Bug 205808 – REGRESSION: [ Mac WK2 ] fast/text/international/unicode-bidi-other-neutrals.html is a flaky failure

• Status: RESOLVED
• Resolution: FIXED
• Component: Text

[Truitt Savell]

fast/text/international/unicode-bidi-other-neutrals.html

Description:
This test is a flaky image failure on Mac wk2. the image appears to have a random 'b' being placed off to the side. this was added in https://trac.webkit.org/changeset/253068/webkit and appears to have been flaky sense then.

History:
https://results.webkit.org/?limit=10000&suite=layout-tests&test=fast%2Ftext%2Finternational%2Funicode-bidi-other-neutrals.html

Diff:
https://build.webkit.org/results/Apple-Catalina-Debug-WK2-Tests/r254052%20(1496)/fast/text/international/unicode-bidi-other-neutrals-diffs.html

[Radar WebKit Bug Importer]

<rdar://problem/58345056>

[Truitt Savell]

marked this test as failing while it is investigated: https://trac.webkit.org/changeset/254060/webkit

[Doug Kelly]

I wonder if this is because of the test's use of the marquee tag -- I wonder if div will reproduce the original issue just the same?

[Alexey Proskuryakov]

I think that stopping the marquee with scrollamount="0" or a large scrolldelay would be another workaround.

[Doug Kelly]

Realistically, there's also no reason this needs to be an image test... Call it my unfamiliarity with the WebKit test infrastructure, but since this test was verifying that a page with other neutrals at the end of a bidi run would not crash/assert, simply using dumpAsText should be sufficient.

Yes, I'll work on a patch. :)

[Doug Kelly]

Created attachment 387110 [details]
Patch

[Maciej Stachowiak]

Comment on attachment 387110 [details]
Patch

r=me

[Jon Lee]

can we check this in?

[WebKit Commit Bot]

Comment on attachment 387110 [details]
Patch

Clearing flags on attachment: 387110

Committed r254705: <https://trac.webkit.org/changeset/254705>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.