Re: Free Form RPG Opinions Wanted

[Douglas Handy <dhandy1@xxxxxxxxxxxxx>]

Joe,

>But there's still the too many extenders issue.  I
>can see that as a problem.  One you'd rarely run into, but a problem
>nonetheless.  

The point was that it was not as simple as reusing the EXACT SAME parsing
already in the compiler, as you suggested.

>How often does this happen in your code, Doug?

I'm still on V5R1, so it has never happened to me yet.  And granted, I don't
expect it ever would in my style of coding.  But that is not the rules IBM must
play by.  They have to take into consideration all kinds of things which don't
matter to me in my little world of only working with USA based companies.

>Beside (and keep this between you and me) but the idea of sticking precision
>rules in the opcode never seemed like a very good idea.  I'd prefer BIFs:
>%HALF(x), %PREC(x: {*DEFAULT}|*RESULT).  That way I have fine-grained
>control over the entire statement.

Well there is %DEC() vs %DECH(), %INT() vs %INTH(), etc

>> And you'd need to add a semi-colon to the end of the statement

>Not true.  Just use the line continuation conventions already used in eval
>statements.  Which is basically, if the next line doesn't have an opcode,
>it's a continuation.

I'll let Hans answer this if he cares too, or else just go back to the archives
where he has explained it in the past.   It is not that simple to avoid
ambiguities.

Doug