Bug 33374 – REGRESSION(r52976): svg/W3C-SVG-1.1/filters-conv-01-f.svg fails on Windows

• Component: SVG
• Status: NEW

[Csaba Osztrogonác]

http://build.webkit.org/results/Windows%20Release%20%28Tests%29/r52980%20%287717%29/results.html
http://build.webkit.org/results/Windows%20Debug%20%28Tests%29/r52980%20%288282%29/results.html

After http://trac.webkit.org/changeset/52976 MAC's generate good output for svg/W3C-SVG-1.1/filters-conv-01-f.svg 
(without DRT sideeffect), so Eric updated the expected file by http://trac.webkit.org/changeset/52980 ,
but unfortunately Windows has still sideeffect.

[Csaba Osztrogonác]

svg/W3C-SVG-1.1/filters-conv-01-f.svg skipped temporarily until fix by http://trac.webkit.org/changeset/52990

[Eric Seidel (no email)]

Adam Roben just confirmed for me on his windows build that running svg/W3C-SVG-1.1/filters-conv-01-f.svg by itself on windows produces 117, a result which does not match the mac when run stand-alone, but does match the mac when run after http/tests/uri/escaped-entity.html as mentioned in bug 32294.

This is not what I expected.  This suggests that the mac checkin is perhaps wrong and that we don't understand this issue.

I'm leaning towards rolling out the run-webkit-tests change and skipped changes.

[Alexey Proskuryakov]

This will likely be handled, or at least made much simpler, by bug 33400.

[Alexey Proskuryakov]

Actually, not. This test just needs platform-specific results - kerning doesn't work with this font on Windows. Mitz tells me we don't need a new bug about that.

[Nikolas Zimmermann]

Changed component to SVG, so it shows up in my all-svg-bugs search.

[Stephen Chenney]

This is Chromium issue http://code.google.com/p/chromium/issues/detail?id=32682

[Nikolas Zimmermann]

Does this still fail on windows? Did anyone investigate with a win machine whats going on there?

[Karl Dubost]

Created attachment 478895 [details]
rendering in safari, firefox, chrome

Is this still revelant for the windows platform?
on macOS there is no issue.