math: correct mPi4 comment

The previous comment mis-stated the number of bits in mPi4.
The correct value is 19*64 + 1 == 1217 bits.

Change-Id: Ife971ff6936ce2d5b81ce663ce48044749d592a0
Reviewed-on: https://go-review.googlesource.com/c/154017
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

diff --git a/src/math/trig_reduce.go b/src/math/trig_reduce.go
index 7bc72e986d30520c54284ca9995888d0d188eb99..6f8eaba9b918497ad7e02b987a9e59b4932e4d30 100644 (file)
--- a/src/math/trig_reduce.go
+++ b/src/math/trig_reduce.go
@@ -68,8 +68,8 @@ func trigReduce(x float64) (j uint64, z float64) {
 
 // mPi4 is the binary digits of 4/pi as a uint64 array,
 // that is, 4/pi = Sum mPi4[i]*2^(-64*i)
-// 19 64-bit digits gives 1153 bits of precision to handle
-// the largest possible float64 exponent.
+// 19 64-bit digits and the leading one bit give 1217 bits
+// of precision to handle the largest possible float64 exponent.
 var mPi4 = [...]uint64{
        0x0000000000000001,
        0x45f306dc9c882a53,