On Jan 2, 2008, at 12:09 PM, Dirk Haun wrote:
Apologies if I'm missing anything obvious, but I'm just sitting here
brooding over an Atom Protocol Exerciser Report and wondering "what
does
it want from me?"
The APE's Mini One / Mini Two test apparently POSTs a second entry for
an already existing one (with the same id). Since RFC5023 states that
POST is for creation and PUT is for modification, I'm wondering what
to
do in that situation.
Oh hell yes, that's a bug. Foo. Apparently enough server
implementations ignore client-provided IDs that nobody noticed
before. I don't think that what the Ape's doing is strictly forbidden
by the RFC, but it's clearly bad practice and a server would
definitely be within its rights in throwing the dupe-ID POSTs on the
floor.
Fixed, thanks. -Tim
