- From: Poul-Henning Kamp <phk@phk.freebsd.dk>
- Date: Sun, 16 Oct 2016 07:08:24 +0000
- To: Willy Tarreau <w@1wt.eu>
- cc: Martin Thomson <martin.thomson@gmail.com>, Matt Menke <mmenke@google.com>, HTTP Working Group <ietf-http-wg@w3.org>
-------- In message <20161016052641.GA28987@1wt.eu>, Willy Tarreau writes: >On Sat, Oct 15, 2016 at 10:26:04PM +0000, Poul-Henning Kamp wrote: >> >And what about #4 consisting in indicating the encoding in the header >> >field *name* instead >> >> That's quite a hack, isn't it ? > >Not that much after all when you think about it, because our parsers >will have to consider the field name to know how to decode the value. At the semantic level, yes. But not at the serdes level. And if we pull off the "data dictionary" idea, we can machine generate parsers which do that for specific cases. But I would hate to make it part of the definition. If nothing else, because it takes the 19 existing headers out of the CS cover. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 phk@FreeBSD.ORG | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence.
Received on Sunday, 16 October 2016 07:08:52 UTC