Should the docstring for `div` clarify floating-point behaviour as in `fld`?

[Sagnac]

At least until div is fixed for floats (#49561).

At the moment the docstrings for div refer only to integers, but there's no type restriction on the inputs which leads to a discrepancy if called with floats:

julia> div(1.0, 0.1, RoundToZero)
9.0

julia> round(1.0 / 0.1, RoundToZero)
10.0

This is of course because decimals such as 0.1 don't have an exact floating-point representation so the result can differ when rounded down; the docstring for fld includes this warning, but div does not, so even simple operations such as 1.0 ÷ 0.1 can lead to unintuitive results if this is not specified.