utf16: add DecodeRune, EncodeRune

R=r
CC=golang-dev
http://codereview.appspot.com/970041

2 files changed, 67 insertions, 4 deletions

diff --git a/src/pkg/utf16/utf16.go b/src/pkg/utf16/utf16.go
index 303162452..372e38a71 100644
--- a/src/pkg/utf16/utf16.go
+++ b/src/pkg/utf16/utf16.go
@@ -18,6 +18,33 @@ const (
 	surrSelf = 0x10000
 )
 
+// IsSurrogate returns true if the specified Unicode code point
+// can appear in a surrogate pair.
+func IsSurrogate(rune int) bool {
+	return surr1 <= rune && rune < surr3
+}
+
+// DecodeRune returns the UTF-16 decoding of a surrogate pair.
+// If the pair is not a valid UTF-16 surrogate pair, DecodeRune returns
+// the Unicode replacement code point U+FFFD.
+func DecodeRune(r1, r2 int) int {
+	if surr1 <= r1 && r1 < surr2 && surr2 <= r2 && r2 < surr3 {
+		return (int(r1)-surr1)<<10 | (int(r2) - surr2) + 0x10000
+	}
+	return unicode.ReplacementChar
+}
+
+// EncodeRune returns the UTF-16 surrogate pair r1, r2 for the given rune.
+// If the rune is not a valid Unicode code point or does not need encoding,
+// EncodeRune returns U+FFFD, U+FFFD.
+func EncodeRune(rune int) (r1, r2 int) {
+	if rune < surrSelf || rune > unicode.MaxRune || IsSurrogate(rune) {
+		return unicode.ReplacementChar, unicode.ReplacementChar
+	}
+	rune -= surrSelf
+	return surr1 + (rune>>10)&0x3ff, surr2 + rune&0x3ff
+}
+
 // Encode returns the UTF-16 encoding of the Unicode code point sequence s.
 func Encode(s []int) []uint16 {
 	n := len(s)
@@ -38,9 +65,9 @@ func Encode(s []int) []uint16 {
 			a[n] = uint16(v)
 			n++
 		default:
-			v -= surrSelf
-			a[n] = uint16(surr1 + (v>>10)&0x3ff)
-			a[n+1] = uint16(surr2 + v&0x3ff)
+			r1, r2 := EncodeRune(v)
+			a[n] = uint16(r1)
+			a[n+1] = uint16(r2)
 			n += 2
 		}
 	}
@@ -57,7 +84,7 @@ func Decode(s []uint16) []int {
 		case surr1 <= r && r < surr2 && i+1 < len(s) &&
 			surr2 <= s[i+1] && s[i+1] < surr3:
 			// valid surrogate sequence
-			a[n] = (int(r)-surr1)<<10 | (int(s[i+1]) - surr2) + 0x10000
+			a[n] = DecodeRune(int(r), int(s[i+1]))
 			i++
 			n++
 		case surr1 <= r && r < surr3:
diff --git a/src/pkg/utf16/utf16_test.go b/src/pkg/utf16/utf16_test.go
index c6e269aad..c0848aa38 100644
--- a/src/pkg/utf16/utf16_test.go
+++ b/src/pkg/utf16/utf16_test.go
@@ -8,6 +8,7 @@ import (
 	"fmt"
 	"reflect"
 	"testing"
+	"unicode"
 )
 
 type encodeTest struct {
@@ -32,6 +33,41 @@ func TestEncode(t *testing.T) {
 	}
 }
 
+func TestEncodeRune(t *testing.T) {
+	for i, tt := range encodeTests {
+		j := 0
+		for _, r := range tt.in {
+			r1, r2 := EncodeRune(r)
+			if r < 0x10000 || r > unicode.MaxRune {
+				if j >= len(tt.out) {
+					t.Errorf("#%d: ran out of tt.out", i)
+					break
+				}
+				if r1 != unicode.ReplacementChar || r2 != unicode.ReplacementChar {
+					t.Errorf("EncodeRune(%#x) = %#x, %#x; want 0xfffd, 0xfffd", r, r1, r2)
+				}
+				j++
+			} else {
+				if j+1 >= len(tt.out) {
+					t.Errorf("#%d: ran out of tt.out", i)
+					break
+				}
+				if r1 != int(tt.out[j]) || r2 != int(tt.out[j+1]) {
+					t.Errorf("EncodeRune(%#x) = %#x, %#x; want %#x, %#x", r, r1, r2, tt.out[j], tt.out[j+1])
+				}
+				j += 2
+				dec := DecodeRune(r1, r2)
+				if dec != r {
+					t.Errorf("DecodeRune(%#x, %#x) = %#x; want %#x", r1, r2, dec, r)
+				}
+			}
+		}
+		if j != len(tt.out) {
+			t.Errorf("#%d: EncodeRune didn't generate enough output", i)
+		}
+	}
+}
+
 type decodeTest struct {
 	in  []uint16
 	out []int