Re: Who the bloody eff still codes CABEQ in this century?!?!?!?

[Dan <dan27649@xxxxxxxxx>]

I like both Scott's and Jon's ideas for "customizing" the compiler.

The problem, of course, is all of the legacy code.  No one's gonna buy the
idea that if I need to make a one-line change to a program that I first
have to gut it of all the CAB*, CAS*, GOTO, and TAG opcodes.

I could see a process that wouldn't need IBM's blessing that removes all
authority to the RPG compiler commands except to a restricted profile, and
force compiles to go through a pre-processor that does the vetting.  Again,
with the legacy code issue, if you had a SCM tool, you could set a base for
existing members and not allow any new uses of "banned" opcodes.  Frankly,
though, if developers and code reviewers can't discipline themselves to
adhere to coding standards, there are other problems that need to be
addressed.

- Dan

On Thu, Jul 20, 2017 at 2:41 PM, Scott Klement <rpg400-l@xxxxxxxxxxxxxxxx>
wrote:

> Maybe the thing IBM should do is add an exit point.  During compile, the
> RPG compiler could call any registered exit programs and pass the opcodes
> to it while checking the source.
>
> Customers like us could then write code to enforce rules like "never use
> CABEQ" or "never use MOVE" or "no gotos allowed" or whatever a shop wants
> to do.