Add custom legalization for floating point constants.

Use the simplest expansion which materializes the bits of the floating
point constant as an integer and then bit-casts to the floating point
type. In the future, we may want to use constant pools instead. Either
way, we need custom legalization.

Also add a legalize_monomorphic() function to the Python targetISA class
which permits the configuration of a default legalization action for
monomorphic instructions, just like legalize_type() does for polymorphic
instructions.

lib/cretonne/src/ir/immediates.rs

@@ -599,6 +599,11 @@ impl Ieee32 {
     pub fn with_float(x: f32) -> Ieee32 {
         Ieee32(unsafe { mem::transmute(x) })
     }
+
+    /// Get the bitwise representation.
+    pub fn bits(self) -> u32 {
+        self.0
+    }
 }
 
 impl Display for Ieee32 {
@@ -630,6 +635,11 @@ impl Ieee64 {
     pub fn with_float(x: f64) -> Ieee64 {
         Ieee64(unsafe { mem::transmute(x) })
     }
+
+    /// Get the bitwise representation.
+    pub fn bits(self) -> u64 {
+        self.0
+    }
 }
 
 impl Display for Ieee64 {