Hi David,
I understand the trouble and will at least improve the Javadocs of @Post and
@Put to mention this extra constraint. This change was made (late in the dev
process I agree) to clarify the situations where several @Post annotations
are defined for the same interface/resource and we need to select the best
one based on the entity media type.
Also, if we want to be able to reuse the interfaces on the client side, then
an input parameter that will get converted to an entity is necessary (unless
there is no entity for the POST method in which case we allow the lack of
input parameter). This was close to an API design bug from my point of view,
hence the late change.
On the positive side, the adjustment for existing code should be really
simple in all cases. For you, just do this:
@Post
public void handleSomePost(Form form) {
doSomethingWithAForm(form);
.....
}
Best regards,
Jerome
2010/7/12 David Fogel <[email protected]>
> Hi Jerome-
>
> We just updated to trunk (revision 6751), and many of our annotated
> @Post methods in our application have stopped working!
>
> Based on commit comments, the relevant changes are sometime in the
> last 24 hours (revisions 6744 and 6745) made to AnnotationInfo.java,
> where your comment was:
> - When selecting an annotated method, if the request has an
> available entity, now the Java methods without an input
> parameter are rejected.
>
> Throughout our current project, we have many annotated methods that
> look like this:
>
> @Post
> public void handleSomePost() { // <-- note no params
> doSomethingWithAForm(new Form(getRequest().getEntity()));
> .....
> }
>
> This has worked fine for us throughout most of the 2.0 process, and I
> haven't seen any documentation that prohibits this.
>
> While I understand the logic behind @Post methods taking a single
> parameter representing the request entity body, it's not at all
> obvious that a Post handler without any params should be considered
> incompatible with a request that contains an entity. Additionally, it
> seems late in the game to change this logic- although if it turns out
> that we're the only people using this pattern, then we can certainly
> make the adjustment- but I'd recommend documenting the
> annotation-processing very thoroughly, especially for this case.
>
> Thanks,
> -Dave Fogel
>
> ------------------------------------------------------
>
> http://restlet.tigris.org/ds/viewMessage.do?dsForumId=4447&dsMessageId=2632190
>
------------------------------------------------------
http://restlet.tigris.org/ds/viewMessage.do?dsForumId=4447&dsMessageId=2632339
