IBM Enterprise COBOL - Add support for BINARY-CHAR, BINARY-SHORT, BINARY-LONG, BINARY-DOUBLE as alternative to COMP-5

Hi,

The COMP-5 USAGE, (or TRUNC(BIN) compilation option), allows binary values ​​to be stored according to the entire storage capacity of the machine's internal representation.

However COMP-5 is a constructor implementation and have portability problems.

We want to implement USAGE that are constructor independent, and the ISO standard defines them: BINARY-CHAR, BINARY-SHORT, BINARY-LONG, BINARY-DOUBLE.

ISO USAGE COMP-5 USAGE Range of values

---------------------- ------------- -----------------------------------------------------------------

BINARY-CHAR SIGNED N/A -128 [-2**7] <= n < 128 [2**7]

BINARY-CHAR UNSIGNED N/A 0 <= n < 256 [2**8]

BINARY-SHORT SIGNED S9(4) COMP-5 -32768 [-2**15] <= n < 32768 [2**15]

BINARY-SHORT UNSIGNED 9(4) COMP-5 0 <= n < 65536 [2**16]

BINARY-LONG SIGNED S9(9) COMP-5 -2147483648 [-2**31] <= n < 2147483648 [2**31]

BINARY-LONG UNSIGNED 9(9) COMP-5 0 <= n < 4294967296 [2**32]

BINARY-DOUBLE SIGNED S9(18) COMP-5 -9223372036854775808 [-2**63] <= n < 9223372036854775808 [2**63]

BINARY-DOUBLE UNSIGNED 9(18) COMP-5 0 <= n < 18446744073709551616 [2**64]

To see if for these USAGE the PICTURE clause remains required.

If present it can define the number of equivalent decimal digits allowed to represent the value.

In the presence of a PICTURE, the SIGNED / UNSIGNED clause would not be required or should be consistent with the PICTURE.

Thanks.