Re: Disturbing Asset Discovery

[DAsmussen@xxxxxxx]

Terri,

Oooooh, don't even get me started on this one!  This has been a problem with 
the AS/Set product since SQL was first introduced to it.  The ADK editor 
doesn't check SQL statement syntax, but the program generator does.  If the 
generator finds a problem with an SQL statement, it just stops creation of 
the given subroutine at that point.  No messages, no errors, no warnings, 
just <STOP> -- doesn't show up in the CA0085O printout, either.

Consider yourself lucky.  At least you didn't encounter this while 
maintaining one of those "500" programs, and have the last two lines (of 600) 
disappear from your subroutine...

Regards!

Dean Asmussen
Enterprise Systems Consulting, Inc.
Fuquay-Varina, NC  USA
E-mail:  DAsmussen@xxxxxxx

"It's what you learn after you know it all that counts." -- John Wooden

In a message dated 3/18/2003 2:24:29 PM Eastern Standard Time, 
THarteau@xxxxxxxxxxxxxxxxxx writes:

> Hi,
>    We are on BPCS V6.04 with that same version of Asset. I made a copy
> of an existing Asset program that has many SQL commands in it. I changed my
> primary file to another one.  It compiled fine and ran with no halts. As I
> was looking at the data, I found some items that did not update correctly.
> I traced it back to a subroutine with an SQL declare followed by the open,
> fetch and a bunch of other calculations. When I looked at the asset source,
> everything was fine.  When I put it in debug, I entered the SR, but there
> was nothing there.
>    What had happened was that in the where portion of the declare, I had
> a field that was no longer defined.  The only RPG code generated for the
> subroutine was BEGSR &ENDSR. No compile error, no run error, no code.  The
> scary part is this SR only affected about 1/4 of the items. I could have
> very easily looked at some items and thought everything was OK, and put it
> into production. Luckily, I'm kind of anal sometimes :)
>    Is it normal that SQL statements are not checked for validity? When I
> fixed the field name, everything was OK.
>