IMHO, we need at least one interop test. So that people can exercise Axis2 M1 code against their favourite server impl or client impl. It will also weed out problems in the HTTP layer like you mentioned (problems sending messages to Axis 1.X server?). I very much like the idea of picking say Whitemesa's web client, running it against our server and seeing that it indeed works. As i like to say, if you have control over both ends then you can do pretty much anything to transfer data and claim anything you want...Interop is hard and we need to start early. My call, let's add at least one standard interop test that other vendors (and Axis 1.X) have and help kick the tires.
thanks, dims On Tue, 22 Feb 2005 08:32:04 +0600, Srinath Perera <[EMAIL PROTECTED]> wrote: > Hi Dims > what should we do about interops? seems it is your call! > Thanks > Srinath > > On Mon, 21 Feb 2005 16:48:12 +0600, Sanjiva Weerawarana > <[EMAIL PROTECTED]> wrote: > > Aren't samples good enough at this point? Its not a big deal to do > > the interop stuff (not all, but some) .. are there doc/lit echo > > type ones? Those are rather trivial to impl on the server and cannot > > fail in Axis2 ;-). Client is a bit more work but also pretty > > damned near impossible to fail! > > > > I'm not happy about doing encoding interop tests when we haven't > > done that bit yet (formally). However, if you are really gung ho > > about it I won't get in the way! > > > > Sanjiva. > > > > ----- Original Message ----- > > From: "Davanum Srinivas" <[EMAIL PROTECTED]> > > To: "Sanjiva Weerawarana" <[EMAIL PROTECTED]> > > Cc: <[email protected]> > > Sent: Monday, February 21, 2005 3:57 PM > > Subject: Re: [Axis2] typos / interop tests? > > > > > Sanjiva, > > > > > > We need SOME evidence we are on the right track? For me it's interop > > > service/client. We can write all the code we want, but if it can't be > > > used then it's of no use. We can put a huge disclaimer in bold, > > > flashing letters to send the right signal that this is just a tech > > > preview thingy. > > > > > > -- dims > > > > > > > > > On Mon, 21 Feb 2005 10:24:07 +0600, Sanjiva Weerawarana > > > <[EMAIL PROTECTED]> wrote: > > > > Interop tests- doing interop tests will give the wrong signal > > > > about M1 IMO. The objective of M1 IMO is to show progress and to > > > > give (potential Axis) developers a consistent starting point. Its > > > > not a statement of "ready-to-use"ness by any means; there may be, > > > > or rather will be, significant/fundamental changes after M1. > > > > > > > > If we start implementing interop tests then we're telling that > > > > we are "production ready"! Also, without encoding support echo-style > > > > interop is really really damned good, esp. on the server side! > > > > > > > > Sanjiva. > > > > > > > > ----- Original Message ----- > > > > From: "Davanum Srinivas" <[EMAIL PROTECTED]> > > > > To: <[email protected]> > > > > Sent: Saturday, February 19, 2005 9:55 AM > > > > Subject: [Axis2] typos / interop tests? > > > > > > > > > Srinath, > > > > > > > > > > Can you please rename: > > > > > - HandlerMetaData.java to HandlerMetadata.java (lowercase 'd') > > > > > - PhaseMetaData.java to PhaseMetadata.java (lowercase 'd') > > > > > - OmStAXBuilderTest.java to OMStAXBuilderTest.java (uppercase 'M') > > > > > > > > > > Do we have any interop tests? (see http://www.whitemesa.net/) > > > > > > > > > > Thanks, > > > > > dims > > > > > > > > > > -- > > > > > Davanum Srinivas - http://webservices.apache.org/~dims/ > > > > > > > > > > > > > > > > > > -- > > > Davanum Srinivas - http://webservices.apache.org/~dims/ > > > > > > -- Davanum Srinivas - http://webservices.apache.org/~dims/
