RE: Binary Bit Map

[Scott Klement <rpg400-l@xxxxxxxxxxxxxxxx>]

> Thanks to all who replied to my original question.  I am at V5R2 so I just
> broke the bit map into 8 fields and used the biton/bitoff opcode.

Consider using %bitand and %bitor to retrieve and set (respectively) the 
bits.  It'll be faster and require less code.

Also, consider coding a constant that corresponds to each bit to make it 
clear what the bit is for.  That'll make your code easier to read.


> I have a somewhat related question.  Some of the fields in the message
> need to be coded as Binary Coded Decimal (BCD).  From what I understand
> this is a unsigned packed numeric.  Odd length fields are padded with a
> leading zero.  So 99,999 would be represented in packed as:   99 99 9F and
> in BCD:  09 99 99.

Assuming the BCD field is big-endian:

Create a data structure that overlays a packed field with an unsigned 
integer field. (make the packed field 15P 0 and the unsigned field 20U 0). 
Maybe overlay it with an 8A character field as well (depending on what 
you're doing.)

To convert from BCD to packed, use the char field to store the BCD field 
into the data structure. Then multiply the unsigned field by 16 and then 
use %bitor (or BITON if like, or even MLLZO) to set the last nibble to 
x'0f'.Read the result from the packed field.

To convert from packed to BCD, store the data by setting the packed field, 
then divide the unsigned field by 16, and retrieve the results from the 
char field.

If it's little-endian, you'd need to convert it to bigendian when 
assigning the BCD to the data structure, and convert it back to little 
endian when extracting the BCD field from the DS.

Make sense?