- From: Daniel Stenberg <daniel@haxx.se>
- Date: Wed, 4 Sep 2013 23:29:19 +0200 (CEST)
- To: Martin Thomson <martin.thomson@gmail.com>
- cc: HTTP Working Group <ietf-http-wg@w3.org>
On Wed, 4 Sep 2013, Martin Thomson wrote: >> The effect seems to risk a little more data waste in the HTTP+upgrade case. > > Yeah, I'm thinking that for https: URLs it's easy. ALPN means that there is > no probing. And you get the option to use flow control. In fact, that's > actually much nicer. Yes I agree that the https method of deciding protocol is nicer and less problematic. But I forsee a future where people will use the same operations like today but only enable http2, and then the URL will remain as-is which often means http://. > For http: URLs, it's a little trickier. You might do a HEAD request for the > upgrade request, followed by a POST after the upgrade completes. That's > round trips still, and if the upgrade fails, it's even worse. Though, I'll > note that it's still the same speed in the success case as the https: URL > case ... I've already been burned in the past trying to do HEAD when a POST is asked for. There are unfortunately too many servers out there that accept POST but refuse HEAD for the same URI so it isn't a way I would like to go either... But again, I don't think this is a big problem. -- / daniel.haxx.se
Received on Wednesday, 4 September 2013 21:29:53 UTC