issue146 Specifiy JAX-RS request methods on annotations #198
Conversation
tck/src/main/java/org/eclipse/microprofile/lra/tck/participant/api/ContextTckResource.java
Show resolved
Hide resolved
| - A JAX-RS method can return a `202 Accepted` HTTP status code. | ||
| If there is no `@Status` method then the response MAY provide | ||
| a status URL in the `HTTP Location` header field so that the | ||
| implementation can discover the final outcome. |
There was a problem hiding this comment.
Ok, During the discussions I always had the impression that the location Header would be made available in the case when there IS @Status method available.
In this scenario, the Location Header becomes an alternative for the @Status method and reinvoking the Compensate method and makes sense.
| This URL, if present, MUST obey the requirements specificed in | ||
| the javadoc for the <<source-Status,Status annotation>>. | ||
| If the developer has not provided an `@Status` method nor a status | ||
| URL then the implementation may reinvoke the `@Compensate` method |
There was a problem hiding this comment.
may? It MUST reinvoke as it otherwise has no way of retrieving the final status.
| the javadoc for the <<source-Status,Status annotation>>. | ||
| If the developer has not provided an `@Status` method nor a status | ||
| URL then the implementation may reinvoke the `@Compensate` method | ||
| (ie it MUST be idempotent). |
There was a problem hiding this comment.
The text after the paragraph describing Compensate says:
Similarly if the resource cannot perform a completion activity immediately.
There was a problem hiding this comment.
I think I've resolved everything here so @rdebusscher I will merge the PR at close of business today unless you have any further issues with it?
#146
This PR clarifies which JAX-RS request methods can be used on annotations.
@rdebusscher I added a note about including an optional status URL in
202 Acceptedresponses to cover the situation where there is no@Statusmethod. If you accept this text then I don't think we need a new issue for it.I also did not change any text that discussed idempotency since we are still discussing that on issue #187