RE: [Fwd: SAWSDL Eclipse tests]

Hello All,

I did further testing and found that although a WSDL document that contains
types and porttype validates cleanly if attrExtensions is the first element
under operation, a binding that references that porttype is flagged with an
error.  The validator does not think the porttype exists.  I tried several
changes to the WSDL, but the error persisted.  Eclipse will successfully
generate a Web Services JavaBean skeleton from this annotated WSDL file
despite the validation error.  Thanks to Richard for identifying this
problem.

Regards,
Joel


public-ws-semann-request@w3.org wrote on 06/15/2007 04:22:11 PM:

>
> Richard, I see you did write the actual error. Sorry, my bad, too much
> email. Joel, can you please check if you also see the problem with the
> WSDL file that Richard sent earlier?
> It looks to me like some bug in the tooling, failing silently on the
> interface and then complaining in the binding.
> Best regards,
> Jacek
>
> On Tue, 2007-06-12 at 18:35 +0200, richard.aillet@fr.thalesgroup.com
> wrote:
> > Dear Jacek,
> >
> > I just check what you evocate. And you are right!
> > If I reverse the order in the portType.operation element by putting the
> > sawsdl:attrExtensions element before the the wsdl:input, the validation
> > (under Eclipse 3.2.2 and WTP 1.5.4) does no more raise error ... at
this
> > line.
> >
> > Unfortunately it raises an error later below in the WSDL at the
"binding"
> > element declaration. The error is :
> >
> > << The '***PortType' port type specified for the '***Binding' binding
is
> > undefined.
> > Check the '***PortType' port type name and ensure it is defined. >>
> >
> > As if the previous annoatetd portType element was not well declared.
> > Curious.
> >
> > Regards,
> > Richard.
> >
> >
> > -----Message d'origine-----
> > De : Jacek Kopecky [mailto:jacek.kopecky@deri.org]
> > Envoyé : mardi 12 juin 2007 17:49
> > À : Pierre.CHATEL@fr.thalesgroup.com;
richard.aillet@fr.thalesgroup.com;
> > Joel Farrell
> > Cc : SAWSDL public list
> > Objet : [Fwd: SAWSDL Eclipse tests]
> >
> >
> > Dear Pierre, Richard and Joel,
> >
> > I wonder whether the problem might be in the wrong ordering of the
> > elements. I just checked the WSDL schema and the schema primer and it
> > seems that the extensibility elements should come first (according to
> > the schema, that is), whereas our spec has two examples with the
reverse
> > order, first wsdl:input and wsdl:output, then sawsdl:attrExtensions.
> >
> > I'm not sure though whether tools would expect the extension elements
to
> > be after the WSDL elements (because the order matters - it specifies
the
> > message exchange pattern), in which case tweaking the schema could be
> > necessary.
> >
> > Richard or Pierre, can you please try the file with the reverse order
of
> > sawsdl and wsdl attributes? Or whether you could tweak the schema used
> > by your tool to also allow extension elements after the WSDL elements?
> >
> > Best regards,
> > Jacek
> >
> > -------- Forwarded Message --------
> > From: Joel Farrell <joelf@us.ibm.com>
> > To: SAWSDL public list <public-ws-semann@w3.org>
> > Subject: SAWSDL Eclipse tests
> > Date: Tue, 12 Jun 2007 09:09:07 -0400
> >
> >
> >
> > I have done some further testing on the following developer tools:
> >
> > IBM WebSphere Integration Developer 6.0.1
> > Eclipse 3.2.1 with Web Tools Platform 1.5.2
> > Eclipse 3.2.2 with Web Tools Platform 1.5.4
> >
> > In each case, WSDL files with attrExtensions under operation can be
> > created and edited without errors. However, in each tool, when an
> > explicit validation is done, the attrExtensions element is flagged as
> > invalid.
> >
> > Eclipse uses wsdl4j to parse WSDL files. The two wsdl4j parsers in the
> > two Eclipse Web Tools Platform levels both use the same xsd file - the
> > one that allows element extension for operation , but they still do not
> > accept the extension element as used.
> >
> > However, in all three tools, I could still successfully generate code
> > from the WSDL file. Documenting this fact may allow eclipse-based
> > attrExtension usage to proceed.
> >
> > Regards,
> > Joel
> >
> >
>
>