sockets - opening and closing

[Dar Scott]

On Friday, May 3, 2002, at 08:33 PM, Shao Sean wrote:

>> I've seen this kind of thing when the server is the same computer.
> i'm on the server machine (server and client same machine) and another
> client in australia (can't physically get further away from me 
> then that ;-)

I noticed that a Revolution TCP client on W2K will time out in 21 
seconds.  Even Australia to Canada should connect in less than 
that, I think.

In my situation, the connection is actually established.  You might 
try checking with repeated netstat -p TCP.  Or use TCPmon to watch 
connections come and go.  Or, if your server is relatively quiet, 
look at traffic with Ethereal or a similar tool.

>
> --
>> Could the server have put you on a black list?
> i don't have any such feature... yet =)

A stab in the dark.
>
> --
>> Are you sure you are closing?  Maybe you have hit the max number of
> closing yes, there's only myself and matt on the server, plus 
> there's no
> internally coded limit to the number of users available to connect

I had to say that; we are on the use-revolution list.
>
> --
>> For what I have seen with the same-computer problem, just dragging
> maybe i'll just throw in a dummy handler before calling the open 
> socket..?

I tried doing all the same things by script and it did not work.  
(Assuming we are seeing the same thing.)

The behavior on W2K is not the same as on OS X, the only other 
platform I checked.  On OS X, an open will get a socket open 
callback upon establishment, a socketClosed upon refusal, or a 
socketClosed in 75 seconds otherwise.  On Windows 2000, an open 
will get a socket open callback upon establishment (except for your 
and my situations ;-), a socket open callback if refused (delayed 
.9 seconds), and an open callback in 21 seconds if nobody is home.

Dar Scott