RESOLVED INVALID 14654
DOM::previousSibling returns incorrect node 
https://bugs.webkit.org/show_bug.cgi?id=14654
Summary DOM::previousSibling returns incorrect node
Marek Sykora
Reported 2007-07-18 02:02:01 PDT
Testcase: <html> <body> <table> <tr><td></td> <td><a onclick="javascript:alert(this.parentNode.previousSibling);" href="">run testcase</a></td></tr> </table> </body> </html> Safari 3.0 returns [object Text], Safari 3.0.2 returns [object HTMLTableCellElement]. Safari 2.0.4 returns [object TD].
Attachments
testcase (172 bytes, text/html)
2007-07-18 02:02 PDT, Marek Sykora 		no flags
Details
View All Add attachment proposed patch, testcase, etc.
Marek Sykora
Comment 1 2007-07-18 02:02:45 PDT
Created attachment 15559 [details] testcase
Alexey Proskuryakov
Comment 2 2007-07-18 10:10:28 PDT
I'm getting identical results from the latest nightly and from Firefox (and that's [object Text]). So, this looks like an intentional bugfix. May I ask you to verify your results with Safari beta 3.0.2? It's quite surprising that there was a regression in this area. Also, you may want to use Hixie's Live DOM Viewer <http://software.hixie.ch/utilities/js/live-dom-viewer/> to visualize DOM trees created by browsers.
mitz
Comment 3 2007-07-18 10:21:30 PDT
The change to allow whitespace-only text nodes in tables was <http://trac.webkit.org/projects/webkit/changeset/21720>.
Marek Sykora
Comment 4 2007-07-19 08:17:30 PDT
(In reply to comment #2) > May I ask you to verify your results with Safari beta 3.0.2? Verified with Safari beta 3.0.2 (522.12) on Tiger and 3.0.2 (522.11.3) on WinXP -- both return [object HTMLTableCellElement].
Alexey Proskuryakov
Comment 5 2007-07-19 08:54:12 PDT
The behavior of current nightlies is the intended one, and it matches Firefox. So, I'm closing this bug. As Mitz points out, this fix was made too late to make it into Safari 3 betas.
Note You need to log in before you can comment on or make changes to this bug.