Frank/Kelvin,
I am under the impression that fixing JIRA Tuscany-153 will get
ChangeSummaryType properties working and allow ChangeSummary-enabled
DataObjects to be passed in web service invocations, correct? I noticed
Kelvin supplied a patch for this recently. Is this patch a complete fix
for the problem or just a stepping-stone towards it? Do you have a rough
ETA when this JIRA might be fixed? Do you think it will be fixed in M2?
BTW, any rough estimates when M2 will be released? Any possibility it
will include a separate SDO distribution similar to the C++ M1 distro?
Folks like myself upgrading from SDO 1.0 would greatly appreciate a
separate distribution. Should I great a JIRA for this issue?
Thanks in advance for the update.
- Ron
Frank Budinsky wrote:
Ron,
This is an area of the spec that is full of holes. The bottom line is that
in SDO 2.0.1 the Java interface commonj.sdo.DataGraph has no defined
relationship to the DataObject that would be created from the complexType
BaseDataGraphType, other than the one vaguely implied by this example.
DataGraphs are not DataObjects in the current spec, Tuscany, or any other
implementation of SDO that I'm aware of. The SDO collaboration group has
been discussing changing this in SDO 2.1. The general feeling is that it
would be good to make DataGraphs proper DataObjects (it would solve a lot
of the other questions like how do you create a DataGraph), but we need to
define the behavior of DataGraph/DataObject methods like getDataGraph,
getChangeSummary, and getRootObject, some of which are duplicates
(declared in both interfaces), and all of which have ambiguous behavior in
the combined DataGraph/DataObject case.
As far as Tuscany is concerned, a prereq is to get ChangeSummaryType
properties to work - which we've started working on, but still need a few
weeks. Once we have that working, we can start implementing DataGraph as a
proper DataObject (ahead of the 2.1 spec). After that, I think you could
do what your asking for.
Short term, I think you're out of luck.
Frank.
Ron Gavlin <[EMAIL PROTECTED]> wrote on 07/10/2006 04:20:32 PM:
I am attempting to use Tuscany SDO to pass a DataGraph as a
parameter to a document-wrapped web service method, similar to the
"Web services and DataGraphs Example" on pages 137-139 in the SDO 2.
01 specification. In my case, however, I am attempting to wrap the
DataGraph within a statically-generated DataObject that represents
the method being invoked. The relevant portions of the schema are as
follows:
<!-- from the spec -->
<element name="companyDataGraph" type="company:CompanyDataGraphType" />
<complexType name="CompanyDataGraphType">
<complexContent>
<extension base="sdo:BaseDataGraphType">
<sequence>
<element name="company" type="company:CompanyType" />
</sequence>
</extension>
</complexContent>
</complexType>
<!-- my addition -->
<xsd:element name="updateCompanies" type="UpdateCompanies" />
<xsd:complexType name="UpdateCompanies">
<xsd:sequence>
<xsd:element ref="companyDatagraph" />
</xsd:sequence>
</xsd:complexType>
Should a full SDO 2.01 implementation support this use case? If so,
how close is Tuscany to being able to support it? In particular,
when I run Tuscany XSD2JavaGenerator, the class generated for
CompanyDataGraphType extends DataObjectImpl and not DataGraphImpl.
Is fixing this simply a matter of enhancing the XSD2JavaGenerator to
recognize "sdo:BaseDataGraphType" extensions and make these classes
extend DataGraphImpl instead of DataObjectImpl? Is serializing/de-
serializing a DataGraph as a child of a DataObject doable?
- Ron
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
