The BBC and Master Computer Public Domain Library

B1C7 Load FWA with Float Variable

Submitted by Steve Fewell

Routine:ain
Name: Copy Float Variable to the FWA
Starting Address: &B1C7
Entry criteria: &2A and &2B point to a 5-byte Floating-Point variable to unpack and load.
        Y = 5 (size of variable).
Exit: The FWA contains the value of the variable.

Description:

Zero the FWA Rounding and FWA Overflow bytes.
Store the 5th byte of the variable (pointed to by (&2A, &2B)) in FWA Mantissa byte 4.
Store the 4th byte of the variable (pointed to by (&2A, &2B)) in FWA Mantissa byte 3.
Store the 3rd byte of the variable (pointed to by (&2A, &2B)) in FWA Mantissa byte 2.
Store the 2nd byte of the variable (pointed to by (&2A, &2B)) in FWA Sign byte.
Store the 1st byte of the variable (pointed to by (&2A, &2B)) in FWA Exponent byte.

If exponent = 0 and sign = 0 and FWA Man 2 = 0 and FWA Man 3 = 0 and FWA Man 4 = 0 then
the number is zero so jump to end of routine (B1F2) to store zero in FWA Mantissa 1 and
exit with A = #&FF (indicating a float result).

If exponent is non-zero, copy sign byte to FWA Mantissa byte 1 (setting the top
bit, as the sign is now unpacked and held separately from the top byte of the mantissa).
Exit with A=#&FF.

Disassembly for the Load FWA with Float Variable routine