In a message to Phil Crown Mike Bilow wrote:

 PC> How about using a timer in the REXX script itself?

 PC>   call time('e') /* start the timer */

 PC>   call time('r') /* returns elapsed secs since time('e') */

 PC>   if result > MAXTIMEALLOWED then ...

MB> This will not achieve what I want.  The goal is to time out a Rexx
MB> function call -- in the RXFTP library, as it happens -- if it takes
MB> too long.  The idea is that I am starting an FTP session and do not
MB> want to hang forever if the FTP server goes away unilaterally. 
MB> Ideally, the RXFTP library would have the necessary timeout facility
MB> built into it, but it doesn't.  I am willing to have the whole Rexx
MB> program die if a timeout expires; I just want to guarantee that the
MB> Rexx program terminates somehow within a specified worst case
MB> maximum amount of time.  If the Rexx program is still running well
MB> enough to make calls to the time() function, then I don't need to
MB> kill it.

Why not launch a separate process for your function and leave the mail
program waiting.. if the function ends normally, the process signals the
program to continue; if it times out, the main program closes in fail mode.


         Cheers,
           Rich

http://www.netside.net/~rveraa/