Re: [jira] Commented: (AXIS2-249) Exessive number & size of files generated from WSDL

[Davanum Srinivas <da...@gmail.com>]

Yes. We imply that we should have our own databinding  :)

-- dims

On 9/30/05, Sanjiva Weerawarana <sa...@opensource.lk> wrote:
> On Fri, 2005-09-30 at 13:47 -0400, Davanum Srinivas wrote:
> > Me Neither. We need to fix it.
> >
> > -- dims
> >
> > On 9/30/05, Tom Jordahl <to...@macromedia.com> wrote:
> > > Simon Fell wrote:
> > > > It seems incredable to me, that you need to generate 20x more files
> > > than
> > > > Axis 1.1, yet its considered acceptible, Am I the only one that thinks
> > > > this is completely out of control ?
> > >
> > > Nope.
>
> Just to be clear .. we can't actually fix this behavior if the user uses
> XMLBeans data binding; its the nature of the beast. However, we must
> offer a better data binding system for users that never got married to
> XMLBeans and its way of doing things.
>
> Sanjiva.
>
>
>


--
Davanum Srinivas : http://wso2.com/ - Oxygenating The Web Service Platform

Re: [jira] Commented: (AXIS2-249) Exessive number & size of files generated from WSDL

[Steve Loughran <st...@apache.org>]

Davanum Srinivas wrote:
> Yes. We imply that we should have our own databinding  :)
> 
> -- dims

-0

I find that Xom and Xpath work quite nicely. The good thing about Xom is 
that it is already highly performance tuned, and you can provide your 
own factory to create a graph of your custom classes, not those of the 
base set. So whereas XmlBeans gives you a graph of java classes that 
map1:1 with the XSD-defined message, Xom gives you a graph of java 
classes that contain the operations that you yourself add in.

Xom is LGPL, to complicate things in Apache-land.

-steve