summaryrefslogtreecommitdiff
path: root/src/pkg/utf8
diff options
context:
space:
mode:
Diffstat (limited to 'src/pkg/utf8')
-rw-r--r--src/pkg/utf8/utf8.go150
-rw-r--r--src/pkg/utf8/utf8_test.go90
2 files changed, 120 insertions, 120 deletions
diff --git a/src/pkg/utf8/utf8.go b/src/pkg/utf8/utf8.go
index ad78f599c..8e373e32d 100644
--- a/src/pkg/utf8/utf8.go
+++ b/src/pkg/utf8/utf8.go
@@ -6,40 +6,40 @@
// This package calls a Unicode character a rune for brevity.
package utf8
-import "unicode" // only needed for a couple of constants
+import "unicode" // only needed for a couple of constants
// Numbers fundamental to the encoding.
const (
- RuneError = unicode.ReplacementChar; // the "error" Rune or "replacement character".
- RuneSelf = 0x80; // characters below Runeself are represented as themselves in a single byte.
- UTFMax = 4; // maximum number of bytes of a UTF-8 encoded Unicode character.
+ RuneError = unicode.ReplacementChar // the "error" Rune or "replacement character".
+ RuneSelf = 0x80 // characters below Runeself are represented as themselves in a single byte.
+ UTFMax = 4 // maximum number of bytes of a UTF-8 encoded Unicode character.
)
const (
- _T1 = 0x00; // 0000 0000
- _Tx = 0x80; // 1000 0000
- _T2 = 0xC0; // 1100 0000
- _T3 = 0xE0; // 1110 0000
- _T4 = 0xF0; // 1111 0000
- _T5 = 0xF8; // 1111 1000
-
- _Maskx = 0x3F; // 0011 1111
- _Mask2 = 0x1F; // 0001 1111
- _Mask3 = 0x0F; // 0000 1111
- _Mask4 = 0x07; // 0000 0111
-
- _Rune1Max = 1<<7 - 1;
- _Rune2Max = 1<<11 - 1;
- _Rune3Max = 1<<16 - 1;
- _Rune4Max = 1<<21 - 1;
+ _T1 = 0x00 // 0000 0000
+ _Tx = 0x80 // 1000 0000
+ _T2 = 0xC0 // 1100 0000
+ _T3 = 0xE0 // 1110 0000
+ _T4 = 0xF0 // 1111 0000
+ _T5 = 0xF8 // 1111 1000
+
+ _Maskx = 0x3F // 0011 1111
+ _Mask2 = 0x1F // 0001 1111
+ _Mask3 = 0x0F // 0000 1111
+ _Mask4 = 0x07 // 0000 0111
+
+ _Rune1Max = 1<<7 - 1
+ _Rune2Max = 1<<11 - 1
+ _Rune3Max = 1<<16 - 1
+ _Rune4Max = 1<<21 - 1
)
func decodeRuneInternal(p []byte) (rune, size int, short bool) {
- n := len(p);
+ n := len(p)
if n < 1 {
return RuneError, 0, true
}
- c0 := p[0];
+ c0 := p[0]
// 1-byte, 7-bit sequence?
if c0 < _Tx {
@@ -55,66 +55,66 @@ func decodeRuneInternal(p []byte) (rune, size int, short bool) {
if n < 2 {
return RuneError, 1, true
}
- c1 := p[1];
+ c1 := p[1]
if c1 < _Tx || _T2 <= c1 {
return RuneError, 1, false
}
// 2-byte, 11-bit sequence?
if c0 < _T3 {
- rune = int(c0&_Mask2)<<6 | int(c1&_Maskx);
+ rune = int(c0&_Mask2)<<6 | int(c1&_Maskx)
if rune <= _Rune1Max {
return RuneError, 1, false
}
- return rune, 2, false;
+ return rune, 2, false
}
// need second continuation byte
if n < 3 {
return RuneError, 1, true
}
- c2 := p[2];
+ c2 := p[2]
if c2 < _Tx || _T2 <= c2 {
return RuneError, 1, false
}
// 3-byte, 16-bit sequence?
if c0 < _T4 {
- rune = int(c0&_Mask3)<<12 | int(c1&_Maskx)<<6 | int(c2&_Maskx);
+ rune = int(c0&_Mask3)<<12 | int(c1&_Maskx)<<6 | int(c2&_Maskx)
if rune <= _Rune2Max {
return RuneError, 1, false
}
- return rune, 3, false;
+ return rune, 3, false
}
// need third continuation byte
if n < 4 {
return RuneError, 1, true
}
- c3 := p[3];
+ c3 := p[3]
if c3 < _Tx || _T2 <= c3 {
return RuneError, 1, false
}
// 4-byte, 21-bit sequence?
if c0 < _T5 {
- rune = int(c0&_Mask4)<<18 | int(c1&_Maskx)<<12 | int(c2&_Maskx)<<6 | int(c3&_Maskx);
+ rune = int(c0&_Mask4)<<18 | int(c1&_Maskx)<<12 | int(c2&_Maskx)<<6 | int(c3&_Maskx)
if rune <= _Rune3Max {
return RuneError, 1, false
}
- return rune, 4, false;
+ return rune, 4, false
}
// error
- return RuneError, 1, false;
+ return RuneError, 1, false
}
func decodeRuneInStringInternal(s string) (rune, size int, short bool) {
- n := len(s);
+ n := len(s)
if n < 1 {
return RuneError, 0, true
}
- c0 := s[0];
+ c0 := s[0]
// 1-byte, 7-bit sequence?
if c0 < _Tx {
@@ -130,83 +130,83 @@ func decodeRuneInStringInternal(s string) (rune, size int, short bool) {
if n < 2 {
return RuneError, 1, true
}
- c1 := s[1];
+ c1 := s[1]
if c1 < _Tx || _T2 <= c1 {
return RuneError, 1, false
}
// 2-byte, 11-bit sequence?
if c0 < _T3 {
- rune = int(c0&_Mask2)<<6 | int(c1&_Maskx);
+ rune = int(c0&_Mask2)<<6 | int(c1&_Maskx)
if rune <= _Rune1Max {
return RuneError, 1, false
}
- return rune, 2, false;
+ return rune, 2, false
}
// need second continuation byte
if n < 3 {
return RuneError, 1, true
}
- c2 := s[2];
+ c2 := s[2]
if c2 < _Tx || _T2 <= c2 {
return RuneError, 1, false
}
// 3-byte, 16-bit sequence?
if c0 < _T4 {
- rune = int(c0&_Mask3)<<12 | int(c1&_Maskx)<<6 | int(c2&_Maskx);
+ rune = int(c0&_Mask3)<<12 | int(c1&_Maskx)<<6 | int(c2&_Maskx)
if rune <= _Rune2Max {
return RuneError, 1, false
}
- return rune, 3, false;
+ return rune, 3, false
}
// need third continuation byte
if n < 4 {
return RuneError, 1, true
}
- c3 := s[3];
+ c3 := s[3]
if c3 < _Tx || _T2 <= c3 {
return RuneError, 1, false
}
// 4-byte, 21-bit sequence?
if c0 < _T5 {
- rune = int(c0&_Mask4)<<18 | int(c1&_Maskx)<<12 | int(c2&_Maskx)<<6 | int(c3&_Maskx);
+ rune = int(c0&_Mask4)<<18 | int(c1&_Maskx)<<12 | int(c2&_Maskx)<<6 | int(c3&_Maskx)
if rune <= _Rune3Max {
return RuneError, 1, false
}
- return rune, 4, false;
+ return rune, 4, false
}
// error
- return RuneError, 1, false;
+ return RuneError, 1, false
}
// FullRune reports whether the bytes in p begin with a full UTF-8 encoding of a rune.
// An invalid encoding is considered a full Rune since it will convert as a width-1 error rune.
func FullRune(p []byte) bool {
- _, _, short := decodeRuneInternal(p);
- return !short;
+ _, _, short := decodeRuneInternal(p)
+ return !short
}
// FullRuneInString is like FullRune but its input is a string.
func FullRuneInString(s string) bool {
- _, _, short := decodeRuneInStringInternal(s);
- return !short;
+ _, _, short := decodeRuneInStringInternal(s)
+ return !short
}
// DecodeRune unpacks the first UTF-8 encoding in p and returns the rune and its width in bytes.
func DecodeRune(p []byte) (rune, size int) {
- rune, size, _ = decodeRuneInternal(p);
- return;
+ rune, size, _ = decodeRuneInternal(p)
+ return
}
// DecodeRuneInString is like DecodeRune but its input is a string.
func DecodeRuneInString(s string) (rune, size int) {
- rune, size, _ = decodeRuneInStringInternal(s);
- return;
+ rune, size, _ = decodeRuneInStringInternal(s)
+ return
}
// RuneLen returns the number of bytes required to encode the rune.
@@ -221,24 +221,24 @@ func RuneLen(rune int) int {
case rune <= _Rune4Max:
return 4
}
- return -1;
+ return -1
}
// EncodeRune writes into p (which must be large enough) the UTF-8 encoding of the rune.
// It returns the number of bytes written.
func EncodeRune(rune int, p []byte) int {
// Negative values are erroneous. Making it unsigned addresses the problem.
- r := uint(rune);
+ r := uint(rune)
if r <= _Rune1Max {
- p[0] = byte(r);
- return 1;
+ p[0] = byte(r)
+ return 1
}
if r <= _Rune2Max {
- p[0] = _T2 | byte(r>>6);
- p[1] = _Tx | byte(r)&_Maskx;
- return 2;
+ p[0] = _T2 | byte(r>>6)
+ p[1] = _Tx | byte(r)&_Maskx
+ return 2
}
if r > unicode.MaxRune {
@@ -246,33 +246,33 @@ func EncodeRune(rune int, p []byte) int {
}
if r <= _Rune3Max {
- p[0] = _T3 | byte(r>>12);
- p[1] = _Tx | byte(r>>6)&_Maskx;
- p[2] = _Tx | byte(r)&_Maskx;
- return 3;
+ p[0] = _T3 | byte(r>>12)
+ p[1] = _Tx | byte(r>>6)&_Maskx
+ p[2] = _Tx | byte(r)&_Maskx
+ return 3
}
- p[0] = _T4 | byte(r>>18);
- p[1] = _Tx | byte(r>>12)&_Maskx;
- p[2] = _Tx | byte(r>>6)&_Maskx;
- p[3] = _Tx | byte(r)&_Maskx;
- return 4;
+ p[0] = _T4 | byte(r>>18)
+ p[1] = _Tx | byte(r>>12)&_Maskx
+ p[2] = _Tx | byte(r>>6)&_Maskx
+ p[3] = _Tx | byte(r)&_Maskx
+ return 4
}
// RuneCount returns the number of runes in p. Erroneous and short
// encodings are treated as single runes of width 1 byte.
func RuneCount(p []byte) int {
- i := 0;
- var n int;
+ i := 0
+ var n int
for n = 0; i < len(p); n++ {
if p[i] < RuneSelf {
i++
} else {
- _, size := DecodeRune(p[i:]);
- i += size;
+ _, size := DecodeRune(p[i:])
+ i += size
}
}
- return n;
+ return n
}
// RuneCountInString is like RuneCount but its input is a string.
@@ -280,10 +280,10 @@ func RuneCountInString(s string) (n int) {
for _ = range s {
n++
}
- return;
+ return
}
// RuneStart reports whether the byte could be the first byte of
// an encoded rune. Second and subsequent bytes always have the top
// two bits set to 10.
-func RuneStart(b byte) bool { return b&0xC0 != 0x80 }
+func RuneStart(b byte) bool { return b&0xC0 != 0x80 }
diff --git a/src/pkg/utf8/utf8_test.go b/src/pkg/utf8/utf8_test.go
index 595efc634..68bfa6a77 100644
--- a/src/pkg/utf8/utf8_test.go
+++ b/src/pkg/utf8/utf8_test.go
@@ -5,15 +5,15 @@
package utf8_test
import (
- "bytes";
- "strings";
- "testing";
- . "utf8";
+ "bytes"
+ "strings"
+ "testing"
+ . "utf8"
)
type Utf8Map struct {
- rune int;
- str string;
+ rune int
+ str string
}
var utf8map = []Utf8Map{
@@ -47,27 +47,27 @@ var utf8map = []Utf8Map{
// strings.Bytes with one extra byte at end
func makeBytes(s string) []byte {
- s += "\x00";
- b := strings.Bytes(s);
- return b[0 : len(s)-1];
+ s += "\x00"
+ b := strings.Bytes(s)
+ return b[0 : len(s)-1]
}
func TestFullRune(t *testing.T) {
for i := 0; i < len(utf8map); i++ {
- m := utf8map[i];
- b := makeBytes(m.str);
+ m := utf8map[i]
+ b := makeBytes(m.str)
if !FullRune(b) {
t.Errorf("FullRune(%q) (rune %04x) = false, want true", b, m.rune)
}
- s := m.str;
+ s := m.str
if !FullRuneInString(s) {
t.Errorf("FullRuneInString(%q) (rune %04x) = false, want true", s, m.rune)
}
- b1 := b[0 : len(b)-1];
+ b1 := b[0 : len(b)-1]
if FullRune(b1) {
t.Errorf("FullRune(%q) = true, want false", b1)
}
- s1 := string(b1);
+ s1 := string(b1)
if FullRuneInString(s1) {
t.Errorf("FullRune(%q) = true, want false", s1)
}
@@ -76,11 +76,11 @@ func TestFullRune(t *testing.T) {
func TestEncodeRune(t *testing.T) {
for i := 0; i < len(utf8map); i++ {
- m := utf8map[i];
- b := makeBytes(m.str);
- var buf [10]byte;
- n := EncodeRune(m.rune, &buf);
- b1 := buf[0:n];
+ m := utf8map[i]
+ b := makeBytes(m.str)
+ var buf [10]byte
+ n := EncodeRune(m.rune, &buf)
+ b1 := buf[0:n]
if !bytes.Equal(b, b1) {
t.Errorf("EncodeRune(%#04x) = %q want %q", m.rune, b1, b)
}
@@ -89,40 +89,40 @@ func TestEncodeRune(t *testing.T) {
func TestDecodeRune(t *testing.T) {
for i := 0; i < len(utf8map); i++ {
- m := utf8map[i];
- b := makeBytes(m.str);
- rune, size := DecodeRune(b);
+ m := utf8map[i]
+ b := makeBytes(m.str)
+ rune, size := DecodeRune(b)
if rune != m.rune || size != len(b) {
t.Errorf("DecodeRune(%q) = %#04x, %d want %#04x, %d", b, rune, size, m.rune, len(b))
}
- s := m.str;
- rune, size = DecodeRuneInString(s);
+ s := m.str
+ rune, size = DecodeRuneInString(s)
if rune != m.rune || size != len(b) {
t.Errorf("DecodeRune(%q) = %#04x, %d want %#04x, %d", s, rune, size, m.rune, len(b))
}
// there's an extra byte that bytes left behind - make sure trailing byte works
- rune, size = DecodeRune(b[0:cap(b)]);
+ rune, size = DecodeRune(b[0:cap(b)])
if rune != m.rune || size != len(b) {
t.Errorf("DecodeRune(%q) = %#04x, %d want %#04x, %d", b, rune, size, m.rune, len(b))
}
- s = m.str + "\x00";
- rune, size = DecodeRuneInString(s);
+ s = m.str + "\x00"
+ rune, size = DecodeRuneInString(s)
if rune != m.rune || size != len(b) {
t.Errorf("DecodeRuneInString(%q) = %#04x, %d want %#04x, %d", s, rune, size, m.rune, len(b))
}
// make sure missing bytes fail
- wantsize := 1;
+ wantsize := 1
if wantsize >= len(b) {
wantsize = 0
}
- rune, size = DecodeRune(b[0 : len(b)-1]);
+ rune, size = DecodeRune(b[0 : len(b)-1])
if rune != RuneError || size != wantsize {
t.Errorf("DecodeRune(%q) = %#04x, %d want %#04x, %d", b[0:len(b)-1], rune, size, RuneError, wantsize)
}
- s = m.str[0 : len(m.str)-1];
- rune, size = DecodeRuneInString(s);
+ s = m.str[0 : len(m.str)-1]
+ rune, size = DecodeRuneInString(s)
if rune != RuneError || size != wantsize {
t.Errorf("DecodeRuneInString(%q) = %#04x, %d want %#04x, %d", s, rune, size, RuneError, wantsize)
}
@@ -133,12 +133,12 @@ func TestDecodeRune(t *testing.T) {
} else {
b[len(b)-1] = 0x7F
}
- rune, size = DecodeRune(b);
+ rune, size = DecodeRune(b)
if rune != RuneError || size != 1 {
t.Errorf("DecodeRune(%q) = %#04x, %d want %#04x, %d", b, rune, size, RuneError, 1)
}
- s = string(b);
- rune, size = DecodeRune(b);
+ s = string(b)
+ rune, size = DecodeRune(b)
if rune != RuneError || size != 1 {
t.Errorf("DecodeRuneInString(%q) = %#04x, %d want %#04x, %d", s, rune, size, RuneError, 1)
}
@@ -147,18 +147,18 @@ func TestDecodeRune(t *testing.T) {
// Check that negative runes encode as U+FFFD.
func TestNegativeRune(t *testing.T) {
- errorbuf := make([]byte, UTFMax);
- errorbuf = errorbuf[0:EncodeRune(RuneError, errorbuf)];
- buf := make([]byte, UTFMax);
- buf = buf[0:EncodeRune(-1, buf)];
+ errorbuf := make([]byte, UTFMax)
+ errorbuf = errorbuf[0:EncodeRune(RuneError, errorbuf)]
+ buf := make([]byte, UTFMax)
+ buf = buf[0:EncodeRune(-1, buf)]
if !bytes.Equal(buf, errorbuf) {
t.Errorf("incorrect encoding [% x] for -1; expected [% x]", buf, errorbuf)
}
}
type RuneCountTest struct {
- in string;
- out int;
+ in string
+ out int
}
var runecounttests = []RuneCountTest{
@@ -170,7 +170,7 @@ var runecounttests = []RuneCountTest{
func TestRuneCount(t *testing.T) {
for i := 0; i < len(runecounttests); i++ {
- tt := runecounttests[i];
+ tt := runecounttests[i]
if out := RuneCountInString(tt.in); out != tt.out {
t.Errorf("RuneCountInString(%q) = %d, want %d", tt.in, out, tt.out)
}
@@ -193,28 +193,28 @@ func BenchmarkRuneCountTenJapaneseChars(b *testing.B) {
}
func BenchmarkEncodeASCIIRune(b *testing.B) {
- buf := make([]byte, UTFMax);
+ buf := make([]byte, UTFMax)
for i := 0; i < b.N; i++ {
EncodeRune('a', buf)
}
}
func BenchmarkEncodeJapaneseRune(b *testing.B) {
- buf := make([]byte, UTFMax);
+ buf := make([]byte, UTFMax)
for i := 0; i < b.N; i++ {
EncodeRune('本', buf)
}
}
func BenchmarkDecodeASCIIRune(b *testing.B) {
- a := []byte{'a'};
+ a := []byte{'a'}
for i := 0; i < b.N; i++ {
DecodeRune(a)
}
}
func BenchmarkDecodeJapaneseRune(b *testing.B) {
- nihon := strings.Bytes("本");
+ nihon := strings.Bytes("本")
for i := 0; i < b.N; i++ {
DecodeRune(nihon)
}
generated by cgit v1.2.3 (git 2.46.0) at 2025-10-27 08:11:35 +0000