Re: Pascal to ADA

[Ole-Hjalmar Kristensen <ole-hjalmar.kristensen@substitute_employer_here.com>]

1. Why do you need this routine at all? 

2. You can accomplish the same in Ada in different ways. two obvious
   solutions are unchecked_conversion or an overlay with an address
   clause between a packed array of bytes and your 32 bit integer.

Overlay version:

type Long_Int is range -2**31 .. 2**31 - 1;
type Byte is mod 2**8;
...

Procedure Add(N: Long_Int; M: Long_Int; R : out Long_Int) is
  Cheat_N : array(0..3) of Byte;
  pragma pack(Cheat_N);
  for Cheat_N'address use N'address;
...
begin
...
end Add;

3. Logical operators also work on modular integers. Ada also has
   arithmetic shift, so you can conveniently access the carry bit.

>>>>> "TG" == Tilman Gloetzner <tilman.gloetzner@etas.co.kr> writes:

    TG> Hello,
    TG> 1) what is a effecient way of rewriting the following code fragement in ADA
    TG> ?  With the absolute keyword, Pascal  effectively casts
    TG> N, M, and R from a 4 byte integer type to an array of 4 bytes by an address
    TG> access.  I guess to implement the same mechanism in
    TG> ADA is not in line with ADA's philosophy.

    TG> 2) How do I mask out bits in a (unsigned) number type, as logical operators
    TG> work only on booleans ?

    TG> Thank you,

    TG> Tilman

    TG> ----------------------------------------------------------------
    TG> PROCEDURE Add32(N:LONGINT;M:LONGINT;VAR R:LONGINT) ;
    TG> VAR
    TG>      _N : ARRAY[0..3] OF BYTE ABSOLUTE N ;
    TG>      _M : ARRAY[0..3] OF BYTE ABSOLUTE M ;
    TG>      _R : ARRAY[0..3] OF BYTE ABSOLUTE R ;
    TG>      A  : BYTE ;
    TG>      B  : WORD ;
    TG>      C  : WORD ; { Carry }
    TG> BEGIN
    TG>      R := 0 ;
    TG>      C := 0 ;
    TG>      FOR A := 0 TO 3 DO BEGIN
    TG>          B := _N[A] + _M[A] + C ;
    TG>          _R[A] := B AND $FF ;
    TG>          C := B SHR 8 ;
    TG>      END ;
    TG> END  ;



-- 
Strange attractors stole my wife