RESOLVED FIXED168341
LayoutTest fast/multicol/column-span-parent-continuation-crash.html is a flaky timeout
https://bugs.webkit.org/show_bug.cgi?id=168341
Summary LayoutTest fast/multicol/column-span-parent-continuation-crash.html is a flak...
Attachments
proposed fix (1.35 KB, patch)
2017-02-18 14:31 PST, Alexey Proskuryakov
no flags
proposed fix (2.73 KB, patch)
2017-02-18 14:36 PST, Alexey Proskuryakov
no flags
Ryan Haddad
Comment 1 2017-02-17 14:18:38 PST
From the diff: CONSOLE MESSAGE: line 20: TypeError: null is not an object (evaluating 'child.parentNode')
Alexey Proskuryakov
Comment 2 2017-02-18 14:31:21 PST
Created attachment 302057 [details] proposed fix
Alexey Proskuryakov
Comment 3 2017-02-18 14:36:45 PST
Created attachment 302059 [details] proposed fix The patch didn't apply because the file had CRLF line endings. This one should work I hope.
Daniel Bates
Comment 4 2017-02-20 13:55:55 PST
Comment on attachment 302059 [details] proposed fix View in context: https://bugs.webkit.org/attachment.cgi?id=302059&action=review > LayoutTests/fast/multicol/column-span-parent-continuation-crash.html:6 > +<style> Is the test intentionally needing the <style> here? Otherwise this markup is invalid by the HTML standard as <style> is expected to be a child of <head>. > LayoutTests/fast/multicol/column-span-parent-continuation-crash.html:20 > + child = document.getElementById('test'); I know you are just moving code. I do not not see the need for |child| to be a global variable. I would add the keywords "var " or "let" to the left hand side of this expression. > LayoutTests/fast/multicol/column-span-parent-continuation-crash.html:22 > + child = document.getElementById('anything'); This file alternatives between using single and double quoted sting literals. We should take this opportunity to pick one style and use throughout this file for consistently.
Alexey Proskuryakov
Comment 5 2017-02-20 14:04:43 PST
Comment on attachment 302059 [details] proposed fix I'm not eager to change more than absolutely needed in this file, as it's a regression test for something that was timing dependent, and I can't confirm that the test is still effective after the changes. As for style fixes, my preference is to not enforce style rules in tests, as minor variations have been known to uncover unexpected bugs that we don't explicitly test for.
WebKit Commit Bot
Comment 6 2017-02-20 14:30:46 PST
Comment on attachment 302059 [details] proposed fix Clearing flags on attachment: 302059 Committed r212678: <http://trac.webkit.org/changeset/212678>
WebKit Commit Bot
Comment 7 2017-02-20 14:30:50 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.