On Saturday 28 May 2005 12:42, Rusty Shackleford wrote: > Browsers don't "listen". They inititiate a connection, process the > requested transaction with the web server, and close the connection. The > simply can't be used to "listen" for an arbitrary connection.
Don't be so closed-minded. Yes, technically you're absolutely correct. However it's trival to simulate 'listening', especially in this kind of situation. > Again, this is not the way to do this. Dozens, or hundreds of clients > constantly hammering a server with "Have you got anything new? No? > OK..." messages every couple of seconds is an excellent example of how > NOT to design a system. Yes, you can get away with it, if the resources > involved are not an issue, but I think it fair to assume that for many > interested in this discussion, resources like bandwidth and CPU usage > ARE issues. Nonsense. Ever hear of poll()? It accepts a timeout. You make your request and let it block for, say 3 seconds plus a random (1-1500) ms. With any kind of modern processor running on the server you will be very hard pressed to kill it with requests for data. > Quite. Alas, that means an instance of SOMETHING on the client that can > listen for, and respond to, arbitrary events. Nah; see above. I am very much against clients having to listen for connections because it introduces all kinds of problems with NAT. Unfortunately it's a reality and has to be worked around and port forwarding and firewall tricks just aren't a good solution. -A. _______________________________________________ Asterisk-Users mailing list [email protected] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
