Thanks Adrian for your feedback. I will start to work on this and hope to find a solution towards the end of the week.
BTW, if the request tag name is the same as the operation name, the server works fine with the fix that I did yesterday. Which means, I already have dropped the need for SOAP Action. But I must resolve this bug before we could call it done. Thanks, Samisa... On 8/3/05, Adrian Dick <[EMAIL PROTECTED]> wrote: > Samisa, > > +1 Keeping a mapping between service/operation and request/response tags > sounds good to me. In addition to Axis Java, I believe this is the > solution for a number of other server implementations in the market place. > > Once this final change is made we need to ensure we have couple of > testcases in which the WSDL contains NO soap action, or "garbage" soap > action. As I believe all our WSDLs currently contain a SOAP action in the > form previously required by our server. > > Adrian > _______________________________________ > Adrian Dick ([EMAIL PROTECTED]) > > "Samisa Abeysinghe" <[EMAIL PROTECTED]> wrote on 03/08/2005 05:24:42: > > > Hi All, > > I was able to come up with a quick fix to drop the SOAP Action > > usage on the server side. > > > > However, this morning, testing with the server, I realized that > > there is a problem in the fix provided. When picking the operation to > > invoke based on the request SOAP message, it is possible that the SOAP > > request element name to be different from the operation name. (In other > > words the input message element name and the operation name can differ > > in the WSDL) > > > > Looking into the generated Java code for WSDLs, it looks that in > > Java, the wsdd keeps track of Request and Response element names and map > > that to the operation name. > > We only have the allowed operation name list as of now in our > > WSDD. What I propose is that we also include the request, response tag > > name mapping against the operation name in WSDD and we can auto generate > > this stuff from WSSL tool. > > > > Any thoughts on this are welcome. > > > > Until this is fixed, we will have some tests failing with C++ server > > side. > > > > Thanks, > > Samisa... > >
