W3C home > Mailing lists > Public > www-dom@w3.org > January to March 2002

Re: getElementsByTagNameNS("*",) should be clarified in DOM L3 Core

From: David Brownell <david-b@pacbell.net>
Date: Wed, 06 Feb 2002 17:45:27 -0800
To: Curt Arnold <curta_ontheroad@yahoo.com>, www-dom@w3.org
Message-id: <04ce01c1af79$1ec95860$6800000a@brownell.org>
> Most implementations interpret '*' to match all
> elements with the specified local name regardless if
> they are namespace qualified or not.

Which IMO is clearly nonconforming behavior on the
part of any such implementations.

>      GNUJAXP has
> interpreted "matches all namespaces" as meaning that
> getElementsByTagName("*","whatever") should not match
> elements that are not namespace qualified.

That is, the issue is:  Does the text in the spec that 'The
special value "*" matches all namespaces" mean what
it says?

It's completely clear from the namespace spec that
there do exist names which are not in a namespace.
For example, the element in the document "<foo/>"
is in no namespace.

> It would be good to resolve the issue (assuming to
> favor the much more common interpretation of matching
> both qualified and unqualified elements) and clarify
> the text in the DOM L3 text instead of issuing a
> errata after DOM L3.

IMO what Curt's proposing is to bless an implementation
bug by changing the behavior of the specification.  While
W3C can of course do that, I'd not encourage it.

- Dave
Received on Wednesday, 6 February 2002 20:47:22 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 20 October 2015 10:46:09 UTC