Rollup merge of rust-lang#47919 - varkor:to_degrees-precision, r=rkru…

…ppe Use constant for 180/π in to_degrees The current `f32|f64.to_degrees` implementation uses a division to calculate `180/π`, which causes a loss of precision. Using a constant is still not perfect (implementing a maximally-precise algorithm would come with a high performance cost), but improves precision with a minimal change. As per the discussion in rust-lang#29944, this fixes rust-lang#29944 (the costs of improving the precision further would not outweigh the gains).

src/libcore/num/f32.rs

@@ -239,7 +239,9 @@ impl Float for f32 {
     /// Converts to degrees, assuming the number is in radians.
     #[inline]
     fn to_degrees(self) -> f32 {
-        self * (180.0f32 / consts::PI)
+        // Use a constant for better precision.
+        const PIS_IN_180: f32 = 57.2957795130823208767981548141051703_f32;
+        self * PIS_IN_180
     }
 
     /// Converts to radians, assuming the number is in degrees.

src/libcore/num/f64.rs

@@ -237,6 +237,9 @@ impl Float for f64 {
     /// Converts to degrees, assuming the number is in radians.
     #[inline]
     fn to_degrees(self) -> f64 {
+        // The division here is correctly rounded with respect to the true
+        // value of 180/π. (This differs from f32, where a constant must be
+        // used to ensure a correctly rounded result.)
         self * (180.0f64 / consts::PI)
     }
 

src/libstd/f32.rs

@@ -1531,6 +1531,7 @@ mod tests {
         assert!(nan.to_degrees().is_nan());
         assert_eq!(inf.to_degrees(), inf);
         assert_eq!(neg_inf.to_degrees(), neg_inf);
+        assert_eq!(1_f32.to_degrees(), 57.2957795130823208767981548141051703);
     }
 
     #[test]