----- Original Message ----- From: "Tom Jordahl" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Monday, April 08, 2002 9:09 AM Subject: Re: Testing with .NET
> > Thanks for the reply Steve. > > This would be very interesting to test .NET client against Axis, did I just hear you volunteer to contribute that? :-) I know that I for one have .NET on my Win2K machine and I would run a test target against Axis if it existed. If you can send some stuff to the list, I will test it out and commit it! well, I was hoping somebody else will run with it, but if not, then I will see what I can do. > My main concern, however, is to preserve the Axis client behaviors >against various .NET services. To ensure that we continue to >interoperate on the client side. The difficulty is that our test suite >generally tests Axis against itself, which can sometimes hide problems >if client and server have the same behavior. There have been various >problems with the single test we currently have (SalesRankAndPrice) that >tests a service at perfectxml.net. But I think we should add more .NET >services on the internet to the suite. If we do add more, I think we >need to split them off in to a separate test target, to allow >disconnected testing. I am interested in having interop working right from the WSDL onwards; having nightly tests of WSDL to code to invocation against different servers ensures that nothing has broken -far better to find out every night than at some intermittent interop workshop. If I did the tests, (and it is still an if) I'd make the .net tests a separate build.xml file, for complete standaloneness. Before I do the tests I need to bring Ant's <csc> task up to date and add the <vbc> and a wrapper to the .net WSDLto(CSC/VB) app
