RE: Working Group Last Call: Bootstrapping WebSockets with HTTP/2

I'm supportive of the general idea of the draft.  It feels like it's been a pretty short period to have it adopted, but perhaps that's okay since the draft itself is relatively straightforward.

Some bookkeeping discussion:

  *   Section 4 says that CONNECT is defined in RFC 7540, Section 8.3, which isn't quite accurate.
     *   7540 just has the description of how CONNECT gets mapped into HTTP/2's framing layer, because it's not quite the same as other methods.  This document probably doesn't need to actually modify 7540, because it's employing the extension framework defined therein.
     *   The CONNECT method itself is actually defined in 7231, section 4.3.6.  You're changing how the authority is interpreted in these requests; are we sure this doesn't need to be marked as modifying 7231?
  *   Section 5 "replaces the GET based request in [RFC6455]" and makes other modifications to the requirements contained therein. This seems to suggest that it updates RFC6455 as well.

Random thoughts:

  *   This really is a protocol switch.  Part of me wants the successful response to be 101, because you are employing a different protocol on that stream.

I have a few editorial nits, but I'll send a PR for those.

-----Original Message-----
From: Mark Nottingham []
Sent: Wednesday, March 28, 2018 2:48 PM
To: HTTP Working Group <>
Cc: Patrick McManus <>
Subject: Working Group Last Call: Bootstrapping WebSockets with HTTP/2

Hi everyone,

Patrick (as editor) has incorporated the discussion from London and believes this is ready for WGLC; there are no open issues.

Please have a look at:

... and bring up any issues on-list or on its issues list; likewise statements of support (or otherwise) for publication on-list would be appreciated.

WGLC will end on 12-4-2018.



Mark Nottingham

Received on Wednesday, 28 March 2018 22:44:17 UTC