Re: Database Trigger and Update problems

[Scott Klement <klemscot@xxxxxxxxxxxx>]

Hi Terry,

> The problem is .... if you change the data to correct values or even
> get a new record and try to update it ... you now get error messages
> cpf5147 ... instruction sequence error and/or cpf5129 that the
> trigger program sort of left something hanging. I do not know how to
> handle this.

In my last message to you, I tried to explain the consequences of escape
messages.  I tried to tell you not to do this unless YOU KNOW THAT the
system is expecting you to use escape messages for this.

>From these errors, it's quite clear to me that the system wasn't expecting
you to kill it's routines in the middle of what they're doing.

> I honestly do not know which program is really the problem.

I must've done a seriously poor job of explaining things then.

You are using the QMHSNDPM API to send an escape message CPF9898 from the
trigger to your update program.  This is killing (and by "killing" I mean
terminating the program) the operating system's database routines in the
middle of what they were doing.

You're leaving the database routines in a state where they think they're
waiting for the trigger program.   When you call those same routines again
they get confused, thinking that they're still waiting for that trigger
program to end!

In my previous message to you, I was worried that this would happen.  I
didn't know for sure, so I warned you with something to the effect of
"don't do things this way unless you know that it's designed to work this
way."  I'm a bit frustrated that, even after you received errors, you
still didn't take heed of my warning.