W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 2013

Re: :scheme and CONNECT method

From: 陈智昌 <willchan@chromium.org>
Date: Tue, 13 Aug 2013 19:53:59 +0200
Message-ID: <CAA4WUYhJPXj9oJMzPFy3ewhscOHJnYZo4z7OP4O85P5dNsPbrg@mail.gmail.com>
To: Martin Thomson <martin.thomson@gmail.com>
Cc: ietf-http-wg@w3.org, Tatsuhiro Tsujikawa <tatsuhiro.t@gmail.com>
On Aug 13, 2013 7:35 PM, "Martin Thomson" <martin.thomson@gmail.com> wrote:
>
> On 13 August 2013 15:42, William Chan (陈智昌) <willchan@chromium.org> wrote:
> > I suggest opening an issue for this.
>
> https://github.com/http2/http2-spec/issues/230

Thanks!

>
> Just spit-balling, but I think that the '*' form is a better way to
> deal with this, i.e.,
>
> :method = CONNECT
> :scheme https
> :host example.org:443
> :path *

All in all, I don't particularly care. That said, I don't know if it's a
big deal, but I think there might be information loss that way. AIUI, the
Host header may be present even for CONNECT. So my only concern is possible
compatibility issues, which would be silly, but maybe possible.

>
> The trick here is that the DATA frames that form the "request body"
> constitute a separate channel that is reliable, in-ordered, etc...
> This channel can be used to perform a TLS handshake and run any
> protocol you like.

Yep, sorry for not spelling this out before.
Received on Tuesday, 13 August 2013 17:54:26 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:14:14 UTC