Too little information is given. The statement that the CLP wrapper
gave more meaningful error messages seems almost to suggest that the
error is now known.? If the trigger program is not being called [debug
and trace would assist to determine that] or fails to activate [manifest
effectively same as not being called], then there will be no "previous
messages in the job log to determine the error." Such error conditions
for which the trigger program could not be activated would typically be
MCH08xx, MCH16xx, MCH34xx. It is complicated to explain why a failure
to activate the program does not leave any prior messages, so I will not
even try.

If the CLLE uses a named activation group instead of *CALLER, then
use *CALLER instead; at least until all other issues are resolved. If
any RCLACTGRP is used in the session, remove that action, then start a
new session; do not reintroduce the request until a better understanding
of the what/how that impacts the trigger program.

Regards, Chuck
--
All comments provided "as is" with no warranties of any kind
whatsoever and may not represent positions, strategies, nor views of my
employer

Martin Hinze wrote:
> Firing an after insert trigger always ended in msg CPF502B (twice)
> and SQL0443. Looking (hard!) for more information was without
> success. Only after i wrapped a CLP round the original CLLE trigger
> program it then gave me more meaningful error messages.
>
> What is wrong with using CLLE-triggers?

Hi Martin,

SQL0443? is it something with too long and to complex? sounds like the
phantastic new query engine, bug or feature is here the question

Dieter Bender

Martin Hinze wrote:

> Firing an after insert trigger always ended in msg
> CPF502B (twice) and SQL0443. Looking (hard!) for more
> information was without success.
> Only after i wrapped a CLP round the original CLLE
> trigger program it then gave me more meaningful error
> messages.
>
> What is wrong with using CLLE-triggers?
>
> Martin

CRPence wrote:
> Too little information is given.
What would be needed?
The statement that the CLP wrapper
> gave more meaningful error messages seems almost to suggest that the
> error is now known.?
There was no error, besides that the clle *pgm object
was used but did definitely not run.
But when the wrapped pgm had an error it was shown now.
If the trigger program is not being called [debug
> and trace would assist to determine that]
believe me i really tried but with no success!
>
> If the CLLE uses a named activation group instead of *CALLER, then use
> *CALLER instead;
I'll give it another try, thx!

Regards, Martin

Hi,

have you tried to debug your trigger program to get more information?

The best way is to use the debugger in iSeries Navigator's Run an SQL
script.
After you started the debugger, just try to insert a row into your
table.
Then step through your CLLE prgoram to make sure where the error
occurs and analyze the joblog after.

Birgitta

On 2 Nov., 21:33, Dieter Bender wrote:
> Hi Martin,
>
> SQL0443? is it something with too long and to complex? sounds like the
> phantastic new query engine, bug or feature is here the question
>
> Dieter Bender
>
SQL0443 means, an error in the trigger program occured.
If you look at the message file QSQLMSG, Message-Id SQL0443, you get
detailed information what SQL0443 means.

Because the trigger is CLLE and not SQL, I assume it has nothing to do
with the SQE!

Not all errors that occur are initiated by the SQE!
(We already had some discussions about this topic.)
Most SQL-problems, our clients had the last time that I analyzed were
caused by the CQE. What I see, almost nobody is really working with
the SQE, because most the legacy applications work with DDS described
logical files with select/omit clauses. When moving to the SQE we
could gain for around 90% of all SQL statements a much better
performance.

In spite of all I'm far off to declare the SQE as perfect. There are
still a lot of (big) bugs. Just yesterday I discovered a new one.

Birgitta

Martin Hinze wrote:
> Firing an after insert trigger always ended in msg CPF502B (twice) and
> SQL0443. Looking (hard!) for more information was without success.
> Only after i wrapped a CLP round the original CLLE
> trigger program it then gave me more meaningful error messages.
>
> What is wrong with using CLLE-triggers?

I don't know - I just used one, and it worked fine. I
wondered if perhaps it had something to do with your
activation group, but I successively created the CLLE
program with DFTACTGRP(*YES), DFTACTGRP(*NO)
ACTGRP(*NEW), and DFTACTGRP(*NO) ACTGRP(MYAG), and the
trigger ran without error in all three cases.

I forgot to mention that the clle was running fine when it was fired
from a 5250 session.

Meanwhile IBM helped me fixing the problem: it was the (missing)
librarylist beeing active for ODBC.
As the CLLE used some servicpgms located in different libs.

Thx to all.
Martin