Re: OMG MOVE sucks!

+1000

Roger Harman
COMMON Certified Application Developer - ILE RPG on IBM i on Power

--

From: RPG400-L <rpg400-l-bounces@xxxxxxxxxxxxxxxxxx> on behalf of John Yeung <gallium.arsenide@xxxxxxxxx>

Sent: Thursday, September 19, 2019 9:50 AM

To: RPG programming on the IBM i (AS/400 and iSeries) <rpg400-l@xxxxxxxxxxxxxxxxxx>

Subject: OMG MOVE sucks!

 


I apologize in advance for the following rant.



I know people like to extol the virtues of "proper" ILE (modules,

prototypes, etc.) and often tend to minimize the value of RPG IV when

used "merely" as RPG III or RPG/400 with slightly massaged syntax.



But damn, some of those "little" niceties turn out to bring huge benefits!



Today's hero of no-"proper"-ILE-knowledge-required features is EVAL. I

just spent an embarrassing amount of time tracking down weird behavior

that boils down to: Someone didn't realize that two character fields

were different length, and they used MOVE when they should have

used... well, they SHOULD have used EVAL, but even in pure RPG III

they should have used MOVEL.



I have no idea why it wasn't a shop standard to always use MOVEL when

assigning character fields (pre-EVAL). I would *hope* that most

programmers came to this convention on their own. I know reference

manuals are supposed to be neutral and matter-of-fact, but really, in

this case, there really should have been guidance *in the manual* that

you NEVER use MOVE if MOVEL will do the job. (I get that not everyone

reads the manual, but if it's in the manual, then at least there's a

better chance that educators and mentors will be exposed to it and

pass it on.)



EVAL also happens to take care of the (P) extender, which is all too

easy to forget (if you know about it at all). So, I hope no one is

using ANY of the MOVE* family anymore. I don't think there are any

excuses left (and haven't been for many years now).



The closest thing to an excuse I would accept (and I wouldn't accept

it, it's just the closest) is that some people are taught that it's a

virtue to preserve the style of any existing code that you're going in

to modify. I can understand why this is desirable. But it presupposes

that the existing code is of a certain minimal level of quality, and

that you should put aside your own *aesthetics* for the sake of

cohesiveness and regularity in the code base. I would contend that we

should consider MOVE* a bug, not merely a style we disagree with. It

should be classified under "if this works, it's only an accident, and

we've been lucky so far".



If you find yourself dealing with a program which is chock full of

MOVE* and you need to add one line but you don't have time to

eradicate MOVE* from the whole program, in my book it is justified to

just use EVAL on that line you're adding and leave the rest of the

program as-is. Yeah, you're introducing an inconsistency in the style,

but at least you're not adding a new bug.



OK, sorry, had to vent a little. Now back to our regularly scheduled

programming.



John Y.