Bug 14927

Summary: use element in SVG does not render after parent is displayed off and on
Product: WebKit Reporter: Jean-Pierre Fiset <bugzilla>
Component: SVGAssignee: Nobody <webkit-unassigned>
Status: RESOLVED FIXED    
Severity: Normal CC: ahayes, bugzilla
Priority: P2 Keywords: HasReduction
Version: 523.x (Safari 3)   
Hardware: Mac   
OS: All   
Attachments:
Description Flags
Simple Test Case
none
First attempt oliver: review+

Description Jean-Pierre Fiset 2007-08-10 08:02:27 PDT
In SVG, if the parent element of a use element is hidden (via display="none") and then made visible again (via display="inline"), the content of the use element fails to render. A similar test case, where the actual elements are used instead of <use> elements, behaves correctly.

This problem appears on Safari 3.0.3 and on WebKit night r24944. This problem was also reproduced on MacIntel and Windows XP.
Comment 1 Jean-Pierre Fiset 2007-08-10 08:03:58 PDT
Created attachment 15909 [details]
Simple Test Case
Comment 2 Rob Buis 2007-08-16 13:04:09 PDT
Created attachment 16000 [details]
First attempt

This should fix it :)
Cheers,

Rob.
Comment 3 Oliver Hunt 2007-08-16 23:47:05 PDT
Comment on attachment 16000 [details]
First attempt

Could you use a rect instead of circle in the test case?

otherwise r=me
Comment 4 Rob Buis 2007-08-16 23:53:06 PDT
(In reply to comment #1)
> Created an attachment (id=15909) [edit]
> Simple Test Case

I forgot to say, these test cases are excellent, keep up the good reports!
Cheers,

Rob.
Comment 5 Rob Buis 2007-08-16 23:55:12 PDT
Landed in r25123.
Comment 6 Jean-Pierre Fiset 2007-08-17 05:38:38 PDT
(In reply to comment #3)
> (From update of attachment 16000 [details] [edit])
> Could you use a rect instead of circle in the test case?
> 
> otherwise r=me
> 

Would you like me to submit a new test case or are you stating that, in the future, I should submit test cases with rectangles as opposed to circles?

What are the ramifications of using <rect> over <circle> in test cases?
Comment 7 Jean-Pierre Fiset 2007-08-17 12:10:26 PDT
(In reply to comment #4)
> (In reply to comment #1)
> > Created an attachment (id=15909) [edit]
> > Simple Test Case
> 
> I forgot to say, these test cases are excellent, keep up the good reports!
> Cheers,
> 
> Rob.
> 

Some credit is due to the organization that pays me to develop an SVG-enabled interactive atlas framework. See <http://nunaliit.org>

Thanks for noticing.

JP
Comment 8 Oliver Hunt 2007-08-18 02:38:25 PDT
Jean-Pierre, it just makes the pixel tests a bit more robust across platforms, my comment was regarding the automated test that Rob included in his patch :D
Comment 9 Oliver Hunt 2007-08-18 02:39:29 PDT
Jean-Pierre, so your test was perfect for the purpose of finding and fixing the source of the bug :D
Comment 10 Rob Buis 2007-08-19 01:26:47 PDT
Landed in r25123.