[Webkit-unassigned] [Bug 221635] Crash in WebCore::RenderThemeMac::paintSearchFieldResultsDecorationPart() from large scale

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=221635

--- Comment #6 from Julian Gonzalez <julian_a_gonzalez at apple.com> ---
(In reply to Tim Horton from comment #5)
> Comment on attachment 419774 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=419774&action=review
> 
> >>>> LayoutTests/platform/mac/editing/style/large-scale-crash-expected.txt:1
> >>>> +* { -webkit-appearance: searchfield-results-decoration; transform: scale(100); writing-mode: vertical-rl; } if (window.testRunner) testRunner.dumpAsText(); onload = () => { document.styleSheets[0].insertRule(`* { all: initial; }`); }; This tests that we do not hit an assertion while rendering the page. (which prints out its style). PASS
> >>> 
> >>> Is this all scrunched on one line for a reason? Usually we format -expected HTML nicely like we do the real test.
> >> 
> >> It might be the side effect of CSS rule here...
> > 
> > OH, I misread, I thought it was a expected.html, not .txt. Sorry :)
> 
> (You could probably make the expected.txt look a little less ridiculous by
> only applying the style to a single element, instead of `*`, but obviously
> would want to make sure it still triggered the crash if you did that)

I tried this a few times (isolating the rule to one element) but I wasn't able to keep the crashing property of the test that way :(

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20210210/38f017cc/attachment.htm>