sockets

[Kenneth Brody]

On 5/18/2012 1:51 PM, Ed Hilovsky wrote:
>
> I think FP needs to be a bit more forthcoming regarding this. You sell a
> socket license. (but you don't sell socket connections) If a server has a
> license limit of 25 connections, and there are 25 connections, when the 26th
> connection is attempted, it will fail.

Can you install that "server" on 25 different computers, as long as each one 
only uses 1 connection?

> But when one of those 25 does a
> socketclose then the 26th ( if it retried) would connect.

And when one of the 25 *clerk/*report sessions exit, the 26th one could gain 
access to the socket functionality.

> You "REQUIRE" a certain method of programming to make your method work.

Huh?

> Just
> because I use a CALL statement, you can't understand that that "process" is
> closed and the socket was also.

Again, we are *not* counting the number of *connections*.  The fact that you 
have closed one of the *connections* is irrelevant.

> Obviously, FP is right and we all are wrong.

When you say "filePro doesn't count X correctly" when the fact is that 
filePro doesn't count "X" at all, but rather "Y".

Again, if you have a 25-user runtime, and your processing uses the main file 
plus 5 others via lookup "connections", are you using 1 runtime license or 6?

[...]

Now, if your point is that filePro *should* count X and not Y, then you'll 
need to discuss that with TPTB.  But, to insist that filePro *is* counting 
X, and doing it incorrectly, then yes, you are wrong.

-- 
Kenneth Brody