Subject: Re: Nicer way to change context-node than xsl:for-each ?
From: Florent Georges <lists@xxxxxxxxxxxx>
Date: Mon, 7 Dec 2009 09:01:52 +0000 (GMT)
|
Mukul Gandhi wrote:
> Florent Georges wrote:
Hi,
> > The first difference
that come to my mind is that the
> > sequence ctor of a function is always
evaluated in temporary
> > output state.
> That's correct I think. The
sequence constructor of
> xsl:function just computes what to return from the
function.
> The function returns, a sequence finally (I think, generically
>
any XDM primitive).
> Whereas instructions in body of, xsl:call-template
actually
> contributes to the output of the template (call-template!).
Well, you can always wrap a xsl:call-template into a
xsl:variable, and call a
function from xsl:sequence. But more
fundamentally, you are never allowed to
use instructions
requiring final output state (namely, xsl:result-document)
from
within a function or anything else called directly or indirectly
from a
function.
Regards,
--
Florent Georges
http://www.fgeorges.org/
|