Re: Compiled program size when /INCLUDEing consts and prototypes

[Charles Wilt <charles.wilt@xxxxxxxxx>]

So playing around with it...

I can defined 119 constants before the 4KB bump occurs...
I can then define another 120 before the next 4KB bump...

Still curious as to what the space is used for...

Wonder how much a PR takes...


Charles


On Thu, Mar 16, 2017 at 10:36 AM, Charles Wilt <charles.wilt@xxxxxxxxx>
wrote:

> So I've got a simple RPGLE program...
>
>       /FREE
>        dsply 'Hello World';
>        *INLR=*ON;
>        return;
>       /END-FREE
>
> Compiled with:
>  OPTION(*NOUNREF) DBGVIEW(*NONE) OPTIMIZE(*FULL) TGTRLS(*CURRENT)
>
> I've also got two /INCLUDE files, one with prototypes and one with
> constant definitions.
>
> I was surprised to find that static storage appears to go up about 16
> bytes per prototype; even though none of the procedures are called.
>
> I was even more surprised to find that program size seems to up 4KB if I
> include both prototypes & constant files; if I only include one of them, it
> doesn't change.  However, if I cut down the number of PR's & CONSTs...then
> the program size goes back down 4KB.
>
> Can anyone (Barbara :) ) provide some insight into this madness?
>
> I'm trying to argue that this style of using PRs
>  /define inc_myfun1_pr
>  /define inc_myfun2_pr
>  /define inc_myfun3_pr
>  /include MYPROTOS
>  /undefine inc_myfun1_pr
>  /undefine inc_myfun2_pr
>  /undefine inc_myfun3_pr
>
> To only include the prototypes of procedures you need to call is silly.
>
> Additionally, I'm arguing for having a STDCONST file of the most used
> constants is a good idea.  Rather than defining the constants in each
> program.
>
> The push-back is that the unneeded PRs & CONSTs would increase program
> size.  I said no they won't, but apparently I was wrong. :)
>
> So I'm trying to understand the behavior so I can counter-argue correctly.
>
> Thanks!
> Charles