W3C home > Mailing lists > Public > w3c-ietf-xmldsig@w3.org > April to June 2002

Re: [XPath Filter 2.0] Problem with Transform model

From: Christian Geuer-Pollmann <geuer-pollmann@nue.et-inf.uni-siegen.de>
Date: Mon, 27 May 2002 11:12:36 +0200
To: Gregor Karlinger <gregor.karlinger@iaik.at>, XMLSigWG <w3c-ietf-xmldsig@w3.org>
Message-ID: <953921.1022497956@pinkpanther>
Hi Gregor,

--On Sonntag, 26. Mai 2002 14:31 +0200 Gregor Karlinger 
<gregor.karlinger@iaik.at> wrote:

> after I implemented the current spec of XPath Filter 2.0[1], I
> started to play around with some akward combinations of input
> node set (I) and expanded evaluated XPath expression (S) and
>  came across the following problem:
>
> If the transform implementation receives an empty input node set
> as its input paramter, there is no way to get access to the
> input document. But this is necessary because the context
> node for the XPath transform has to be initialized with the
> root node of the input document (according to section 3.3
> of [1]).

Absolutely right. If you don't have at least one node which works as an 
"anchor" into the to-be-signed document, you can't proceed (well, it'd be 
possible to create nasty hacks which do not only pass the node set 
information to the transform but also the owner document).

> A possible solution for this problem could be to explicitely
> forbid an emtpy input for the XPath Filter 2.0 transform. Any
> thought?

For me, +1. That'd be a good addition/clarification to the spec. Empty node 
sets should be an error.


Regards,
Christian
Received on Monday, 27 May 2002 05:07:25 GMT

This archive was generated by hypermail 2.2.0 + w3c-0.29 : Thursday, 13 January 2005 12:10:15 GMT