Re: RPG400-L Digest, Vol 2, Issue 38

DeLong, Eric wrote:

Doc,

OK, not sure I agree with your assessment of where linker information should
be kept, but that's a small point and I'll not argue....

Well, it's not a small point-- but it's not down to me to be arbiter elegantum. You have something that works, everyone else can use it, and agrees to it. Sounds good to me.

Service programs.

Great, ain't they?

That to me is the most direct and logical approach to ILE development, as it
separates the programmer (somewhat) from the teduim of mass rebuild.  Throw
in a little Binder source, and you can modify underlying code without
requiring mass recompiles (unless your prototypes change, which hopefully is
not happening too often).

No arguement there.

 I freely admit my lack of experience in "unstable
environments" as you describe, so I may just not understand what you're
saying. However, I still can't see how TMKMAKE makes things better, only how
it makes it more "Unix-like".  Perhaps that's an advantage to some, but not
for me.  Sorry if I'm just being dense......

I know it'll be heresey for me to say this, but "UNIX-Like" doesn't always equal "bad". There are a lot of good things that I've grown to appreciate on Unix systems over the course of the last 10 years since I've been working with them. One of them is the concept of a .lib file... which is analgous to a service program, which we both agree are good things (in a very non-Martha Stewart like way).

The hardest part of using make is actually learning how the stupid make scripts work. Nightmare. Bad ones are nearly as arcane as RPG code riddled with left-hand indicators, CABxx, and ITER...LEAVE blocks <grin>.

The point is, that TMKMAKE is an often-overlooked tool that can do wonders for a big-scale project that has rapid development turnover.

-Doc