Re: POST for an exisiting id?

["Thomas Broyer" <t.broyer@xxxxxxxxx>]

2008/1/3, Dirk Haun:
>
> Thomas Broyer wrote:
> >
> >I don't know what the APE waits for, but 400 is fine. 409 (Conflict)
> >or 422 (Unprocessable entity; from RFC 2518) might feel better but 400
> >is good enough.
>
> I'm not fluent in Ruby, but looking at the source for the APE (def
> test_sorting), I get the impression that this test is actually trying to
> post 3 separate new entries and then do something with them. So it
> expects the second post to succeed.
>
> So this is either a bug in the APE (forgetting to provide different IDs
> for the 3 entries) or a problem on our end (are we expected to auto-
> assign new IDs when one already exists?).

Some implementations overwrite atom:ids (i.e. they don't trust what
they're sent), but that's certainly not a must: you're free to use the
provided atom:id and/or then reject duplicates with a 4xx.
So if you're right about what the APE's doing, I'd say it's a bug in the APE.

Tim?

-- 
Thomas Broyer