math/big: make ErrNaN actually implement the error interface (oversight)

There was no way to get to the error message before.

Change-Id: I4aa9d3d9f468c33f9996295bafcbed097de0389f
Reviewed-on: https://go-review.googlesource.com/8660Reviewed-by: Alan Donovan <adonovan@google.com>

src/math/big/float.go

@@ -71,6 +71,11 @@ type ErrNaN struct {
 	msg string
 }
 
+// ErrNan implements the error interface.
+func (err ErrNaN) Error() string {
+	return err.msg
+}
+
 // NewFloat allocates and returns a new Float set to x,
 // with precision 53 and rounding mode ToNearestEven.
 // NewFloat panics with ErrNaN if x is a NaN.

src/math/big/float_test.go

@@ -12,6 +12,9 @@ import (
 	"testing"
 )
 
+// Verify that ErrNaN implements the error interface.
+var _ error = ErrNaN{}
+
 func (x *Float) uint64() uint64 {
 	u, acc := x.Uint64()
 	if acc != Exact {