Keepalive in client api.

• From: Anselm Baird_Smith <abaird@www43.inria.fr>
• Date: Mon, 30 Jun 1997 10:03:01 +0200 (MET DST)
• To: Alexandre Rafalovitch <arafalov@socs.uts.EDU.AU>
• Cc: www-jigsaw@w3.org
• Message-Id: <199706300803.KAA19456@www43.inria.fr>

Alexandre Rafalovitch writes:
 > Hi,
 > 
 > Does anybody understand how keepalive works on Client site api. I tried to
 > trace my way through all the Factories and choices and got completely
 > lost.
 >
 > I am trying to understand 2 things:
 > 1) How does client site figures that it can use keep-alive
 > 2) Does it support http1.0 keep-alive as well as http1.1 and if it does,
 > do I have to do anything special to activate it? 
 > 
 > Basically, I want to use proper keep-alive whichever version of http, the
 > server I am talking to is using.

Hi Alex,

Jigsaw client side API does this for free (well, for the complexity
you saw in the code). As far as I remember:

For each new queried server, the HttpManager will create an instance
of some HttpServer (the most usefull one being HttpBasicServer). That
basic server negotiates with the manager to check if it can keeep the
connection open (the manager will close least recently used
connections when some limit is reached).

Does this answer the question ?
Anselm.

Received on Monday, 30 June 1997 04:03:32 UTC