Bug 15214

Summary: SVG link with target="_top" opens new window
Product: WebKit Reporter: Jeff Schiller <jeffschiller>
Component: SVGAssignee: Nobody <webkit-unassigned>
Status: RESOLVED FIXED    
Severity: Normal Keywords: HasReduction
Priority: P2    
Version: 523.x (Safari 3)   
Hardware: PC   
OS: Windows XP   
URL: http://www.codedread.com/test-safari-link.svg
Attachments:
Description Flags
Simple test case
none
a fix rwlbuis: review+

Description Jeff Schiller 2007-09-14 08:44:52 PDT
Safari 3.0.3 Beta

Clicking the link opens a new window (target="_top" should replace the top-most frame as Firefox, Opera and IE+ASV3 do.
Comment 1 Jeff Schiller 2007-09-14 08:45:35 PDT
Created attachment 16292 [details]
Simple test case
Comment 2 David Kilzer (:ddkilzer) 2007-09-18 08:10:37 PDT
Confirmed difference in behavior between Firefox 2.0.0.6 and Opera 9.22 versus a local debug build of WebKit r25545 with Safari 3 Public Beta v. 3.0.3 (522.12.1) on Mac OS X 10.4.10 (8R218).

Comment 3 Eric Seidel (no email) 2007-09-25 07:22:23 PDT
As part of fixing this, I made sure we passed all the SVG WG test cases.  This one is relevant to this code change, but the test seems wrong:
http://www.w3.org/Graphics/SVG/Test/20061213/htmlEmbedHarness/full-linking-a-07-t.html

specifically:
The bottom-most (blue) arrows links to the same external SVG file, but with xlink:show="replace". Both the left and the right blue arrows should produce the image of the linkingToc-t.svg in a new frame.

The left arrow should not open a new window, as xlink:replace behaves like target="_self" according to the xlink spec.
http://www.w3.org/TR/xlink/#show-att
Comment 4 Eric Seidel (no email) 2007-09-25 07:23:51 PDT
One remaining question is "who wins?" when you have conflicting attributes, such as target="_self" and xlink:show="new".  Currently I've coded things so that target always wins.
Comment 5 Eric Seidel (no email) 2007-09-25 07:28:18 PDT
Created attachment 16386 [details]
a fix
Comment 6 Rob Buis 2007-09-25 07:58:03 PDT
Comment on attachment 16386 [details]
a fix

r=me
Comment 7 Eric Seidel (no email) 2007-09-25 13:15:36 PDT
Landed on the feature branch as r25729.

You should be able to verify once the build shows up at:
http://nightly.webkit.org/builds/mac-feature-branch/1