- From: Poul-Henning Kamp <phk@phk.freebsd.dk>
- Date: Thu, 26 Jan 2012 11:20:31 +0000
- To: Willy Tarreau <w@1wt.eu>
- cc: Amos Jeffries <squid3@treenet.co.nz>, ietf-http-wg@w3.org
In message <20120126103013.GC8887@1wt.eu>, Willy Tarreau writes: >On Thu, Jan 26, 2012 at 09:48:06AM +0000, Poul-Henning Kamp wrote: >That doesn't change the fact that the recipients must decompress what >they receive to understand what it talks about if the sender is using >compression in a non-null mode. If the client don't want compression, it can just refrain from initiating a compressed session, and I'm prefectly happy giving servers a way to announce their abilities. > - user agent : no need to compress it, just specify a new non-abusive format Actually, it should go away, it is totally counter to the idea that your content would depend on the other end. if anything, replace it by a client-header that says something intelligent about rendering abilities and area: Render: text, graphics=color, area=1024x768, javascript Render: text=color, javascript > - cookies : the largest ones are those handling an encrypted context so you > won't compress them They're usually hexdumped, so I would compress them close to 50%. >You can report information on chunk boundaries. It would obivously have to be at chunk boundaries to remain in sync. -- 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 Thursday, 26 January 2012 11:21:06 UTC