Re: New Version Notification for draft-thomson-http-replay-00.txt

On 25 July 2017 at 22:27, Benjamin Kaduk <> wrote:
>> * 0-RTT request with no early-data header: postpone request processing
>> until ClientFinished
> That seems permissible, but sending 4NN is also fine and requires less state
> on the server.

This is our "legacy" case.  That is, clients that do 0-RTT today
without this new mechanism.  The cost to sending 4NN is that it causes
the client to use generic 400 processing, which probably doesn't
include a retry.  That's not the end of the world, though it probably
means that the response is a terminal failure as opposed to a
recoverable one.

Received on Wednesday, 26 July 2017 00:27:20 UTC