Re: [xslt] key() in match pattern of xsl:key

["Joel E. Denny" <jdenny ces clemson edu>]

On Thu, 8 Dec 2005, Kasimier Buchcik wrote:

> On Thu, 2005-12-08 at 07:12 -0500, Joel E. Denny wrote:

> > The EXSLT idea isn't working out well. For a large test case, calling one 
> > key from the match pattern of another key allows my stylesheet to run in 
> > about 1 sec.  After converting the called key to an EXSLT user function, 
> > it takes 25 secs.  I should've realized it would be much worse given the 
> > large data set.

For anyone following closely, I should probably clarify that the 
stylesheet I'm talking about above is not the same stylesheet I posted 
earlier, which was just a toy example.  Of course, they both use the same 
idea.

> Joel, I have 0% percent knowledge of this issue, but I tested you case
> with Stylus Studio (I think it uses Xalan for XSLT) and the result was
> the one you expected. So understand this as additional info without any
> intentions from me.

Thanks for doing this.  This is good to know.

Could you clarify one point? My use of the word `expected' in the original 
post was a bad precedent as it's ambiguous. I gave two sets of output 
produced by libxslt: (1) the desired output, which is produced when 
`keyAnotB' comes before `keyB', and (2) some undesired output, which is 
produced when `keyAnotB' comes after `keyB'.  Did you happen to try both?  
Are you saying the output always agreed with libxslt or was always the 
desired output?

> I recommend asking the people of the XSL list [1] about how the most
> XSTL processors have implemented this. Would be good to go the de facto
> way here if this is an 'implementation defined' issue.
> 
> [1] xsl-list lists mulberrytech com

Thanks for the suggestion. I will pursue this.

Joel