- From: Willy Tarreau <w@1wt.eu>
- Date: Sun, 6 Aug 2017 07:02:19 +0200
- To: Julian Reschke <julian.reschke@gmx.de>
- Cc: Adam Roach <adam@nostrum.com>, Kazuho Oku <kazuhooku@gmail.com>, The IESG <iesg@ietf.org>, draft-ietf-httpbis-early-hints@ietf.org, Mark Nottingham <mnot@mnot.net>, httpbis-chairs@ietf.org, HTTP Working Group <ietf-http-wg@w3.org>
On Sat, Aug 05, 2017 at 10:07:08PM +0200, Julian Reschke wrote: > On 2017-08-05 21:26, Adam Roach wrote: > > On 8/3/17 9:31 PM, Kazuho Oku wrote: > > > Consider the case where a proxy that cannot correctly handle an > > > informational response is involved. If the client sets a HTTP header > > > indicating that it is capable of receiving 103, the proxy will simply > > > pass through the header. Therefore, it would become a false signal to > > > the server. > > > > > > Thanks for the explanation. This seems reasonable. I'm still a bit > > uneasy about the implied encouragement to user-agent-string sniffing, > > but the intended status of "experimental" makes me worry about it much > > less than I would otherwise. Hopefully, we can learn from this > > experiment how pervasive incorrect handing of 100-class responses in > > general is for HTTP clients. > > Quite pervasive. See, for instance, > <http://bugs.java.com/bugdatabase/view_bug.do?bug_id=8170305>. At least fixing broken existing code is not hard thanks to 100 having been present for a long time, for example here : http://git.haproxy.org/?p=haproxy.git;a=commit;h=a14ad72 :-) Willy
Received on Sunday, 6 August 2017 05:02:56 UTC