Operators applying on integer numeric types (signed or unsigned) and the INTEGER promotion

[MrSwiss]

From fxm's list, with FBC 0.24.0 it becomes (the first time) clear, the penalty (we're all paying)
for all the added OO and ByRef etc. stuff (that most of us don't use anyway).
... UPCASTING EVERYTHING ... (resulting in slow down of the running code), whether it makes
sense, or not.

This is not the case in 0.24.0 -- as the results clearly show!

[FreeFox]

Just wondering whether in the days of version 0.24.0 the Integer type was only 2 bytes rather than 4 ?

If not, it seems anomalous that byte x byte = integer but short x short = short.

[fxm]

For the fbc 0.24.0, the (U)Integer size is 4 bytes.
That's why I qualified the coercion results as surprising!