Hi Yin,

The goal for XPointer isn't to implement all of XPath, just a subset which
is streamable so that it can work with the XInclude processor which is
streaming. I'm believe the support in Xalan requires that the document has
been loaded into a DTM or DOM and so wouldn't be suitable for this purpose.

I think extending the existing XPath support in Xerces would be a better
idea.

Thanks.

Michael Glavassevich
XML Parser Development
IBM Toronto Lab
E-mail: [email protected]
E-mail: [email protected]

Yin Lei <[email protected]> wrote on 03/16/2010 03:42:07 AM:

> Hi,
> Thank you for your reply. Now i want to implement XPointer xpointer()
>   which was one important part of XInclude for Xerces, so, i have to
> implements all the XPath 1.0 support for Xerces.For me,there are two
> ways to go:
>
> 1. improve existing XPath support in Xerces;
>
> 2.implement a brand new XPath support for Xerces
>
> May be proposal 1 is better, but for me, i am so familiar with
> Xalan,and Xalan implements all the XPath 1.0 already, if i transfer
> Xalan XPath 1.0 implemention to Xerces (may be need some
> change),this transfer job will be faster and much more convenient.
>
> So,You guys are experienced Xerces developers and open source
> pionneer,i need more professional advises from you. Thank you!
> 2010/3/16 Michael Glavassevich <[email protected]>
> Hi,
>
> The existing XPath class in Xerces was developed to support the
> subset of XPath required for XML Schema identity constraints [1].
> This is what it's used for within Xerces. It was never intended to
> implement all of XPath.
>
> Thanks.
>
> [1] http://www.w3.org/TR/xmlschema-1/#coss-identity-constraint
>
> Michael Glavassevich
> XML Parser Development
> IBM Toronto Lab
> E-mail: [email protected]
> E-mail: [email protected]
>
> xunlong gui <[email protected]> wrote on 03/15/2010 11:35:38 PM:
>
> > hi guys,
>
> >
> > Xerces's XPath class seems nearly useless, it can only work with
> > simple xpath like "first/second" or ".//first/secnod", if i use it
> > to parse complex xpath such as "first/seco...@third='fourth']", it
> > will throw exception. So, i want to check if it is unfinished?If i
> > want to use Xerces to parse my XPath expression, should I implement
> > XPath class by myself? Thank you

Reply via email to