Re: [xsl] extract xpath locator

[Nicolas Mazziotta <nicolas.mazziotta@xxxxxxxxx>]

I works!
Thanks a lot for the fast answer.
Sorry it was in the archive.

N. Mazziotta

On Thu, 12 Aug 2004 10:25:27 +0100
David Carlisle <davidc@xxxxxxxxx> wrote:

> 
> The archives of this list (and possibly the faq) have a very complete
> implementation of this from Dimitre if I recall that handles all kinds
> of nodes, comments text etc, but basically you want to use
> [<xsl:value-of
> select="1+count($node/preceding-sibling::*[local-name(.)=local-name($node)
>                                    and namespace-uri(.)= and namespace-uri($node)])"/>]
> 
> or simpler don't write out 
> <xsl:value-of select="name($node)">
> just write out a *
> then you can do
> [<xsl:value-of
> select="1+count($node/preceding-sibling::*)"/>]
> 
> ie produce
> 
> /*[1]/*[2]
> 
> rather than
> "/doc/i[2]".
> 
> 
> > I have tried to "count(preceding-sibling::$name)"
> 
> You can never use a variable to hold a bit of Xpath syntax, this would
> be like having x="+ 2" in C and hoping that 1 x was a) legal and b)
> evaluated to  3. Variables can only be bound to values.
> 
> David
> 
> 
> ________________________________________________________________________
> This e-mail has been scanned for all viruses by Star Internet. The
> service is powered by MessageLabs. For more information on a proactive
> anti-virus service working around the clock, around the globe, visit:
> http://www.star.net.uk
> ________________________________________________________________________