RE: set a pointer to const

-----Original Message-----
From: rpg400-l-admin@midrange.com [mailto:rpg400-l-admin@midrange.com]On
Behalf Of Hans Boldt

Steve asked:

>> any plans to allow a DS and pointer to be declared as const?
>>
>>  d pAddr            s               *   const
>>  d Addr             ds                  likeds( Address )
>>  d                                      const
>>  d                                      based( pAddr )
>>

Hans did not mince words and replied:

>No.

Hans further said:

>What exactly are you trying to accomplish here with that type of
>syntax?  If you want to modify the value of a parameter, then pass
>it by reference.  If you want to modify the value of a parameter
>but not change the parameter passed in the call, then pass it by
>value. If you're just interested in the value passed, then just
>reference it normally.

I have C and C++ code that parses command strings and does some specialized
scanning that I would like to port from visual C++ to ILE. These functions
accept pointers as arguments, some that are pointers to const data.  In some
cases prototyping as %string works, but other times simply passing a pointer
is best.

In general I would like to find out if mix and matching C++ and RPG
procedures works better than coding entirely in RPG. But the RPG limitation
of setting a pointer to only non const variables will not work well with C++
which makes heavy use of const references and pointers.

Basically, if RPG provides the pointer, why restrict it to only non const
variables?  Esp since const is such a good modular programming practice and
the solution of qualifying a pointer and DS as const appears to be very
doable.

Does the IBM RPG dept accept customer suggestions and requests?  Is there an
email address that a question can be posted to?  The decision to not extend
const to pointers and data structs is wrong and I would like the IBM RPG
person in charge to confirm and respond to my objections.

thank you,

Steve Richter