On 04/12/2010 03:25 AM, Henry Zongaro wrote:
Hi, David.
Thanks for your responses. I just had one point on which I wanted to
get clarification. I asked:
> 1. What is the form of the input accepted by PsychoPath? Is it
> restricted to taking a DOM (Level 2?) input tree - whose nodes might
> implement the ItemPSVI interface of the Xerces XML Schema API?
and you reponded:
David Carver <d_a_car...@yahoo.com> wrote on 04/11/2010 09:48:22 AM:
> For XML Schema Awarness which PsychoPath supports, it requires
> Xerces-J. For all other DOM's that are given to it, it defaults to
> just the XPath core functionality. It currently requires a DOM
> Level 3 implementation, which the vast majority of DOM's out there
> now support.
So does that mean input must always be in the form of a DOM tree?
As of right now, yes, it processes a DOM tree. I've run it against,
both the Xerces-J DOM which is it's reference it uses for tests, and the
eclipse WTP DOM. Both seem to work. If given another DOM, I would
expect it to work as well.
However, would love to get some resources working on bringing in other
structures and working with them. As Jesper said currently we don't do
a SAX or StaX based walk through. However, we haven't seen any problems
with performance yet, with the DOM structure we are using.
One thing to note about PsychoPath's history is that it originally
started out as a College research project, and sat on Sourceforge for
about 4 or 5 years. It's primary goal was to first and foremost
provide an open-source XPath Schema Aware engine. I talked to the
original authors and manage to bring it over to eclipse. We spent the
last 6 to 8 months just bringing it up to compliance with the changes
that occured from those early drafts and the current W3C Test Suite.
So we are at a point where we are about 99.9% compliant with the
specification, now it's time to look at enhancing and allowing the
engine to read and accept different inputs.
Again, patches and enhancements would be welcome to support other input
mechanisms.
Dave
