[webkit-dev] [webkit-changes] [52980] trunk/LayoutTests

[Eric Seidel]

I've reduced #3 down to the 7 tests which seem required to make it fail:
https://bugs.webkit.org/show_bug.cgi?id=33372

Given that currently it requires 7 tests to make it fail, it seems we
should continue to skip this one for now.

-eric

On Sat, Jan 9, 2010 at 10:45 AM, Alexey Proskuryakov <ap at webkit.org> wrote:
>
> 09.01.2010, в 06:37, Nikolas Zimmermann написал(а):
>
>>> 1. http/tests/uri/escaped-entity.html is disabled because it affects svg/W3C-SVG-1.1/filters-conv-01-f.svg
>>>
>>> The effect is slightly slightly different text rendering - kerning gets disabled in "Vertical blur" text. This is caused by trying to display a rare Unicode symbol in the former test, which magically affects another test much later.
>>>
>>> The way to resolve this for this is to make a standalone test case, file a bug, and then to remove the weird character from HTTP test output. It's not clear if this is a WebKit or CG bug.
>>>
>>> 2. svg/W3C-SVG-1.1/filters-conv-01-f.svg fails on Windows
>>>
>>> Closely related to the above. We probably have wrong results landed, in fact.
>>
>> #1 should be fixed by Dan's patch at https://bugs.webkit.org/show_bug.cgi?id=33400 which seems to be the root of the filters-conv-01-f.svg problem on Mac.
>
> Correct, that fixes the root cause. In the meanwhile, I changed escaped-entity.html to not dump the offending character - that's not what the test checks for anyway - and re-enabled the test.
>
> #2 just needs new results landed for Windows, which is tracked by <https://bugs.webkit.org/show_bug.cgi?id=33374>. Eric volunteered to find out what breaks #3, which is tracked by <https://bugs.webkit.org/show_bug.cgi?id=33372>. The difference is significant - it looks just like the wrong output that we used to produce.
>
> - WBR, Alexey Proskuryakov
>
> _______________________________________________
> webkit-dev mailing list
> webkit-dev at lists.webkit.org
> http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
>