pub fn float_slot(ty: Type) -> Option<usize>Expand description
Which of the two float widths a type is, or nothing for anything that is not a float.
Two rather than slot’s four, and a table of its own rather than more entries in that one,
because a float and an int of the same width are not the same term to any rule: they are in
different register files and every instruction that touches them is a different instruction. A
long double is none of them, since it is on the x87 stack rather than in a vector register
and nothing here is written about that stack.