You guys may be able to speed some of this up by leveraging the PsychoPath XPath 2.0 processor that is at Eclipse.
It is already being used by Xerces-J for XSD 1.1 Assertion support. Why re-invent the wheel if you can leverage something that already exists. http://wiki.eclipse.org/PsychoPathXPathProcessor PsychoPath already is XML Schema Aware, leverages Xerces-J 2 and passes about 99.9% of the Xpath 2.0 test suite. Dave On 03/26/2010 09:25 PM, ustbcoder wrote: > Hi Michael, > Starting with utility parts is regular way to implement a new > recommendation specification. But i think, functions is a very > practical part. If i finish it first,may be some guys can go it into > service. Months ago,i was working with a XSLT project based on Xalan, > i need some XPath 2.0 functions,but i have to finished them by Xalan > extension mechanism one by one, it is a boring work, so ,i think, if > some one implements these XPath 2.0 functions already, how beautiful > the world is! > But i want to say that implementing functions is just a pre-handler > job for Xalan's XSLT 2.0 implemention job, it must be done sooner or > later, so i have started it,it won't go wrong. Then start with which > parts,schema change,sequences change or some utility parts, i am not > very sure, so,i need your help,all of you can give me some advises. > 2010-03-27 > ------------------------------------------------------------------------ > ustbcoder > ------------------------------------------------------------------------ > *发件人：* Michael Ludwig > *发送时间：* 2010-03-27 04:39:44 > *收件人：* xalan-j-us...@xml.apache.org > *抄送：* > *主题：* Re: Re: implement XPath 2.0 function for xalan > xunlong gui schrieb am 26.03.2010 um 23:27:15 (+0800): > > 1.implements all the XSLT 2.0 functions which do not need sequences > > and XSLT 2.0 schema support > Leaving aside (W3C XML) schema awareness, the new type system of XPath > 2.0 seems the most fundamental change from 1.0 to 2.0, so I wonder > whether focussing on the utility parts would be a good start for any > effort to go to 2.0? > -- > Michael Ludwig
