W3C home > Mailing lists > Public > public-xmlsec@w3.org > December 2010

ACTION-742: Propose text regarding 6.7.1.1 and use of c14n2

From: Cantor, Scott E. <cantor.2@osu.edu>
Date: Fri, 10 Dec 2010 20:13:00 +0000
To: "public-xmlsec@w3.org (public-xmlsec@w3.org)" <public-xmlsec@w3.org>
Message-ID: <7EE86E89365CA94F8E7B8251F926071001D00B@CIO-KRC-D1MBX01.osuad.osu.edu>
We resolved on the last call that we should require that the "subtree with exclusions" output model for selection (and input to 2.0 c14n) should guarantee specific "context" information about the surrounding XML document state so that particular c14n algorithms can choose whether/how to deal with the information. Specifically this applies to xml:space and xml:base (and I would probably argue xml:lang also).

In the particular case of C14N 2.0, inherited xml:base and xml:lang are ignored, but xml:space is needed because of the whitespace trimming option.

I suggest the final bullet of 6.7.1.1 be changed to:

Each subtree (except when the subtree is of a complete document) includes the following additional information:

    * The set of namespace declarations in scope (i.e., inherited from the ancestors of the subtree).
    * The inherited values of the xml:space, xml:lang, and xml:base attributes.
    * If there are multiple ancestor elements with xml:base defined, they must be combined in accordance with [Ref] (what's the reference here? C14N 2.0?)

     Implementations MAY omit the additional context information, in total or in part, if they have specific knowledge of the requirements of the relevant canonicalization algorithm in use. For example, Canonical XML 2.0 requires only namespace declarations and the inherited value of xml:space. In the absence of such knowledge, all the information MUST be maintained and supplied.

-- Scott
Received on Friday, 10 December 2010 20:13:37 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 10 December 2010 20:13:38 GMT