Re: Fail generating classes: java.lang.IllegalStateException

[Nacho G. Mac Dowell]

Hi Carolina, I was taking a look at mpeg7_tva.xsd and I can see a 
potential problem for a proper binding and this is a recursion problem. 
Lets take a closer look to TermDefinitionType (which apparently is 
causing the problem):

 <complexType name="TermDefinitionType">
   <complexContent>
     <extension base="mpeg7:TermDefinitionBaseType">
       <sequence>
         <element name="Term" minOccurs="0" maxOccurs="unbounded">
           <complexType>
             <complexContent>
               <extension base="mpeg7:TermDefinitionType">
                 <attribute name="relation" 
type="mpeg7:termRelationQualifierType" use="optional" default="NT"/>
               </extension>
             </complexContent>
           </complexType>
         </element>
       </sequence>
     </extension>
   </complexContent>
 </complexType>

The schema is extending a recursive element before it is actually 
processed. Although this is a correct schema, from a binding perspective 
it seems to me as a dead-end. We could "translate" this schema to an 
equivalent one like this:

 <complexType name="TermDefinitionType">
   <complexContent>
     <extension base="mpeg7:TermDefinitionBaseType">
       <sequence>
         <element name="Term" minOccurs="0" maxOccurs="unbounded" 
type="mpeg7:TermDefinitionTypeNode">
         </element>
       </sequence>
     </extension>
   </complexContent>
 </complexType>

 <complexType name="TermDefinitionTypeNode">
   <complexContent>
     <extension base="mpeg7:TermDefinitionBaseType">
       <attribute name="relation" 
type="mpeg7:termRelationQualifierType" use="optional" default="NT" />
     </extension>
   </complexContent>
 </complexType>

This definition would validate any xml that validated the former. 
Further investigation would be required to have a valid proof, but I 
think it is clear enough that it would produce the same xml. I found 
very interesting taking a look at this schema and I wanted to ask you if 
I can post a question in the w3c mailing list with this example. IMHO 
recursion-limits are not well defined in xml-schema. The schema 
specification just states a couple of restrictions to recursion which I 
find insufficient.
I remember when recursion was developped for JaxMe the following comment 
arised:

If the child is of the same type, but has been extended: what would 
happen?
Good question, which i delegate for later. ;-)
Here we are! The bad news is that IMHO the schema provided would be very 
hard (impossible?) to bind. The good news is that there will always be 
an isomorphic schema. So, to the point now: The schema will now compile 
with another change: anyType doesn't seem to be well supported 
(Jochen?). There is another workaround for this: just eliminate the 
restriction and it will support anyType (strange enough). So where 
you've got:

   <complexType name="Mpeg7BaseType" abstract="true">
       <complexContent>
           <restriction base="anyType" />
       </complexContent>
   </complexType>
just type: <complexType name="Mpeg7BaseType" abstract="true"/> which is 
semantically identical (as far as I know anyType is the default type for 
a complexType).

Now your schema compiles! You have to be aware though that there is no 
root element so no xml document will conform to this schema.

An obviest drawback of all this is that the schema you are using seems 
to be a public one... so you may dislike "adapting it for JaxMe".

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