adler32: speed up ~40% by avoiding bounds checks

before & after:
adler32.BenchmarkGolden	  100000	     14747 ns/op
adler32.BenchmarkGolden	  200000	      8761 ns/op

Found by profiling PNG encoding.

R=rsc, bradfitzwork, eds
CC=golang-dev
https://golang.org/cl/4441073

src/pkg/hash/adler32/adler32.go

@@ -43,8 +43,8 @@ func (d *digest) Size() int { return Size }
 
 // Add p to the running checksum a, b.
 func update(a, b uint32, p []byte) (aa, bb uint32) {
-	for i := 0; i < len(p); i++ {
+	for _, pi := range p {
-		a += uint32(p[i])
+		a += uint32(pi)
 		b += a
 		// invariant: a <= b
 		if b > (0xffffffff-255)/2 {

src/pkg/hash/adler32/adler32_test.go

@@ -5,6 +5,7 @@
 package adler32
 
 import (
+	"bytes"
 	"io"
 	"testing"
 )
@@ -61,3 +62,16 @@ func TestGolden(t *testing.T) {
 		}
 	}
 }
+
+func BenchmarkGolden(b *testing.B) {
+	b.StopTimer()
+	c := New()
+	var buf bytes.Buffer
+	for _, g := range golden {
+		buf.Write([]byte(g.in))
+	}
+	b.StartTimer()
+	for i := 0; i < b.N; i++ {
+		c.Write(buf.Bytes())
+	}
+}