[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [wsbpel] Issue - 264 - <correlationSet> specifications on <invoke>with @initiate="no" & @pattern="response" should be invalid
Folks, So, here is the problem. Alex Previous Stated: --------------------------------------------------------------------- Hi Prasad, I am not sure that is an invalid case. That is more like a valid usecase of check whether the response is consistent with a previously initiated CS. If the response is inconsistent, correlationViolation fault would be thrown. Similar situation applies to the combination of @pattern="request" and @initiate="no" on an <invoke>. An initiated CS serves two functions: (1) message correlation (2) message content consistency Does it make sense to you? Thanks! Regards, Alex Yiu ----------------------------------------------------------- So is it the *assumption* then that the message that is the "response" for an <invoke> *will* be delivered to correct (matching) instance of the process *independent* of any <CorrelationSet>s that are specified in the message? If so, where is that stated in the spec? Or do the usual <CorrelationSet> value matching rules apply for the incoming Message to delivered to the proper instance? If not, why not? When it is an asynchronous exchange do we not have the same issue of matching the response of an invoke with correct instance, if correlation sets are not utilized? In the following example, does the "Invoice" correlation on the response need to match for the message to be delivered ? We do not state anywhere clearly, on an <invoke> if the <CorrelationSet>s in the reply must match or not for message to be delivered? This needs to be clarified. <invoke .....> <correlations> <correlation set="PurchaseOrder" initiate="yes" pattern="request"/> <correlation set="Invoice" initiate="no" pattern="response"/> </correlations> </invoke> Regards, Prasad -------- Original Message --------
+1 to Danny. This does not seem to be an issue to me either. Regards, Alex Yiu Danny van der Rijn wrote: I disagree. The correlation can be used to enforce a restraint via a correlationViolation fault. |
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]