Re: squid rewrite

From: Dean Gaudet <dgaudet-list-squid-dev@dont-contact.us>
Date: Tue, 10 Jun 1997 01:41:33 -0700 (PDT)

On Tue, 10 Jun 1997, Stewart Forster wrote:
> This is a problem, BUT squid is not really CPU bound, so the problem
> is not as big as it seems.

Actually the only reason I brought this point up was because I expect that
you may run into cpu-bound problems when you delve into http/1.1. For
example you may have to do unchunking if the client is 1.0 and the origin
is 1.1. Or you may have to chunk if the client is 1.1 and the origin is
1.0. You may want to allow for inflation (only jigsaw supports deflation
at the moment, but it's "easy" to add to apache, especially if you
pre-deflate the files).

> Uh. BIG problem here. You can't handover connected sockets between
> two already running processes.

As others have pointed out there is a method of doing this... and if you
look at NCSA HTTPD source code you'll see they've implemented it across a
wide variety of platforms. It's not that efficient though afaik.

Dean
Received on Tue Jul 29 2003 - 13:15:41 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:11:20 MST