Re: Sequence numbers (was: RDi 9.5 announcement)

On Thu, Sep 3, 2015 at 10:07 AM, Buck Calabro <kc2hiz@xxxxxxxxx> wrote:

In RDi, one can disable displaying the sequence numbers:

This was never contested, and in fact completely supports my "let the
editor handle it" stance.

and
error messages should report the ordinal line number accordingly.

In RDi, clicking on an error message automatically brings you to the
line in error.

Again, more support of my position.

You can see in RPG IV compiler listings and dumps that IBM had
precisely this in mind.

I believe the compiler-generated line numbers are there to accommodate
I- and O-specifications 'generated' from externally defined files, as
well as source included via /COPY or /INCLUDE.

Ah. Good point. I keep forgetting about these. Here we have an example
of something in the RPG language that isn't terribly well suited for
either stored sequence numbers or purely ordinal module-level line
numbers.

It's not insurmountable, of course. In most of the programming world,
purely ordinal numbers are qualified by the source file (source member
in IBM i parlance). For RPG (and other languages which have the
semantics of "literally insert some other chunk of source code right
here, as though I had typed it into the same member"), this would have
the undesirable effect of not being a single "number space" for the
entire module.

So, not ideal, but the situation is not helped in any way by keeping
around stored sequence numbers. My assessment of the (utter lack of)
value of stored sequence numbers stands.

John Y.