W3C home > Mailing lists > Public > w3c-ietf-xmldsig@w3.org > April to June 2000

RE: Schemas and DTD for X509Data

From: Brian LaMacchia <bal@microsoft.com>
Date: Wed, 14 Jun 2000 23:46:28 -0700
Message-ID: <39ADCF833E74D111A2D700805F1951EF255DE194@RED-MSG-06>
To: "'Joseph M. Reagle Jr.'" <reagle@w3.org>, gregor.karlinger@iaik.at
Cc: XML <w3c-ietf-xmldsig@w3.org>, Barb Fox <bfox@exchange.microsoft.com>
Gregor is correct, the X509Data schema is incorrect as it currently stands
(Joseph, looks like you swapped the upper bounds on the first choice w/ the
sequence).  I believe this is the correct schema:

   <element name='X509Data'> 
     <complexType content='elementOnly'> 
       <choice minOccurs='1' maxOccurs='1'>				***
changed
         <sequence minOccurs='1' maxOccurs='unbounded'> 		***
changed
           <choice minOccurs='1' maxOccurs='1'> 
             <element ref='ds:X509IssuerSerial'/>
             <element name='X509SKI' type='string'/> <!-- should this be
binary? -->
             <element name='X509SubjectName' type='string'/> 
           </choice>  
         </sequence>
         <element name='X509Certificate' type='ds:CryptoBinary'
minOccurs='1' maxOccurs='1'/> **changed
         <element name='X509CRL' type='ds:CryptoBinary' minOccurs='1'
maxOccurs='1'/> 	    **changed
       </choice>
     </complexType>
   </element>

The maxOccurs of the first choice should be 1, not unbounded, and the
maxOccurs of the sequence should be unbounded, not 1.  Also, the "zero or
more" qualifiers for X509Certificate and X509CRL became "exactly one".  This
change means that an individual X509Data clause may be one of:
a) a collection of X509IssuerSerial, X509SKI and X509SubjectName elements,
b) exactly one X509Certificate element, or
c) exactly one X509CRL

Given this schema, I believe the declaration in the immediately preceeding
paragraph in the draft is now correct:

Multiple declarations about a single certificate (e.g., a X509SubjectName
and X509IssuerSerial element) MUST be grouped inside a single X509Data
element; 
	[this is covered by the sequence of choice of... --bal]
multiple declarations about the same key but different certificates (related
to that single key) MUST be grouped within a single KeyInfo element but
multiple X509Data elements. 
	[this is covered by the outer choice --bal]
For example, the following block contains two pointers to certificate-A
(issuer/serial number & SKI) and a single reference to certificate-B
(Subject Name):

I believe this paragraph and the scheme correctly capture the desire of the
working group as expressed in Victoria.  I would point out that there's one
unfortuate side-effect of this syntax, which is that expressing a cert chain
gets a little more verbose than it needs to be.  A three-cert chain will
look like this:

<KeyInfo>
  <X509Data>
    <X509Certificate>
	Base64 data for cert 1
    </X509Certificate>
  </X509Data>
  <X509Data>
    <X509Certificate>
	Base64 data for cert 2
    </X509Certificate>
  </X509Data>
  <X509Data>
    <X509Certificate>
	Base64 data for cert 3
    </X509Certificate>
  </X509Data>
</KeyInfo>

So you end up with two levels of tags for each cert.

I agree with making SKI a ds:CryptoBinary; it's just a OCTET STRING.

   						--bal

-----Original Message-----
From: Joseph M. Reagle Jr. [mailto:reagle@w3.org]
Sent: Tuesday, June 13, 2000 11:10 AM
To: gregor.karlinger@iaik.at
Cc: XML; Barb Fox; Brian LaMacchia
Subject: Re: Schemas and DTD for X509Data



Ah... this is the tricky bit that I might've goofed up from Brian when I
translated that section's clarifications into the new schema syntax.

At 07:57 PM 6/13/00 +0200, Gregor Karlinger wrote:
 >I think there is a contradiction between the textual description of
 >section 4.4.4 and the corresponding Schema and DTD definitions:
 >
 >The text (and the provided example) permit multiple declarations inside
 >a single X509Data Element, such as X509IssuerSerial and X509SKI. But as
 >you can see in the following Schema and DTD definitions, such multiple
 >declarations are not possible:

Doesn't the <choice/> above it permit it?

 >   <element name='X509Data'>
 >     <complexType content='elementOnly'>
 >       <choice minOccurs='1' maxOccurs='unbounded'>
 >         <sequence minOccurs='1' maxOccurs='1'>
 >                                            ^
 >                                            I
 >--------------------------------------------I

Oh, and below you quoted some text that included a comment from me. Was
X509SKI supposed to be string or binary?

 >           <choice minOccurs='1' maxOccurs='1'>
 >             <element ref='ds:X509IssuerSerial'/>
 >             <element name='X509SKI' type='string'/> <!-- should this be
 >binary? -->
 >             <element name='X509SubjectName' type='string'/>
 >           </choice>


_________________________________________________________
Joseph Reagle Jr.   
W3C Policy Analyst                mailto:reagle@w3.org
IETF/W3C XML-Signature Co-Chair   http://www.w3.org/People/Reagle/
Received on Thursday, 15 June 2000 02:48:24 GMT

This archive was generated by hypermail 2.2.0 + w3c-0.29 : Thursday, 13 January 2005 12:10:09 GMT