On Sun, 2 Sep 2001, robert burrell donkin wrote:
> Date: Sun, 2 Sep 2001 21:57:13 +0100
> From: robert burrell donkin <[EMAIL PROTECTED]>
> Reply-To: [EMAIL PROTECTED]
> To: [EMAIL PROTECTED]
> Subject: (Digester) Namespace logic
>
> (you might have guessed that i'm working on test cases...)
>
> i'm a little confused about the (intended) matching rules namespace logic.
> i'd say that the current behaviour is probably not correct - but maybe
> that's what's intended :)
>
> should a rule with a null namespace be matched when a non-null namespace
> is passed in? in other words does a null namespace for a rule imply that
> it matches only elements without a namespace or does a null namespace
> match any namespace. (of course, this is assuming that the rule would
> qualify otherwise)
>
My thinking was that a Rule with no associated namespace URI must be
matched literally against the qualified Name (i.e. including any prefix).
In other words, an element
<foo:bar ... />
would be matched by pattern "foo:bar" in such a Rule. This also is -- or
at least should be :-) -- backwards compatible with the way pattern
matching worked before namespace awareness was introduced.
>
> if i know which is correct i can prepare documentation patches and a code
> patch (if that's needed).
> (this is easy for me to do at the moment since i'm trying to prepare a
> separate (and extended) test case RulesBase which can be extended to act
> as a basis for testing extensions of RulesBase)
>
Thanks ... I will be happy to check in the results of your efforts.
> - robert
>
Craig
