Bug 104722 - -webkit-svg-shadow radius changes don't cause children's boundaries to be recomputed
Summary: -webkit-svg-shadow radius changes don't cause children's boundaries to be rec...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: SVG (show other bugs)
Version: 528+ (Nightly build)
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Tim Horton
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2012-12-11 14:48 PST by Tim Horton
Modified: 2012-12-11 17:16 PST (History)
6 users (show)

See Also:

Attachments
repro (928 bytes, text/html)
2012-12-11 14:48 PST, Tim Horton 		no flags Details
patch (6.03 KB, patch)
2012-12-11 16:24 PST, Tim Horton 		no flags Details | Formatted Diff | Diff
patch (8.25 KB, patch)
2012-12-11 17:07 PST, Tim Horton 		simon.fraser: review+
Details | Formatted Diff | Diff
Show Obsolete (1) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Tim Horton 2012-12-11 14:48:34 PST 
Created attachment 178884 [details]
repro

There's a note in SVGRenderStyle::diff():

// Shadow changes require relayouts, as they affect the repaint rects.

And it then returns StyleDifferenceLayout.

Then, RenderSVGModelObject::styleWillChange() (and friends) marks the boundaries dirty when it gets that StyleDifferenceLayout.

But! We also need to our children's boundaries as dirty, since the parent's shadow is baked into them.

I have a patch to recursively dirty children's boundaries if the old and/or new style has a -webkit-svg-shadow which fixes the attached testcase. Not sure if that hammer is too big, though.

<rdar://problem/12821080>
Comment 1 Tim Horton 2012-12-11 16:24:59 PST 
Created attachment 178912 [details]
patch

Next, to write a test.
Comment 2 Tim Horton 2012-12-11 17:07:24 PST 
Created attachment 178924 [details]
patch
Comment 3 Tim Horton 2012-12-11 17:16:15 PST 
http://trac.webkit.org/changeset/137393