Bug 104945 – Layout Test fast/events/mouse-cursor-image-set.html is flaky

• Component: Tools / Tests
• Status: RESOLVED FIXED

[Eric Seidel (no email)]

The following layout test is flaky on all (chromium?) platforms

fast/events/mouse-cursor-image-set.html
http://trac.webkit.org/browser/trunk/LayoutTests/fast/events/mouse-cursor-image-set.html

http://test-results.appspot.com/dashboards/flakiness_dashboard.html#showExpectations=true&tests=fast%2Fevents%2Fmouse-cursor-image-set.html

This test has only been around for 7 days.

[Eric Seidel (no email)]

I'm not sure if it's an image loading issue:
 TEST CASE: Implicit hot-spot at (5,4) physical for 1x and (28,3) physical for 2x
-Cursor Info: type=Custom hotSpot=5,4 image=25x25
+Cursor Info: type=Custom hotSpot=0,0 image=0x0

it always seems to fail the same subtest.

[Eric Seidel (no email)]

I think it's this typo here:
 { url: 'resources/greenbox-hotspot35-4.cur' },

the '3'.

The loading system seems fragile.  Maybe we could read those images from the CSS values?

[Rick Byers]

I'll look into this.  Yes if the test fails to reliably preload one of the images then it can get output like this.  Ideally I'd be able to assert that an image was already preloaded or something.  And yes I thought about trying to combine the lists to avoid duplication and risk of inconsistency but I didn't come up with anything that was clean and simple.  I'll re-evaluate.  Let me know if you have any suggestions.

[Rick Byers]

Dammit, and yes this is just a typo or copy-paste error with 'hotspot35-4' that should be 'hotspot5-4', dammit - sorry about that.

I'll submit a trivial fix now, and file a separate bug to improve the infrastructure to make such mismatches impossible (which I probably won't get to until next week since I'm travelling this week).

[Rick Byers]

Underlying issue tracked in bug 104952

[Rick Byers]

Created attachment 179363 [details]
Patch

[Eric Seidel (no email)]

Comment on attachment 179363 [details]
Patch

Thank you Rick and Beth!

[Rick Byers]

(In reply to comment #7)
> (From update of attachment 179363 [details])
> Thank you Rick and Beth!

Thank you for noticing and ccing me!  I wish we had a way for test owners to be auto-notified of any failures.

[WebKit Review Bot]

Comment on attachment 179363 [details]
Patch

Clearing flags on attachment: 179363

Committed r137684: <http://trac.webkit.org/changeset/137684>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.