- From: Mark Nottingham <mnot@mnot.net>
- Date: Wed, 28 Jan 2015 19:42:06 +1100
- To: "Julian F. Reschke" <julian.reschke@gmx.de>
- Cc: Martin Thomson <martin.thomson@gmail.com>, Michael Sweet <msweet@apple.com>, Bence Béky <bnc@chromium.org>, Erik Nygren <erik@nygren.org>, HTTP <ietf-http-wg@w3.org>
> On 28 Jan 2015, at 7:39 pm, Julian Reschke <julian.reschke@gmx.de> wrote: > > On 2015-01-28 09:33, Mark Nottingham wrote: >> On 15 Jan 2015, at 5:24 pm, Julian Reschke <julian.reschke@gmx.de> wrote: >>> >>> On 2015-01-15 04:37, Martin Thomson wrote: >>>> On 14 January 2015 at 17:52, Michael Sweet <msweet@apple.com> wrote: >>>>> Haven't looked at the changed text, but the maximum length of a FQDN is 255 octets... >>>> >>>> Indeed it is. But origin includes <scheme>://<fqdn>[:<port>]. A >>>> 9-bit length would suffice... >>> >>> Stop it. Let's not mix binary and text in the same frame. I'll add a "origin" parameter to the header field syntax. >> >> I'm personally -0.5 about that; putting the origin in the header field syntax when it isn't useful in the header field's semantics is bound to cause confusion. >> >> My .02 - decide on a generous bit length and we're done this one. >> >> Cheers, > > Having two different approaches in the same frame (binary and text) makes me shudder... Well, the frame *is* binary; nothing's going to change that. And, the content *is* text -- likewise. The question is whether there's one text field or two... Cheers, -- Mark Nottingham https://www.mnot.net/
Received on Wednesday, 28 January 2015 08:42:41 UTC