Closed Bug 1524774 Opened 10 months ago Closed 9 months ago

Intermittent scrolling/frame-scrolling-attr-2.html == scrolling/frame-scrolling-attr-ref.html | image comparison, max difference: 1, number of differing pixels: 1


(Core :: Layout, defect, P5)




Tracking Status
firefox67 --- fixed


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



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

Waiting for reply on bug 1524031

There are 115 failures associated to this bug in the last 7 days. These are occurring on Linux, asan and debug builds.

:svoisen can you take a look?

Flags: needinfo?(svoisen)
Whiteboard: [stockwell needswork]

dholbert: Any thoughts on this?

Flags: needinfo?(dholbert)

It looks like a subtle antialiasing difference on the corner of a scrollbar.

Not a user-noticeable issue, and possibly not even our bug (could be an underlying GTK rendering bug).

I'll land a fuzzy annotation to allow for this 1px difference.

Flags: needinfo?(svoisen)

Unfortunately this test already has an unrelated "fuzzy-if" rule:

fuzzy-if(asyncPan&&!layersGPUAccelerated,0-140,0-6818) == iframe-scrolling-attr-2.html iframe-scrolling-attr-ref.html

I always have a hard time remembering & reasoning about reftest annotation precedence (there are cases where it doesn't do what you'd expect, particularly when you're mixing fails-if / skip-if with fuzzy-if). But since I'm just adding "fuzzy-if" alongside another fuzzy-if, I'm pretty sure it's straightforward and ordering/precedence doesn't matter.

(I have a patch ready to land on inbound; I'm just waiting for my local rebuild of inbound tip to complete, to sanity-check that the reftest still runs and that I didn't mess up the syntax.)

Pushed by
Add fuzzy annotation to allow for subtle 1px difference on scrollbar corner in reftest frame-scrolling-attr-2.html. (no review, test-manifest-tweak only)

This should stop being reported as a failure, but please reopen & needinfo me if there are still somehow instances of this after comment 15 is merged around.

Flags: needinfo?(dholbert)
Assignee: nobody → dholbert
Closed: 9 months ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla67

This is still happening.

Recent failure:

[task 2019-02-14T09:00:22.274Z] 09:00:22 INFO - REFTEST TEST-START | file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/scrolling/frame-scrolling-attr-2.html == file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/scrolling/frame-scrolling-attr-ref.html
[task 2019-02-14T09:00:22.274Z] 09:00:22 INFO - REFTEST TEST-LOAD | file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/scrolling/frame-scrolling-attr-2.html | 38 / 98 (38%)
[task 2019-02-14T09:00:22.953Z] 09:00:22 INFO - REFTEST TEST-UNEXPECTED-FAIL | file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/scrolling/frame-scrolling-attr-2.html == file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/scrolling/frame-scrolling-attr-ref.html | image comparison, max difference: 1, number of differing pixels: 1
[task 2019-02-14T09:00:22.955Z] 09:00:22 INFO - REFTEST IMAGE 1 (TEST): data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAyAAAAPoCAYAAAAmy
[task 2019-02-14T09:00:22.956Z] 09:00:22 INFO - REFTEST IMAGE 2 (REFERENCE): data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAyAAAAPoC
[task 2019-02-14T09:00:22.956Z] 09:00:22 INFO - REFTEST INFO | Saved log: START file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/scrolling/frame-scrolling-attr-2.html
[task 2019-02-14T09:00:22.960Z] 09:00:22 INFO - REFTEST INFO | Saved log: [CONTENT] OnDocumentLoad triggering AfterOnLoadScripts
[task 2019-02-14T09:00:22.962Z] 09:00:22 INFO - REFTEST INFO | Saved log: [CONTENT] FlushRendering generated paint for window data:text/html,<body style='overflow:scroll'>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>X<br>
[task 2019-02-14T09:00:22.963Z] 09:00:22 INFO - REFTEST INFO | Saved log: Initializing canvas snapshot
[task 2019-02-14T09:00:22.964Z] 09:00:22 INFO - REFTEST INFO | Saved log: DoDrawWindow 0,0,800,1000
[task 2019-02-14T09:00:22.966Z] 09:00:22 INFO - REFTEST INFO | Saved log: [CONTENT] RecordResult fired
[task 2019-02-14T09:00:22.967Z] 09:00:22 INFO - REFTEST INFO | Saved log: RecordResult fired
[task 2019-02-14T09:00:22.969Z] 09:00:22 INFO - REFTEST INFO | Saved log: RecordResult fired
[task 2019-02-14T09:00:22.970Z] 09:00:22 INFO - REFTEST TEST-END | file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/scrolling/frame-scrolling-attr-2.html == file:///builds/worker/workspace/build/tests/reftest/tests/layout/reftests/scrolling/frame-scrolling-attr-ref.html

Flags: needinfo?(dholbert)
Resolution: FIXED → ---
Target Milestone: mozilla67 → ---

Thanks. That's because I'm silly and annotated the wrong test. (I annotated iframe-scrolling-attr-2.html, with an "i", rather than frame-scrolling-attr-2.html)

Given bug 1525568 (a similar 2px intermittent failure on Mac) and given the existing annotation on the "-1" version of this test, I'm just going to push a non-platform-specific annotation that allows 2px to differ by 1 per color channel).

(waiting on inbound to reopen before I can land the patch.)

See Also: → 1525568
Pushed by
Backed out changeset 6d917b008b34 because it annotated the wrong test
Annotate frame-scrolling-attr-2.html as having 2 slightly fuzzy pixels. (no review, test-annotation-only)
Blocks: 1525568
Flags: needinfo?(dholbert)
Closed: 9 months ago9 months ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla67
Blocks: 1528141
You need to log in before you can comment on or make changes to this bug.