Hi all!

As you saw, PMP now supports floating point for PIC16 processors.

There's a main issue with this: the / operator is no more a synonym of the DIV operator for division. Now the / operator generates always a real divide for PIC16.

This would give strange things not always visible:

var

  I, J, K : integer;

...

K := I / J;

This code generates a floating point divide followed by an integer truncation.

So now PMP warns this explicitly to avoid stealth unoptimized code.

PMP team leader