Hello Stephan,
I think the intialization of messages of which structures are
ascertained at runtime might be a source of incompatibility. If an outgoing
message has an element with "maxOccurs" attribute being "unbounded", and you
can decide how many times the element will occur in the message only at
runtime, then there is no standard way to initialize the message as per
WS-BPEL 2.0.
In ODE 2.0, it can be done using a proprietary extension function to
XPath ode:insert-as-last-into($var, $item). For example:
<assign>
<copy>
<from>
<literal>
<s0_tns:qJoinResponse />
</literal>
</from>
<to>$output0.body</to>
</copy>
</assign>
<assign>
<copy>
<from>
<literal>
<s0_tns:result >
<s0_tns:y >
</s0_tns:y>
</s0_tns:result>
</literal>
</from>
<to variable='_elem1'/>
</copy>
</assign>
<forEach counterName='_i13' parallel='no'>
<startCounterValue>1</startCounterValue>
<finalCounterValue>$_i12</finalCounterValue> <!-- -- The value of
$_i12 is decided at runtime -->
<scope>
<assign>
<copy>
<from
expressionLanguage='urn:oasis:names:tc:wsbpel:2.0:sublang:xpath2.0'>ode:insert-as-last-into($output0.body,
$_elem1)</from>
<to>$output0.body</to>
</copy>
</assign>
</scope>
</forEach>
Regards,
Wenfeng
=======On 2009-07-23 at 16:52:49 Stephan Kiesling wrote: =======
>Hello everyone!
>
>I'm working on a Compliance Test Suite for BPEL engines. In other words:
>I'm trying to measure interoperability of an engine according to the
>standard.
>
>Actually I'm trying to get some informations about known bugs or
>incompatibilities between different engines.
>
>I'm mailing this because I want to make a survey about these issues and
>I hope that you can tell me anything.
>
>Many thanks in advance!
= = = = = = = = = = = = = = = = = = = =
ZHAO Wenfeng (赵文峰)
Ph.D. Candidate
http://www.bupt.edu.cn
