Hi Eran:
Just a few clarifications...
Eran Chinthaka wrote:
Glen Daniels wrote:
I'm not worried about the DOM dependency either. Schema processing is
typically (not always, but usually) going to be a design time thing
where squeaking every bit of performance out isn't as important.
Its not about performance I'm talking here. Think of a situation where
you have the whole WSDL in an OM tree. Do you want to get the schema
element from it, build DOM and pump it to XmlSchema. I hope you won't
say yes, if you care at least a lil bit about elegance.
Of course ideally not, but see the other thread (DOM and OM) also -
don't we have to pump a LLOM through a LLOM->DOM converter if we want to
use DOOM?
I *really* don't want to get into another whole "Yet Another XML Factory
Abstraction" thing unless we truly need it. And DOM is a pretty
convenient
Since when you started to say DOM has a convenient API? I for one hates it.
I can remember you were hating DOM apis a lot, during the first Axis2
f2f and was very much impressed with JDom API.
To be clear, I didn't mean to say DOM is a *good* API. :) What I meant
by "convenient" is that it's standard and it's available everywhere.
XmlSchema seems like a library with a lot of potential usage both inside
and outside Apache - and since it doesn't have the huge need for
performance and integration with Axis that our WS-* extensions do, I
think it's actually probably a good thing that it doesn't depend on OM.
That doesn't mean I don't want it to work well with OM, just that I'm
wary about a dependency.
(and standard, as Ajith points out) API for walking around an
XML tree with lots of cross-references.
Anyway, I'm ok to go ahead with DOOM as Sanjiva suggested.
+1, but see other thread for question about LLOM -> DOM performance.
--Glen
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
