Oliver,
Axis client should throw the fault specified in the WSDL if the Soap:Fault comes from an Axis server. The client side exception will NOT contain the data that we in the server side exception, this is not implemented yet. See the test in src/test/wsdl/faults. The basic problem is there is no specification as to how the client knows where the fault name lives and where the data lives in the soap:fault that is returned from a web service. There is now way currently to tie the info from the WSDL to an actual implementation class. Axis currently embeds the actual java class name of the exception in the faultDetails element, but this is non-standard (i.e. we made this up). -- Tom Jordahl Macromedia -----Original Message----- From: Oliver Suciu [mailto:[EMAIL PROTECTED]] Sent: Friday, April 19, 2002 4:30 PM To: [EMAIL PROTECTED] Subject: Re: .Net ignores <wsdl:fault>? Yes, but wsdl.exe doesn't generate the exception classes, so I can't catch an exception more specific than SoapException. It actually doesn't matter cuz I don't even get exceptions propagated properly to an Axis client -- looks like the current state of the art is not ready yet for <fault>... -- Oliver PS: Just found out that C# has no equivalent to the Java "throws" declaration -- interesting... Scott Seely wrote: > > At this point in time, what .NET will do is receive the data and make it > visible to the client code as XML through the details element (IIRC). > > -----Original Message----- > From: Oliver Suciu [mailto:[EMAIL PROTECTED]] > Sent: Friday, April 19, 2002 10:49 AM > To: [EMAIL PROTECTED]; [EMAIL PROTECTED] > Subject: .Net ignores <wsdl:fault>? > > Hi all, > > (Sorry, this one is somewhat geared towards .Net...) > > I've got an Axis service that throws an exception, and WSDL2Java > correctly generates a <wsdl:fault> for the <wsdl:operation>. > > However, the .Net wsdl.exe seems to completely ignore the <fault>, > and no corresponding client-side code is generated. > > Has anybdoy got this working? > Anybdoy looked into .Net interop with respect to <fault>? > > Thanks, > > -- Oliver
