Re: Trigger program with LR?

[Scott Klement <rpg400-l@xxxxxxxxxxxxxxxx>]

Hi Titus,

> Should a trigger program wirtten in RPGLE end with LR on?

I always end with *INLR = *OFF because the program will re-start faster
which is important for triggers in my environment.

> When ending with LR, I think it will be loaded in main storage for every
> call. The will cost a lot of resources.

That depends on the activation group.  If it's running in the default
activation group, then yes -- *INLR may unload it from memory.  If it's
created with ACTGRP(*NEW) then it'll always completely end, no matter how
you specify *INLR, and if it's in a named activation group, then turning
on *INLR will only cause the variables to be reinitialized -- the program
will remain in memory until the actgrp is reclaimed.

As I mentioned, I prefer to have my program remain active and not have to
reinitialize the variables because it's the fastest way to do things -- so
I specify ACTGRP('QILE') and *INLR=*OFF.

> But when I do not end with LR, who ends the trigger program after doing
> all updates to a physical file like an SQL UPDATE would do?

The trigger program will end when you sign off or when the activation
group is reclaimed.  In most cases it does not make sense to unload a
trigger program until you sign off.  If you want to unload/reload it for
testing purposes, then compile it with ACTGRP('MYACTGRP') and *INLR=*OFF,
and then use RCLACTGRP MYACTGRP from the command-line when you want to
unload it.