Sebastien,

Venkat's patch includes a test showing what I'd hoped to get working in
opening the JIRA.   In r557916 you can see the portType and binding WSDL
files, both the same TNS and the latter doing a WSDL import of the former.

Haven't had a chance to look at his changes.... but I wanted to clarify that
I wasn't suggesting being able to directly use schema defs from an imported
WSDL, but rather only
the portType definition.

It was in some code in WSDLDocumentProcessor processing schema defs which
happened to be where we broke down before this patch.. which is how that got
brought up.

On a related note, I happened to notice that
org.apache.axis2.wsdl.util.WSDL4JImportedWSDLHelper seems to exist with the
goal of aggregating info across WSDL imports but haven't gotten a chance to
look at this further (or even test out the r557916  patch actually).

Thanks,
Scott



On 7/20/07, Jean-Sebastien Delfino <[EMAIL PROTECTED]> wrote:

Venkata Krishnan wrote:
> Hi,
>
> I am taking a look at
> https://issues.apache.org/jira/browse/TUSCANY-1426 and to get to the
> crux of that problem I am first trying to see if wsdl imports work.
>
> Here is what I am attempting.  There is B.wsdl that imports A.wsdl,
> where B.wsdl has the bindings and services defined and A.wsdl has the
> types, messages and porttypes defined.
>
> But this does not seem to work.  It ends up that while Tuscany WSDL
> Interface processor is dealing with B.wsdl it does not seem to find
> the elements  (remember they are defined inside inline xsd in A.wsdl).
>
> Has this been tried out before?  Could somebody please point me to
> samples or itests that have this working in the trunk?
>
> Related to this, but slightly different question is...
> - Do we expect WSDL4J to be resolving the imports and handing us over
> an aggregated wsdl definition or are we going to use the model  and
> artifact resolution mechanisms within Tuscany to do this ?
>
> Thanks
>
> - Venkat
>

Importing A.wsdl in B.wsdl does not make XSD elements from an XML schema
inlined in A.wsdl available to B.wsdl. XSD elements are only made
available by XSD imports. However, I'm not sure I completely understand
your scenario as I'm guessing from your description. Can you post the
actual A.wsdl and B.wsdl?

With respect to the artifact resolution mechanism:
- if wsdlLocation  and schemaLocation are specified in a WSDL or XSD, we
should honor them
- if they are not present, and in XML documents that do not use these
mechanisms (e.g. SCA .composite files) we should use the Tuscany
namespace based resolution mechanism, find XML artifacts by namespace in
the current contribution and the set of the namespaces imported in it by
SCA <imports> (in sca-contribution.xml)

This is my understanding of how artifact resolution should work in SCA
contributions, described in sections 1.10.2.1 and 1.10.5 of the SCA
assembly spec 1.0.

--
Jean-Sebastien


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


Reply via email to