-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
frank kolmann wrote:
The WDSCi verifier is worse than useless with the modular/SQL code.
I use PDM/SEU for all my compilations. I need to eat crow every time I do a
compile.
I dont think I am the only one with this problem, I see many references on
the net to a change
done to the 'beloved by me[sic];' SQL RPG precompiler at V5R3.
I agree that the verifier, and even the compiler diagnostics for
SQLRPGLE modules are fairly worthless in WDSC 7. I usually open the
source member for display (PDM option 5), open the compile listing with
option 15, and type *ERR on the command line, just like 1986. It's easy
enough, if a little kludgey, to locate the errors that way, then fix
them in WDSC.
IBM did indeed fix the SQL RPG precompiler back there somewhere around
V5R3. I ended up having to make all host variables globally unique in
each module. I've never had a problem across modules though. Actually, I
don't understand how that's possible, since each module is compiled
independently. The linking step (CRTPGM or CRTSRVPGM) doesn't know
anything about variables unless they're exported. Are you perhaps
exporting *ALL or something? I'm not sure if that's possible either.
Never tried it. At any rate, I put a two character prefix, based on the
method name, on each variable with method scope, and that's working for
me, although the names are a little clumsy. Retrofitting all the old
code that got broken by the "fix" was sort of a pain. There is a ray of
hope, I understand. With V6.0, the precompiler is supposed to allow
local scoping of host variables. Apparently, that was implemented by
doing some slight of hand that moves the values to global storage, but
maybe it will help some. I hope to get to V6.0 some day.
I've been fighting the battle against PDM too, with no enthusiasm from
the programmers. I don't understand how anyone could be so afraid of
learning something (sort of) new that they would prefer PDM to an
integrated IDE, even if it does have a few warts. Of course the DEMs are
to blame for the bad economy too, provided you're willing to ignore 8
years with that last idiot. Maybe logic is not a good motivator of human
behavior.
- --
Pete Hall
pete@xxxxxxxxxxxxxx
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
iEYEARECAAYFAkrrj58ACgkQXczQcKdXKg7CggCgvWB4PnNGFnUrELx2IHUeXhls
my0AniEPc8bBEEqs50hH8IegUvZBm9zQ
=cJ5B
-----END PGP SIGNATURE-----
As an Amazon Associate we earn from qualifying purchases.