371552 - setAttributeNode behavior when node was created with createAttributeNS

Status: RESOLVED WORKSFORME

(Core :: DOM: Core & HTML, defect)

Description

[Robert Sayre]

Attachment: sample xml document

Comment 1

[Robert Sayre]

Attachment: test case

Comment 2

[Robert Sayre]

works in Opera 9

Comment 3

[Robert Sayre]

works in WebKit trunk, but not in released Safari

Comment 4

[Robert Sayre]

This was discovered while adding dom-level2-core to the mochitest run. I tested Fx 1.0.8, Fx 1.5, and Fx 2.0.0.2, and none of them passed, so it's not a regression.

Comment 5

[Alex Vincent [:WeirdAl]]

I believe this is INVALID.  You're mixing DOM Level 1 (non-namespace-aware methods) with DOM Level 2 (namespace-aware methods).

    attribute = doc.createAttributeNS("http://www.w3.org/DOM","domestic");
    newAttribute = element.setAttributeNode(attribute);

That should probably be setAttributeNodeNS.
http://www.w3.org/TR/2004/REC-DOM-Level-3-Core-20040407/core.html#ID-ElSetAtNodeNS

Comment 6

[Alex Vincent [:WeirdAl]]

Attachment: testcase using setAttributeNodeNS

Comment 7

[Robert Sayre]

(In reply to comment #5)
> I believe this is INVALID.  You're mixing DOM Level 1 (non-namespace-aware
> methods) with DOM Level 2 (namespace-aware methods).

Nice catch. But it wasn't me! :) It was the W3C test suite. I'll note that the spec says "don't do that", but it doesn't specify what should happen if you do.

Is this a bug in Opera, WebKit, and the W3C test suite? Let's CC Anne and see what he thinks.

Comment 8

[Robert Sayre]

Filed WebKit bug:

http://bugs.webkit.org/show_bug.cgi?id=12881

Comment 9

[Anne (:annevk)]

I think having setAttributeNodeNS doesn't make much sense... And if it needs to be there for compatibility it should just map to setAttributeNode.

Comment 10

[Arkadiusz Michalski (Spirit)]

After:
https://github.com/whatwg/dom/issues/117
https://bugzilla.mozilla.org/show_bug.cgi?id=1227458 

this bug should not be closed too?