The BBC and Master Computer Public Domain Library

9F3B '*' operator - multiplication

Submitted by Steve Fewell

Description:

[9F3B:] Check the value type (in A).
If variable type = 0 then issue a Type Mismatch error.

If variable type is negative then the first value is a Float, so jump to &9F23, which
does the following:
* Push FWA to Stack
* Get the second value from the expression handler (level 6) [A012]
* If second value is a String then generate a Type Mismatch error. [96DD]
* If second value is an Integer then convert it to a Float [96DD]
* Pop the first Float value from Stack to argp (&4A, &4B)
* Call the Floating-Point Multiplication routine to multiply the values (&A6A6)
* Set A = #&FF, as we have a Floating-Point result.
* Jump to 9FCA (Expression handler level 5) to check for further *,/,MOD or DIV operators.

Otherwise, the value is Integer.
If the IWA byte 4 or IWA byte 3 is zero or the top bit of IWA byte 2 is set,
then we are dealing with a large number, so we will need to use Floating-Point multiplication
(as the result will most likely not be an integer), so call 9F20. [Note: What happens if the IWA
contains &10FFFFFF, it looks like Floating-Point multiplication will not be used for this!].
9F20 does the following:
* Convert the Integer in the IWA to a Float (FWA).
* Push FWA to Stack
* Get the second value from the expression handler (level 6) [A012]
* If second value is a String then generate a Type Mismatch error. [96DD]
* If second value is an Integer then convert it to a Float [96DD]
* Pop the first Float value from Stack to argp (&4A, &4B)
* Call the Floating-Point Multiplication routine to multiply the values (&A6A6)
* Set A = #&FF, as we have a Floating-Point result.
* Jump to 9FCA (Expression handler level 5) to check for further *,/,MOD or DIV operators.

Otherwise, the first value (multiplicand) is an integer which isn't too large,
so push the first value to the stack and call the Expression Handler level 6 routine
(&A00F) to get the second value to multiply (the multiplier).
If the second value is a string then generate a Type Mismatch error.

If the second value is a Float then jump to &9F15 to do the following:
* Pop the first value (the Integer) from the Stack
* Push the second value (the Float) to the Stack
* Convert the Integer to a Float (and put it in the FWA).
* Pop the second Float value from Stack to argp (&4A, &4B)
* Call the Floating-Point Multiplication routine to multiply the values (&A6A6)
* Set A = #&FF, as we have a Floating-Point result.
* Jump to 9FCA (Expression handler level 5) to check for further *,/,MOD or DIV operators.

Otherwise, check whether the second value is too large for Integer multiplication
routine (it's too large if IWA byte 3 or 4 is not 0, or top bit of byte 2 is 1).
If the second value is too large for Integer Multiplication then jump to &9F12
to do the following:
* Convert the second value (which is Integer) to a Float
* Pop the first value (also an Integer) from the Stack
* Push the second value (the Float) to the Stack
* Convert the Integer to a Float (and put it in the FWA).
* Pop the second Float value from Stack to argp (&4A, &4B)
* Call the Floating-Point Multiplication routine to multiply the values (&A6A6)
* Set A = #&FF, as we have a Floating-Point result.
* Jump to 9FCA (Expression handler level 5) to check for further *,/,MOD or DIV operators.

Otherwise, both Integers are in the required range, so do the Integer Multiplication
routine (&9F64).
This routine automatically jumps back to 9FCA on completion, to check for further *,/,
MOD or DIV operators.

Disassembly for the '*' Operator - Multiplication routine