Brian De Pradine wrote: > >> >> I agree with you about your claim on invalid messages being sent by >> Axis2, if that is happening now. If that is the case, we need to fix it >> in the sending logic and not in dispatchers or message context. > > The issue here is not with the sending logic, but with the inbound > processing logic. The WS-addressing final inbound handler must be able > to understand that the absence of a was:To header, in a message that > includes other mandatory ws-addressing headers, is semantically > equivalent to the presence of the header, > <wsa:To>http://www.w3.org/2005/08/addressing/anonymous</wsa:To>. Once > this logic is available then the handler will be spec compliant, and we > will be able to add the capability to omit optional headers.
messageContext.to will always be either of transport URL or the value of wsa:To address. If you want to check the presence of a wsa:To header, then you need to look it in the envelope and not in the message contecxt. > As an alternative, there is an existing property called > org.apache.axis2.Constants.Configuration.TRANSPORT_IN_URL that already > holds the inbound transport URL. If some code wants to use the transport > URL then maybe we should consider stipulating that the code must always > use this property to perform the lookup? On the other way round, if some one wants to look at the addressing To value set it as a property to message context or look in envelope. Presence of the above TRANSPORT_IN_URL is not a must. But presence of messagecontext.to is a must. -- Chinthaka
signature.asc
Description: OpenPGP digital signature
