Re: [xml] Canonicalization (C14N) of an XML element

[Daniel Veillard <veillard redhat com>]

On Tue, Feb 08, 2005 at 11:30:44AM -0800, Aleksey Sanin wrote:
> Yes. You have to add all the nodes that you want to get
> canonicalized. However, in this particular case there is
> a better way. The xmlC14NExecute function allows one to
> provide a callback that would be called to determine if
> a given node is "visible" (i.e. included in c14n output)
> or not. To canonicalize just a given node (with all the
> attributes and children nodes) you can write a callback
> function that would return "true" for your node, its attributes,
> namespaces and children nodes and "false" for all other
> nodes in the document.

  Aleksey,

 I don't know if you have followed the -slightly heated- thread
about xml:id and C14N starting at
  http://lists.w3.org/Archives/Public/public-xml-id/2005Feb/0038.html

basically the c14n spec expect propagation if all xml:* attributes
while it is an error to propagate those for xml:base and possibly for 
xml:id.
It seems to me this is the code in c14n.c after line 982. I'm of the
opinion that the C14N spec is broken in that it was assuming cascading
behaviour for all xml:* attributes and while this is true for xml:space
and xml:lang it is not true for new attributes like xml:base and xml:id.
As an implementor what do you think ? My take is that the c14n spec and
the code need fixing, but I would not go forward without your input,
(it's your code :-), and can potentially generate some interop problems
for documents using xml:base and xml:id, which may raise bug reports on
your side.

  No need to rush but it's a problem,

Daniel

-- 
Daniel Veillard      | Red Hat Desktop team http://redhat.com/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/