- From: Roy T. Fielding <fielding@gbiv.com>
- Date: Thu, 10 Dec 2020 09:59:09 -0800
- To: Willy Tarreau <w@1wt.eu>
- Cc: Julian Reschke <julian.reschke@gmx.de>, HTTP Working Group <ietf-http-wg@w3.org>
> On Dec 10, 2020, at 7:35 AM, Willy Tarreau <w@1wt.eu> wrote: > > Hi Julian, > > On Thu, Dec 10, 2020 at 02:05:35PM +0100, Julian Reschke wrote: >> So, if you have >> >> Accept: text/html;level=1;q=0.5 >> >> "level" is a parameter of the media type "text/html". >> >> On the other hand, in >> >> Accept: text/html;q=0.5;level=1 >> >> "level" is an accept-extension. >> >> Has anyone ever *seen* this in use? > > Wow, I'm pretty sure I never noticed this. Even the q= I've probably > seen it more often in specs than in traces, so the two combined should > be extremely rare.. > >> 0) leave things as they are >> >> 1) note that this is not in use, advise not to send it, and advise >> recipients to ignore it (essentially deprecating it) >> >> 2) kill it completely >> >> 3) ...? > > Not having ever met it I could be in favor of 2 but if it used to be > valid it's a bit harsh for past implementers who tried hard to comply > with the specs so probably 1 would be more suitable (like we did for > the chunk extensions). > > Just my two cents, > Willy FWIW, I have never seen it used in practice. IIRC, I added accept-ex due to a hypothetical question about how to distinguish between the media type parameters and other global parameters we might add in the future to be like q. We never added such parameters, and I doubt that we could do so now without breaking someone. So, I'm +1 to remove them. But if anyone has seen them in use, please let us know. Cheers, ....Roy
Received on Thursday, 10 December 2020 17:59:31 UTC