401112 - Filter invalidation problems

Status: RESOLVED FIXED

(Core :: SVG, defect, P3)

Description

[tor]

There are a number of cases where invalidation doesn't account for the space occupied by the filter.

Comment 1

[tor]

Attachment: test transformation change - click on rect, should move with no artifacts

Comment 2

[tor]

Attachment: test of filtered object removal - click on rect, should vanish without artifacts

Comment 3

[tor]

Attachment: test of visibility change - click on rect, should vanish without artifacts

Comment 4

[tor]

Attachment: test of adding filter to object - click on rect, should switch to blurred black rect

Comment 5

[tor]

Attachment: test of removing filter from object - click on blurred black rect, should be replaced by red rect

Comment 6

[tor]

Attachment: fix filter invalidation problems

Comment 8

[Jonathan Watt [:jwatt]]

Comment on attachment 286173 [details] [diff] [review]
fix filter invalidation problems

r=jwatt

>   /*
>    * Figures out the worst case invalidation area for a frame, taking
>-   * into account filters.  Empty return if no filter in the hierarchy.
>+   * into account filters.

s/into account filters/filters into account/  ?

Comment 9

[tor]

Checked in with suggested comment rewording.  Testcases not checked in as reftests because we haven't figured out how to use the reftest framework to test paint invalidation issues.

Comment 10

[Paul Stone]

It looks like the last testcase, 'test of removing filter from object - click on blurred black rect, should be replaced by black rect' has regressed. 

When I click the blurred rect, I see a red rect with a blurred black border around it.