Bug 205808

Summary:

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

Product:

WebKit

Reporter:

Truitt Savell <tsavell>

Component:

Text

Assignee:

Nobody <webkit-unassigned>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

commit-queue, dougk, jonlee, mjs, mmaxfield, rniwa, webkit-bot-watchers-bugzilla, webkit-bug-importer

Priority:

Keywords:

InRadar

Version:

WebKit Nightly Build

Hardware:

Unspecified

OS:

Unspecified

Attachments:

Description	Flags
Patch	none

Truitt Savell

Reported 2020-01-06 10:03:03 PST

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

Attachments
Patch (3.17 KB, patch) 2020-01-08 10:12 PST, Doug Kelly	no flags	Details Formatted Diff Diff
View All Add attachment proposed patch, testcase, etc.

Radar WebKit Bug Importer

Comment 1 2020-01-06 10:03:21 PST

<rdar://problem/58345056>

Truitt Savell

Comment 2 2020-01-06 10:08:49 PST

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

Doug Kelly

Comment 3 2020-01-06 10:23:03 PST

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

Comment 4 2020-01-06 18:50:50 PST

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

Doug Kelly

Comment 5 2020-01-07 17:48:44 PST

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

Comment 6 2020-01-08 10:12:29 PST

Created attachment 387110 [details] Patch

Maciej Stachowiak

Comment 7 2020-01-08 13:55:03 PST

Comment on attachment 387110 [details] Patch r=me

Jon Lee

Comment 8 2020-01-10 11:55:28 PST

can we check this in?

WebKit Commit Bot

Comment 9 2020-01-16 13:32:04 PST

Comment on attachment 387110 [details] Patch Clearing flags on attachment: 387110 Committed r254705: <https://trac.webkit.org/changeset/254705>

WebKit Commit Bot

Comment 10 2020-01-16 13:32:05 PST

All reviewed patches have been landed. Closing bug.

Note You need to log in before you can comment on or make changes to this bug.