In my tests RLOCK() with SET REPROCESS TO 0 and on error routine in place was much faster than with SET REPROCESS TO 0 SECONDS.

>Hello!
>
>Did you ever experience any problems using "SET REPROCESS TO 0 SECONDS" ?
>This command seems to be MUCH faster than "SET REPROCESS TO 0", which tries to lock the record only one time.
>Is there a difference... i dont see it. Microsoft says, that the value should be higher than 0 if you use the seconds command.
>
>My problem was to determine how many users are using out programm in their network. Every user creates a record when he starts his programm and locks it with LOCK() directly after he created it.
>I'm trying to count the locked records in this table and found out, that i could only use the LOCK() command to check if a record is locked. Because ISRLOCKED() did not check if there was somebody locking it from another client.
>
>Tried the "SET REPROCESS TO 1" and found out, that it almost uses 1 second to give me a .F. if somebody locked it. That would be 400 seconds when 400 users where in our Application. After that our User could start working.
>
>Then tried the "SET REPROCESS TO 0 Seconds" and found out, that it uses much less than 1 second to give me a .F. if somebody locked it.
>Is there any command beeing faster than LOCK() to check if a record is locked?
>
>Greetings