Re: Alissa Cooper's Yes on draft-ietf-httpbis-http2-16: (with COMMENT)

On 23 January 2015 at 18:55, Nicholas Hurley <hurley@todesschaf.org> wrote:

> I would argue that the amount of interoperability we have going on here
> indicates that the state machine is far from broken. Large and complex (as
> most protocol state machines are), perhaps, but definitely not broken.


The interoperability has been for the most part not on this part of the
state machine.   This part of the state machine has probably hardly been
tested outside of unit tests, as I don't think many connections will be
testing the "short period" of frame handling on the other side of a reset.

Getting protocols to work is the easy part.  Getting them to fail nicely is
the hard part.

I fail to see how adding the reset state adds complexity to implementations
that are already implementing the "short period" behavioural change.   It
will only add complexity to those that are not implementing that part of
the spec, which is exactly the sort of interoperability problem that would
be good to solve.





-- 
Greg Wilkins <gregw@intalio.com>  @  Webtide - *an Intalio subsidiary*
http://eclipse.org/jetty HTTP, SPDY, Websocket server and client that scales
http://www.webtide.com  advice and support for jetty and cometd.