Matt Sergeant wrote:
And the "value" attribute for xsp:attribute.
Funny, I thought about this originally, too, but there's no such animal. XSP's 'attribute' element has the same behavior as XSLT's[1]. That is, the value for the newly created attribute may appear only as the text contents of the 'attribute' element and there is no 'value' (or similar) attribute.
Or... where you suggesting that we add a 'value' attribute? (FWIW, I say leave it as it is now and keep conceptual compatability w/ XSLT).
But I do worry slightly that this is the thin end of the wedge... So we should either make sure the AVT function is fully exposed so other taglibs can (ab)use it, or decide now it should remain private.
Hmmm... I can certianly see where providing access to the AVT behavior could be helpful/useful to taglibrarians. OTOH, the more we bolt on to the current XSP.pm, the harder we make it to refactor later without breaking back-compat (and it *will* be refactored, my friend, oh, yes. and with extreme prejudice :-). I cast a tenative vote for "make it public". Other thoughts?
-kip
[1] http://www.w3.org/TR/xslt#creating-attributes
