Hi, I have not looked into the SOAP specs in detail but may I suggest a solution. I'm sure SOAP 1.1 and 1.2 has similar properties and behaviors even though they are named diffrently. So my guess is we can have a common model (one of our own) that has both the features and have two implementations of that set of interfaces for SOAP 1.1 and 1.2 (or whatever comes later even :)) The internal engine works with our model and wouldn't see the diff.
thoughts ? On Thu, 3 Mar 2005 11:49:59 +0600, Eran Chinthaka <[EMAIL PROTECTED]> wrote: > > > Hi, > > > > We implemented a SOAP layer on top of OM using the SOAPBuilder. But it was > in accordance with SOAP 1.1. But now I think its time to support SOAP 1.2 as > well. There we have a problem as some of the things are bit different in 1.2 > compared to 1.1. (see here > http://www.w3.org/TR/2003/REC-soap12-part0-20030624/#L4697) > > > > For example; SOAP fault element has many difference between these two. > > And there are some name changes as well. For example actor has changed to > role. So I have a problem which method to use. Is it getRole() or getActor() > I should use ? or put both ? > > > > When u think about handlers accessing the SOAP message using OM, one might > write there code to comply with SOAP 1.1 and whilst other with SOAP 1.2. > > > > If you put a method like getRole, well that can be used to return actor for > SOAP 1.1. But is it ok ? > > > > Plus, there are some specific information available in SOAP faults. In 1.2 > Code and Reason are mandatory and the element hierarchy there is different > to (of more informative) than SOAP 1.1. > > > > So how do we support both â.. ?? > > > > Option 1 : Having two SOAP builders for SOAP 1.1 and SOAP 1.2. But then > handler writers will have to have a big if.. else .. > > Option 2 : Support SOAP 1.2 only and if there is something missing, user > (handler writer or any other accessing SOAP), must take care of that. > > Option 3 : ââ.. ?? > > > > I'd like to see a very convenient API for user to manipulate the SOAP > message using OM. > > > > Comments/ Thoughts .. > > > > > > Regards, > > Eran Chinthaka > > -- Ajith Ranabahu
