Re: [xsl] xsl:result-document and Saxon serializer problems

Subject: Re: [xsl] xsl:result-document and Saxon serializer problems
From: Kenneth Stephen <marvin.the.cynical.robot@xxxxxxxxx>
Date: Sun, 21 Aug 2005 15:59:33 -0500
Michael,

    Thanks for the answer. Your efforts to enlighten the mortals among
us (this particular mortal, at any rate) are appreciated. One more
question....

On 8/21/05, Michael Kay <mike@xxxxxxxxxxxx> wrote:
> Nothing has changed here: except that the exclude-result-prefixes attribute
> is now available on any XSLT element, and applies to all LREs lexically
> contained within that element. Previously it could be specified only on the
> xsl:stylesheet element or on the literal result element itself. But e-r-p
> has always affected LREs only.
>
> It might be true that a dynamic scope would sometimes be more useful than a
> static scope, but that's not the way it is (and it's not the way it was in
> 1.0.)
> >
> >     Also, I didnt understand your point (b). I did want it to apply
> > only to everything within that result-document element.
>
> Fine, I was trying to work out the thinking behind your incorrect code and
I
> obviously got it wrong. But the code was still incorrect: the only element
> child of xsl:result-document is a call-template instruction, and e-r-p has
> no effect on a call-template instruction. I guessed that you had written it
> because you thought e-r-p might somehow be passed via the call-template to
> the template being called as a sort of implicit parameter.
>

    Actually my thinking was more on the lines of " in XSLT 1.0,
defining the e-r-p attribute at the xsl:stylesheet level appeared to
make it take effect on the nodes in the result tree. There was no
"passing" around involved. My hope was that the same kind of mechanism
would apply to the sequence constructed by xsl:call-template. However,
that is neither here nor there now.

    The question that now concerns me is how to remove the img-parms
namespace from the output file. I dont understand why the serializer
thinks it is necessary to define that namespace, and I dont know how
to convince it that it isnt necessary. Any ideas?

Thanks,
Kenneth

Current Thread