Hi Stephen,
some other users have pointed us to this problem before that had been
included in the 0.9.9 release of castor. As far as I know this is fixed
in CVS and we intend to build a new bugfix release till end of this
month. You may want to have a look at issues 1152, 1153 and 1216 which
had recently been resolved.
Having said that we would appreciate if you could test the CVS version
to verify if your problem is also fixed.
Regards
Ralf
Stephen Dennison schrieb:
> This was a tough one to explain, so if the following isn't clear, feel
> free to ask me to elucidate further.
>
>
> I have a schema that has a simpleType that contains an attribute that
> has a ref to an attribute group. In castor 0.9.5.2 this resulted in
> code being generated for the attribute as though it were a simpletype,
> but in castor 0.9.9 it just calls the field an Object and doesn't
> generate code for it.
>
>
> Here's an example of the setup:
> I have one schema that contains something like this:
> ...
> <xsd:attributeGroup name="someStuff">
> <xsd:attribute ref="classification" use="required"/>
> <xsd:attribute ref="markings" use="optional"/>
> </xsd:attributeGroup>
> <xsd:attribute name="classification>
> <xsd:simpleType>
> <xsd:restriction base="xsd:string">
> <xsd:enumeration value="one"/>
> <xsd:enumeration value="two"/>
> <xsd:enumeration value="three"/>
> </xsd:restriction>
> </xsd:simpleType>
> </xsd:attribute>
> <xsd:attribute name="markings">
> ...
> </xsd:attribute>
> ...
>
> And I have another schema that has something like this:
> ...
> <xsd:complexType name="someComplexType">
> <xsd:attributeGroup ref="theCorrectNS:someStuff"/>
> </xsd:complexType>
> ...
>
> So, in the 0.9.5.2, A class was generated for
> someComplextTypeSomeStuffType that would allow the values one, two, or
> three where one was defined as a static
> someComplextTypeStuffType("one"), two was a static
> someComplextTypeStuffType("two"), etc. (not sure of the actual
> case/capitalization here, but you get the idea)
> This class was referenced by the someComplextType class. like this:
> ...
> public class someComplexType implements java.io.Serializable{
> ...
> someComplexTypeSomeStuff _classficiation;
> ...
> }
>
> In castor 0.9.9, the same class looks like this:
> ...
> public class someComplexType implements java.io.Serializable{
> ...
> java.lang.Object _classification;
> ...
> }
>
>
> For now, I can get a compatible class generated if I make simpleTypes
> that restrict base type xsd:string, name them what I expect should
> have been generated, put my enumerated values in them, and manually
> add all of the attributes to someComplexType. The problem with this
> is that I have to locate every place where we are using an attribute
> group and manually build the simple types by inspecting all of the
> schemas that we include.
>
>
>
> So, is this a bug, or is there a setting somewhere that I don't know
> about that I need to turn on, or a syntax fix that I need, in order to
> get backwards-compatible classes generated?
>
> -------------------------------------------------
> If you wish to unsubscribe from this list, please
> send an empty message to the following address:
>
> [EMAIL PROTECTED]
> -------------------------------------------------
-------------------------------------------------
If you wish to unsubscribe from this list, please
send an empty message to the following address:
[EMAIL PROTECTED]
-------------------------------------------------
