Auto merge of rust-lang#3688 - tgross35:float-cast-test-refactor, r=RalfJung

Refactor float casting tests, add `f16` and `f128`

This is an attempt to remove the magic from a lot of the numbers tested, which makes things easier when adding `f16` and `f128`. A nice side effect is that these tests now cover all int <-> float conversions with the same amount of tests.

Author: bors

src/tools/miri/src/intrinsics/mod.rs

@@ -392,10 +392,10 @@ pub trait EvalContextExt<'tcx>: crate::MiriInterpCxExt<'tcx> {
                         bug!("float_finite: non-float input type {}", x.layout.ty)
                     };
                     Ok(match fty {
-                        FloatTy::F16 => unimplemented!("f16_f128"),
+                        FloatTy::F16 => x.to_scalar().to_f16()?.is_finite(),
                         FloatTy::F32 => x.to_scalar().to_f32()?.is_finite(),
                         FloatTy::F64 => x.to_scalar().to_f64()?.is_finite(),
-                        FloatTy::F128 => unimplemented!("f16_f128"),
+                        FloatTy::F128 => x.to_scalar().to_f128()?.is_finite(),
                     })
                 };
                 match (float_finite(&a)?, float_finite(&b)?) {