Path: blob/main/vendor/golang.org/x/text/encoding/japanese/shiftjis.go
2893 views
// Copyright 2013 The Go Authors. All rights reserved.1// Use of this source code is governed by a BSD-style2// license that can be found in the LICENSE file.34package japanese56import (7"unicode/utf8"89"golang.org/x/text/encoding"10"golang.org/x/text/encoding/internal"11"golang.org/x/text/encoding/internal/identifier"12"golang.org/x/text/transform"13)1415// ShiftJIS is the Shift JIS encoding, also known as Code Page 932 and16// Windows-31J.17var ShiftJIS encoding.Encoding = &shiftJIS1819var shiftJIS = internal.Encoding{20&internal.SimpleEncoding{shiftJISDecoder{}, shiftJISEncoder{}},21"Shift JIS",22identifier.ShiftJIS,23}2425type shiftJISDecoder struct{ transform.NopResetter }2627func (shiftJISDecoder) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {28r, size := rune(0), 029loop:30for ; nSrc < len(src); nSrc += size {31switch c0 := src[nSrc]; {32case c0 < utf8.RuneSelf:33r, size = rune(c0), 13435case 0xa1 <= c0 && c0 < 0xe0:36r, size = rune(c0)+(0xff61-0xa1), 13738case (0x81 <= c0 && c0 < 0xa0) || (0xe0 <= c0 && c0 < 0xfd):39if c0 <= 0x9f {40c0 -= 0x7041} else {42c0 -= 0xb043}44c0 = 2*c0 - 0x214546if nSrc+1 >= len(src) {47if !atEOF {48err = transform.ErrShortSrc49break loop50}51r, size = '\ufffd', 152goto write53}54c1 := src[nSrc+1]55switch {56case c1 < 0x40:57r, size = '\ufffd', 1 // c1 is ASCII so output on next round58goto write59case c1 < 0x7f:60c0--61c1 -= 0x4062case c1 == 0x7f:63r, size = '\ufffd', 1 // c1 is ASCII so output on next round64goto write65case c1 < 0x9f:66c0--67c1 -= 0x4168case c1 < 0xfd:69c1 -= 0x9f70default:71r, size = '\ufffd', 272goto write73}74r, size = '\ufffd', 275if i := int(c0)*94 + int(c1); i < len(jis0208Decode) {76r = rune(jis0208Decode[i])77if r == 0 {78r = '\ufffd'79}80}8182case c0 == 0x80:83r, size = 0x80, 18485default:86r, size = '\ufffd', 187}88write:89if nDst+utf8.RuneLen(r) > len(dst) {90err = transform.ErrShortDst91break loop92}93nDst += utf8.EncodeRune(dst[nDst:], r)94}95return nDst, nSrc, err96}9798type shiftJISEncoder struct{ transform.NopResetter }99100func (shiftJISEncoder) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {101r, size := rune(0), 0102loop:103for ; nSrc < len(src); nSrc += size {104r = rune(src[nSrc])105106// Decode a 1-byte rune.107if r < utf8.RuneSelf {108size = 1109110} else {111// Decode a multi-byte rune.112r, size = utf8.DecodeRune(src[nSrc:])113if size == 1 {114// All valid runes of size 1 (those below utf8.RuneSelf) were115// handled above. We have invalid UTF-8 or we haven't seen the116// full character yet.117if !atEOF && !utf8.FullRune(src[nSrc:]) {118err = transform.ErrShortSrc119break loop120}121}122123// func init checks that the switch covers all tables.124switch {125case encode0Low <= r && r < encode0High:126if r = rune(encode0[r-encode0Low]); r>>tableShift == jis0208 {127goto write2128}129case encode1Low <= r && r < encode1High:130if r = rune(encode1[r-encode1Low]); r>>tableShift == jis0208 {131goto write2132}133case encode2Low <= r && r < encode2High:134if r = rune(encode2[r-encode2Low]); r>>tableShift == jis0208 {135goto write2136}137case encode3Low <= r && r < encode3High:138if r = rune(encode3[r-encode3Low]); r>>tableShift == jis0208 {139goto write2140}141case encode4Low <= r && r < encode4High:142if r = rune(encode4[r-encode4Low]); r>>tableShift == jis0208 {143goto write2144}145case encode5Low <= r && r < encode5High:146if 0xff61 <= r && r < 0xffa0 {147r -= 0xff61 - 0xa1148goto write1149}150if r = rune(encode5[r-encode5Low]); r>>tableShift == jis0208 {151goto write2152}153}154err = internal.ErrASCIIReplacement155break156}157158write1:159if nDst >= len(dst) {160err = transform.ErrShortDst161break162}163dst[nDst] = uint8(r)164nDst++165continue166167write2:168j1 := uint8(r>>codeShift) & codeMask169j2 := uint8(r) & codeMask170if nDst+2 > len(dst) {171err = transform.ErrShortDst172break loop173}174if j1 <= 61 {175dst[nDst+0] = 129 + j1/2176} else {177dst[nDst+0] = 193 + j1/2178}179if j1&1 == 0 {180dst[nDst+1] = j2 + j2/63 + 64181} else {182dst[nDst+1] = j2 + 159183}184nDst += 2185continue186}187return nDst, nSrc, err188}189190191