On Wed, 2006-10-11 at 13:31 -0400, Ajith Ranabahu wrote: > Hmm... > The instnace XML is indeed valid but it seems the way we handle things > (atleast the assumption we made) is not correct. Right now the > xsd:anyType generates an OMElement as the field (which essentially > assumes that anyType will be a complex type!). But as this example > shows it can be a primitive or derived type and our deserialization > will fail. > The most important issue here is that even if we manage to get the > deser right, the generating of OMElement would not be right in this > case. (How can you assign an integer or string to an OMElement ??) > Would java.lang.Object be a good choice ?
OK I must be missing something: if there's xs:anyType on an xs:element declaration then it can always be represented by an OMElement. Its an *element*! Also, IIRC xs:anyType cannot be a simple type- if its xs:anySimpleType then it can be any simple type but if you want it to be either a simple type or a complex type I think you have to say xs:ur-type or something like that. But that's not the problem here - you should be able to make this work just right with an element. If you want to set its value to 173 then you have to create a new text node containing the string value "173" and stick that underneath the element. What's the problem? Sanjiva. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
