[xmlsec] namespace definition significant ?

[Roland Hedberg]

Hmm, looking at the XML again I'm not so sure anymore.
Sorry, I might have misled you, not giving you the whole picture.

The fact is that what I'm verifying is a SAML Response where the Assertion is signed.
The whole document contains a xmlns:xsi specification, namely at the top, in the Response element.

But if you only look at the Assertion element by itself there is none.

The Assertion element is a child to the Response element, hence it doesn't have to have the xmlns:xsi specification since a parent has it.
The same goes for the Attributes elements that exist below the Assertion element.
But this is only if you look at the Reponse as a XML document.

Does the fact that the Assertion element is a signed element force the inclusion of a xmlns:xsi specification in the Assertion tree ?
Ignoring what is defined in unsigned parent elements ?

Phrased differently *MUST* the Assertion element be self contained ? 

On Apr 13, 2011, at 15:42, Aleksey Sanin wrote:

> Yes.
> 
> http://www.w3.org/TR/xml-c14n
> 
> Aleksey
> 
> 
> On 4/13/11 6:41 AM, Roland Hedberg wrote:
>> Hi!
>> 
>> Trying to find out why a signature verification failed.
>> So, I compared what I got and what xmlsec1 has as predigest data.
>> 
>> Nothing that I could see except for the fact that xmlsec1 in the predigest data has add xmlns specifications for xsi.
>> 
>> <ns1:Attribute FriendlyName="eduPersonEntitlement" Name="urn:oid:1.3.6.1.4.1.5923.1.1.1.7" NameFormat="urn:oasis:names:tc:SAML:2.0:attrname-format:uri"><ns1:AttributeValue xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:type="xs:string">foo</ns1:AttributeValue></ns1:Attribute>
>> 
>> The original was:
>> 
>> <ns1:Attribute FriendlyName="eduPersonEntitlement" Name="urn:oid:1.3.6.1.4.1.5923.1.1.1.7" NameFormat="urn:oasis:names:tc:SAML:2.0:attrname-format:uri"><ns1:AttributeValue xsi:type="xs:string">foo</ns1:AttributeValue></ns1:Attribute>
>> 
>> Is this significant ??
>> 
>> --Roland
>> 
>> _______________________________________________
>> xmlsec mailing list
>> xmlsec at aleksey.com
>> http://www.aleksey.com/mailman/listinfo/xmlsec

--Roland