Re: AW: AW: AW: Status of Implementations

At 10:20 10/6/2000 +0200, Gregor Karlinger wrote:
>I DO think that inserting whitespace text inside SignedInfo will break the
>signature, since
>
>* even a validating parser must not remove insignificant white space during
>   parsing. Instead, he must inform the "application" whether white space is
>   insignificant or not.
>
>* We have not defined any rules in the XMLSignature spec (which is the
>   norm for the "application") how a XMLSignature application must treat
>   whitespace in the SignedInfo element and its descendants.
>
>* And finally, canonical XML does not remove whitespace text.

Hrmmm... well my immediate goal was to make your examples a bit more pretty 
like Kent's example from the last version [1]; this has white space 
indenting in SignedInfo. However, Tidy started as an HTML processor and 
white space isn't important in that context for element content. But I think 
your right, in our context it  does matter. Consequently, this means that 
while you might verify Kent's Signature and Kent might verify your 
Signature, your two applications (when signing the same document with the 
same key and everything else equal) will yield a different SignatureValue 
depending on how you output the XML.

Regardless, I restored your tested examples (if I grabbed the right ones out 
of my attach directory! <smile>) to [2], and if someone provides me with 
pretty and tested examples, I'll include them.

[1] 
http://www.w3.org/TR/2000/WD-xmldsig-core-20000918/signature-example-rsa.xml
[2] http://www.w3.org/Signature/Drafts/WD-xmldsig-core-latest/

__
Joseph Reagle Jr.
W3C Policy Analyst                mailto:reagle@w3.org
IETF/W3C XML-Signature Co-Chair   http://www.w3.org/People/Reagle/

Received on Friday, 6 October 2000 10:47:32 UTC