Scott Klement wrote:
As such, I recommend a simple subprocedure like this:
P getUser B
D PI 10a
D myUser s 10a inz(*user)
c return myUser
And I can have a static variable in the service program, to hold the
user profile from the last time the censorship routine's initialization
was called: if it matches what getUser returns, it skips the
(potentially rather expensive) initialization to that user, but if it
doesn't match, it reinitializes.
The whole issue of prestart jobs (that might actually remain active long
enough to have the censorship information changed from under them) still
kills DETERMINISTIC, but there may yet be some hope.