The real problem is the comparison of unsigned and signed variables,
right? Do we really need them? I think, no. All what we wand is the
possibility to use unsigned values for some external functions, for
large values, where signed variables are negative. We want to calculate
with high values and pass them back to such a external function. We
does not want to compare signed with unsigned variables.
As a result, we wish to use unsigned variables with the math-library,
math-operators and a few people could use the bit-operators. If there
is a need to compare unsigned with signed variables, the solution could
be for the compiler, to interpret all unsigned variables as signed, if they
are be mixed. Other PB-Function could interpret signed variables as
unsigned or generate a compiler-error.
If someone needs to compare a unsigned with a signed variable, he could
write his own workaround like it is done today. Because then, signed and
unsigned variables would be mixed up, the compiler could interpret the
unsigned variable as signed and so it needs the "& $FF" workaround.
This solution would have no performance issues. It is also a little tricky,
but that's the same behavior today. all users who need this features
aren't beginners. And for all beginners, which doesn't get it, can ask
like they do it the whole day.
I hope you think over your decision again
MFG PMV