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

Re: Namespace Injection in DSig 2.0

From: Meiko Jensen <Meiko.Jensen@ruhr-uni-bochum.de>
Date: 3 Sep 2010 10:41:39 +0200
Message-ID: <4C80B4C3.30507@ruhr-uni-bochum.de>
To: "Scott Cantor" <cantor.2@osu.edu>
Cc: "'XMLSec WG Public List'" <public-xmlsec@w3.org>
Scott, all,

thanks for enlightening me. In that case, we're still not having any
progress in terms of fending namespace injection. I've reviewed the last
communications we had on this, and I ended up with the impression that
we had a tendency towards approaches #3 + #4 of my proposal in
http://lists.w3.org/Archives/Public/public-xmlsec/2010May/0027.html ,
however, I found no explicit resolution on this. Maybe we still have to
make a decision here?

Either way, seems like my Action-538 is still open :(

best regards

Meiko



Scott Cantor schrieb:
>> as far as I understood the QNameAware parameter is set "manually" by the
>> signature generator. Hence, he can choose this parameter to contain all
>>     
> the
>   
>> prefixed elements and attributes he used in the selection XPaths. Thus,
>> there is no automated logic involved on how to determine the
>>     
> QNames/prefixes
>   
>> from an XPath; this is up to the developer. What did I get wrong here?
>>     
>
> You're making the same mistake Frederick did at one point (and we'll need to
> add some text so people don't confuse things). The parameter is not
> identifying what actual prefixes or names are *in* the content, it's
> identifying the attributes and elements that themselves contain the QNames
> content. It's up a level from what you're talking about.
>
> Like the ID proposal, it's about injecting signer awareness of the content
> model rather than of the contents of the document. The former can (in some
> applications) be known/invariant, or configured by a deployer. The latter is
> transaction-specific.
>
> As an example:
>
> xsi:type="foo:Bar"
>
> The QNameAware reference is to xsi:type, not to foo:Bar. Anything you put
> into xsi:type can then be dealt with at runtime, simply because the signer
> knows that xsi:type always contains a QName. Doesn't matter what any
> particular QName happened to be.
>
> -- Scott
>
>
>
>   

-- 
Dipl.-Inf. Meiko Jensen
Chair for Network and Data Security 
Horst Görtz Institute for IT-Security 
Ruhr University Bochum, Germany
_____________________________
Universitätsstr. 150, Geb. ID 2/411
D-44801 Bochum, Germany
Phone: +49 (0) 234 / 32-26796
Telefax: +49 (0) 234 / 32-14347
http:// www.nds.rub.de
Received on Friday, 3 September 2010 08:42:08 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 3 September 2010 08:42:08 GMT