xml-soap-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Matthew J. Duftler" <duft...@watson.ibm.com>
Subject RE: Problems with Apache 2.2 clients and MS SOAP SDK servers using complex types
Date Thu, 05 Jul 2001 19:58:31 GMT
Hi Michele,

The current code in the CVS tree now defaults to using the 2001 Schema URIs.
Please try your code with the latest CVS tree, or one of the nightly builds.


> -----Original Message-----
> From: Michele Costabile [mailto:mico@zucchetti.com]
> Sent: Thursday, July 05, 2001 12:56 PM
> To: soap-dev@xml.apache.org
> Subject: Problems with Apache 2.2 clients and MS SOAP SDK servers using
> complex types
> Hi, I tried for three days to get my request right before writing.
> I am writing a client for a  service generated with Microsoft SOAP SDK
> toolkit 2.0. It is one of the samples that come with the SDK: an address
> book (yet another).
> I tried BeanSerializer first and manually wrote a bean that
> should have been
> compatible with the schema in the WSDL file. There have been any sort of
> problems and I always had an error from the request loader on the server
> side.
> I think most of the problems arise because the Apache toolkit 2.2
> I am using
> types everything and uses the 1999 Schema. The WSDL file from MS uses
> 2000/10.
> I used a SOAPMappingRegistry with a 2000 URI, but the xsi and xsd
> namespaces
> declared in the envelope do not change, local name spaces are generated
> instead.
> I finally found a workaround cutting and pasting the requests
> that come from
> the microsoft clients into a sink.write in the type serializer. It is
> very low lever, but it works.
> I avoid even using generateStructureHeader because otherwise conflicts
> between 1999 and 2000/10 schemas arise again.
> All in all I have proved that one can write a Java client to a MS SOAP SDK
> service, but it boils down to printing a request from scrach.
> Why is the default xsd namespace 1999?
> Where could I try to look to find a better solution (or to patch the
> sources)?

View raw message