Re: Thoughts on MTOM testing

I'm curious, because at least one person has asked me:  am I correct in 
assuming that many of these tests will be in practice deployed through the 
existing soapbuilders testing community (and perhaps in occasional private 
1-to-1 experiments as well.)  Stated slightly differently, your note could 
be taken to imply that we in W3C are about to specifically support some 
sort of interop event.   Am I right in guessing that we are primarily 
involved just in discussions to ensure that the tests run over at 
soapbuilders will meet our needs for CR evaluation?  Many thanks.

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------








"Martin Gudgin" <mgudgin@microsoft.com>
Sent by: xml-dist-app-request@w3.org
07/28/04 11:00 AM

 
        To:     <xml-dist-app@w3.org>
        cc:     (bcc: Noah Mendelsohn/Cambridge/IBM)
        Subject:        Thoughts on MTOM testing



I promised to post some thoughts on what kind of tests we could run
during the CR period for MTOM/XOP. Here they are;

1.               In terms of simplicity, I think an echo test would be 
easiest,
but with a little twist. If the media type of the request message is
application/soap+xml then the response should be MTOM. And vice-versa.
This will test both receiving and sending MTOM messages.

2.               We could define an element that contains the optimized 
data, so
that it's easy to write code to find that element, rather than having to
look for a whole bunch of them. This will just make implementing the
echo test easier.

3.               We should test messages with a single binary part.

4.               We should test messages with multiple binary parts.

5.               We could test xmlmime:content-type by specifying it on a 
request
message of application/soap+xml and checking that the Content-Type
header of the corresponding MIME part is set correctly in the response
message.

6.               For failure cases, we could test request messages that 
have
binary parts missing ( dangling include ) and messages that use
Content-Location ( not allowed, MUST use Content-ID ).

Proposed xml for single binary part messages:

<soap:Envelope xmlns:soap='http://www.w3.org/2003/05/soap-envelope' 
               xmlns:xm='http://www.w3.org/2004/06/xmlmime' >
 <soap:Body>
  <p:EchoTest xmlns:p='http://example.org/test' >
   <p:Data xm:content-type='image/jpeg'
>ufzZppdWP+APreQS5N3QMbmer9Qb5jJvBJOhIUqXQNBqketslkPtf3VTkouE5IP3Iwc0s4y
3vNPJaaEEtR6Wdw==</p:Data>
  </p:EchoTest>
 </soap:Body>
</soap:Envelope>

And for multiple binary part messages:

<soap:Envelope xmlns:soap='http://www.w3.org/2003/05/soap-envelope' 
               xmlns:xm='http://www.w3.org/2004/06/xmlmime' >
 <soap:Body>
  <p:EchoTest xmlns:p='http://example.org/test' >
   <p:Data xm:content-type='image/jpeg'
>ufzZppdWP+APreQS5N3QMbmer9Qb5jJvBJOhIUqXQNBqketslkPtf3VTkouE5IP3Iwc0s4y
3vNPJaaEEtR6Wdw==</p:Data>
   <p:Data xm:content-type='application/octet-stream'
>rJF4LVZatSi6Xep5LRPvpC8UaIDbCO4JDm6j9qLDN7noGVA/Gdl+MzpQT4a0CgccwrTyDhL
BOaOibx/ot76G9g==</p:Data>
   <p:Data xm:content-type='audio/wav'
>DyJf8o/ih/WcQn5aV7zcrFJAy7TuDGMGjYgQ82rrwvegAhcAnTayWMFs9rTI1TeCIH4L4Z3
lEQO/WGUG0pOKsA==</p:Data>
  </p:EchoTest>
 </soap:Body>
</soap:Envelope>

Does this make sense? What other tests should we do?

Gudge

Received on Wednesday, 28 July 2004 11:14:44 UTC