Re: [Editorial Errata Reported] RFC7540 (4871) from Patrick McManus on 2016-12-01 (ietf-http-wg@w3.org from October to December 2016)

From: Patrick McManus <pmcmanus@mozilla.com>
Date: Thu, 1 Dec 2016 10:21:00 -0500
To: Martin Thomson <martin.thomson@gmail.com>
Cc: Kazuho Oku <kazuhooku@gmail.com>, Cory Benfield <cory@lukasa.co.uk>, RFC Errata System <rfc-editor@rfc-editor.org>, Mike Belshe <mike@belshe.com>, Roberto Peon <fenix@google.com>, Ben Campbell <ben@nostrum.com>, Alissa Cooper <alissa@cooperw.in>, Alexey Melnikov <aamelnikov@fastmail.fm>, Patrick McManus <pmcmanus@mozilla.com>, Mark Nottingham <mnot@mnot.net>, HTTP Working Group <ietf-http-wg@w3.org>
Message-ID: <CAOdDvNqw98X5ttGFVE-HkwL30bRj2XEnv9WxZ8N5y3HeAzzyxA@mail.gmail.com>

On Thu, Dec 1, 2016 at 4:33 AM, Martin Thomson <martin.thomson@gmail.com>
wrote:

> Yeah, Patrick decided that he would use weights to fix that.  In your
> example, all the HTML would be dependent on an empty node that has
> weight 1, which would be adjacent to the JS/CSS.  In theory the HTML
> does get some share.
>

and I have to say that in general I have a preference for "some small
share" systems rather than strict ordering. It tends to turn priority
inversion problems from deadlocks into performance corner cases. Much safer.

> I think that much more complexity and we might as well just upload
> javascript for the server to run.
>

:) - that's the kind of thinking that gave us PAC.

Received on Thursday, 1 December 2016 15:21:32 UTC