performance/browser_tabopen.js | 0 unexpected reflows - true == true - 
[task 2022-01-24T14:39:13.052Z] 14:39:13     INFO - comparing 4 frames
[task 2022-01-24T14:39:13.053Z] 14:39:13     INFO - Buffered messages finished
[task 2022-01-24T14:39:13.054Z] 14:39:13     INFO - TEST-UNEXPECTED-FAIL | browser/base/content/test/performance/browser_tabopen.js | unexpected changed rect: ({x1:213, x2:230, y1:4, y2:39, w:18, h:36}), window width: 1280 - 
[task 2022-01-24T14:39:13.054Z] 14:39:13     INFO - Stack trace:
[task 2022-01-24T14:39:13.056Z] 14:39:13     INFO - chrome://mochikit/content/browser-test.js:test_ok:1375
[task 2022-01-24T14:39:13.057Z] 14:39:13     INFO - chrome://mochitests/content/browser/browser/base/content/test/performance/head.js:reportUnexpectedFlicker/rects<:633
[task 2022-01-24T14:39:13.058Z] 14:39:13     INFO - chrome://mochitests/content/browser/browser/base/content/test/performance/head.js:reportUnexpectedFlicker:624
[task 2022-01-24T14:39:13.059Z] 14:39:13     INFO - chrome://mochitests/content/browser/browser/base/content/test/performance/head.js:withPerfObserver:693
INFO - data:image/png;base64
INFO - data:image/png;base64
 INFO - Not taking screenshot here: see the one that was previously logged
[task 2022-01-24T14:39:13.238Z] 14:39:13     INFO - TEST-UNEXPECTED-FAIL | browser/base/content/test/performance/browser_tabopen.js | should have 0 unknown flickering areas - Got 1, expected +0
[task 2022-01-24T14:39:13.239Z] 14:39:13     INFO - Stack trace:
[task 2022-01-24T14:39:13.240Z] 14:39:13     INFO - chrome://mochikit/content/browser-test.js:test_is:1403
[task 2022-01-24T14:39:13.241Z] 14:39:13     INFO - chrome://mochitests/content/browser/browser/base/content/test/performance/head.js:reportUnexpectedFlicker:650
[task 2022-01-24T14:39:13.242Z] 14:39:13     INFO - chrome://mochitests/content/browser/browser/base/content/test/performance/head.js:withPerfObserver:693
[task 2022-01-24T14:39:13.242Z] 14:39:13     INFO - Leaving test bound 
[task 2022-01-24T14:39:13.243Z] 14:39:13     INFO - GECKO(4279) | MEMORY STAT | vsize 20984129MB | residentFast 1888MB
[task 2022-01-24T14:39:13.244Z] 14:39:13     INFO - TEST-OK | browser/base/content/test/performance/browser_tabopen.js | took 1491ms
Flags: needinfo?(emilio)
Regressed by: 1751484

Dao, do you know how to best deal with this? This is not totally unexpected, and before it was happening just racily because of the delay in setting visibility... Should I just whitelist the rect?

Flags: needinfo?(dao+bmo)

So this is from browser_tabopen.js's usage of withPerfObserver which takes an allowed list of differing-rects and complains if there are changes that don't match that list.

The data URI screenshots show Firefox...
(1) just as the new tab is starting to appear (its "x" is barely visible, nothing else)
(2) after the new tab has finished opening.

This seems fine. The differences between the two are entirely contained in the tab-strip (in the area that the new tab occupies).

And browser_tabopen.js seems to intend to allow differences there, per its comment // We expect all changes to be within the tab strip.

So: seems like we likely need to adjust/add to the allowed-rect-list here.

Flags: needinfo?(emilio)
Flags: needinfo?(dao+bmo)
Assignee: nobody → emilio
Status: NEW → ASSIGNED

From looking at the screenshot I think this should do, but lmk if you
think otherwise or I got the wrong padding here... Also dump more
information for the next time the test fails.

Depends on D136783

Pushed by ealvarez@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/393e2d7c6bd5
Refactor condition for readability in browser_tabopen.js. r=florian

Yes, that's expected, see the other two patches.

Flags: needinfo?(emilio)

Note (given that the fixes are pending review): if the failure rate is painfully high and is a burden on sheriffs, I think it'd be fine to backout bug 1751484 for now (which should stop the failures), and we can re-land that bug's patch once we've got a full stack of r+'d patches here.

Summary: High frequency Linux browser/base/content/test/performance/browser_tabopen.js | unexpected changed rect: ({x1:213, x2:230, y1:4, y2:39, w:18, h:36}), window width: 1280 - → High frequency Linux/Windows browser/base/content/test/performance/browser_tabopen.js | unexpected changed rect: ({x1:213, x2:230, y1:4, y2:39, w:18, h:36}), window width: 1280 -
Has Regression Range: --- → yes
Whiteboard: [stockwell disable-recommended]
Whiteboard: [stockwell disable-recommended]
Pushed by ealvarez@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/279369667f3a
Dump some debug info in test. r=Gijs

Whether the overflow happens intentionally or not, it wasn't caused by
the regressing patch so I think we should probably just do this for now.

Pushed by ealvarez@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/086074bf3060
Account for tab close icon size in browser_tabopen.js. r=florian
Attachment #9260464 - Attachment is obsolete: true
Status: ASSIGNED → RESOLVED
Closed: 2 years ago
Resolution: --- → FIXED
Target Milestone: --- → 98 Branch
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: