W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 2014

Re: 1xx (informational) responses affect on stream management

From: Martin Thomson <martin.thomson@gmail.com>
Date: Tue, 5 Aug 2014 14:43:20 -0700
Message-ID: <CABkgnnWd3-w1QUJrjLsyY6dXcAGPty5oOG=1yYUo28oNDHx3dg@mail.gmail.com>
To: Simone Bordet <simone.bordet@gmail.com>
Cc: Matthew Cox <macox@microsoft.com>, "ietf-http-wg@w3.org" <ietf-http-wg@w3.org>
On 5 August 2014 14:30, Simone Bordet <simone.bordet@gmail.com> wrote:
> I don't understand this answer in the context of the original
> question, can you please expand ?

Matthew seems concerned about having to enforce frame cardinality
rules based on frame contents.  This is a legitimate concern.
Previously, a request or response was one HEADERS, some number of DATA
and maybe another HEADERS.  Simple.

With the 1xx change, now you can have 0..n HEADERS frames before that
sequence and you can't know if they are valid until a) you look at
:status, and b) you receive the next frame (because a non 1xx status
is only valid if it is followed by DATA, or maybe trailers for the
particularly cunning or perverse).

All I'm suggesting is that you can divide the problem into two: look
at the frames, if the count is good pass them on.  Then, at the HTTP
layer, when you are looking for other things, you can check :status,
along with all the other pseudo-header stuff.
Received on Tuesday, 5 August 2014 21:43:48 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 30 March 2016 09:57:09 UTC