Re: Netscape V3.0.x has persistent connection troubles with Squid 1.2

--MimeMultipartBoundary
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Have we considered using the OPTION method (http/1.1) to interrogate
peer capabilities? (I know this question is rather tangential to the
thread, but it's nearly dawn and my brain doesn't work in a terribly
straightforward way when it gets too close to sunlight)

Being able to query a peer fairly simply for capabilities would seem to
be a fairly handy thing. Having done so, you can assure yourself of all
sorts of things, like the peer's ability to handle persistant
connections, content-encoding capabilities, binary headers, and all of
that.

Though, right at this moment, as far as patching around the user-agent
goes, apache looks out for the rogue netscape version(s?), so it won't
do persistant connections for anything that carries that user-agent. If
it won't do it, is there much to be gained if we do? (short answer:
maybe...persistant connections are more efficient, but the client isn't
hugely widespread [<- wild-assed guess. I don't know how widespread it
might be])

So, assuming we work around the problem of cache-to-cache persistant
transfers...if there is only one cache in the chain, the odds are that
the bad client versions won't get any persistancy anyway. Only if there
are two or more persistance-capable caches in the request-chain.

Of course, in the setup I run through there are about 3-7 squids before
you hit the outside world.

D

Alex Rousskov wrote:
>
> On Sun, 24 May 1998, Stewart Forster wrote:
>
> > The patch I sent simply ignores the Keep-Alive request directive
> > coming from a Netscape 3 client, but allows persistant connections between
> > all other kinds of caches/proxies/browsers.
>
> Hm, perhaps I've missed something, but if your decision to ignore
> "keep-alive" is based on User-Agent header alone, then _all_ proxies in the
> request processing chain will ignore "keep-alive" because User-Agent is
> passed through by a proxy unchanged. You might distinguish between
> Proxy-Connection and Connection headers somehow (assuming clients never send
> Proxy-Connection), but I doubt it is a robust solution.
>
> Alex.

-- 
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GAT d- s++: a C++++$ UL++++B+++S+++C++H++U++V+++$ P+++$ L+++ E-
W+++(--)$ N++ w++$>--- t+ 5++ X+() R+ tv b++++ DI+++ e- h-@ 
------END GEEK CODE BLOCK------
--MimeMultipartBoundary--