At 15:36 05/11/2002 -0700, you wrote:
> Their XPath implementation is broken since it doesn't treat sibling text nodes
> as if they were merged. IIRC, there's a normalize method you can call on the
> document node to merge all the text nodes. Someone more familiar with MSXML
> will have to comment. Note that a Google search for msxml merge text nodes
> turned up
> http://msdn.microsoft.com/library/default.asp?url=/library/en-us/xmlsdk30/htm/xmmthnormaliz
e.asp
> pretty quickly. For MSXML questions, always check the docs at MS first! :)

Thanks. However, I am using their command-line interface (msxsl), which 
does not apply this function to the parsed input, it seems...

Perhaps I will switch to saxon one day... Just to go around all these 
problems ! However, I have a big performance issue (my XSL stylesheets are 
somewhat complex... and XML files quite big).

> Your entity definition is not the cause of the problem, but it will cause
> problems of its own.

Again, thanks a lot. However, I think the best thing to do today is not 
having this kind of entities at all, since I cannot parse an instance just 
for well-formedness. I need a DTD that defines at least these entities, and 
most XML editors (document one, that is) now can output UTF-8 directly.

For this DTD, however, I do not have much control (to my regret, believe me).

I still need a rule for text() nodes grouping, and my attempt at this 
matter has been unsuccessful, despite the helpful hits of Jarno Elovirta...



--
Paul TERRAY - Documents Management Systems Engineer
email : terray@[...].fr 


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list