On Sep 29, 2008, at 10:45 AM, Reto Bachmann-Gmür wrote:
> Marc Hadley wrote:
>> It sounds like you are doing the correct thing. The generic type is
>> either the declared return type of the method
> Ok, so the follow-up question: is this the declared return type of
> the method in the root-resource class or of the method in the
> superclass or interface from where the annotations are taken (for
> the case where the method in the root-resource inherits the
> annotations but declares a more concrete return type)?
>
Good question. I can see arguments for either, what do others think ?
>> or the type returned by a GenericEntity instance.
>> Unless I'm misunderstanding, I think this is somewhat orthogonal to
>> issue 46 since that will only come into play during step 4 of
>> section 4.2.2 (sorting to find the most appropriate writer).
>> Writers can reject generic types they don't support in step 3 and
>> then won't be considered in step 4.
> I think to select the best of the possible combination of resource-
> method and MessageBodyWriter satisfying a request, we should know
> which writer will be selected before actually invoking a method.
>
Hmm, not sure if I understand you. You need to call the isWritable
method at least once for each distinct combination of type, generic
type, annotations and supported media type to determine if the writer
is able to serialize the entity.
Marc.
>> On Sep 29, 2008, at 10:04 AM, Reto Bachmann-Gmür wrote:
>>> Hello
>>>
>>> Rereading section 4.2.2 I'm insecure if its correct to look at the
>>> declared return type of a method (for those that are not Response
>>> or GenericEntity) for selecting the MessageBodyWriter.
>>>
>>> Currently I'm passing the type of the object returned at runtime
>>> as the "type" argument to Providers.getMessageBodyWriter and the
>>> declared return type as genericType.
>>>
>>> Does "for other return types it is the returned object." (section
>>> 4.2.2) means that only the class of the object returned at runtime
>>> is to be considered?
>>>
>>> Issue 46 I think would be easier to solve looking at the declared
>>> return-type only, but currently it would be most useful to
>>> consider both.
>>>
>>> Cheers,
>>> reto
>>>
>> ---
>> Marc Hadley <marc.hadley at sun.com>
>> CTO Office, Sun Microsystems.
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe_at_jsr311.dev.java.net
>> For additional commands, e-mail: users-help_at_jsr311.dev.java.net
>
>
---
Marc Hadley <marc.hadley at sun.com>
CTO Office, Sun Microsystems.