RE: message contents -- was Minutes - WS-Addressing Test Suite TF call 5 Jan 2006

Hi Paul:

The contents of the notify/echo string are a) easy to change, and b) not
things that are really part of an implementation - they're part of the
test suite itself.  As such it shouldn't be much trouble to have
everyone send the same pattern, should it?  Why make life difficult for
ourselves?

What I'm after here is a clear and simple way to look at a given message
exchange and know for sure which test was being executed.  This is going
to make writing validating endpoints easier (otherwise you need to
configure the endpoint before the messages are received - so I'm not
exactly sure how you can say the tests should work without being keyed
on the test number?), it's going to make debugging easier (not that
we'll need any of that of course :)), and it's going to make reading the
log files easier, not to mention writing up the results.

Honestly I don't care if we use a fixed string format or use a SOAP
header, but I feel pretty strongly that it's a mistake to skip this kind
of thing.

--Glen 

> -----Original Message-----
> From: paul.downey@bt.com [mailto:paul.downey@bt.com] 
> Sent: Monday, January 09, 2006 11:04 AM
> To: Glen Daniels; public-ws-addressing-tests@w3.org
> Subject: message contents -- was Minutes - WS-Addressing Test 
> Suite TF call 5 Jan 2006
> 
> Hi Glen,
> 
> sending the test number is optional, given the tests should 
> work without being keyed on the test number.
> 
> However some of the defaulting and explicit values in 
> responses could be legitimately keyed on the test number, and 
> I think that's OK for a joke canned interaction, but not for 
> a test with a real implementation.
> 
> I think we can identify the fault and test number as things 
> stand using the following regex expressions:
> 
>   /^$/                      
>         - receiver must fault
> 
>   /^fault/                  
>         - receiver must fault
> 
>   /^.*(test[0-9]{4,4}).*$/  
>         - $1 contains the test number
> 
> Adding the '[' ']' is OK for you as a sender, but I think you 
> can do what you want without requiring everybody to send text 
> in a different order or including the '[' ']' eye-catchers..
> 
> Sound OK, or am I missing something important?
> 
> Paul
> 
> 
> 
> -----Original Message-----
> From: public-ws-addressing-tests-request@w3.org on behalf of 
> Glen Daniels
> Sent: Mon 1/9/2006 3:19 PM
> To: public-ws-addressing-tests@w3.org
> Subject: RE: Minutes - WS-Addressing Test Suite TF call 5 Jan 2006
>  
> 
> 
> Could we actually go one step further here?  It would be nice 
> if receiver implementations knew what test was actually being 
> run in a standard way, so that they could do validation correctly.
> 
> Can we all standardize on strings like this:
> 
> [test1100] anything here
> 
> In other words, "[" + test name + "]" + string
> 
> I think this will end up making our lives (and log files) MUCH easier.
> We could either keep the idea of faulting on something like this:
> 
> [test1133] fault
> 
> Or just fault on an empty "string":
> 
> [test1133]
> 
> Thoughts?
> --G 
> 
> > -----Original Message-----
> > From: public-ws-addressing-tests-request@w3.org
> > [mailto:public-ws-addressing-tests-request@w3.org] On 
> Behalf Of David 
> > Illsley
> > Sent: Friday, January 06, 2006 6:47 AM
> > To: paul.downey@bt.com
> > Cc: public-ws-addressing-tests@w3.org;
> > public-ws-addressing-tests-request@w3.org
> > Subject: Re: Minutes - WS-Addressing Test Suite TF call 5 Jan 2006
> > 
> > 
> > Having agreed on the call last night that the testname 
> should be moved 
> > into the echo string itself rather than be an attribute (because 
> > having it as an attribute changes the required schema) I 
> realised why 
> > it is that way... so that messages which attempt to 
> generate a fault 
> > by sending the empty string still have the test string in them 
> > somewhere.
> > 
> > I've spoken to Paul and agreed that we'll modify the operation 
> > semantics for notify and echo so that either an empty text 
> string or a 
> > text string beginning with the string 'fault' should cause a fault. 
> > This shouldn't break any existing implementations of the test suite 
> > and should be minimal work to add recognition of 'fault'. 
> We will then 
> > be free to trigger faults using strings like fault-test1133
> > 
> > David
> > 
> > David Illsley
> > Web Services Development
> > IBM Hursley Park, SO21 2JN
> > +44 (0)1962 815049 (Int. 245049)
> > david.illsley@uk.ibm.com
> > 
> > 
> > 
> > <paul.downey@bt.com>
> > Sent by: public-ws-addressing-tests-request@w3.org
> > 
> > 05/01/2006 21:55 To
> > <public-ws-addressing-tests@w3.org>
> > cc
> > Subject
> > Minutes - WS-Addressing Test Suite TF call 5 Jan 2006
> > 
> > 	
> > 
> > 
> > 
> > 
> > 
> > Present: 
> >  David Illsley (IBM)
> >  Mike Vernal (Microsoft)
> >  Mark Nottingham (WG Chair)
> >  Giovanni Boschi
> >  Kim Palko
> >  Glen Daniels (Sonic and Apache)
> >  Paul Downey (BT)
> > 
> > 
> > agenda+ Action items
> > 
> > ACTION: pauld to flag which tests are for CORE, SOAP, WSDL CR
> > (PENDING)
> > 
> > (not discussed!)
> > 
> > -----
> > 
> > Glen concerned about lack of non-anonymous test cases and 
> combinations 
> > thereof:
> > 
> > http://www.w3.org/2002/ws/addr/testsuite/issues/#test15
> > 
> > Glen: to supply non-anon ReplyTo/FaultTo test cases based upon 
> > existing operations.
> > 
> > These shouldn't impact those implementing the current operations as 
> > services.
> > 
> > Implementations are not required to generate each and every 
> test case 
> > message - though expected to process and respond correctly 
> to messages 
> > sent to them. We can use a canned client to send specific requests.
> > 
> > ACTION: Paul to make his canned implementation available
> > 
> > -----
> > 
> > agenda+ Dispatching by Action
> > 
> > GED (v) Action, specific testing required in this round of CR?
> > http://lists.w3.org/Archives/Public/public-ws-addressing-tests
> > /2006Jan/0007.html
> > 
> > Consensus on the call: we're happy with the Status Quo of the tests 
> > for Vancouver - i.e. echoIn and echoOut message bodies enabling 
> > processing keyed by Action and/or GED.
> > 
> > WSDL CR testing may include tests to prove dispatching on 
> the action, 
> > should that be deemed appropriate.
> > 
> > -----
> > 
> > agenda+ Features not covered by tests
> > 
> > http://www.w3.org/2002/ws/addr/testsuite/features/
> > 
> > * wsa:RelatesTo/@RelationshipType defaulting
> > 
> > http://www.w3.org/2002/ws/addr/testsuite/features/#core07
> > 
> > David suggests an echo followed by a notify as one way of proving a 
> > relationship.
> > 
> > requiring defaulting  of RelationshipType is tricky
> > - as with the notify tests, implementations may not be able to 
> > generate all of the messages.
> > 
> > ACTION: Paul to write a RelatesTo test
> > 
> > Again a 'canned' server may be used to send defaulted and 
> undefaulted 
> > replies to test clients.
> > 
> > * optional faults
> > 
> > http://www.w3.org/2002/ws/addr/testsuite/features/#soap07
> > 
> > Lack of test cases - Glen raises it would be interesting to 
> know who 
> > had implemented what error codes - Paul remains concerned 
> how many are 
> > testable or implicitly imply order of processing.
> > 
> > ACTION: Paul to enumerate missing fault code tests
> > ACTION: Paul to elicit which fault codes have been implemented
> > 
> > -----
> > 
> > agenda+ Other Issues
> > 
> > http://www.w3.org/2002/ws/addr/testsuite/issues/
> > 
> > ....
> > 
> > -----
> > 
> > agenda+ The interoperability Event
> > 
> > http://lists.w3.org/Archives/Public/public-ws-addressing/2005D
> > ec/0034.html
> > http://www.w3.org/2002/ws/addr/testsuite#status
> > 
> > Unclear if running implementations on laptops will be available 
> > outside the hotel / room. Ideally endpoints should be publicly 
> > available.
> > 
> > 
> > -----
> > 
> > agenda+ Status of implementations
> > 
> > IBM, Microsoft, Axis2 confirmed on the call, Axis C/C++, Sonic, Sun 
> > and JBOSS possibly ready for Vancouver.
> > 
> > Paul asks that implementations should be able to optionally 
> send the 
> > test case number as content (not as additional elements or 
> attributes 
> > which would require the WSDL changing the Schema element from an 
> > xs:string into ComplexType.) as described:
> > http://www.w3.org/2002/ws/addr/testsuite/#operations
> > 
> > -----
> > Previous meetings:
> > http://lists.w3.org/Archives/Public/public-ws-addressing-tests
> > /2005Dec/0033.html
> > http://lists.w3.org/Archives/Public/public-ws-addressing-tests
> > /2005Dec/0016.html
> > http://lists.w3.org/Archives/Public/public-ws-addressing-tests
> > /2005Dec/0003.html
> > http://lists.w3.org/Archives/Public/public-ws-addressing-tests
> > /2005Nov/0001.html
> > http://lists.w3.org/Archives/Public/public-ws-addressing-tests
> > /2005Oct/0004.html
> > 
> > 
> > 
> > 
> 
> 
> 
>

Received on Monday, 9 January 2006 16:14:55 UTC