Bug 106126

Summary: SVG tests have weird rounding issues, flaking between 0.00 and -0.00
Product: WebKit Reporter: Terry Anderson <tdanderson>
Component: SVGAssignee: Nobody <webkit-unassigned>
Status: NEW ---    
Severity: Normal CC: dglazkov, fmalita, macpherson, pdr, schenney, wjmaclean, zimmermann
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: Unspecified   
OS: Unspecified   

Description Terry Anderson 2013-01-04 13:04:18 PST 
The following layout tests are flaky on XP:

svg/custom/feComponentTransfer-Table.svg (Pass/ImageOnlyFailure)
svg/custom/embedding-external-svgs.xhtml (Pass/Failure)

Probable cause: Unknown
Comment 1 Terry Anderson 2013-01-04 13:16:58 PST 
Updated TestExpectations: http://trac.webkit.org/changeset/138844
Comment 2 Dimitri Glazkov (Google) 2013-01-07 10:10:08 PST 
Also:
svg/overflow/overflow-on-inner-svg-element.svg
svg/dom/css-transforms.xhtml
Comment 3 Philip Rogers 2013-01-07 10:16:54 PST 
Adding some more SVG friends to this; any idea what might be up here?
Comment 4 W. James MacLean 2013-01-07 11:42:10 PST 
(In reply to comment #3)
> Adding some more SVG friends to this; any idea what might be up here?

Is it perhaps due to Skia doing bit-wise compares during matrix equality tests, meaning -0 and 0 fail to be considered equal?
Comment 5 Philip Rogers 2013-01-21 21:15:22 PST 
*** Bug 106387 has been marked as a duplicate of this bug. ***