Better justification for the use of f64::to_bits

Timothée Haudebourg · Timothée Haudebourg · commit 48393aba4075 · 2021-10-20T14:24:58.000+02:00
diff --git a/src/number.rs b/src/number.rs
@@ -54,7 +54,7 @@ impl core::hash::Hash for N {
             N::PosInt(i) => i.hash(h),
             N::NegInt(i) => i.hash(h),
             N::Float(f) => {
-                // Using `f64::to_bits` here is fine since any float values are always finite.
+                // Using `f64::to_bits` here is fine since any float values are never `Nan`.
                 f.to_bits().hash(h);
             }
         }

Original file line number	Diff line number	Diff line change
`@@ -54,7 +54,7 @@ impl core::hash::Hash for N {`
`54`	`54`	`N::PosInt(i) => i.hash(h),`
`55`	`55`	`N::NegInt(i) => i.hash(h),`
`56`	`56`	`N::Float(f) => {`
`57`		- // Using `f64::to_bits` here is fine since any float values are always finite.
	`57`	+ // Using `f64::to_bits` here is fine since any float values are never `Nan`.
`58`	`58`	`f.to_bits().hash(h);`
`59`	`59`	`}`
`60`	`60`	`}`