Re: AW: [sdo] Issue 118 - Proposal for review

I agree with you Ron. I was only answering Blaise's questions to help him 
understand how an implementation (presumably Oracle's) "might" implement 
type-specific wrappers. Our (IBM's) implementation is also not using 
type-specific wrapper classes and as I said in my last reply to Blaise, 
the user shouldn't need to know or care about the type of the actual 
wrapper class or classes. All that matters, as far as the spec is 
concerned, is that this works:

DataObject intWrapperDO = dataFactory.create("commonj.sdo", "Int");
intWrapperDO.setInt("value", 7);
xmlHelper.save(intWrapperDO, "urn:example", "root", System.out);

The implementation, knowing it's a wapper, would serialize it correctly:

<root xmlns="urn:example">7</root>

The rest is implementation-dependent.

Frank.





"Barack, Ron" <> 
09/08/2008 05:02 PM

To
"Blaise Doughan" <>, <>
cc

Subject
AW: [sdo] Issue 118 - Proposal for review






Hi Everyone,
 
Sorry, when writing the agenda I somehow oversaw Frank's email describing 
his proposal.
 
I'm a little confused by the notion of the XSD generation of wrapper 
types, and by the creation of wrapper types through the API, as in the 
code samples below.  This is *much* more detailed that the original 
propsal.
 
We were thinking of implementing this with a single Wrapper type, and 
creating "value" properties... having the infrastructure create an 
arbitrary number of wrapper types (one for each simple type in SDO, and 
one for each simple type in the user's context) just sound like a mess to 
me.  Other implementations might decide differently.  I liked the original 
proposal, as it allowed for both approaches.  In both approaches 
Wrapper.isInstance(someObject) returns the same value.  And 
someObject.get("value") returns the simple value.  I don't see a use case 
for creating wrapper types in a portable way through an API.  And I sure 
don't see a use-case for generating XSD that includes wrapper types.  Are 
these directions where others want to go?
 
 
Best Regards,
Ron
 

________________________________

Von: Blaise Doughan [mailto:]
Gesendet: Mo 08.09.2008 21:28
An: 
Betreff: Re: [sdo] Issue 118 - Proposal for review



Hi Frank,

In order to support the proposed solution for SDO-118, we (Oracle) would
also require wording in the spec that indicates the XSD generation from
wrapper types is special, and is not required to look like the following:

  <xsd:complexType name="VendorIntWrapper">
    <xsd:complexContent>
      <xsd:extension base="sdo:DataTypeWrapper">
        <xsd:sequence>
          <xsd:element name="value" type="xsd:int"/>
        </xsd:sequence>
      </xsd:extension>
    </xsd:complexContent>
  </xsd:complexType>

-Blaise


Frank Budinsky wrote: