FloatToInt currently supports 3 modes:

• eRoundDown - this is eq to C floor (or C++ std::floor): it finds the closest integer literally <= than the given float.
• eRoundUp - this is eq to C ceil: it finds the closest integer >= given float.
• eRoundNearest - this is eq to C round: it rounds to the closest integer.

What is missing here is "truncate": the one that finds a closest integer which is less by magnitude. We may call this eRoundTruncate (unless someone suggests a better name).

The difference between floor and trunc is that for the floor (eRoundDown):
+1.5 = +1
-1.5 = -2
and for trunc (eRoundTruncate):
+1.5 = +1
-1.5 = -1

NOTE: also, the FloatToInt implementation in the engine should be rewritten to use libc or std C++ functions. done

EDIT: In fact, we might also add a mode opposite to "truncate", which finds nearest integer greater by magnitude. The C/C++ do not have the ready function for this, but it may be implemented like

return x < 0 ? floor(x) : ceil(x);