Bug 81029 – css3/filters/custom/custom-filter-property-computed-style.html passes in browser, fails in DRT

• Component: CSS
• Status: RESOLVED INVALID

[Tim Horton]

Running this test in DRT, I see:

Custom with vertex shader : custom(url(vertex.shader))
FAIL removeBaseURL(filterStyle.getPropertyValue('-webkit-filter')) should be custom(url(vertex.shader) none, 1 1 filter-box). Was none.
FAIL jsWrapperClass(filterRule) should be CSSValueList. Was CSSPrimitiveValue.
FAIL jsWrapperClass(filterRule.__proto__) should be CSSValueListPrototype. Was CSSPrimitiveValuePrototype.
FAIL jsWrapperClass(filterRule.constructor) should be CSSValueListConstructor. Was CSSPrimitiveValueConstructor.
FAIL filterRule.length should be 1 (of type number). Was undefined (of type undefined).
FAIL subRule.operationType should be 12. Threw exception TypeError: 'undefined' is not an object (evaluating 'subRule.operationType')
FAIL removeBaseURL(subRule.cssText) should be custom(url(vertex.shader) none, 1 1 filter-box). Threw exception TypeError: 'undefined' is not an object (evaluating 'subRule.cssText')

etc.

Running this test in Safari, I see:

Custom with vertex shader : custom(url(vertex.shader))
PASS removeBaseURL(filterStyle.getPropertyValue('-webkit-filter')) is 'custom(url(vertex.shader) none, 1 1 filter-box)'
PASS jsWrapperClass(filterRule) is 'CSSValueList'
PASS jsWrapperClass(filterRule.__proto__) is 'CSSValueListPrototype'
PASS jsWrapperClass(filterRule.constructor) is 'CSSValueListConstructor'
PASS filterRule.length is 1
PASS subRule.operationType is WebKitCSSFilterValue.CSS_FILTER_CUSTOM
PASS removeBaseURL(subRule.cssText) is 'custom(url(vertex.shader) none, 1 1 filter-box)'

etc.

[Radar WebKit Bug Importer]

<rdar://problem/11040673>

[Tim Horton]

Created attachment 131708 [details]
patch to land failing expectations

[Tim Horton]

TestFailures suggests it might be flaky; I've never seen it pass.

[Tim Horton]

Landed failing results in http://trac.webkit.org/changeset/110612

[Jessie Berlin]

This either passes under WKTR, or is just flakey:

http://build.webkit.org/results/Lion%20Intel%20Debug%20(WebKit2%20Tests)/r110645%20(4872)/css3/filters/custom/custom-filter-property-computed-style-pretty-diff.html

[Tim Horton]

(In reply to comment #5)
> This either passes under WKTR, or is just flakey:
> 
> http://build.webkit.org/results/Lion%20Intel%20Debug%20(WebKit2%20Tests)/r110645%20(4872)/css3/filters/custom/custom-filter-property-computed-style-pretty-diff.html

Passing with WKTR and not DRT would vaguely match the fact that it passes in Safari (with WK2) and not DRT.

[Tim Horton]

(In reply to comment #6)
> (In reply to comment #5)
> > This either passes under WKTR, or is just flakey:
> > 
> > http://build.webkit.org/results/Lion%20Intel%20Debug%20(WebKit2%20Tests)/r110645%20(4872)/css3/filters/custom/custom-filter-property-computed-style-pretty-diff.html
> 
> Passing with WKTR and not DRT would vaguely match the fact that it passes in Safari (with WK2) and not DRT.

Nope, it passes in a WK1 Safari window too.

[Jessie Berlin]

Landed expected passing results for mac-wk2 in http://trac.webkit.org/changeset/110855.