Case of attribute names in HTML documents for L1 Core tests (DOMTS Bug 236)

http://lists.w3.org/Archives/Public/www-dom-ts/2003Jun/0011.html started a
fairly lengthy thread, however some of my statements were based on an improper
understanding of the problem.

To summarize my current understanding of the issue:

At issue are the L1 Core tests hc_attrname, hc_documentcreatedattribute,
hc_elementgetattributenode and hc_namednodemapgetnameditem.  Unlike originally
supposed, none of the tests involved case expectations for tag names.  

Before the proposed modifications, these tests expected attribute names in all
documents to be returned in lower case which matched the behavior of Internet
Explorer, Mozilla and Safari.

Konqueror 3.1.1 and Opera 7 implemented the behavior specified in the L2 HTML
(http://www.w3.org/TR/DOM-Level-2-HTML/html.html#ID-5353782642) and returned the
attribute names in upper case for HTML documents resulting in these
implementations failing these four tests.

The L1 Spec does make an explicit statement that in HTML documents tag names
should be returned as upper case, but does not appear to make a statement about
attribute names.

It seems, on this issue, that all five of the mentioned browsers
behavior are be conformant with DOM L1, but only Konqueror and Opera are conformant
with DOM L2.

The attached patch modifies the existing L1 Core tests so that case-insensitive
comparisons are made on attribute names for HTML documents and lowercase is
expected for XML, XHTML and SVG documents.

In addition, versions of the tests that expect upper-case attribute names in
HTML documents were added to L2 Core along with the supporting test documents.

The WG may want to consider an L2 HTML errata that would make the existing
IE/Mozilla/Safari behavior of lower-casing attribute names normative.

p.s. - I've committed the patch and have marked the defect as FIXED.  However as with
all changes, the defect is not closed until reviewed and should not be seen as final.


http://www.w3.org/Bugs/Public/show_bug.cgi?id=236

Received on Friday, 27 June 2003 03:02:07 UTC