Repository: saintfish/chardet Branch: master Commit: 5e3ef4b5456d Files: 26 Total size: 1.2 MB Directory structure: gitextract_f_ghewj2/ ├── 2022.go ├── AUTHORS ├── LICENSE ├── README.md ├── detector.go ├── detector_test.go ├── example_test.go ├── icu-license.html ├── multi_byte.go ├── recognizer.go ├── single_byte.go ├── testdata/ │ ├── 8859_1_da.html │ ├── 8859_1_de.html │ ├── 8859_1_en.html │ ├── 8859_1_es.html │ ├── 8859_1_fr.html │ ├── 8859_1_pt.html │ ├── big5.html │ ├── euc_jp.html │ ├── euc_kr.html │ ├── gb18030.html │ ├── shift_jis.html │ ├── utf8.html │ └── utf8_bom.html ├── unicode.go └── utf8.go ================================================ FILE CONTENTS ================================================ ================================================ FILE: 2022.go ================================================ package chardet import ( "bytes" ) type recognizer2022 struct { charset string escapes [][]byte } func (r *recognizer2022) Match(input *recognizerInput) (output recognizerOutput) { return recognizerOutput{ Charset: r.charset, Confidence: r.matchConfidence(input.input), } } func (r *recognizer2022) matchConfidence(input []byte) int { var hits, misses, shifts int input: for i := 0; i < len(input); i++ { c := input[i] if c == 0x1B { for _, esc := range r.escapes { if bytes.HasPrefix(input[i+1:], esc) { hits++ i += len(esc) continue input } } misses++ } else if c == 0x0E || c == 0x0F { shifts++ } } if hits == 0 { return 0 } quality := (100*hits - 100*misses) / (hits + misses) if hits+shifts < 5 { quality -= (5 - (hits + shifts)) * 10 } if quality < 0 { quality = 0 } return quality } var escapeSequences_2022JP = [][]byte{ {0x24, 0x28, 0x43}, // KS X 1001:1992 {0x24, 0x28, 0x44}, // JIS X 212-1990 {0x24, 0x40}, // JIS C 6226-1978 {0x24, 0x41}, // GB 2312-80 {0x24, 0x42}, // JIS X 208-1983 {0x26, 0x40}, // JIS X 208 1990, 1997 {0x28, 0x42}, // ASCII {0x28, 0x48}, // JIS-Roman {0x28, 0x49}, // Half-width katakana {0x28, 0x4a}, // JIS-Roman {0x2e, 0x41}, // ISO 8859-1 {0x2e, 0x46}, // ISO 8859-7 } var escapeSequences_2022KR = [][]byte{ {0x24, 0x29, 0x43}, } var escapeSequences_2022CN = [][]byte{ {0x24, 0x29, 0x41}, // GB 2312-80 {0x24, 0x29, 0x47}, // CNS 11643-1992 Plane 1 {0x24, 0x2A, 0x48}, // CNS 11643-1992 Plane 2 {0x24, 0x29, 0x45}, // ISO-IR-165 {0x24, 0x2B, 0x49}, // CNS 11643-1992 Plane 3 {0x24, 0x2B, 0x4A}, // CNS 11643-1992 Plane 4 {0x24, 0x2B, 0x4B}, // CNS 11643-1992 Plane 5 {0x24, 0x2B, 0x4C}, // CNS 11643-1992 Plane 6 {0x24, 0x2B, 0x4D}, // CNS 11643-1992 Plane 7 {0x4e}, // SS2 {0x4f}, // SS3 } func newRecognizer_2022JP() *recognizer2022 { return &recognizer2022{ "ISO-2022-JP", escapeSequences_2022JP, } } func newRecognizer_2022KR() *recognizer2022 { return &recognizer2022{ "ISO-2022-KR", escapeSequences_2022KR, } } func newRecognizer_2022CN() *recognizer2022 { return &recognizer2022{ "ISO-2022-CN", escapeSequences_2022CN, } } ================================================ FILE: AUTHORS ================================================ Sheng Yu (yusheng dot sjtu at gmail dot com) ================================================ FILE: LICENSE ================================================ Copyright (c) 2012 chardet Authors Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. Partial of the Software is derived from ICU project. See icu-license.html for license of the derivative portions. ================================================ FILE: README.md ================================================ # chardet chardet is library to automatically detect [charset](http://en.wikipedia.org/wiki/Character_encoding) of texts for [Go programming language](http://golang.org/). It's based on the algorithm and data in [ICU](http://icu-project.org/)'s implementation. ## Documentation and Usage See [pkgdoc](https://pkg.go.dev/github.com/saintfish/chardet) ================================================ FILE: detector.go ================================================ // Package chardet ports character set detection from ICU. package chardet import ( "errors" "sort" ) // Result contains all the information that charset detector gives. type Result struct { // IANA name of the detected charset. Charset string // IANA name of the detected language. It may be empty for some charsets. Language string // Confidence of the Result. Scale from 1 to 100. The bigger, the more confident. Confidence int } // Detector implements charset detection. type Detector struct { recognizers []recognizer stripTag bool } // List of charset recognizers var recognizers = []recognizer{ newRecognizer_utf8(), newRecognizer_utf16be(), newRecognizer_utf16le(), newRecognizer_utf32be(), newRecognizer_utf32le(), newRecognizer_8859_1_en(), newRecognizer_8859_1_da(), newRecognizer_8859_1_de(), newRecognizer_8859_1_es(), newRecognizer_8859_1_fr(), newRecognizer_8859_1_it(), newRecognizer_8859_1_nl(), newRecognizer_8859_1_no(), newRecognizer_8859_1_pt(), newRecognizer_8859_1_sv(), newRecognizer_8859_2_cs(), newRecognizer_8859_2_hu(), newRecognizer_8859_2_pl(), newRecognizer_8859_2_ro(), newRecognizer_8859_5_ru(), newRecognizer_8859_6_ar(), newRecognizer_8859_7_el(), newRecognizer_8859_8_I_he(), newRecognizer_8859_8_he(), newRecognizer_windows_1251(), newRecognizer_windows_1256(), newRecognizer_KOI8_R(), newRecognizer_8859_9_tr(), newRecognizer_sjis(), newRecognizer_gb_18030(), newRecognizer_euc_jp(), newRecognizer_euc_kr(), newRecognizer_big5(), newRecognizer_2022JP(), newRecognizer_2022KR(), newRecognizer_2022CN(), newRecognizer_IBM424_he_rtl(), newRecognizer_IBM424_he_ltr(), newRecognizer_IBM420_ar_rtl(), newRecognizer_IBM420_ar_ltr(), } // NewTextDetector creates a Detector for plain text. func NewTextDetector() *Detector { return &Detector{recognizers, false} } // NewHtmlDetector creates a Detector for Html. func NewHtmlDetector() *Detector { return &Detector{recognizers, true} } var ( NotDetectedError = errors.New("Charset not detected.") ) // DetectBest returns the Result with highest Confidence. func (d *Detector) DetectBest(b []byte) (r *Result, err error) { var all []Result if all, err = d.DetectAll(b); err == nil { r = &all[0] } return } // DetectAll returns all Results which have non-zero Confidence. The Results are sorted by Confidence in descending order. func (d *Detector) DetectAll(b []byte) ([]Result, error) { input := newRecognizerInput(b, d.stripTag) outputChan := make(chan recognizerOutput) for _, r := range d.recognizers { go matchHelper(r, input, outputChan) } outputs := make([]recognizerOutput, 0, len(d.recognizers)) for i := 0; i < len(d.recognizers); i++ { o := <-outputChan if o.Confidence > 0 { outputs = append(outputs, o) } } if len(outputs) == 0 { return nil, NotDetectedError } sort.Sort(recognizerOutputs(outputs)) dedupOutputs := make([]Result, 0, len(outputs)) foundCharsets := make(map[string]struct{}, len(outputs)) for _, o := range outputs { if _, found := foundCharsets[o.Charset]; !found { dedupOutputs = append(dedupOutputs, Result(o)) foundCharsets[o.Charset] = struct{}{} } } if len(dedupOutputs) == 0 { return nil, NotDetectedError } return dedupOutputs, nil } func matchHelper(r recognizer, input *recognizerInput, outputChan chan<- recognizerOutput) { outputChan <- r.Match(input) } type recognizerOutputs []recognizerOutput func (r recognizerOutputs) Len() int { return len(r) } func (r recognizerOutputs) Less(i, j int) bool { return r[i].Confidence > r[j].Confidence } func (r recognizerOutputs) Swap(i, j int) { r[i], r[j] = r[j], r[i] } ================================================ FILE: detector_test.go ================================================ package chardet_test import ( "github.com/saintfish/chardet" "io" "os" "path/filepath" "testing" ) func TestDetector(t *testing.T) { type file_charset_language struct { File string IsHtml bool Charset string Language string } var data = []file_charset_language{ {"utf8.html", true, "UTF-8", ""}, {"utf8_bom.html", true, "UTF-8", ""}, {"8859_1_en.html", true, "ISO-8859-1", "en"}, {"8859_1_da.html", true, "ISO-8859-1", "da"}, {"8859_1_de.html", true, "ISO-8859-1", "de"}, {"8859_1_es.html", true, "ISO-8859-1", "es"}, {"8859_1_fr.html", true, "ISO-8859-1", "fr"}, {"8859_1_pt.html", true, "ISO-8859-1", "pt"}, {"shift_jis.html", true, "Shift_JIS", "ja"}, {"gb18030.html", true, "GB-18030", "zh"}, {"euc_jp.html", true, "EUC-JP", "ja"}, {"euc_kr.html", true, "EUC-KR", "ko"}, {"big5.html", true, "Big5", "zh"}, } textDetector := chardet.NewTextDetector() htmlDetector := chardet.NewHtmlDetector() buffer := make([]byte, 32<<10) for _, d := range data { f, err := os.Open(filepath.Join("testdata", d.File)) if err != nil { t.Fatal(err) } defer f.Close() size, _ := io.ReadFull(f, buffer) input := buffer[:size] var detector = textDetector if d.IsHtml { detector = htmlDetector } result, err := detector.DetectBest(input) if err != nil { t.Fatal(err) } if result.Charset != d.Charset { t.Errorf("Expected charset %s, actual %s", d.Charset, result.Charset) } if result.Language != d.Language { t.Errorf("Expected language %s, actual %s", d.Language, result.Language) } } } ================================================ FILE: example_test.go ================================================ package chardet_test import ( "fmt" "github.com/saintfish/chardet" ) var ( zh_gb18030_text = []byte{ 71, 111, 202, 199, 71, 111, 111, 103, 108, 101, 233, 95, 176, 108, 181, 196, 210, 187, 214, 214, 190, 142, 215, 103, 208, 205, 163, 172, 129, 75, 176, 108, 208, 205, 163, 172, 178, 162, 190, 223, 211, 208, 192, 172, 187, 248, 187, 216, 202, 213, 185, 166, 196, 220, 181, 196, 177, 224, 179, 204, 211, 239, 209, 212, 161, 163, 10, } ) func ExampleTextDetector() { detector := chardet.NewTextDetector() result, err := detector.DetectBest(zh_gb18030_text) if err == nil { fmt.Printf( "Detected charset is %s, language is %s", result.Charset, result.Language) } // Output: // Detected charset is GB-18030, language is zh } ================================================ FILE: icu-license.html ================================================ ICU License - ICU 1.8.1 and later

ICU License - ICU 1.8.1 and later

COPYRIGHT AND PERMISSION NOTICE

Copyright (c) 1995-2012 International Business Machines Corporation and others

All rights reserved.

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, provided that the above copyright notice(s) and this permission notice appear in all copies of the Software and that both the above copyright notice(s) and this permission notice appear in supporting documentation.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT OF THIRD PARTY RIGHTS. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR HOLDERS INCLUDED IN THIS NOTICE BE LIABLE FOR ANY CLAIM, OR ANY SPECIAL INDIRECT OR CONSEQUENTIAL DAMAGES, OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.

Except as contained in this notice, the name of a copyright holder shall not be used in advertising or otherwise to promote the sale, use or other dealings in this Software without prior written authorization of the copyright holder.


All trademarks and registered trademarks mentioned herein are the property of their respective owners.

================================================ FILE: multi_byte.go ================================================ package chardet import ( "errors" "math" ) type recognizerMultiByte struct { charset string language string decoder charDecoder commonChars []uint16 } type charDecoder interface { DecodeOneChar([]byte) (c uint16, remain []byte, err error) } func (r *recognizerMultiByte) Match(input *recognizerInput) (output recognizerOutput) { return recognizerOutput{ Charset: r.charset, Language: r.language, Confidence: r.matchConfidence(input), } } func (r *recognizerMultiByte) matchConfidence(input *recognizerInput) int { raw := input.raw var c uint16 var err error var totalCharCount, badCharCount, singleByteCharCount, doubleByteCharCount, commonCharCount int for c, raw, err = r.decoder.DecodeOneChar(raw); len(raw) > 0; c, raw, err = r.decoder.DecodeOneChar(raw) { totalCharCount++ if err != nil { badCharCount++ } else if c <= 0xFF { singleByteCharCount++ } else { doubleByteCharCount++ if r.commonChars != nil && binarySearch(r.commonChars, c) { commonCharCount++ } } if badCharCount >= 2 && badCharCount*5 >= doubleByteCharCount { return 0 } } if doubleByteCharCount <= 10 && badCharCount == 0 { if doubleByteCharCount == 0 && totalCharCount < 10 { return 0 } else { return 10 } } if doubleByteCharCount < 20*badCharCount { return 0 } if r.commonChars == nil { confidence := 30 + doubleByteCharCount - 20*badCharCount if confidence > 100 { confidence = 100 } return confidence } maxVal := math.Log(float64(doubleByteCharCount) / 4) scaleFactor := 90 / maxVal confidence := int(math.Log(float64(commonCharCount)+1)*scaleFactor + 10) if confidence > 100 { confidence = 100 } if confidence < 0 { confidence = 0 } return confidence } func binarySearch(l []uint16, c uint16) bool { start := 0 end := len(l) - 1 for start <= end { mid := (start + end) / 2 if c == l[mid] { return true } else if c < l[mid] { end = mid - 1 } else { start = mid + 1 } } return false } var eobError = errors.New("End of input buffer") var badCharError = errors.New("Decode a bad char") type charDecoder_sjis struct { } func (charDecoder_sjis) DecodeOneChar(input []byte) (c uint16, remain []byte, err error) { if len(input) == 0 { return 0, nil, eobError } first := input[0] c = uint16(first) remain = input[1:] if first <= 0x7F || (first > 0xA0 && first <= 0xDF) { return } if len(remain) == 0 { return c, remain, badCharError } second := remain[0] remain = remain[1:] c = c<<8 | uint16(second) if (second >= 0x40 && second <= 0x7F) || (second >= 0x80 && second <= 0xFE) { } else { err = badCharError } return } var commonChars_sjis = []uint16{ 0x8140, 0x8141, 0x8142, 0x8145, 0x815b, 0x8169, 0x816a, 0x8175, 0x8176, 0x82a0, 0x82a2, 0x82a4, 0x82a9, 0x82aa, 0x82ab, 0x82ad, 0x82af, 0x82b1, 0x82b3, 0x82b5, 0x82b7, 0x82bd, 0x82be, 0x82c1, 0x82c4, 0x82c5, 0x82c6, 0x82c8, 0x82c9, 0x82cc, 0x82cd, 0x82dc, 0x82e0, 0x82e7, 0x82e8, 0x82e9, 0x82ea, 0x82f0, 0x82f1, 0x8341, 0x8343, 0x834e, 0x834f, 0x8358, 0x835e, 0x8362, 0x8367, 0x8375, 0x8376, 0x8389, 0x838a, 0x838b, 0x838d, 0x8393, 0x8e96, 0x93fa, 0x95aa, } func newRecognizer_sjis() *recognizerMultiByte { return &recognizerMultiByte{ "Shift_JIS", "ja", charDecoder_sjis{}, commonChars_sjis, } } type charDecoder_euc struct { } func (charDecoder_euc) DecodeOneChar(input []byte) (c uint16, remain []byte, err error) { if len(input) == 0 { return 0, nil, eobError } first := input[0] remain = input[1:] c = uint16(first) if first <= 0x8D { return uint16(first), remain, nil } if len(remain) == 0 { return 0, nil, eobError } second := remain[0] remain = remain[1:] c = c<<8 | uint16(second) if first >= 0xA1 && first <= 0xFE { if second < 0xA1 { err = badCharError } return } if first == 0x8E { if second < 0xA1 { err = badCharError } return } if first == 0x8F { if len(remain) == 0 { return 0, nil, eobError } third := remain[0] remain = remain[1:] c = c<<0 | uint16(third) if third < 0xa1 { err = badCharError } } return } var commonChars_euc_jp = []uint16{ 0xa1a1, 0xa1a2, 0xa1a3, 0xa1a6, 0xa1bc, 0xa1ca, 0xa1cb, 0xa1d6, 0xa1d7, 0xa4a2, 0xa4a4, 0xa4a6, 0xa4a8, 0xa4aa, 0xa4ab, 0xa4ac, 0xa4ad, 0xa4af, 0xa4b1, 0xa4b3, 0xa4b5, 0xa4b7, 0xa4b9, 0xa4bb, 0xa4bd, 0xa4bf, 0xa4c0, 0xa4c1, 0xa4c3, 0xa4c4, 0xa4c6, 0xa4c7, 0xa4c8, 0xa4c9, 0xa4ca, 0xa4cb, 0xa4ce, 0xa4cf, 0xa4d0, 0xa4de, 0xa4df, 0xa4e1, 0xa4e2, 0xa4e4, 0xa4e8, 0xa4e9, 0xa4ea, 0xa4eb, 0xa4ec, 0xa4ef, 0xa4f2, 0xa4f3, 0xa5a2, 0xa5a3, 0xa5a4, 0xa5a6, 0xa5a7, 0xa5aa, 0xa5ad, 0xa5af, 0xa5b0, 0xa5b3, 0xa5b5, 0xa5b7, 0xa5b8, 0xa5b9, 0xa5bf, 0xa5c3, 0xa5c6, 0xa5c7, 0xa5c8, 0xa5c9, 0xa5cb, 0xa5d0, 0xa5d5, 0xa5d6, 0xa5d7, 0xa5de, 0xa5e0, 0xa5e1, 0xa5e5, 0xa5e9, 0xa5ea, 0xa5eb, 0xa5ec, 0xa5ed, 0xa5f3, 0xb8a9, 0xb9d4, 0xbaee, 0xbbc8, 0xbef0, 0xbfb7, 0xc4ea, 0xc6fc, 0xc7bd, 0xcab8, 0xcaf3, 0xcbdc, 0xcdd1, } var commonChars_euc_kr = []uint16{ 0xb0a1, 0xb0b3, 0xb0c5, 0xb0cd, 0xb0d4, 0xb0e6, 0xb0ed, 0xb0f8, 0xb0fa, 0xb0fc, 0xb1b8, 0xb1b9, 0xb1c7, 0xb1d7, 0xb1e2, 0xb3aa, 0xb3bb, 0xb4c2, 0xb4cf, 0xb4d9, 0xb4eb, 0xb5a5, 0xb5b5, 0xb5bf, 0xb5c7, 0xb5e9, 0xb6f3, 0xb7af, 0xb7c2, 0xb7ce, 0xb8a6, 0xb8ae, 0xb8b6, 0xb8b8, 0xb8bb, 0xb8e9, 0xb9ab, 0xb9ae, 0xb9cc, 0xb9ce, 0xb9fd, 0xbab8, 0xbace, 0xbad0, 0xbaf1, 0xbbe7, 0xbbf3, 0xbbfd, 0xbcad, 0xbcba, 0xbcd2, 0xbcf6, 0xbdba, 0xbdc0, 0xbdc3, 0xbdc5, 0xbec6, 0xbec8, 0xbedf, 0xbeee, 0xbef8, 0xbefa, 0xbfa1, 0xbfa9, 0xbfc0, 0xbfe4, 0xbfeb, 0xbfec, 0xbff8, 0xc0a7, 0xc0af, 0xc0b8, 0xc0ba, 0xc0bb, 0xc0bd, 0xc0c7, 0xc0cc, 0xc0ce, 0xc0cf, 0xc0d6, 0xc0da, 0xc0e5, 0xc0fb, 0xc0fc, 0xc1a4, 0xc1a6, 0xc1b6, 0xc1d6, 0xc1df, 0xc1f6, 0xc1f8, 0xc4a1, 0xc5cd, 0xc6ae, 0xc7cf, 0xc7d1, 0xc7d2, 0xc7d8, 0xc7e5, 0xc8ad, } func newRecognizer_euc_jp() *recognizerMultiByte { return &recognizerMultiByte{ "EUC-JP", "ja", charDecoder_euc{}, commonChars_euc_jp, } } func newRecognizer_euc_kr() *recognizerMultiByte { return &recognizerMultiByte{ "EUC-KR", "ko", charDecoder_euc{}, commonChars_euc_kr, } } type charDecoder_big5 struct { } func (charDecoder_big5) DecodeOneChar(input []byte) (c uint16, remain []byte, err error) { if len(input) == 0 { return 0, nil, eobError } first := input[0] remain = input[1:] c = uint16(first) if first <= 0x7F || first == 0xFF { return } if len(remain) == 0 { return c, nil, eobError } second := remain[0] remain = remain[1:] c = c<<8 | uint16(second) if second < 0x40 || second == 0x7F || second == 0xFF { err = badCharError } return } var commonChars_big5 = []uint16{ 0xa140, 0xa141, 0xa142, 0xa143, 0xa147, 0xa149, 0xa175, 0xa176, 0xa440, 0xa446, 0xa447, 0xa448, 0xa451, 0xa454, 0xa457, 0xa464, 0xa46a, 0xa46c, 0xa477, 0xa4a3, 0xa4a4, 0xa4a7, 0xa4c1, 0xa4ce, 0xa4d1, 0xa4df, 0xa4e8, 0xa4fd, 0xa540, 0xa548, 0xa558, 0xa569, 0xa5cd, 0xa5e7, 0xa657, 0xa661, 0xa662, 0xa668, 0xa670, 0xa6a8, 0xa6b3, 0xa6b9, 0xa6d3, 0xa6db, 0xa6e6, 0xa6f2, 0xa740, 0xa751, 0xa759, 0xa7da, 0xa8a3, 0xa8a5, 0xa8ad, 0xa8d1, 0xa8d3, 0xa8e4, 0xa8fc, 0xa9c0, 0xa9d2, 0xa9f3, 0xaa6b, 0xaaba, 0xaabe, 0xaacc, 0xaafc, 0xac47, 0xac4f, 0xacb0, 0xacd2, 0xad59, 0xaec9, 0xafe0, 0xb0ea, 0xb16f, 0xb2b3, 0xb2c4, 0xb36f, 0xb44c, 0xb44e, 0xb54c, 0xb5a5, 0xb5bd, 0xb5d0, 0xb5d8, 0xb671, 0xb7ed, 0xb867, 0xb944, 0xbad8, 0xbb44, 0xbba1, 0xbdd1, 0xc2c4, 0xc3b9, 0xc440, 0xc45f, } func newRecognizer_big5() *recognizerMultiByte { return &recognizerMultiByte{ "Big5", "zh", charDecoder_big5{}, commonChars_big5, } } type charDecoder_gb_18030 struct { } func (charDecoder_gb_18030) DecodeOneChar(input []byte) (c uint16, remain []byte, err error) { if len(input) == 0 { return 0, nil, eobError } first := input[0] remain = input[1:] c = uint16(first) if first <= 0x80 { return } if len(remain) == 0 { return 0, nil, eobError } second := remain[0] remain = remain[1:] c = c<<8 | uint16(second) if first >= 0x81 && first <= 0xFE { if (second >= 0x40 && second <= 0x7E) || (second >= 0x80 && second <= 0xFE) { return } if second >= 0x30 && second <= 0x39 { if len(remain) == 0 { return 0, nil, eobError } third := remain[0] remain = remain[1:] if third >= 0x81 && third <= 0xFE { if len(remain) == 0 { return 0, nil, eobError } fourth := remain[0] remain = remain[1:] if fourth >= 0x30 && fourth <= 0x39 { c = c<<16 | uint16(third)<<8 | uint16(fourth) return } } } err = badCharError } return } var commonChars_gb_18030 = []uint16{ 0xa1a1, 0xa1a2, 0xa1a3, 0xa1a4, 0xa1b0, 0xa1b1, 0xa1f1, 0xa1f3, 0xa3a1, 0xa3ac, 0xa3ba, 0xb1a8, 0xb1b8, 0xb1be, 0xb2bb, 0xb3c9, 0xb3f6, 0xb4f3, 0xb5bd, 0xb5c4, 0xb5e3, 0xb6af, 0xb6d4, 0xb6e0, 0xb7a2, 0xb7a8, 0xb7bd, 0xb7d6, 0xb7dd, 0xb8b4, 0xb8df, 0xb8f6, 0xb9ab, 0xb9c9, 0xb9d8, 0xb9fa, 0xb9fd, 0xbacd, 0xbba7, 0xbbd6, 0xbbe1, 0xbbfa, 0xbcbc, 0xbcdb, 0xbcfe, 0xbdcc, 0xbecd, 0xbedd, 0xbfb4, 0xbfc6, 0xbfc9, 0xc0b4, 0xc0ed, 0xc1cb, 0xc2db, 0xc3c7, 0xc4dc, 0xc4ea, 0xc5cc, 0xc6f7, 0xc7f8, 0xc8ab, 0xc8cb, 0xc8d5, 0xc8e7, 0xc9cf, 0xc9fa, 0xcab1, 0xcab5, 0xcac7, 0xcad0, 0xcad6, 0xcaf5, 0xcafd, 0xccec, 0xcdf8, 0xceaa, 0xcec4, 0xced2, 0xcee5, 0xcfb5, 0xcfc2, 0xcfd6, 0xd0c2, 0xd0c5, 0xd0d0, 0xd0d4, 0xd1a7, 0xd2aa, 0xd2b2, 0xd2b5, 0xd2bb, 0xd2d4, 0xd3c3, 0xd3d0, 0xd3fd, 0xd4c2, 0xd4da, 0xd5e2, 0xd6d0, } func newRecognizer_gb_18030() *recognizerMultiByte { return &recognizerMultiByte{ "GB-18030", "zh", charDecoder_gb_18030{}, commonChars_gb_18030, } } ================================================ FILE: recognizer.go ================================================ package chardet type recognizer interface { Match(*recognizerInput) recognizerOutput } type recognizerOutput Result type recognizerInput struct { raw []byte input []byte tagStripped bool byteStats []int hasC1Bytes bool } func newRecognizerInput(raw []byte, stripTag bool) *recognizerInput { input, stripped := mayStripInput(raw, stripTag) byteStats := computeByteStats(input) return &recognizerInput{ raw: raw, input: input, tagStripped: stripped, byteStats: byteStats, hasC1Bytes: computeHasC1Bytes(byteStats), } } func mayStripInput(raw []byte, stripTag bool) (out []byte, stripped bool) { const inputBufferSize = 8192 out = make([]byte, 0, inputBufferSize) var badTags, openTags int32 var inMarkup bool = false stripped = false if stripTag { stripped = true for _, c := range raw { if c == '<' { if inMarkup { badTags += 1 } inMarkup = true openTags += 1 } if !inMarkup { out = append(out, c) if len(out) >= inputBufferSize { break } } if c == '>' { inMarkup = false } } } if openTags < 5 || openTags/5 < badTags || (len(out) < 100 && len(raw) > 600) { limit := len(raw) if limit > inputBufferSize { limit = inputBufferSize } out = make([]byte, limit) copy(out, raw[:limit]) stripped = false } return } func computeByteStats(input []byte) []int { r := make([]int, 256) for _, c := range input { r[c] += 1 } return r } func computeHasC1Bytes(byteStats []int) bool { for _, count := range byteStats[0x80 : 0x9F+1] { if count > 0 { return true } } return false } ================================================ FILE: single_byte.go ================================================ package chardet // Recognizer for single byte charset family type recognizerSingleByte struct { charset string hasC1ByteCharset string language string charMap *[256]byte ngram *[64]uint32 } func (r *recognizerSingleByte) Match(input *recognizerInput) recognizerOutput { var charset string = r.charset if input.hasC1Bytes && len(r.hasC1ByteCharset) > 0 { charset = r.hasC1ByteCharset } return recognizerOutput{ Charset: charset, Language: r.language, Confidence: r.parseNgram(input.input), } } type ngramState struct { ngram uint32 ignoreSpace bool ngramCount, ngramHit uint32 table *[64]uint32 } func newNgramState(table *[64]uint32) *ngramState { return &ngramState{ ngram: 0, ignoreSpace: false, ngramCount: 0, ngramHit: 0, table: table, } } func (s *ngramState) AddByte(b byte) { const ngramMask = 0xFFFFFF if !(b == 0x20 && s.ignoreSpace) { s.ngram = ((s.ngram << 8) | uint32(b)) & ngramMask s.ignoreSpace = (s.ngram == 0x20) s.ngramCount++ if s.lookup() { s.ngramHit++ } } s.ignoreSpace = (b == 0x20) } func (s *ngramState) HitRate() float32 { if s.ngramCount == 0 { return 0 } return float32(s.ngramHit) / float32(s.ngramCount) } func (s *ngramState) lookup() bool { var index int if s.table[index+32] <= s.ngram { index += 32 } if s.table[index+16] <= s.ngram { index += 16 } if s.table[index+8] <= s.ngram { index += 8 } if s.table[index+4] <= s.ngram { index += 4 } if s.table[index+2] <= s.ngram { index += 2 } if s.table[index+1] <= s.ngram { index += 1 } if s.table[index] > s.ngram { index -= 1 } if index < 0 || s.table[index] != s.ngram { return false } return true } func (r *recognizerSingleByte) parseNgram(input []byte) int { state := newNgramState(r.ngram) for _, inChar := range input { c := r.charMap[inChar] if c != 0 { state.AddByte(c) } } state.AddByte(0x20) rate := state.HitRate() if rate > 0.33 { return 98 } return int(rate * 300) } var charMap_8859_1 = [256]byte{ 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xAA, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xB5, 0x20, 0x20, 0x20, 0x20, 0xBA, 0x20, 0x20, 0x20, 0x20, 0x20, 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0x20, 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0xDF, 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0x20, 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0xFF, } var ngrams_8859_1_en = [64]uint32{ 0x206120, 0x20616E, 0x206265, 0x20636F, 0x20666F, 0x206861, 0x206865, 0x20696E, 0x206D61, 0x206F66, 0x207072, 0x207265, 0x207361, 0x207374, 0x207468, 0x20746F, 0x207768, 0x616964, 0x616C20, 0x616E20, 0x616E64, 0x617320, 0x617420, 0x617465, 0x617469, 0x642061, 0x642074, 0x652061, 0x652073, 0x652074, 0x656420, 0x656E74, 0x657220, 0x657320, 0x666F72, 0x686174, 0x686520, 0x686572, 0x696420, 0x696E20, 0x696E67, 0x696F6E, 0x697320, 0x6E2061, 0x6E2074, 0x6E6420, 0x6E6720, 0x6E7420, 0x6F6620, 0x6F6E20, 0x6F7220, 0x726520, 0x727320, 0x732061, 0x732074, 0x736169, 0x737420, 0x742074, 0x746572, 0x746861, 0x746865, 0x74696F, 0x746F20, 0x747320, } var ngrams_8859_1_da = [64]uint32{ 0x206166, 0x206174, 0x206465, 0x20656E, 0x206572, 0x20666F, 0x206861, 0x206920, 0x206D65, 0x206F67, 0x2070E5, 0x207369, 0x207374, 0x207469, 0x207669, 0x616620, 0x616E20, 0x616E64, 0x617220, 0x617420, 0x646520, 0x64656E, 0x646572, 0x646574, 0x652073, 0x656420, 0x656465, 0x656E20, 0x656E64, 0x657220, 0x657265, 0x657320, 0x657420, 0x666F72, 0x676520, 0x67656E, 0x676572, 0x696765, 0x696C20, 0x696E67, 0x6B6520, 0x6B6B65, 0x6C6572, 0x6C6967, 0x6C6C65, 0x6D6564, 0x6E6465, 0x6E6520, 0x6E6720, 0x6E6765, 0x6F6720, 0x6F6D20, 0x6F7220, 0x70E520, 0x722064, 0x722065, 0x722073, 0x726520, 0x737465, 0x742073, 0x746520, 0x746572, 0x74696C, 0x766572, } var ngrams_8859_1_de = [64]uint32{ 0x20616E, 0x206175, 0x206265, 0x206461, 0x206465, 0x206469, 0x206569, 0x206765, 0x206861, 0x20696E, 0x206D69, 0x207363, 0x207365, 0x20756E, 0x207665, 0x20766F, 0x207765, 0x207A75, 0x626572, 0x636820, 0x636865, 0x636874, 0x646173, 0x64656E, 0x646572, 0x646965, 0x652064, 0x652073, 0x65696E, 0x656974, 0x656E20, 0x657220, 0x657320, 0x67656E, 0x68656E, 0x687420, 0x696368, 0x696520, 0x696E20, 0x696E65, 0x697420, 0x6C6963, 0x6C6C65, 0x6E2061, 0x6E2064, 0x6E2073, 0x6E6420, 0x6E6465, 0x6E6520, 0x6E6720, 0x6E6765, 0x6E7465, 0x722064, 0x726465, 0x726569, 0x736368, 0x737465, 0x742064, 0x746520, 0x74656E, 0x746572, 0x756E64, 0x756E67, 0x766572, } var ngrams_8859_1_es = [64]uint32{ 0x206120, 0x206361, 0x20636F, 0x206465, 0x20656C, 0x20656E, 0x206573, 0x20696E, 0x206C61, 0x206C6F, 0x207061, 0x20706F, 0x207072, 0x207175, 0x207265, 0x207365, 0x20756E, 0x207920, 0x612063, 0x612064, 0x612065, 0x61206C, 0x612070, 0x616369, 0x61646F, 0x616C20, 0x617220, 0x617320, 0x6369F3, 0x636F6E, 0x646520, 0x64656C, 0x646F20, 0x652064, 0x652065, 0x65206C, 0x656C20, 0x656E20, 0x656E74, 0x657320, 0x657374, 0x69656E, 0x69F36E, 0x6C6120, 0x6C6F73, 0x6E2065, 0x6E7465, 0x6F2064, 0x6F2065, 0x6F6E20, 0x6F7220, 0x6F7320, 0x706172, 0x717565, 0x726120, 0x726573, 0x732064, 0x732065, 0x732070, 0x736520, 0x746520, 0x746F20, 0x756520, 0xF36E20, } var ngrams_8859_1_fr = [64]uint32{ 0x206175, 0x20636F, 0x206461, 0x206465, 0x206475, 0x20656E, 0x206574, 0x206C61, 0x206C65, 0x207061, 0x20706F, 0x207072, 0x207175, 0x207365, 0x20736F, 0x20756E, 0x20E020, 0x616E74, 0x617469, 0x636520, 0x636F6E, 0x646520, 0x646573, 0x647520, 0x652061, 0x652063, 0x652064, 0x652065, 0x65206C, 0x652070, 0x652073, 0x656E20, 0x656E74, 0x657220, 0x657320, 0x657420, 0x657572, 0x696F6E, 0x697320, 0x697420, 0x6C6120, 0x6C6520, 0x6C6573, 0x6D656E, 0x6E2064, 0x6E6520, 0x6E7320, 0x6E7420, 0x6F6E20, 0x6F6E74, 0x6F7572, 0x717565, 0x72206C, 0x726520, 0x732061, 0x732064, 0x732065, 0x73206C, 0x732070, 0x742064, 0x746520, 0x74696F, 0x756520, 0x757220, } var ngrams_8859_1_it = [64]uint32{ 0x20616C, 0x206368, 0x20636F, 0x206465, 0x206469, 0x206520, 0x20696C, 0x20696E, 0x206C61, 0x207065, 0x207072, 0x20756E, 0x612063, 0x612064, 0x612070, 0x612073, 0x61746F, 0x636865, 0x636F6E, 0x64656C, 0x646920, 0x652061, 0x652063, 0x652064, 0x652069, 0x65206C, 0x652070, 0x652073, 0x656C20, 0x656C6C, 0x656E74, 0x657220, 0x686520, 0x692061, 0x692063, 0x692064, 0x692073, 0x696120, 0x696C20, 0x696E20, 0x696F6E, 0x6C6120, 0x6C6520, 0x6C6920, 0x6C6C61, 0x6E6520, 0x6E6920, 0x6E6F20, 0x6E7465, 0x6F2061, 0x6F2064, 0x6F2069, 0x6F2073, 0x6F6E20, 0x6F6E65, 0x706572, 0x726120, 0x726520, 0x736920, 0x746120, 0x746520, 0x746920, 0x746F20, 0x7A696F, } var ngrams_8859_1_nl = [64]uint32{ 0x20616C, 0x206265, 0x206461, 0x206465, 0x206469, 0x206565, 0x20656E, 0x206765, 0x206865, 0x20696E, 0x206D61, 0x206D65, 0x206F70, 0x207465, 0x207661, 0x207665, 0x20766F, 0x207765, 0x207A69, 0x61616E, 0x616172, 0x616E20, 0x616E64, 0x617220, 0x617420, 0x636874, 0x646520, 0x64656E, 0x646572, 0x652062, 0x652076, 0x65656E, 0x656572, 0x656E20, 0x657220, 0x657273, 0x657420, 0x67656E, 0x686574, 0x696520, 0x696E20, 0x696E67, 0x697320, 0x6E2062, 0x6E2064, 0x6E2065, 0x6E2068, 0x6E206F, 0x6E2076, 0x6E6465, 0x6E6720, 0x6F6E64, 0x6F6F72, 0x6F7020, 0x6F7220, 0x736368, 0x737465, 0x742064, 0x746520, 0x74656E, 0x746572, 0x76616E, 0x766572, 0x766F6F, } var ngrams_8859_1_no = [64]uint32{ 0x206174, 0x206176, 0x206465, 0x20656E, 0x206572, 0x20666F, 0x206861, 0x206920, 0x206D65, 0x206F67, 0x2070E5, 0x207365, 0x20736B, 0x20736F, 0x207374, 0x207469, 0x207669, 0x20E520, 0x616E64, 0x617220, 0x617420, 0x646520, 0x64656E, 0x646574, 0x652073, 0x656420, 0x656E20, 0x656E65, 0x657220, 0x657265, 0x657420, 0x657474, 0x666F72, 0x67656E, 0x696B6B, 0x696C20, 0x696E67, 0x6B6520, 0x6B6B65, 0x6C6520, 0x6C6C65, 0x6D6564, 0x6D656E, 0x6E2073, 0x6E6520, 0x6E6720, 0x6E6765, 0x6E6E65, 0x6F6720, 0x6F6D20, 0x6F7220, 0x70E520, 0x722073, 0x726520, 0x736F6D, 0x737465, 0x742073, 0x746520, 0x74656E, 0x746572, 0x74696C, 0x747420, 0x747465, 0x766572, } var ngrams_8859_1_pt = [64]uint32{ 0x206120, 0x20636F, 0x206461, 0x206465, 0x20646F, 0x206520, 0x206573, 0x206D61, 0x206E6F, 0x206F20, 0x207061, 0x20706F, 0x207072, 0x207175, 0x207265, 0x207365, 0x20756D, 0x612061, 0x612063, 0x612064, 0x612070, 0x616465, 0x61646F, 0x616C20, 0x617220, 0x617261, 0x617320, 0x636F6D, 0x636F6E, 0x646120, 0x646520, 0x646F20, 0x646F73, 0x652061, 0x652064, 0x656D20, 0x656E74, 0x657320, 0x657374, 0x696120, 0x696361, 0x6D656E, 0x6E7465, 0x6E746F, 0x6F2061, 0x6F2063, 0x6F2064, 0x6F2065, 0x6F2070, 0x6F7320, 0x706172, 0x717565, 0x726120, 0x726573, 0x732061, 0x732064, 0x732065, 0x732070, 0x737461, 0x746520, 0x746F20, 0x756520, 0xE36F20, 0xE7E36F, } var ngrams_8859_1_sv = [64]uint32{ 0x206174, 0x206176, 0x206465, 0x20656E, 0x2066F6, 0x206861, 0x206920, 0x20696E, 0x206B6F, 0x206D65, 0x206F63, 0x2070E5, 0x20736B, 0x20736F, 0x207374, 0x207469, 0x207661, 0x207669, 0x20E472, 0x616465, 0x616E20, 0x616E64, 0x617220, 0x617474, 0x636820, 0x646520, 0x64656E, 0x646572, 0x646574, 0x656420, 0x656E20, 0x657220, 0x657420, 0x66F672, 0x67656E, 0x696C6C, 0x696E67, 0x6B6120, 0x6C6C20, 0x6D6564, 0x6E2073, 0x6E6120, 0x6E6465, 0x6E6720, 0x6E6765, 0x6E696E, 0x6F6368, 0x6F6D20, 0x6F6E20, 0x70E520, 0x722061, 0x722073, 0x726120, 0x736B61, 0x736F6D, 0x742073, 0x746120, 0x746520, 0x746572, 0x74696C, 0x747420, 0x766172, 0xE47220, 0xF67220, } func newRecognizer_8859_1(language string, ngram *[64]uint32) *recognizerSingleByte { return &recognizerSingleByte{ charset: "ISO-8859-1", hasC1ByteCharset: "windows-1252", language: language, charMap: &charMap_8859_1, ngram: ngram, } } func newRecognizer_8859_1_en() *recognizerSingleByte { return newRecognizer_8859_1("en", &ngrams_8859_1_en) } func newRecognizer_8859_1_da() *recognizerSingleByte { return newRecognizer_8859_1("da", &ngrams_8859_1_da) } func newRecognizer_8859_1_de() *recognizerSingleByte { return newRecognizer_8859_1("de", &ngrams_8859_1_de) } func newRecognizer_8859_1_es() *recognizerSingleByte { return newRecognizer_8859_1("es", &ngrams_8859_1_es) } func newRecognizer_8859_1_fr() *recognizerSingleByte { return newRecognizer_8859_1("fr", &ngrams_8859_1_fr) } func newRecognizer_8859_1_it() *recognizerSingleByte { return newRecognizer_8859_1("it", &ngrams_8859_1_it) } func newRecognizer_8859_1_nl() *recognizerSingleByte { return newRecognizer_8859_1("nl", &ngrams_8859_1_nl) } func newRecognizer_8859_1_no() *recognizerSingleByte { return newRecognizer_8859_1("no", &ngrams_8859_1_no) } func newRecognizer_8859_1_pt() *recognizerSingleByte { return newRecognizer_8859_1("pt", &ngrams_8859_1_pt) } func newRecognizer_8859_1_sv() *recognizerSingleByte { return newRecognizer_8859_1("sv", &ngrams_8859_1_sv) } var charMap_8859_2 = [256]byte{ 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xB1, 0x20, 0xB3, 0x20, 0xB5, 0xB6, 0x20, 0x20, 0xB9, 0xBA, 0xBB, 0xBC, 0x20, 0xBE, 0xBF, 0x20, 0xB1, 0x20, 0xB3, 0x20, 0xB5, 0xB6, 0xB7, 0x20, 0xB9, 0xBA, 0xBB, 0xBC, 0x20, 0xBE, 0xBF, 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0x20, 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0xDF, 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0x20, 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0x20, } var ngrams_8859_2_cs = [64]uint32{ 0x206120, 0x206279, 0x20646F, 0x206A65, 0x206E61, 0x206E65, 0x206F20, 0x206F64, 0x20706F, 0x207072, 0x2070F8, 0x20726F, 0x207365, 0x20736F, 0x207374, 0x20746F, 0x207620, 0x207679, 0x207A61, 0x612070, 0x636520, 0x636820, 0x652070, 0x652073, 0x652076, 0x656D20, 0x656EED, 0x686F20, 0x686F64, 0x697374, 0x6A6520, 0x6B7465, 0x6C6520, 0x6C6920, 0x6E6120, 0x6EE920, 0x6EEC20, 0x6EED20, 0x6F2070, 0x6F646E, 0x6F6A69, 0x6F7374, 0x6F7520, 0x6F7661, 0x706F64, 0x706F6A, 0x70726F, 0x70F865, 0x736520, 0x736F75, 0x737461, 0x737469, 0x73746E, 0x746572, 0x746EED, 0x746F20, 0x752070, 0xBE6520, 0xE16EED, 0xE9686F, 0xED2070, 0xED2073, 0xED6D20, 0xF86564, } var ngrams_8859_2_hu = [64]uint32{ 0x206120, 0x20617A, 0x206265, 0x206567, 0x20656C, 0x206665, 0x206861, 0x20686F, 0x206973, 0x206B65, 0x206B69, 0x206BF6, 0x206C65, 0x206D61, 0x206D65, 0x206D69, 0x206E65, 0x20737A, 0x207465, 0x20E973, 0x612061, 0x61206B, 0x61206D, 0x612073, 0x616B20, 0x616E20, 0x617A20, 0x62616E, 0x62656E, 0x656779, 0x656B20, 0x656C20, 0x656C65, 0x656D20, 0x656E20, 0x657265, 0x657420, 0x657465, 0x657474, 0x677920, 0x686F67, 0x696E74, 0x697320, 0x6B2061, 0x6BF67A, 0x6D6567, 0x6D696E, 0x6E2061, 0x6E616B, 0x6E656B, 0x6E656D, 0x6E7420, 0x6F6779, 0x732061, 0x737A65, 0x737A74, 0x737AE1, 0x73E967, 0x742061, 0x747420, 0x74E173, 0x7A6572, 0xE16E20, 0xE97320, } var ngrams_8859_2_pl = [64]uint32{ 0x20637A, 0x20646F, 0x206920, 0x206A65, 0x206B6F, 0x206D61, 0x206D69, 0x206E61, 0x206E69, 0x206F64, 0x20706F, 0x207072, 0x207369, 0x207720, 0x207769, 0x207779, 0x207A20, 0x207A61, 0x612070, 0x612077, 0x616E69, 0x636820, 0x637A65, 0x637A79, 0x646F20, 0x647A69, 0x652070, 0x652073, 0x652077, 0x65207A, 0x65676F, 0x656A20, 0x656D20, 0x656E69, 0x676F20, 0x696120, 0x696520, 0x69656A, 0x6B6120, 0x6B6920, 0x6B6965, 0x6D6965, 0x6E6120, 0x6E6961, 0x6E6965, 0x6F2070, 0x6F7761, 0x6F7769, 0x706F6C, 0x707261, 0x70726F, 0x70727A, 0x727A65, 0x727A79, 0x7369EA, 0x736B69, 0x737461, 0x776965, 0x796368, 0x796D20, 0x7A6520, 0x7A6965, 0x7A7920, 0xF37720, } var ngrams_8859_2_ro = [64]uint32{ 0x206120, 0x206163, 0x206361, 0x206365, 0x20636F, 0x206375, 0x206465, 0x206469, 0x206C61, 0x206D61, 0x207065, 0x207072, 0x207365, 0x2073E3, 0x20756E, 0x20BA69, 0x20EE6E, 0x612063, 0x612064, 0x617265, 0x617420, 0x617465, 0x617520, 0x636172, 0x636F6E, 0x637520, 0x63E320, 0x646520, 0x652061, 0x652063, 0x652064, 0x652070, 0x652073, 0x656120, 0x656920, 0x656C65, 0x656E74, 0x657374, 0x692061, 0x692063, 0x692064, 0x692070, 0x696520, 0x696920, 0x696E20, 0x6C6120, 0x6C6520, 0x6C6F72, 0x6C7569, 0x6E6520, 0x6E7472, 0x6F7220, 0x70656E, 0x726520, 0x726561, 0x727520, 0x73E320, 0x746520, 0x747275, 0x74E320, 0x756920, 0x756C20, 0xBA6920, 0xEE6E20, } func newRecognizer_8859_2(language string, ngram *[64]uint32) *recognizerSingleByte { return &recognizerSingleByte{ charset: "ISO-8859-2", hasC1ByteCharset: "windows-1250", language: language, charMap: &charMap_8859_2, ngram: ngram, } } func newRecognizer_8859_2_cs() *recognizerSingleByte { return newRecognizer_8859_1("cs", &ngrams_8859_2_cs) } func newRecognizer_8859_2_hu() *recognizerSingleByte { return newRecognizer_8859_1("hu", &ngrams_8859_2_hu) } func newRecognizer_8859_2_pl() *recognizerSingleByte { return newRecognizer_8859_1("pl", &ngrams_8859_2_pl) } func newRecognizer_8859_2_ro() *recognizerSingleByte { return newRecognizer_8859_1("ro", &ngrams_8859_2_ro) } var charMap_8859_5 = [256]byte{ 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0x20, 0xFE, 0xFF, 0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7, 0xD8, 0xD9, 0xDA, 0xDB, 0xDC, 0xDD, 0xDE, 0xDF, 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, 0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7, 0xD8, 0xD9, 0xDA, 0xDB, 0xDC, 0xDD, 0xDE, 0xDF, 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, 0x20, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0x20, 0xFE, 0xFF, } var ngrams_8859_5_ru = [64]uint32{ 0x20D220, 0x20D2DE, 0x20D4DE, 0x20D7D0, 0x20D820, 0x20DAD0, 0x20DADE, 0x20DDD0, 0x20DDD5, 0x20DED1, 0x20DFDE, 0x20DFE0, 0x20E0D0, 0x20E1DE, 0x20E1E2, 0x20E2DE, 0x20E7E2, 0x20EDE2, 0xD0DDD8, 0xD0E2EC, 0xD3DE20, 0xD5DBEC, 0xD5DDD8, 0xD5E1E2, 0xD5E220, 0xD820DF, 0xD8D520, 0xD8D820, 0xD8EF20, 0xDBD5DD, 0xDBD820, 0xDBECDD, 0xDDD020, 0xDDD520, 0xDDD8D5, 0xDDD8EF, 0xDDDE20, 0xDDDED2, 0xDE20D2, 0xDE20DF, 0xDE20E1, 0xDED220, 0xDED2D0, 0xDED3DE, 0xDED920, 0xDEDBEC, 0xDEDC20, 0xDEE1E2, 0xDFDEDB, 0xDFE0D5, 0xDFE0D8, 0xDFE0DE, 0xE0D0D2, 0xE0D5D4, 0xE1E2D0, 0xE1E2D2, 0xE1E2D8, 0xE1EF20, 0xE2D5DB, 0xE2DE20, 0xE2DEE0, 0xE2EC20, 0xE7E2DE, 0xEBE520, } func newRecognizer_8859_5(language string, ngram *[64]uint32) *recognizerSingleByte { return &recognizerSingleByte{ charset: "ISO-8859-5", language: language, charMap: &charMap_8859_5, ngram: ngram, } } func newRecognizer_8859_5_ru() *recognizerSingleByte { return newRecognizer_8859_5("ru", &ngrams_8859_5_ru) } var charMap_8859_6 = [256]byte{ 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, 0xC8, 0xC9, 0xCA, 0xCB, 0xCC, 0xCD, 0xCE, 0xCF, 0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7, 0xD8, 0xD9, 0xDA, 0x20, 0x20, 0x20, 0x20, 0x20, 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, } var ngrams_8859_6_ar = [64]uint32{ 0x20C7E4, 0x20C7E6, 0x20C8C7, 0x20D9E4, 0x20E1EA, 0x20E4E4, 0x20E5E6, 0x20E8C7, 0xC720C7, 0xC7C120, 0xC7CA20, 0xC7D120, 0xC7E420, 0xC7E4C3, 0xC7E4C7, 0xC7E4C8, 0xC7E4CA, 0xC7E4CC, 0xC7E4CD, 0xC7E4CF, 0xC7E4D3, 0xC7E4D9, 0xC7E4E2, 0xC7E4E5, 0xC7E4E8, 0xC7E4EA, 0xC7E520, 0xC7E620, 0xC7E6CA, 0xC820C7, 0xC920C7, 0xC920E1, 0xC920E4, 0xC920E5, 0xC920E8, 0xCA20C7, 0xCF20C7, 0xCFC920, 0xD120C7, 0xD1C920, 0xD320C7, 0xD920C7, 0xD9E4E9, 0xE1EA20, 0xE420C7, 0xE4C920, 0xE4E920, 0xE4EA20, 0xE520C7, 0xE5C720, 0xE5C920, 0xE5E620, 0xE620C7, 0xE720C7, 0xE7C720, 0xE8C7E4, 0xE8E620, 0xE920C7, 0xEA20C7, 0xEA20E5, 0xEA20E8, 0xEAC920, 0xEAD120, 0xEAE620, } func newRecognizer_8859_6(language string, ngram *[64]uint32) *recognizerSingleByte { return &recognizerSingleByte{ charset: "ISO-8859-6", language: language, charMap: &charMap_8859_6, ngram: ngram, } } func newRecognizer_8859_6_ar() *recognizerSingleByte { return newRecognizer_8859_6("ar", &ngrams_8859_6_ar) } var charMap_8859_7 = [256]byte{ 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xA1, 0xA2, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xDC, 0x20, 0xDD, 0xDE, 0xDF, 0x20, 0xFC, 0x20, 0xFD, 0xFE, 0xC0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, 0xF0, 0xF1, 0x20, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, 0xFA, 0xFB, 0xDC, 0xDD, 0xDE, 0xDF, 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0x20, } var ngrams_8859_7_el = [64]uint32{ 0x20E1ED, 0x20E1F0, 0x20E3E9, 0x20E4E9, 0x20E5F0, 0x20E720, 0x20EAE1, 0x20ECE5, 0x20EDE1, 0x20EF20, 0x20F0E1, 0x20F0EF, 0x20F0F1, 0x20F3F4, 0x20F3F5, 0x20F4E7, 0x20F4EF, 0xDFE120, 0xE120E1, 0xE120F4, 0xE1E920, 0xE1ED20, 0xE1F0FC, 0xE1F220, 0xE3E9E1, 0xE5E920, 0xE5F220, 0xE720F4, 0xE7ED20, 0xE7F220, 0xE920F4, 0xE9E120, 0xE9EADE, 0xE9F220, 0xEAE1E9, 0xEAE1F4, 0xECE520, 0xED20E1, 0xED20E5, 0xED20F0, 0xEDE120, 0xEFF220, 0xEFF520, 0xF0EFF5, 0xF0F1EF, 0xF0FC20, 0xF220E1, 0xF220E5, 0xF220EA, 0xF220F0, 0xF220F4, 0xF3E520, 0xF3E720, 0xF3F4EF, 0xF4E120, 0xF4E1E9, 0xF4E7ED, 0xF4E7F2, 0xF4E9EA, 0xF4EF20, 0xF4EFF5, 0xF4F9ED, 0xF9ED20, 0xFEED20, } func newRecognizer_8859_7(language string, ngram *[64]uint32) *recognizerSingleByte { return &recognizerSingleByte{ charset: "ISO-8859-7", hasC1ByteCharset: "windows-1253", language: language, charMap: &charMap_8859_7, ngram: ngram, } } func newRecognizer_8859_7_el() *recognizerSingleByte { return newRecognizer_8859_7("el", &ngrams_8859_7_el) } var charMap_8859_8 = [256]byte{ 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xB5, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, 0xFA, 0x20, 0x20, 0x20, 0x20, 0x20, } var ngrams_8859_8_I_he = [64]uint32{ 0x20E0E5, 0x20E0E7, 0x20E0E9, 0x20E0FA, 0x20E1E9, 0x20E1EE, 0x20E4E0, 0x20E4E5, 0x20E4E9, 0x20E4EE, 0x20E4F2, 0x20E4F9, 0x20E4FA, 0x20ECE0, 0x20ECE4, 0x20EEE0, 0x20F2EC, 0x20F9EC, 0xE0FA20, 0xE420E0, 0xE420E1, 0xE420E4, 0xE420EC, 0xE420EE, 0xE420F9, 0xE4E5E0, 0xE5E020, 0xE5ED20, 0xE5EF20, 0xE5F820, 0xE5FA20, 0xE920E4, 0xE9E420, 0xE9E5FA, 0xE9E9ED, 0xE9ED20, 0xE9EF20, 0xE9F820, 0xE9FA20, 0xEC20E0, 0xEC20E4, 0xECE020, 0xECE420, 0xED20E0, 0xED20E1, 0xED20E4, 0xED20EC, 0xED20EE, 0xED20F9, 0xEEE420, 0xEF20E4, 0xF0E420, 0xF0E920, 0xF0E9ED, 0xF2EC20, 0xF820E4, 0xF8E9ED, 0xF9EC20, 0xFA20E0, 0xFA20E1, 0xFA20E4, 0xFA20EC, 0xFA20EE, 0xFA20F9, } var ngrams_8859_8_he = [64]uint32{ 0x20E0E5, 0x20E0EC, 0x20E4E9, 0x20E4EC, 0x20E4EE, 0x20E4F0, 0x20E9F0, 0x20ECF2, 0x20ECF9, 0x20EDE5, 0x20EDE9, 0x20EFE5, 0x20EFE9, 0x20F8E5, 0x20F8E9, 0x20FAE0, 0x20FAE5, 0x20FAE9, 0xE020E4, 0xE020EC, 0xE020ED, 0xE020FA, 0xE0E420, 0xE0E5E4, 0xE0EC20, 0xE0EE20, 0xE120E4, 0xE120ED, 0xE120FA, 0xE420E4, 0xE420E9, 0xE420EC, 0xE420ED, 0xE420EF, 0xE420F8, 0xE420FA, 0xE4EC20, 0xE5E020, 0xE5E420, 0xE7E020, 0xE9E020, 0xE9E120, 0xE9E420, 0xEC20E4, 0xEC20ED, 0xEC20FA, 0xECF220, 0xECF920, 0xEDE9E9, 0xEDE9F0, 0xEDE9F8, 0xEE20E4, 0xEE20ED, 0xEE20FA, 0xEEE120, 0xEEE420, 0xF2E420, 0xF920E4, 0xF920ED, 0xF920FA, 0xF9E420, 0xFAE020, 0xFAE420, 0xFAE5E9, } func newRecognizer_8859_8(language string, ngram *[64]uint32) *recognizerSingleByte { return &recognizerSingleByte{ charset: "ISO-8859-8", hasC1ByteCharset: "windows-1255", language: language, charMap: &charMap_8859_8, ngram: ngram, } } func newRecognizer_8859_8_I_he() *recognizerSingleByte { r := newRecognizer_8859_8("he", &ngrams_8859_8_I_he) r.charset = "ISO-8859-8-I" return r } func newRecognizer_8859_8_he() *recognizerSingleByte { return newRecognizer_8859_8("he", &ngrams_8859_8_he) } var charMap_8859_9 = [256]byte{ 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xAA, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xB5, 0x20, 0x20, 0x20, 0x20, 0xBA, 0x20, 0x20, 0x20, 0x20, 0x20, 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0x20, 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0x69, 0xFE, 0xDF, 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0x20, 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0xFF, } var ngrams_8859_9_tr = [64]uint32{ 0x206261, 0x206269, 0x206275, 0x206461, 0x206465, 0x206765, 0x206861, 0x20696C, 0x206B61, 0x206B6F, 0x206D61, 0x206F6C, 0x207361, 0x207461, 0x207665, 0x207961, 0x612062, 0x616B20, 0x616C61, 0x616D61, 0x616E20, 0x616EFD, 0x617220, 0x617261, 0x6172FD, 0x6173FD, 0x617961, 0x626972, 0x646120, 0x646520, 0x646920, 0x652062, 0x65206B, 0x656469, 0x656E20, 0x657220, 0x657269, 0x657369, 0x696C65, 0x696E20, 0x696E69, 0x697220, 0x6C616E, 0x6C6172, 0x6C6520, 0x6C6572, 0x6E2061, 0x6E2062, 0x6E206B, 0x6E6461, 0x6E6465, 0x6E6520, 0x6E6920, 0x6E696E, 0x6EFD20, 0x72696E, 0x72FD6E, 0x766520, 0x796120, 0x796F72, 0xFD6E20, 0xFD6E64, 0xFD6EFD, 0xFDF0FD, } func newRecognizer_8859_9(language string, ngram *[64]uint32) *recognizerSingleByte { return &recognizerSingleByte{ charset: "ISO-8859-9", hasC1ByteCharset: "windows-1254", language: language, charMap: &charMap_8859_9, ngram: ngram, } } func newRecognizer_8859_9_tr() *recognizerSingleByte { return newRecognizer_8859_9("tr", &ngrams_8859_9_tr) } var charMap_windows_1256 = [256]byte{ 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x81, 0x20, 0x83, 0x20, 0x20, 0x20, 0x20, 0x88, 0x20, 0x8A, 0x20, 0x9C, 0x8D, 0x8E, 0x8F, 0x90, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x98, 0x20, 0x9A, 0x20, 0x9C, 0x20, 0x20, 0x9F, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xAA, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xB5, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xC0, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, 0xC8, 0xC9, 0xCA, 0xCB, 0xCC, 0xCD, 0xCE, 0xCF, 0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0x20, 0xD8, 0xD9, 0xDA, 0xDB, 0xDC, 0xDD, 0xDE, 0xDF, 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, 0x20, 0x20, 0x20, 0x20, 0xF4, 0x20, 0x20, 0x20, 0x20, 0xF9, 0x20, 0xFB, 0xFC, 0x20, 0x20, 0xFF, } var ngrams_windows_1256 = [64]uint32{ 0x20C7E1, 0x20C7E4, 0x20C8C7, 0x20DAE1, 0x20DDED, 0x20E1E1, 0x20E3E4, 0x20E6C7, 0xC720C7, 0xC7C120, 0xC7CA20, 0xC7D120, 0xC7E120, 0xC7E1C3, 0xC7E1C7, 0xC7E1C8, 0xC7E1CA, 0xC7E1CC, 0xC7E1CD, 0xC7E1CF, 0xC7E1D3, 0xC7E1DA, 0xC7E1DE, 0xC7E1E3, 0xC7E1E6, 0xC7E1ED, 0xC7E320, 0xC7E420, 0xC7E4CA, 0xC820C7, 0xC920C7, 0xC920DD, 0xC920E1, 0xC920E3, 0xC920E6, 0xCA20C7, 0xCF20C7, 0xCFC920, 0xD120C7, 0xD1C920, 0xD320C7, 0xDA20C7, 0xDAE1EC, 0xDDED20, 0xE120C7, 0xE1C920, 0xE1EC20, 0xE1ED20, 0xE320C7, 0xE3C720, 0xE3C920, 0xE3E420, 0xE420C7, 0xE520C7, 0xE5C720, 0xE6C7E1, 0xE6E420, 0xEC20C7, 0xED20C7, 0xED20E3, 0xED20E6, 0xEDC920, 0xEDD120, 0xEDE420, } func newRecognizer_windows_1256() *recognizerSingleByte { return &recognizerSingleByte{ charset: "windows-1256", language: "ar", charMap: &charMap_windows_1256, ngram: &ngrams_windows_1256, } } var charMap_windows_1251 = [256]byte{ 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x90, 0x83, 0x20, 0x83, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x9A, 0x20, 0x9C, 0x9D, 0x9E, 0x9F, 0x90, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x9A, 0x20, 0x9C, 0x9D, 0x9E, 0x9F, 0x20, 0xA2, 0xA2, 0xBC, 0x20, 0xB4, 0x20, 0x20, 0xB8, 0x20, 0xBA, 0x20, 0x20, 0x20, 0x20, 0xBF, 0x20, 0x20, 0xB3, 0xB3, 0xB4, 0xB5, 0x20, 0x20, 0xB8, 0x20, 0xBA, 0x20, 0xBC, 0xBE, 0xBE, 0xBF, 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0xFF, 0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF, 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0xFF, } var ngrams_windows_1251 = [64]uint32{ 0x20E220, 0x20E2EE, 0x20E4EE, 0x20E7E0, 0x20E820, 0x20EAE0, 0x20EAEE, 0x20EDE0, 0x20EDE5, 0x20EEE1, 0x20EFEE, 0x20EFF0, 0x20F0E0, 0x20F1EE, 0x20F1F2, 0x20F2EE, 0x20F7F2, 0x20FDF2, 0xE0EDE8, 0xE0F2FC, 0xE3EE20, 0xE5EBFC, 0xE5EDE8, 0xE5F1F2, 0xE5F220, 0xE820EF, 0xE8E520, 0xE8E820, 0xE8FF20, 0xEBE5ED, 0xEBE820, 0xEBFCED, 0xEDE020, 0xEDE520, 0xEDE8E5, 0xEDE8FF, 0xEDEE20, 0xEDEEE2, 0xEE20E2, 0xEE20EF, 0xEE20F1, 0xEEE220, 0xEEE2E0, 0xEEE3EE, 0xEEE920, 0xEEEBFC, 0xEEEC20, 0xEEF1F2, 0xEFEEEB, 0xEFF0E5, 0xEFF0E8, 0xEFF0EE, 0xF0E0E2, 0xF0E5E4, 0xF1F2E0, 0xF1F2E2, 0xF1F2E8, 0xF1FF20, 0xF2E5EB, 0xF2EE20, 0xF2EEF0, 0xF2FC20, 0xF7F2EE, 0xFBF520, } func newRecognizer_windows_1251() *recognizerSingleByte { return &recognizerSingleByte{ charset: "windows-1251", language: "ar", charMap: &charMap_windows_1251, ngram: &ngrams_windows_1251, } } var charMap_KOI8_R = [256]byte{ 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xA3, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xA3, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0xC0, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, 0xC8, 0xC9, 0xCA, 0xCB, 0xCC, 0xCD, 0xCE, 0xCF, 0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7, 0xD8, 0xD9, 0xDA, 0xDB, 0xDC, 0xDD, 0xDE, 0xDF, 0xC0, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, 0xC8, 0xC9, 0xCA, 0xCB, 0xCC, 0xCD, 0xCE, 0xCF, 0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7, 0xD8, 0xD9, 0xDA, 0xDB, 0xDC, 0xDD, 0xDE, 0xDF, } var ngrams_KOI8_R = [64]uint32{ 0x20C4CF, 0x20C920, 0x20CBC1, 0x20CBCF, 0x20CEC1, 0x20CEC5, 0x20CFC2, 0x20D0CF, 0x20D0D2, 0x20D2C1, 0x20D3CF, 0x20D3D4, 0x20D4CF, 0x20D720, 0x20D7CF, 0x20DAC1, 0x20DCD4, 0x20DED4, 0xC1CEC9, 0xC1D4D8, 0xC5CCD8, 0xC5CEC9, 0xC5D3D4, 0xC5D420, 0xC7CF20, 0xC920D0, 0xC9C520, 0xC9C920, 0xC9D120, 0xCCC5CE, 0xCCC920, 0xCCD8CE, 0xCEC120, 0xCEC520, 0xCEC9C5, 0xCEC9D1, 0xCECF20, 0xCECFD7, 0xCF20D0, 0xCF20D3, 0xCF20D7, 0xCFC7CF, 0xCFCA20, 0xCFCCD8, 0xCFCD20, 0xCFD3D4, 0xCFD720, 0xCFD7C1, 0xD0CFCC, 0xD0D2C5, 0xD0D2C9, 0xD0D2CF, 0xD2C1D7, 0xD2C5C4, 0xD3D120, 0xD3D4C1, 0xD3D4C9, 0xD3D4D7, 0xD4C5CC, 0xD4CF20, 0xD4CFD2, 0xD4D820, 0xD9C820, 0xDED4CF, } func newRecognizer_KOI8_R() *recognizerSingleByte { return &recognizerSingleByte{ charset: "KOI8-R", language: "ru", charMap: &charMap_KOI8_R, ngram: &ngrams_KOI8_R, } } var charMap_IBM424_he = [256]byte{ /* -0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -A -B -C -D -E -F */ /* 0- */ 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 1- */ 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 2- */ 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 3- */ 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 4- */ 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, 0x49, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 5- */ 0x40, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, 0x59, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 6- */ 0x40, 0x40, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 7- */ 0x40, 0x71, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x00, 0x40, 0x40, /* 8- */ 0x40, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 9- */ 0x40, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, 0x98, 0x99, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* A- */ 0xA0, 0x40, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, 0xA8, 0xA9, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* B- */ 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* C- */ 0x40, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* D- */ 0x40, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, 0x98, 0x99, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* E- */ 0x40, 0x40, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, 0xA8, 0xA9, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* F- */ 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, } var ngrams_IBM424_he_rtl = [64]uint32{ 0x404146, 0x404148, 0x404151, 0x404171, 0x404251, 0x404256, 0x404541, 0x404546, 0x404551, 0x404556, 0x404562, 0x404569, 0x404571, 0x405441, 0x405445, 0x405641, 0x406254, 0x406954, 0x417140, 0x454041, 0x454042, 0x454045, 0x454054, 0x454056, 0x454069, 0x454641, 0x464140, 0x465540, 0x465740, 0x466840, 0x467140, 0x514045, 0x514540, 0x514671, 0x515155, 0x515540, 0x515740, 0x516840, 0x517140, 0x544041, 0x544045, 0x544140, 0x544540, 0x554041, 0x554042, 0x554045, 0x554054, 0x554056, 0x554069, 0x564540, 0x574045, 0x584540, 0x585140, 0x585155, 0x625440, 0x684045, 0x685155, 0x695440, 0x714041, 0x714042, 0x714045, 0x714054, 0x714056, 0x714069, } var ngrams_IBM424_he_ltr = [64]uint32{ 0x404146, 0x404154, 0x404551, 0x404554, 0x404556, 0x404558, 0x405158, 0x405462, 0x405469, 0x405546, 0x405551, 0x405746, 0x405751, 0x406846, 0x406851, 0x407141, 0x407146, 0x407151, 0x414045, 0x414054, 0x414055, 0x414071, 0x414540, 0x414645, 0x415440, 0x415640, 0x424045, 0x424055, 0x424071, 0x454045, 0x454051, 0x454054, 0x454055, 0x454057, 0x454068, 0x454071, 0x455440, 0x464140, 0x464540, 0x484140, 0x514140, 0x514240, 0x514540, 0x544045, 0x544055, 0x544071, 0x546240, 0x546940, 0x555151, 0x555158, 0x555168, 0x564045, 0x564055, 0x564071, 0x564240, 0x564540, 0x624540, 0x694045, 0x694055, 0x694071, 0x694540, 0x714140, 0x714540, 0x714651, } func newRecognizer_IBM424_he(charset string, ngram *[64]uint32) *recognizerSingleByte { return &recognizerSingleByte{ charset: charset, language: "he", charMap: &charMap_IBM424_he, ngram: ngram, } } func newRecognizer_IBM424_he_rtl() *recognizerSingleByte { return newRecognizer_IBM424_he("IBM424_rtl", &ngrams_IBM424_he_rtl) } func newRecognizer_IBM424_he_ltr() *recognizerSingleByte { return newRecognizer_IBM424_he("IBM424_ltr", &ngrams_IBM424_he_ltr) } var charMap_IBM420_ar = [256]byte{ /* -0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -A -B -C -D -E -F */ /* 0- */ 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 1- */ 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 2- */ 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 3- */ 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 4- */ 0x40, 0x40, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, 0x49, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 5- */ 0x40, 0x51, 0x52, 0x40, 0x40, 0x55, 0x56, 0x57, 0x58, 0x59, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 6- */ 0x40, 0x40, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 7- */ 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, /* 8- */ 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x8A, 0x8B, 0x8C, 0x8D, 0x8E, 0x8F, /* 9- */ 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, 0x98, 0x99, 0x9A, 0x9B, 0x9C, 0x9D, 0x9E, 0x9F, /* A- */ 0xA0, 0x40, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, 0xA8, 0xA9, 0xAA, 0xAB, 0xAC, 0xAD, 0xAE, 0xAF, /* B- */ 0xB0, 0xB1, 0xB2, 0xB3, 0xB4, 0xB5, 0x40, 0x40, 0xB8, 0xB9, 0xBA, 0xBB, 0xBC, 0xBD, 0xBE, 0xBF, /* C- */ 0x40, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x40, 0xCB, 0x40, 0xCD, 0x40, 0xCF, /* D- */ 0x40, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, 0x98, 0x99, 0xDA, 0xDB, 0xDC, 0xDD, 0xDE, 0xDF, /* E- */ 0x40, 0x40, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, 0xA8, 0xA9, 0xEA, 0xEB, 0x40, 0xED, 0xEE, 0xEF, /* F- */ 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0xFB, 0xFC, 0xFD, 0xFE, 0x40, } var ngrams_IBM420_ar_rtl = [64]uint32{ 0x4056B1, 0x4056BD, 0x405856, 0x409AB1, 0x40ABDC, 0x40B1B1, 0x40BBBD, 0x40CF56, 0x564056, 0x564640, 0x566340, 0x567540, 0x56B140, 0x56B149, 0x56B156, 0x56B158, 0x56B163, 0x56B167, 0x56B169, 0x56B173, 0x56B178, 0x56B19A, 0x56B1AD, 0x56B1BB, 0x56B1CF, 0x56B1DC, 0x56BB40, 0x56BD40, 0x56BD63, 0x584056, 0x624056, 0x6240AB, 0x6240B1, 0x6240BB, 0x6240CF, 0x634056, 0x734056, 0x736240, 0x754056, 0x756240, 0x784056, 0x9A4056, 0x9AB1DA, 0xABDC40, 0xB14056, 0xB16240, 0xB1DA40, 0xB1DC40, 0xBB4056, 0xBB5640, 0xBB6240, 0xBBBD40, 0xBD4056, 0xBF4056, 0xBF5640, 0xCF56B1, 0xCFBD40, 0xDA4056, 0xDC4056, 0xDC40BB, 0xDC40CF, 0xDC6240, 0xDC7540, 0xDCBD40, } var ngrams_IBM420_ar_ltr = [64]uint32{ 0x404656, 0x4056BB, 0x4056BF, 0x406273, 0x406275, 0x4062B1, 0x4062BB, 0x4062DC, 0x406356, 0x407556, 0x4075DC, 0x40B156, 0x40BB56, 0x40BD56, 0x40BDBB, 0x40BDCF, 0x40BDDC, 0x40DAB1, 0x40DCAB, 0x40DCB1, 0x49B156, 0x564056, 0x564058, 0x564062, 0x564063, 0x564073, 0x564075, 0x564078, 0x56409A, 0x5640B1, 0x5640BB, 0x5640BD, 0x5640BF, 0x5640DA, 0x5640DC, 0x565840, 0x56B156, 0x56CF40, 0x58B156, 0x63B156, 0x63BD56, 0x67B156, 0x69B156, 0x73B156, 0x78B156, 0x9AB156, 0xAB4062, 0xADB156, 0xB14062, 0xB15640, 0xB156CF, 0xB19A40, 0xB1B140, 0xBB4062, 0xBB40DC, 0xBBB156, 0xBD5640, 0xBDBB40, 0xCF4062, 0xCF40DC, 0xCFB156, 0xDAB19A, 0xDCAB40, 0xDCB156, } func newRecognizer_IBM420_ar(charset string, ngram *[64]uint32) *recognizerSingleByte { return &recognizerSingleByte{ charset: charset, language: "ar", charMap: &charMap_IBM420_ar, ngram: ngram, } } func newRecognizer_IBM420_ar_rtl() *recognizerSingleByte { return newRecognizer_IBM420_ar("IBM420_rtl", &ngrams_IBM420_ar_rtl) } func newRecognizer_IBM420_ar_ltr() *recognizerSingleByte { return newRecognizer_IBM420_ar("IBM420_ltr", &ngrams_IBM420_ar_ltr) } ================================================ FILE: testdata/8859_1_da.html ================================================ Forside - Virk.dk
Til forsiden


AnstSelv vejviser indberetninger

 

Cookies

På Virk.dk bruger vi cookies bl.a. til at forbedre vores indhold og struktur via din og andres faktiske brug af Virk.dk.

Læs mere om cookies og brugen af dem her
 

Vejledninger på Virk.dk

Vidste du, at du på Virk.dk kan finde vejledninger og guides, der fører dig igennem dine forpligtelser over for det offentlige i forskellige situationer? Du kan via e-læringsfilm fx lære om, hvilke pligter du skal registrere ift. SKAT og ATP, få hjælp til hvordan du bruger NemRefusion og Regnskab 2.0 eller få svar på hvad Næringsbasen er.

Se alle vejledninger på Virk.dk

Du kan logge ind med Digital signatur eller NemID

Viden og værktøjer
Startvkst


Virk.dk l Kontakt l Vilkår l Cookies


================================================ FILE: testdata/8859_1_de.html ================================================ infopark.de Startseite | Anbieter von: Infopark Cloud Express | CMS und CRM on-demand. | Infopark powers great websites.
infopark Logo

Wir untersttzen diesen Browser nicht mehr.

Sie mchten die Webseiten der Infopark AG mit dem Internet Explorer 6 besuchen.
Wir untersttzen diesen Browser nicht mehr. Bitte machen Sie einen Browser-Update.

Vielen Dank - Ihr Infopark-Team

Transparent
Transparent_right

Cloud-Lsungen

referenzprojekte-video

Great websites on-demand.
Entdecken Sie unsere Cloud-Services auf Basis von Ruby on Rails. CMS und CRM fr komplexe Web-Projekte.

Auszeichnungen

Infoparks gewonnene Awards

Der Infopark Cloud Express wurde bereits mehrfach bei nationalen und paneuropischen Wettbewerben ausgezeichnet.

Infopark Cloud Express

infopark-cloud-express

Nachrichten

Infopark powers great websites Infopark powers great websites.
================================================ FILE: testdata/8859_1_en.html ================================================ Amazon.com: Online Shopping for Electronics, Apparel, Computers, Books, DVDs & more
A different version of this web site containing similar content optimized for screen readers and mobile devices may be found at the web address: www.amazon.com/access
Textbooks Amazon Textbooks

 
 
Kindle, from $79 Kindle: Back to School. From $79

Sign in
Fall's New Shape View Looks The Amazon Clothing Store White Hot Denim

What Other Customers Are Looking At Right Now

The Hunger Games
Jennifer Lawrence, Josh Hutcherson, ...
Blu-ray
$39.99 $19.99
Monster Shooter: The Lost Levels

Gamelion Studios
$0.99 $0.00

The Best Prices on the Most Laptops

================================================ FILE: testdata/8859_1_es.html ================================================ MercadoLibre Mxico - Donde comprar y vender de todo.

MercadoLibre. Donde compras y vendes de todo.

6 meses sin intereses con Bancomer

Otros medios de pago

Publicidad

Nuevo Nokia Lumia 900
Enva Mails, chatea y Navega. Adquirelo Ahora!


================================================ FILE: testdata/8859_1_fr.html ================================================ Orange : tlphones, forfaits, Internet, actualit, sport, video
Accueil
bienvenue

  • vente flash
  • Xbox 360
  • open mini

boutique internet, mobile, internet + mobile

accessoires

BOUTIQUE ACCESSOIRESles équipements PC, Livebox et iPhonedécouvrir

Clé 3G+ / tablette

Clé 3G+découvrez nos offres pour profiter d'internet partout découvrir

fixe

découvrir toute l'offre fixeles offres, services et téléphones fixesdécouvrir
 

nos engagements

  • livraison gratuite

    La livraison des produits intervient sous 3 jours ouvrés compter de la validation de votre commande

  • paiement sécurisé

    Pour les paiements par carte bancaire, nous utilisons la norme SSL, standard le plus répandu

  • satisfait ou remboursé

    La boutique Orange vous ouvre cette possibilité pendant les 14 jours qui suivent la réception de votre colis.

  • SAV

    En 48h maximum, Orange vous échange votre mobile l'adresse de votre choix contre un matériel identique

  • conditions générales de vente
 

Dcouvrez l’univers Mobile Orange avec une large gamme de tlphones mobiles, et un grand choix d’offres mobiles : mobicarte, forfaits bloqus, forfaits Origami avec illimit et cl 3G+. Retrouvez l’univers Internet d’Orange avec les offres ADSL, TV, fixe et fibre optique mais galement la Livebox et tous ses accessoires high tech. Accdez Internet en dplacement depuis votre ordinateur portable avec la Cl 3G+ avec les pass, forfaits bloqus ou illimits Orange.

Clients Orange

pour profiter de tous vos services : messagerie, espace client,
changer de mobile ...
identifiez-vous
une question ? un problème ? consultez l'assistance

pour parcourir orange.fr

Le cadeau de PoutineLe cadeau de Poutine
accédez au portail

et suivez l'actualité, consultez la météo, visionnez le programme TV, interrogez l'assistance, comparez et achetez au meilleur prix, trouvez un correspondant avec le 118712 ...

suivi de commande Suivez en direct l'état d'avancement de votre commande.
Adoptez le kit oreillette
Adoptez le kit oreilletteLes bons conseils pour réduire son exposition aux ondesen savoir plus
================================================ FILE: testdata/8859_1_pt.html ================================================ O Jogo
Anterior Prximo
  • CLASSIFICAES
  • PRXIMA JORNADA
  • RESULTADOS
Liga ZON Sagres
01ª jornada
Liga ZON Sagres
01ª jornada
  • Olhanense
    17/8
    Estoril
  • Rio Ave
    18/8
    Martimo
  • Benfica
    18/8
    Braga
  • Nacional
    19/8
    V. Setbal
  • Paos de Ferreira
    19/8
    Moreirense
  • Gil Vicente
    19/8
    FC Porto
  • V. Guimares
    19/8
    Sporting
  • Beira-Mar
    20/8
    Acadmica
Liga ZON Sagres
01ª jornada
  • Olhanense
    17/8
    Estoril
  • Rio Ave
    18/8
    Martimo
  • Benfica
    18/8
    Braga
  • Nacional
    19/8
    V. Setbal
  • Paos de Ferreira
    19/8
    Moreirense
  • Gil Vicente
    19/8
    FC Porto
  • V. Guimares
    19/8
    Sporting
  • Beira-Mar
    20/8
    Acadmica

Dos quatro primeiros do ltimo campeonato, qual est a realizar a melhor pr-poca?

Votar
================================================ FILE: testdata/big5.html ================================================ ɹqlGu״IsDC

ɹql

`
YɷsDG

================================================ FILE: testdata/euc_jp.html ================================================ ڳŷԾShopping is Entertainment! 󥿡ͥåȺ̿䡢Υ饤󥷥åԥ󥰥ߥ˥ƥ
  • Edy
  • ġС
  • ֥å
  • ڷ
  • ȥ٥
  • ݡ
    • 2
    • 3
    • 9
    • 10
    • 11
    • 15
    • 17
    • 18
    • 19
    • 21
    • 22
    • 23
    • 24
    • 25
    • 26
    • 27
    • 28
    • 29
    • 32
    • 35
    • 36
    • 45
    • 46
    • 47
    • 48
    • 51
    • 54
    • 55
    • 56
    • 58
    • 59
    • 60
    • 63
    ӥ 褯Ȥӥ 褯Ȥӥ 
     
  • 󥭥
  • ̵
  • Ʊ
  • 쥸åȥ
  • ġС
  • ޻Ծ
  • Ծ
  • ֥å
  • ʥ
  • Żҥ֥å
  • ڷ
  • ȥ٥
  • ݡ
  • ŻԾ
  • ꥢ륿󥭥
  • ݥȥå׻Ծ
  • ̵å
ǽ Page  / 

ޤǤΤʸǺûˤϤ

û

Ϥʤäݥ5ܡ

Υ

ϰΤͤ餪䤤碌¿

ͤ餪礻¿

ʡ

ڵš

ڽܤ̣С

Racouponʥ顦ݥ 816()

ȥɥ816()

ۥåȥƥ

Pick Upܾ

ܻԾ

ŷ󤷤󥷥åԥ󥰥ӥ
Copyright (c) 1997-2012 Rakuten, Inc. All Rights Reserved.

close

Welcome International Customers

Singapore
Choose your location from a list of our global sites
ƹڡ˥ʱѸ
Japan
ܳŷԾ˥
View Rakuten in Japanese
ܳŷԾ˥
Always show Rakuten in Japanese

How to search for international shipping items
ʳоݾʤõˤĤơ

See more countries

================================================ FILE: testdata/euc_kr.html ================================================ Ʈ

ģҽĺ αϸ ģ ҽ ֽϴ


ε~
ε~
================================================ FILE: testdata/gb18030.html ================================================ ٶһ£֪
|¼ע

  ҳ ֪ MP3ͼ Ƭ Ƶ ͼ

ٿĿhao123 | >>

ѰٶΪҳѰٶΪҳѰٶӵ

ٶƹ | ư | ڰٶ | About Baidu

©2012 Baidu ʹðٶǰض ICP֤030173

================================================ FILE: testdata/shift_jis.html ================================================ dti services
@
 
 


dti servicesł́ASAXs[hAmA_bg[ɂql̂j[Yɂo܂悤AӂmȃT[rX񋟒vĂ܂B

dqiGNgjbNR}[XjAIC񋟁AC^NeBufBAioʐMjBC^[lbgɑ\Rs[^[ƃlbg[N̏[͎̐傫ςĂ܂B

dti services̃T[o[́AC^[lbgrWlXɑΉĂ܂BC[R}[XARecT[rXA摜zMAzMAXg[~ȎIAvtFbVi[YȊ‹񋟂܂B

Ђ̃eNjJ́A펞24ԑ̐őΉĂ܂̂łȂ鎖ԂɂvɑΉvĂ܂BX̃eiXSĂC܂Bɐp񋟂Ă܂̂ŁAGh[U[lɂ͑ł͑̊邱Ƃ̂łȂAC^[lbguEWO񑩒v܂B

ICNƉƂƂ܂ŁAdti servicesł͂qlITj[Yɂ肨ł悤AX^C̃zXeBO\[VȂтɊǗT[rX񋟂Ă܂BLxȒmAЂITX^btƊȃCtŁAql̊mȎѐx܂B

Ђ̃zXeBOAȂтIPANZX\[V́A̒ǐȂƎ̃X^C񋟒v܂BrWlXj[YAZpʂ܂ŁAql100% mȃT[rXA_A萫Ƃ΂₢Ή񑩂Ă܂B

EFuzXeBOǂ߂ɁÂ鑤ʂ𓖎ЃX^bt₷v܂B߂Ă̕łASĂp܂BЃEFuzXeBO̊{AEFuzXeBÔp@܂ŁAe؁AŠmɁAT[rX^S߂Ă񋟒v܂B

Ђ͂K͂̊ƂT|[gv܂B

@ICƂ߂Ďn߂ցAȒPŌoϓIȃEFuzXeBO񋟒v܂B
@荂\ȃptH[}XKvƂƗlɂ́Aĝ鍂^zXeB
@@ O񋟒vĂ܂B

܂AGITpɂΉ”\Ȉ肵W[\[V񋟒v܂BāARP[VAуf[^Z^[̃j[YɑΉ”\ȊǗT[rXƍō̎{݂pӒvĂ܂B
@

================================================ FILE: testdata/utf8.html ================================================ UTF-8 - Wikipedia, the free encyclopedia

UTF-8

From Wikipedia, the free encyclopedia
Jump to: navigation, search

UTF-8 (UCS Transformation Format—8-bit[1]) is a variable-width encoding that can represent every character in the Unicode character set. It was designed for backward compatibility with ASCII and to avoid the complications of endianness and byte order marks in UTF-16 and UTF-32.

UTF-8 has become the dominant character encoding for the World-Wide Web, accounting for more than half of all Web pages.[2][3][4] The Internet Engineering Task Force (IETF) requires all Internet protocols to identify the encoding used for character data, and the supported character encodings must include UTF-8.[5] The Internet Mail Consortium (IMC) recommends that all e-mail programs be able to display and create mail using UTF-8.[6] UTF-8 is also increasingly being used as the default character encoding in operating systems, programming languages, APIs, and software applications.[citation needed]

UTF-8 encodes each of the 1,112,064[7] code points in the Unicode character set using one to four 8-bit bytes (termed "octets" in the Unicode Standard). Code points with lower numerical values (i.e. earlier code positions in the Unicode character set, which tend to occur more frequently in practice) are encoded using fewer bytes.[8] The first 128 characters of Unicode, which correspond one-to-one with ASCII, are encoded using a single octet with the same binary value as ASCII, making valid ASCII text valid UTF-8-encoded Unicode as well.

The official IANA code for the UTF-8 character encoding is UTF-8.[9]

Contents

[edit] History

By early 1992 the search was on for a good byte-stream encoding of multi-byte character sets. The draft ISO 10646 standard contained a non-required annex called UTF-1 that provided a byte-stream encoding of its 32-bit code points. This encoding was not satisfactory on performance grounds, but did introduce the notion that bytes in the range of 0–127 continue representing the ASCII characters in UTF, thereby providing backward compatibility with ASCII.

In July 1992, the X/Open committee XoJIG was looking for a better encoding. Dave Prosser of Unix System Laboratories submitted a proposal for one that had faster implementation characteristics and introduced the improvement that 7-bit ASCII characters would only represent themselves; all multibyte sequences would include only bytes where the high bit was set.

In August 1992, this proposal was circulated by an IBM X/Open representative to interested parties. Ken Thompson of the Plan 9 operating system group at Bell Labs then made a crucial modification to the encoding to allow it to be self-synchronizing, meaning that it was not necessary to read from the beginning of the string to find code point boundaries. Thompson's design was outlined on September 2, 1992, on a placemat in a New Jersey diner with Rob Pike. The following days, Pike and Thompson implemented it and updated Plan 9 to use it throughout, and then communicated their success back to X/Open.[10]

UTF-8 was first officially presented at the USENIX conference in San Diego, from January 25–29, 1993.

In November 2003 UTF-8 was restricted by RFC 3629 to four bytes in order to match the constraints of the UTF-16 character encoding.

[edit] Description

The design of UTF‑8 is most easily seen in the table of the scheme as originally proposed by Dave Prosser and subsequently modified by Ken Thompson (the x's are replaced by the bits of the code point):

Bits Last code point Byte 1 Byte 2 Byte 3 Byte 4 Byte 5 Byte 6
  7 U+007F 0xxxxxxx
11 U+07FF 110xxxxx 10xxxxxx
16 U+FFFF 1110xxxx 10xxxxxx 10xxxxxx
21 U+1FFFFF 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
26 U+3FFFFFF 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
31 U+7FFFFFFF 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx

The salient features of the above scheme are as follows:

  1. One-byte codes are used only for the ASCII values 0 through 127. In this case the UTF-8 code has the same value as the ASCII code. The high-order bit of these codes is always 0.
  2. Codepoints larger than 127 are represented by multi-byte sequences, composed of a leading byte and one or more continuation bytes. The leading byte has two or more high-order 1s followed by a 0, while continuation bytes all have '10' in the high-order position.
  3. The remaining bits of the encoding are used for the bits of the codepoint being encoded, padded with high-order 0s if necessary. The number of bytes in the encoding is the minimum required to hold all the significant bits of the codepoint.
  4. The number of high-order 1s in the leading byte of a multi-byte sequence indicates the number of bytes in the sequence, so that the length of the sequence can be determined without examining the continuation bytes.
  5. Single bytes, leading bytes, and continuation bytes do not share values. This makes the scheme "self synchronizing", allowing the start of a character to be found by backing up at most five bytes (three bytes in actual UTF‑8 as explained below).

The first 128 characters (US-ASCII) need one byte. The next 1,920 characters need two bytes to encode. This covers the remainder of almost all Latin-derived alphabets, and also Greek, Cyrillic, Coptic, Armenian, Hebrew, Arabic, Syriac and Tāna alphabets, as well as Combining Diacritical Marks. Three bytes are needed for characters in the rest of the Basic Multilingual Plane (which contains virtually all characters in common use). Four bytes are needed for characters in the other planes of Unicode, which include less common CJK characters and various historic scripts and mathematical symbols.

The original specification covered numbers up to 31 bits (the original limit of the Universal Character Set). In November 2003 UTF-8 was restricted by RFC 3629 to end at U+10FFFF, in order to match the constraints of the UTF-16 character encoding. This removed all 5- and 6-byte sequences, and about half of the 4-byte sequences.

[edit] Examples

Let us consider how to encode the Euro sign, €.

  1. The Unicode code point for "€" is U+20AC.
  2. According to the scheme table above, this will take three bytes to encode, as it is between U+07FF and U+FFFF.
  3. Hexadecimal 20AC is binary 0010000010101100. The two leading zeros are added because, as the scheme table shows, a three-byte encoding needs exactly sixteen bits from the code point.
  4. Because it is a three-byte encoding, the leading byte starts with three 1s, then a 0 (1110...)
  5. The remaining bits of this byte are taken from the code point (11100010), leaving ...000010101100.
  6. Each of the continuation bytes starts with 10 and takes six bits of the code point (so 10000010, then 10101100).

The three bytes 11100010 10000010 10101100 can be more concisely written in hexadecimal, as E2 82 AC.

The following table summarises this conversion, as well as others with different lengths in UTF-8. The colors indicate how bits from the code point are distributed among the UTF-8 bytes. Additional bits added by the UTF-8 encoding process are shown in black.

Character Binary code point Binary UTF-8 Hexadecimal UTF-8
$ U+0024 0100100 00100100 24
¢ U+00A2 00010100010 11000010 10100010 C2 A2
U+20AC 0010000010101100 11100010 10000010 10101100 E2 82 AC
𤭢 U+24B62 000100100101101100010 11110000 10100100 10101101 10100010 F0 A4 AD A2

[edit] Overlong encodings

The standard specifies that the correct encoding of a codepoint use only the minimum number of bytes required to hold the significant bits of the codepoint. Longer encodings are called overlong and are not valid UTF-8 representations of the codepoint. This rule maintains a one-to-one correspondence between codepoints and their valid encodings, so that there is a unique valid encoding for each codepoint. Allowing multiple encodings would make testing for string equality difficult to define.

In principle, it would be possible to inflate the number of bytes in an encoding by padding the codepoint with leading 0s. To encode the Euro sign € from the above example in four bytes instead of three, it could be padded with leading 0s until it was 21 bits long—000000010000010101100. The leading byte prefix for a four-byte encoding is 11110, and so the complete, overlong encoding is 11110000 10000010 10000010 10101100 (or F0 82 82 AC in hexadecimal).

Although overlong encodings are forbidden in UTF-8, at least one derivative makes use of the form. Modified UTF-8 requires the Unicode codepoint U+0000 (the NUL character) to be encoded in the overlong form 11000000 10000000 (hex C0 80), rather than 00000000 (hex 00). This allows the byte 00 to be used as a string terminator.

[edit] Codepage layout

UTF-8
_0 _1 _2 _3 _4 _5 _6 _7 _8 _9 _A _B _C _D _E _F
 
0_
 
NUL
0000
0
SOH
0001
1
STX
0002
2
ETX
0003
3
EOT
0004
4
ENQ
0005
5
ACK
0006
6
BEL
0007
7
BS
0008
8
HT
0009
9
LF
000A
10
VT
000B
11
FF
000C
12
CR
000D
13
SO
000E
14
SI
000F
15
 
1_
 
DLE
0010
16
DC1
0011
17
DC2
0012
18
DC3
0013
19
DC4
0014
20
NAK
0015
21
SYN
0016
22
ETB
0017
23
CAN
0018
24
EM
0019
25
SUB
001A
26
ESC
001B
27
FS
001C
28
GS
001D
29
RS
001E
30
US
001F
31
 
2_
 
SP
0020
32
!
0021
33
"
0022
34
#
0023
35
$
0024
36
%
0025
37
&
0026
38
'
0027
39
(
0028
40
)
0029
41
*
002A
42
+
002B
43
,
002C
44
-
002D
45
.
002E
46
/
002F
47
 
3_
 
0
0030
48
1
0031
49
2
0032
50
3
0033
51
4
0034
52
5
0035
53
6
0036
54
7
0037
55
8
0038
56
9
0039
57
:
003A
58
;
003B
59
<
003C
60
=
003D
61
>
003E
62
?
003F
63
 
4_
 
@
0040
64
A
0041
65
B
0042
66
C
0043
67
D
0044
68
E
0045
69
F
0046
70
G
0047
71
H
0048
72
I
0049
73
J
004A
74
K
004B
75
L
004C
76
M
004D
77
N
004E
78
O
004F
79
 
5_
 
P
0050
80
Q
0051
81
R
0052
82
S
0053
83
T
0054
84
U
0055
85
V
0056
86
W
0057
87
X
0058
88
Y
0059
89
Z
005A
90
[
005B
91
\
005C
92
]
005D
93
^
005E
94
_
005F
95
 
6_
 
`
0060
96
a
0061
97
b
0062
98
c
0063
99
d
0064
100
e
0065
101
f
0066
102
g
0067
103
h
0068
104
i
0069
105
j
006A
106
k
006B
107
l
006C
108
m
006D
109
n
006E
110
o
006F
111
 
7_
 
p
0070
112
q
0071
113
r
0072
114
s
0073
115
t
0074
116
u
0075
117
v
0076
118
w
0077
119
x
0078
120
y
0079
121
z
007A
122
{
007B
123
|
007C
124
}
007D
125
~
007E
126
DEL
007F
127
 
8_
 

+00
128

+01
129

+02
130

+03
131

+04
132

+05
133

+06
134

+07
135

+08
136

+09
137

+0A
138

+0B
139

+0C
140

+0D
141

+0E
142

+0F
143
 
9_
 

+10
144

+11
145

+12
146

+13
147

+14
148

+15
149

+16
150

+17
151

+18
152

+19
153

+1A
154

+1B
155

+1C
156

+1D
157

+1E
158

+1F
159
 
A_
 

+20
160

+21
161

+22
162

+23
163

+24
164

+25
165

+26
166

+27
167

+28
168

+29
169

+2A
170

+2B
171

+2C
172

+2D
173

+2E
174

+2F
175
 
B_
 

+30
176

+31
177

+32
178

+33
179

+34
180

+35
181

+36
182

+37
183

+38
184

+39
185

+3A
186

+3B
187

+3C
188

+3D
189

+3E
190

+3F
191
 
2-byte
C_
 
2-byte
inval

(0000)
192
2-byte
inval

(0040)
193
Latin-1
0080
194
Latin-1
00C0
195
Latin
Ext-A

0100
196
Latin
Ext-A

0140
197
Latin
Ext-B

0180
198
Latin
Ext-B

01C0
199
Latin
Ext-B

0200
200
IPA
0240
201
IPA
0280
202
Spaci
Modif

02C0
203
Combi
Diacr

0300
204
Combi
Diacr

0340
205
Greek
0380
206
Greek
03C0
207
 
2-byte
D_
 
Cyril
0400
208
Cyril
0440
209
Cyril
0480
210
Cyril
04C0
211
Cyril
0500
212
Armen
0540
213
Hebrew
0580
214
Hebrew
05C0
215
Arabic
0600
216
Arabic
0640
217
Arabic
0680
218
Arabic
06C0
219
Syriac
0700
220
Arabic
0740
221
Thaana
0780
222
N'Ko
07C0
223
 
3-byte
E_
 
Indic
0800*
224
Misc.
1000
225
Symbol
2000
226
Kana
CJK

3000
227
CJK
4000
228
CJK
5000
229
CJK
6000
230
CJK
7000
231
CJK
8000
232
CJK
9000
233
Asian
A000
234
Hangul
B000
235
Hangul
C000
236
Hangul
Surr

D000
237
Priv Use
E000
238
Forms
F000
239
 
4-byte
F_
 
Ancient
Sym,CJK

10000*
240
unall
40000
241
unall
80000
242
Tags
Priv

C0000
243
Priv
Use

100000
244
4-byte
inval

140000
245
4-byte
inval

180000
246
4-byte
inval

1C0000
247
5-byte
inval

200000*
248
5-byte
inval

1000000
249
5-byte
inval

2000000
250
5-byte
inval

3000000
251
6-byte
inval

4000000*
252
6-byte
inval

40000000
253


254


255

Legend: Yellow cells are control characters, blue cells are punctuation, purple cells are digits and green cells are ASCII letters.

Orange cells with a large dot are continuation bytes. The hexadecimal number shown after a "+" plus sign is the value of the 6 bits they add.

White cells are the start bytes for a sequence of multiple bytes, the length shown at the left edge of the row. The text shows the Unicode blocks encoded by sequences starting with this byte, and the hexadecimal code point shown in the cell is the lowest character value encoded using that start byte. When a start byte could form both overlong and valid encodings, the lowest non-overlong-encoded codepoint is shown, marked by an asterisk "*".

Red cells must never appear in a valid UTF-8 sequence. The first two (C0 and C1) could only be used for overlong encoding of basic ASCII characters. The remaining red cells indicate start bytes of sequences that could only encode numbers larger than the 0x10FFFF limit of Unicode. The byte 244 (hex 0xF4) could also encode some values greater than 0x10FFFF; such a sequence is also invalid.

[edit] Invalid byte sequences

Not all sequences of bytes are valid UTF-8. A UTF-8 decoder should be prepared for:

  • the red invalid bytes in the above table
  • an unexpected continuation byte
  • a start byte not followed by enough continuation bytes
  • a sequence that decodes to a value that should use a shorter sequence (an "overlong form")
  • A 4-byte sequence (starting with 0xF4) that decodes to a value greater than U+10FFFF

Many earlier decoders would happily try to decode these. Carefully crafted invalid UTF-8 could make them either skip or create ASCII characters such as NUL, slash, or quotes. Invalid UTF-8 has been used to bypass security validations in high profile products including Microsoft's IIS web server[11] and Apache's Tomcat servlet container.[12]

RFC 3629 states "Implementations of the decoding algorithm MUST protect against decoding invalid sequences."[13] The Unicode Standard requires decoders to "...treat any ill-formed code unit sequence as an error condition. This guarantees that it will neither interpret nor emit an ill-formed code unit sequence."

Many UTF-8 decoders throw exceptions on encountering errors,[14] since such errors suggest the input is not a UTF-8 string at all. This can turn what would otherwise be harmless errors (producing a message such as "no such file") into a denial of service bug. For instance, Python 3.0 would exit immediately if the command line or environment variables contained invalid UTF-8,[15] so it was impossible for any Python program to detect and recover from such an error.

An increasingly popular option is to detect errors with a separate API, and for converters to translate the first byte to a replacement and continue parsing with the next byte. Popular replacements are:

  • The replacement character "�" (U+FFFD)
  • The invalid Unicode code points U+DC80..U+DCFF where the low 8 bits are the byte's value.
  • Interpret the bytes according to ISO-8859-1 or CP1252.

Replacing errors is "lossy": more than one UTF-8 string converts to the same Unicode result. Therefore the original UTF-8 should be stored, and translation should only be used when displaying the text to the user.

[edit] Invalid code points

According to the UTF-8 definition (RFC 3629) the high and low surrogate halves used by UTF-16 (U+D800 through U+DFFF) are not legal Unicode values, and the UTF-8 encoding of them is an invalid byte sequence and thus should be treated as described above.

Whether an actual application should do this with surrogate halves is debatable.[who?] Allowing them allows lossless storage of invalid UTF-16, and allows CESU encoding (described below) to be decoded. There are other code points that are far more important to detect and reject, such as the reversed-BOM U+FFFE, or the C1 controls, caused by improper conversion of CP1252 text or double-encoding of UTF-8. These are invalid in HTML.

[edit] Official name and variants

The official name is "UTF-8". All letters are upper-case, and the name is hyphenated. This spelling is used in all the documents relating to the encoding.

Alternatively, the name "utf-8" may be used by all standards conforming to the Internet Assigned Numbers Authority (IANA) list (which include CSS, HTML, XML, and HTTP headers),[16] as the declaration is case insensitive.[17]

Other descriptions that omit the hyphen or replace it with a space, such as "utf8" or "UTF 8", are not accepted as correct by the governing standards.[18] Despite this, most agents such as browsers can understand them, and so standards intended to describe existing practice (such as HTML5) may effectively require their recognition.

MySQL omits the hyphen in the following query:

SET NAMES 'utf8'

[edit] Derivatives

The following implementations show slight differences from the UTF-8 specification. They are incompatible with the UTF-8 specification.

[edit] CESU-8

Many programs added UTF-8 conversions for UCS-2 data and did not alter this UTF-8 conversion when UCS-2 was replaced with the surrogate-pair using UTF-16. In such programs each half of a UTF-16 surrogate pair is encoded as its own 3-byte UTF-8 encoding, resulting in 6-byte sequences rather than 4 bytes for characters outside the Basic Multilingual Plane. Oracle and MySQL databases use this, as well as Java and Tcl as described below, and probably many Windows programs where the programmers were unaware of the complexities of UTF-16. Although this non-optimal encoding is generally not deliberate, a supposed benefit is that it preserves UTF-16 binary sorting order when CESU-8 is binary sorted.

[edit] Modified UTF-8

In Modified UTF-8,[19] the null character (U+0000) is encoded as 0xC0,0x80; this is not valid UTF-8[20] because it is not the shortest possible representation. Modified UTF-8 strings never contain any actual null bytes but can contain all Unicode code points including U+0000,[21] which allows such strings (with a null byte appended) to be processed by traditional null-terminated string functions.

All known Modified UTF-8 implementations also treat the surrogate pairs as in CESU-8.

In normal usage, the Java programming language supports standard UTF-8 when reading and writing strings through InputStreamReader and OutputStreamWriter. However it uses Modified UTF-8 for object serialization,[22] for the Java Native Interface,[23] and for embedding constant strings in class files.[24] Tcl also uses the same modified UTF-8[25] as Java for internal representation of Unicode data, but uses strict CESU-8 for external data.

[edit] Extending from 31 bit to 36 bit range

Extending the accepted input pattern from 6 bytes to 7 bytes would allow over 70 billion code points to be encoded;[26] however, this would require an initial byte value of 0xFE to be accepted as a 7-byte sequence indicator (see under Advantages in section "Compared to single-byte encodings").

[edit] Byte order mark

Many Windows programs (including Windows Notepad) add the bytes 0xEF, 0xBB, 0xBF at the start of any document saved as UTF-8. This is the UTF-8 encoding of the Unicode byte order mark (BOM), and is commonly referred to as a UTF-8 BOM, even though it is not relevant to byte order. The BOM can also appear if another encoding with a BOM is translated to UTF-8 without stripping it. Older text editors may display the BOM as "" at the start of the document.

The Unicode Standard neither requires nor recommends the use of the BOM for UTF-8.[27] The presence of the UTF-8 BOM may cause interoperability problems with existing software that could otherwise handle UTF-8; for example:

  • Programming language parsers not explicitly designed for UTF-8 can often handle UTF-8 in string constants and comments, but cannot parse the BOM at the start of the file.
  • Programs that identify file types by leading characters may fail to identify the file if a BOM is present even if the user of the file could skip the BOM. An example is the Unix shebang syntax. Another example is Internet Explorer which will render pages in standards mode only when it starts with a document type declaration.

If compatibility with existing programs is not important, the BOM could be used to identify UTF-8 encoding. Because checking if text is valid UTF-8 is very reliable (the majority of random byte sequences are not valid UTF-8) such use should not be necessary. Programs that insert information at the start of a file will break this identification (one example is offline browsers that add the originating URL to the start of the file).

Unofficially, UTF-8-BOM or UTF-8-NOBOM are sometimes used to refer to text files which contain or lack a BOM. In Japan especially, "UTF-8 encoding without BOM" is sometimes called "UTF-8N".[citation needed]

[edit] Advantages and disadvantages

[edit] General

[edit] Advantages

  • The ASCII characters are represented by themselves as single bytes that do not appear anywhere else, which makes UTF-8 work with the majority of existing APIs that take bytes strings but only treat a small number of ASCII codes specially. This removes the need to write a new Unicode version of every API, and makes it much easier to convert existing systems to UTF-8 than any other Unicode encoding.
  • UTF-8 is the only encoding for XML entities that does not require a BOM or an indication of the encoding.[28]
  • UTF-8 and UTF-16 are the standard encodings for Unicode text in HTML documents, with UTF-8 as the preferred and most used encoding.
  • UTF-8 strings can be fairly reliably recognized as such by a simple heuristic algorithm.[29] The probability of a random string of bytes which is not pure ASCII being valid UTF-8 is 3.9% for a two-byte sequence,[30] and decreases exponentially for longer sequences. ISO/IEC 8859-1 is even less likely to be mis-recognized as UTF-8: the only non-ASCII characters in it would have to be in sequences starting with either an accented letter or the multiplication symbol and ending with a symbol. This is an advantage that most other encodings do not have, causing errors (mojibake) if the receiving application isn't told and can't guess the correct encoding. Even word-based UTF-16 can be mistaken for byte encodings (like in the "bush hid the facts" bug).
  • Sorting of UTF-8 strings as arrays of unsigned bytes will produce the same results as sorting them based on Unicode code points.
  • Other byte-based encodings can pass through the same API. This means, however, that the encoding must be identified. Because the other encodings are unlikely to be valid UTF-8, a reliable way to implement this is to assume UTF-8 and switch to a legacy encoding only if several invalid UTF-8 byte sequences are encountered.

[edit] Disadvantages

  • A UTF-8 parser that is not compliant with current versions of the standard might accept a number of different pseudo-UTF-8 representations and convert them to the same Unicode output. This provides a way for information to leak past validation routines designed to process data in its eight-bit representation.[31]

[edit] Compared to single-byte encodings

[edit] Advantages

  • UTF-8 can encode any Unicode character, avoiding the need to figure out and set a "code page" or otherwise indicate what character set is in use, and allowing output in multiple scripts at the same time. For many scripts there have been more than one single-byte encoding in usage, so even knowing the script was insufficient information to display it correctly.
  • The bytes 0xFE and 0xFF do not appear, so a valid UTF-8 stream never matches the UTF-16 byte order mark and thus cannot be confused with it. The absence of 0xFF (0377) also eliminates the need to escape this byte in Telnet (and FTP control connection).

[edit] Disadvantages

  • UTF-8 encoded text is larger than the appropriate single-byte encoding except for plain ASCII characters. In the case of scripts which used 8-bit character sets with non-Latin scripts encoded in the upper half (such as most Cyrillic and Greek alphabet code pages), characters in UTF-8 will be double the size. For some scripts such as Thai and Hindi's Devanagari, characters will be triple the size (this has caused objections in India and other countries).
  • It is possible in UTF-8 (or any other multi-byte encoding) to split or truncate a string in the middle of a character, which may result in an invalid string. This will not happen in correct handling of UTF-8.
  • If the code points are all the same size, measurements of a fixed number of them is easy. Due to ASCII-era documentation where "character" is used as a synonym for "byte" this is often considered important. However, by measuring string positions using bytes instead of "characters" most algorithms can be easily and efficiently adapted for UTF-8[citation needed].
  • Some software such as text editors will refuse to correctly display or interpret UTF-8 unless it starts with a Byte Order Mark. This has the effect of making it impossible to use UTF-8 with any older software than can handle ASCII-like encodings but cannot handle the byte order mark. This is considered an incorrect implementation of the text editor, not the older software.

[edit] Compared to other multi-byte encodings

[edit] Advantages

  • UTF-8 uses the codes 0–127 only for the ASCII characters. This means that UTF-8 is an ASCII extension and can with limited change be supported by software that supports an ASCII extension and handles non-ASCII characters as free text.
  • UTF-8 can encode any Unicode character. Files in different scripts can be displayed correctly without having to choose the correct code page or font. For instance Chinese and Arabic can be supported (in the same text) without special codes inserted or manual settings to switch the encoding.
  • UTF-8 is "self-synchronizing": character boundaries are easily found when searching either forwards or backwards. If bytes are lost due to error or corruption, one can always locate the beginning of the next character and thus limit the damage. Many multi-byte encodings are much harder to resynchronize.
  • Any byte oriented string searching algorithm can be used with UTF-8 data, since the sequence of bytes for a character cannot occur anywhere else. Some older variable-length encodings (such as Shift JIS) did not have this property and thus made string-matching algorithms rather complicated. In Shift JIS the end byte of a character and the first byte of the next character could look like another legal character, something that can't happen in UTF-8.
  • Efficient to encode using simple bit operations. UTF-8 does not require slower mathematical operations such as multiplication or division (unlike the obsolete UTF-1 encoding).

[edit] Disadvantages

  • For certain scripts UTF-8 will take more space than an older multi-byte encoding. East Asian scripts generally have two bytes per character in their multi-byte encodings yet take three bytes per character in UTF-8.


[edit] Compared to UTF-16

[edit] Advantages

  • A text byte stream cannot be losslessly converted to UTF-16, due to the possible presence of errors in the byte stream encoding. This causes unexpected and often severe problems attempting to use existing data in a system that uses UTF-16 as an internal encoding. Results are security bugs, DoS if bad encoding throws an exception, and data loss when different byte streams convert to the same UTF-16. Due to the ASCII compatibility and high degree of pattern recognition in UTF-8, random byte streams can be passed losslessly through a system using it, as interpretation can be deferred until display.
  • Converting to UTF-16 while maintaining compatibility with existing programs (such as was done with Windows) requires every API and data structure that takes a string to be duplicated. Invalid encodings make the duplicated APIs not exactly map to each other, often making it impossible to do some action with one of them.
  • Characters outside the basic multilingual plane are not a special case. UTF-16 is often mistaken to be the obsolete constant-length UCS-2 encoding, leading to code that works for most text but suddenly fails for non-BMP characters.[32]
  • Text encoded in UTF-8 is often smaller than (or the same size as) the same text encoded in UTF-16.
    • This is always true for text using only code points below U+0800 (which includes all modern European scripts), as each code point's UTF-8 encoding is one or two bytes then.
    • Even if text contains code points between U+0800 and U+FFFF, it might contain so many code points below U+0080 (which UTF-8 encodes in one byte) that the UTF-8 encoding is still smaller. As HTML markup and line terminators are code points below U+0080, most HTML source is smaller if encoded in UTF-8 even for Asian scripts.
    • Non-BMP characters (U+10000 and above) are encoded in UTF-8 in four bytes, the same size as in UTF-16.
  • Most communication and storage was designed for a stream of bytes. A UTF-16 string must use a pair of bytes for each code unit:
    • The order of those two bytes becomes an issue and must be specified in the UTF-16 protocol, such as with a byte order mark.
    • If an odd number of bytes is missing from UTF-16, the whole rest of the string will be meaningless text. Any bytes missing from UTF-8 will still allow the text to be recovered accurately starting with the next character after the missing bytes. If any partial character is removed the corruption is always recognizable.

[edit] Disadvantages

  • Characters U+0800 through U+FFFF use three bytes in UTF-8, but only two in UTF-16. As a result, text in (for example) Chinese, Japanese or Hindi could take more space in UTF-8 if there are more of these characters than there are ASCII characters. This happens for pure text,[33] but rarely for HTML documents. For example, both the Japanese UTF-8 and the Hindi Unicode articles on Wikipedia take more space in UTF-16 than in UTF-8 .[34]

[edit] See also

[edit] References

  1. ^ [|The Unicode Consortium]. "Chapter 2. General Structure". The Unicode Standard (6.0 ed.). Mountain View, California, USA: The Unicode Consortium. ISBN 978-1-936213-01-6. http://www.unicode.org/versions/Unicode6.0.0/. . RFC 3629 also refers to UTF-8 as "UCS transformation format". Also commonly known as "Unicode Transformation Format".
  2. ^ Mark Davis (28 January 2010). "Unicode nearing 50% of the web". Official Google Blog. Google. http://googleblog.blogspot.com/2010/01/unicode-nearing-50-of-web.html. Retrieved 5 December 2010. 
  3. ^ "UTF-8 Usage Statistics". BuiltWith. http://trends.builtwith.com/encoding/UTF-8. Retrieved 2011-03-28. 
  4. ^ "Usage of character encodings for websites". W3Techs. http://w3techs.com/technologies/overview/character_encoding/all. Retrieved 2010-03-30. 
  5. ^ Alvestrand, H. (1998). "IETF Policy on Character Sets and Languages". RFC 2277. Internet Engineering Task Force. 
  6. ^ "Using International Characters in Internet Mail". Internet Mail Consortium. August 1, 1998. http://www.imc.org/mail-i18n.html. Retrieved 2007-11-08. 
  7. ^ Not all of the 1,112,064 possible code points have been assigned characters; many are reserved for future use, and some are reserved for private use, while still others are specified as permanently undefined.
  8. ^ More precisely, the number of bytes used to encode a character at a given code point is a monotonically increasing function of the numerical value of the code point.
  9. ^ Internet Assigned Numbers Authority (4 November 2010). "CHARACTER SETS". IANA. http://www.iana.org/assignments/character-sets. Retrieved 5 December 2010. 
  10. ^ Pike, Rob (2003-04-03). "UTF-8 history". http://www.cl.cam.ac.uk/~mgk25/ucs/utf-8-history.txt. 
  11. ^ Marin, Marvin (2000-10-17). "Web Server Folder Traversal MS00-078". http://www.sans.org/resources/malwarefaq/wnt-unicode.php. 
  12. ^ "National Vulnerability Database - Summary for CVE-2008-2938". http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2008-2938. 
  13. ^ Yergeau, F. (2003). "UTF-8, a transformation format of ISO 10646". RFC 3629. Internet Engineering Task Force 
  14. ^ Examples: UTF8 (Java Class Library API) or java.nio.charset.CharsetDecoder.decode
  15. ^ "Non-decodable Bytes in System Character Interfaces". http://www.python.org/dev/peps/pep-0383/. 
  16. ^ W3C: Setting the HTTP charset parameter notes that the IANA list is used for HTTP
  17. ^ Internet Assigned Numbers Authority Character Sets
  18. ^ RFC 3629 UTF-8 see chapter 8. MIME registration, first paragraph
  19. ^ "Java SE 6 documentation for Interface java.io.DataInput, subsection on Modified UTF-8". Sun Microsystems. 2008. http://java.sun.com/javase/6/docs/api/java/io/DataInput.html#modified-utf-8. Retrieved 2009-05-22. 
  20. ^ "[...] the overlong UTF-8 sequence C0 80 [...]", "[...] the illegal two-octet sequence C0 80 [...]""Request for Comments 3629: "UTF-8, a transformation format of ISO 10646"". 2003. http://www.apps.ietf.org/rfc/rfc3629.html#page-5. Retrieved 2009-05-22. 
  21. ^ "[...] Java virtual machine UTF-8 strings never have embedded nulls.""The Java Virtual Machine Specification, 2nd Edition, section 4.4.7: "The CONSTANT_Utf8_info Structure"". Sun Microsystems. 1999. http://java.sun.com/docs/books/jvms/second_edition/html/ClassFile.doc.html#7963. Retrieved 2009-05-24. 
  22. ^ "[...] encoded in modified UTF-8.""Java Object Serialization Specification, chapter 6: Object Serialization Stream Protocol, section 2: Stream Elements". Sun Microsystems. 2005. http://java.sun.com/javase/6/docs/platform/serialization/spec/protocol.html#8299. Retrieved 2009-05-22. 
  23. ^ "The JNI uses modified UTF-8 strings to represent various string types.""Java Native Interface Specification, chapter 3: JNI Types and Data Structures, section: Modified UTF-8 Strings". Sun Microsystems. 2003. http://java.sun.com/j2se/1.5.0/docs/guide/jni/spec/types.html#wp16542. Retrieved 2009-05-22. 
  24. ^ "[...] differences between this format and the "standard" UTF-8 format.""The Java Virtual Machine Specification, 2nd Edition, section 4.4.7: "The CONSTANT_Utf8_info Structure"". Sun Microsystems. 1999. http://java.sun.com/docs/books/jvms/second_edition/html/ClassFile.doc.html#7963. Retrieved 2009-05-23. 
  25. ^ "In orthodox UTF-8, a NUL byte(\x00) is represented by a NUL byte. [...] But [...] we [...] want NUL bytes inside [...] strings [...]""Tcler's Wiki: UTF-8 bit by bit (Revision 6)". 2009-04-25. http://wiki.tcl.tk/_/revision?N=1211&V=6. Retrieved 2009-05-22. 
  26. ^ Computational Methods in Linguistic Research, 2004
  27. ^ The Unicode Standard - Chapter 2, see chapter 2.6 page 30 bottom.
  28. ^ W3.org
  29. ^ W3 FAQ: Multilingual Forms: a Perl regular expression to validate a UTF-8 string)
  30. ^ There are 256 × 256 − 128 × 128 not-pure-ASCII two-byte sequences, and of those, only 1920 encode valid UTF-8 characters (the range U+0080 to U+07FF), so the proportion of valid not-pure-ASCII two-byte sequences is 3.9%. Note that this assumes that control characters pass as ASCII; without the control characters, the percentage proportions drop somewhat).
  31. ^ Tools.ietf.org
  32. ^ "Should UTF-16 be considered harmful?". Stackoverflow.com. http://stackoverflow.com/questions/1049947/should-utf-16-be-considered-harmful. Retrieved 2010-09-13. 
  33. ^ Although the difference may not be great: the 2010-11-22 version of hi:यूनिकोड (Unicode in Hindi), when the pure text was pasted to Notepad, generated 19 KB when saved as UTF-16 and 22 KB when saved as UTF-8.
  34. ^ The 2010-10-27 version of ja:UTF-8 generated 169 KB when converted with Notepad to UTF-16, and only 101 KB when converted back to UTF-8. The 2010-11-22 version of hi:यूनिकोड (Unicode in Hindi) required 119 KB in UTF-16 and 76 KB in UTF-8.

[edit] External links

There are several current definitions of UTF-8 in various standards documents:

  • RFC 3629 / STD 63 (2003), which establishes UTF-8 as a standard Internet protocol element
  • The Unicode Standard, Version 6.0, §3.9 D92, §3.10 D95 (2011)
  • ISO/IEC 10646:2003 Annex D (2003)

They supersede the definitions given in the following obsolete works:

  • ISO/IEC 10646-1:1993 Amendment 2 / Annex R (1996)
  • The Unicode Standard, Version 5.0, §3.9 D92, §3.10 D95 (2007)
  • The Unicode Standard, Version 4.0, §3.9–§3.10 (2003)
  • The Unicode Standard, Version 2.0, Appendix A (1996)
  • RFC 2044 (1996)
  • RFC 2279 (1998)
  • The Unicode Standard, Version 3.0, §2.3 (2000) plus Corrigendum #1 : UTF-8 Shortest Form (2000)
  • Unicode Standard Annex #27: Unicode 3.1 (2001)

They are all the same in their general mechanics, with the main differences being on issues such as allowed range of code point values and safe handling of invalid input.

Personal tools
Namespaces

Variants
Actions
Navigation
Interaction
Toolbox
Print/export
Languages
================================================ FILE: testdata/utf8_bom.html ================================================  UTF-8 - Wikipedia, the free encyclopedia

UTF-8

From Wikipedia, the free encyclopedia
Jump to: navigation, search

UTF-8 (UCS Transformation Format—8-bit[1]) is a variable-width encoding that can represent every character in the Unicode character set. It was designed for backward compatibility with ASCII and to avoid the complications of endianness and byte order marks in UTF-16 and UTF-32.

UTF-8 has become the dominant character encoding for the World-Wide Web, accounting for more than half of all Web pages.[2][3][4] The Internet Engineering Task Force (IETF) requires all Internet protocols to identify the encoding used for character data, and the supported character encodings must include UTF-8.[5] The Internet Mail Consortium (IMC) recommends that all e-mail programs be able to display and create mail using UTF-8.[6] UTF-8 is also increasingly being used as the default character encoding in operating systems, programming languages, APIs, and software applications.[citation needed]

UTF-8 encodes each of the 1,112,064[7] code points in the Unicode character set using one to four 8-bit bytes (termed "octets" in the Unicode Standard). Code points with lower numerical values (i.e. earlier code positions in the Unicode character set, which tend to occur more frequently in practice) are encoded using fewer bytes.[8] The first 128 characters of Unicode, which correspond one-to-one with ASCII, are encoded using a single octet with the same binary value as ASCII, making valid ASCII text valid UTF-8-encoded Unicode as well.

The official IANA code for the UTF-8 character encoding is UTF-8.[9]

Contents

[edit] History

By early 1992 the search was on for a good byte-stream encoding of multi-byte character sets. The draft ISO 10646 standard contained a non-required annex called UTF-1 that provided a byte-stream encoding of its 32-bit code points. This encoding was not satisfactory on performance grounds, but did introduce the notion that bytes in the range of 0–127 continue representing the ASCII characters in UTF, thereby providing backward compatibility with ASCII.

In July 1992, the X/Open committee XoJIG was looking for a better encoding. Dave Prosser of Unix System Laboratories submitted a proposal for one that had faster implementation characteristics and introduced the improvement that 7-bit ASCII characters would only represent themselves; all multibyte sequences would include only bytes where the high bit was set.

In August 1992, this proposal was circulated by an IBM X/Open representative to interested parties. Ken Thompson of the Plan 9 operating system group at Bell Labs then made a crucial modification to the encoding to allow it to be self-synchronizing, meaning that it was not necessary to read from the beginning of the string to find code point boundaries. Thompson's design was outlined on September 2, 1992, on a placemat in a New Jersey diner with Rob Pike. The following days, Pike and Thompson implemented it and updated Plan 9 to use it throughout, and then communicated their success back to X/Open.[10]

UTF-8 was first officially presented at the USENIX conference in San Diego, from January 25–29, 1993.

In November 2003 UTF-8 was restricted by RFC 3629 to four bytes in order to match the constraints of the UTF-16 character encoding.

[edit] Description

The design of UTF‑8 is most easily seen in the table of the scheme as originally proposed by Dave Prosser and subsequently modified by Ken Thompson (the x's are replaced by the bits of the code point):

Bits Last code point Byte 1 Byte 2 Byte 3 Byte 4 Byte 5 Byte 6
  7 U+007F 0xxxxxxx
11 U+07FF 110xxxxx 10xxxxxx
16 U+FFFF 1110xxxx 10xxxxxx 10xxxxxx
21 U+1FFFFF 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
26 U+3FFFFFF 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
31 U+7FFFFFFF 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx

The salient features of the above scheme are as follows:

  1. One-byte codes are used only for the ASCII values 0 through 127. In this case the UTF-8 code has the same value as the ASCII code. The high-order bit of these codes is always 0.
  2. Codepoints larger than 127 are represented by multi-byte sequences, composed of a leading byte and one or more continuation bytes. The leading byte has two or more high-order 1s followed by a 0, while continuation bytes all have '10' in the high-order position.
  3. The remaining bits of the encoding are used for the bits of the codepoint being encoded, padded with high-order 0s if necessary. The number of bytes in the encoding is the minimum required to hold all the significant bits of the codepoint.
  4. The number of high-order 1s in the leading byte of a multi-byte sequence indicates the number of bytes in the sequence, so that the length of the sequence can be determined without examining the continuation bytes.
  5. Single bytes, leading bytes, and continuation bytes do not share values. This makes the scheme "self synchronizing", allowing the start of a character to be found by backing up at most five bytes (three bytes in actual UTF‑8 as explained below).

The first 128 characters (US-ASCII) need one byte. The next 1,920 characters need two bytes to encode. This covers the remainder of almost all Latin-derived alphabets, and also Greek, Cyrillic, Coptic, Armenian, Hebrew, Arabic, Syriac and Tāna alphabets, as well as Combining Diacritical Marks. Three bytes are needed for characters in the rest of the Basic Multilingual Plane (which contains virtually all characters in common use). Four bytes are needed for characters in the other planes of Unicode, which include less common CJK characters and various historic scripts and mathematical symbols.

The original specification covered numbers up to 31 bits (the original limit of the Universal Character Set). In November 2003 UTF-8 was restricted by RFC 3629 to end at U+10FFFF, in order to match the constraints of the UTF-16 character encoding. This removed all 5- and 6-byte sequences, and about half of the 4-byte sequences.

[edit] Examples

Let us consider how to encode the Euro sign, €.

  1. The Unicode code point for "€" is U+20AC.
  2. According to the scheme table above, this will take three bytes to encode, as it is between U+07FF and U+FFFF.
  3. Hexadecimal 20AC is binary 0010000010101100. The two leading zeros are added because, as the scheme table shows, a three-byte encoding needs exactly sixteen bits from the code point.
  4. Because it is a three-byte encoding, the leading byte starts with three 1s, then a 0 (1110...)
  5. The remaining bits of this byte are taken from the code point (11100010), leaving ...000010101100.
  6. Each of the continuation bytes starts with 10 and takes six bits of the code point (so 10000010, then 10101100).

The three bytes 11100010 10000010 10101100 can be more concisely written in hexadecimal, as E2 82 AC.

The following table summarises this conversion, as well as others with different lengths in UTF-8. The colors indicate how bits from the code point are distributed among the UTF-8 bytes. Additional bits added by the UTF-8 encoding process are shown in black.

Character Binary code point Binary UTF-8 Hexadecimal UTF-8
$ U+0024 0100100 00100100 24
¢ U+00A2 00010100010 11000010 10100010 C2 A2
U+20AC 0010000010101100 11100010 10000010 10101100 E2 82 AC
𤭢 U+24B62 000100100101101100010 11110000 10100100 10101101 10100010 F0 A4 AD A2

[edit] Overlong encodings

The standard specifies that the correct encoding of a codepoint use only the minimum number of bytes required to hold the significant bits of the codepoint. Longer encodings are called overlong and are not valid UTF-8 representations of the codepoint. This rule maintains a one-to-one correspondence between codepoints and their valid encodings, so that there is a unique valid encoding for each codepoint. Allowing multiple encodings would make testing for string equality difficult to define.

In principle, it would be possible to inflate the number of bytes in an encoding by padding the codepoint with leading 0s. To encode the Euro sign € from the above example in four bytes instead of three, it could be padded with leading 0s until it was 21 bits long—000000010000010101100. The leading byte prefix for a four-byte encoding is 11110, and so the complete, overlong encoding is 11110000 10000010 10000010 10101100 (or F0 82 82 AC in hexadecimal).

Although overlong encodings are forbidden in UTF-8, at least one derivative makes use of the form. Modified UTF-8 requires the Unicode codepoint U+0000 (the NUL character) to be encoded in the overlong form 11000000 10000000 (hex C0 80), rather than 00000000 (hex 00). This allows the byte 00 to be used as a string terminator.

[edit] Codepage layout

UTF-8
_0 _1 _2 _3 _4 _5 _6 _7 _8 _9 _A _B _C _D _E _F
 
0_
 
NUL
0000
0
SOH
0001
1
STX
0002
2
ETX
0003
3
EOT
0004
4
ENQ
0005
5
ACK
0006
6
BEL
0007
7
BS
0008
8
HT
0009
9
LF
000A
10
VT
000B
11
FF
000C
12
CR
000D
13
SO
000E
14
SI
000F
15
 
1_
 
DLE
0010
16
DC1
0011
17
DC2
0012
18
DC3
0013
19
DC4
0014
20
NAK
0015
21
SYN
0016
22
ETB
0017
23
CAN
0018
24
EM
0019
25
SUB
001A
26
ESC
001B
27
FS
001C
28
GS
001D
29
RS
001E
30
US
001F
31
 
2_
 
SP
0020
32
!
0021
33
"
0022
34
#
0023
35
$
0024
36
%
0025
37
&
0026
38
'
0027
39
(
0028
40
)
0029
41
*
002A
42
+
002B
43
,
002C
44
-
002D
45
.
002E
46
/
002F
47
 
3_
 
0
0030
48
1
0031
49
2
0032
50
3
0033
51
4
0034
52
5
0035
53
6
0036
54
7
0037
55
8
0038
56
9
0039
57
:
003A
58
;
003B
59
<
003C
60
=
003D
61
>
003E
62
?
003F
63
 
4_
 
@
0040
64
A
0041
65
B
0042
66
C
0043
67
D
0044
68
E
0045
69
F
0046
70
G
0047
71
H
0048
72
I
0049
73
J
004A
74
K
004B
75
L
004C
76
M
004D
77
N
004E
78
O
004F
79
 
5_
 
P
0050
80
Q
0051
81
R
0052
82
S
0053
83
T
0054
84
U
0055
85
V
0056
86
W
0057
87
X
0058
88
Y
0059
89
Z
005A
90
[
005B
91
\
005C
92
]
005D
93
^
005E
94
_
005F
95
 
6_
 
`
0060
96
a
0061
97
b
0062
98
c
0063
99
d
0064
100
e
0065
101
f
0066
102
g
0067
103
h
0068
104
i
0069
105
j
006A
106
k
006B
107
l
006C
108
m
006D
109
n
006E
110
o
006F
111
 
7_
 
p
0070
112
q
0071
113
r
0072
114
s
0073
115
t
0074
116
u
0075
117
v
0076
118
w
0077
119
x
0078
120
y
0079
121
z
007A
122
{
007B
123
|
007C
124
}
007D
125
~
007E
126
DEL
007F
127
 
8_
 

+00
128

+01
129

+02
130

+03
131

+04
132

+05
133

+06
134

+07
135

+08
136

+09
137

+0A
138

+0B
139

+0C
140

+0D
141

+0E
142

+0F
143
 
9_
 

+10
144

+11
145

+12
146

+13
147

+14
148

+15
149

+16
150

+17
151

+18
152

+19
153

+1A
154

+1B
155

+1C
156

+1D
157

+1E
158

+1F
159
 
A_
 

+20
160

+21
161

+22
162

+23
163

+24
164

+25
165

+26
166

+27
167

+28
168

+29
169

+2A
170

+2B
171

+2C
172

+2D
173

+2E
174

+2F
175
 
B_
 

+30
176

+31
177

+32
178

+33
179

+34
180

+35
181

+36
182

+37
183

+38
184

+39
185

+3A
186

+3B
187

+3C
188

+3D
189

+3E
190

+3F
191
 
2-byte
C_
 
2-byte
inval

(0000)
192
2-byte
inval

(0040)
193
Latin-1
0080
194
Latin-1
00C0
195
Latin
Ext-A

0100
196
Latin
Ext-A

0140
197
Latin
Ext-B

0180
198
Latin
Ext-B

01C0
199
Latin
Ext-B

0200
200
IPA
0240
201
IPA
0280
202
Spaci
Modif

02C0
203
Combi
Diacr

0300
204
Combi
Diacr

0340
205
Greek
0380
206
Greek
03C0
207
 
2-byte
D_
 
Cyril
0400
208
Cyril
0440
209
Cyril
0480
210
Cyril
04C0
211
Cyril
0500
212
Armen
0540
213
Hebrew
0580
214
Hebrew
05C0
215
Arabic
0600
216
Arabic
0640
217
Arabic
0680
218
Arabic
06C0
219
Syriac
0700
220
Arabic
0740
221
Thaana
0780
222
N'Ko
07C0
223
 
3-byte
E_
 
Indic
0800*
224
Misc.
1000
225
Symbol
2000
226
Kana
CJK

3000
227
CJK
4000
228
CJK
5000
229
CJK
6000
230
CJK
7000
231
CJK
8000
232
CJK
9000
233
Asian
A000
234
Hangul
B000
235
Hangul
C000
236
Hangul
Surr

D000
237
Priv Use
E000
238
Forms
F000
239
 
4-byte
F_
 
Ancient
Sym,CJK

10000*
240
unall
40000
241
unall
80000
242
Tags
Priv

C0000
243
Priv
Use

100000
244
4-byte
inval

140000
245
4-byte
inval

180000
246
4-byte
inval

1C0000
247
5-byte
inval

200000*
248
5-byte
inval

1000000
249
5-byte
inval

2000000
250
5-byte
inval

3000000
251
6-byte
inval

4000000*
252
6-byte
inval

40000000
253


254


255

Legend: Yellow cells are control characters, blue cells are punctuation, purple cells are digits and green cells are ASCII letters.

Orange cells with a large dot are continuation bytes. The hexadecimal number shown after a "+" plus sign is the value of the 6 bits they add.

White cells are the start bytes for a sequence of multiple bytes, the length shown at the left edge of the row. The text shows the Unicode blocks encoded by sequences starting with this byte, and the hexadecimal code point shown in the cell is the lowest character value encoded using that start byte. When a start byte could form both overlong and valid encodings, the lowest non-overlong-encoded codepoint is shown, marked by an asterisk "*".

Red cells must never appear in a valid UTF-8 sequence. The first two (C0 and C1) could only be used for overlong encoding of basic ASCII characters. The remaining red cells indicate start bytes of sequences that could only encode numbers larger than the 0x10FFFF limit of Unicode. The byte 244 (hex 0xF4) could also encode some values greater than 0x10FFFF; such a sequence is also invalid.

[edit] Invalid byte sequences

Not all sequences of bytes are valid UTF-8. A UTF-8 decoder should be prepared for:

  • the red invalid bytes in the above table
  • an unexpected continuation byte
  • a start byte not followed by enough continuation bytes
  • a sequence that decodes to a value that should use a shorter sequence (an "overlong form")
  • A 4-byte sequence (starting with 0xF4) that decodes to a value greater than U+10FFFF

Many earlier decoders would happily try to decode these. Carefully crafted invalid UTF-8 could make them either skip or create ASCII characters such as NUL, slash, or quotes. Invalid UTF-8 has been used to bypass security validations in high profile products including Microsoft's IIS web server[11] and Apache's Tomcat servlet container.[12]

RFC 3629 states "Implementations of the decoding algorithm MUST protect against decoding invalid sequences."[13] The Unicode Standard requires decoders to "...treat any ill-formed code unit sequence as an error condition. This guarantees that it will neither interpret nor emit an ill-formed code unit sequence."

Many UTF-8 decoders throw exceptions on encountering errors,[14] since such errors suggest the input is not a UTF-8 string at all. This can turn what would otherwise be harmless errors (producing a message such as "no such file") into a denial of service bug. For instance, Python 3.0 would exit immediately if the command line or environment variables contained invalid UTF-8,[15] so it was impossible for any Python program to detect and recover from such an error.

An increasingly popular option is to detect errors with a separate API, and for converters to translate the first byte to a replacement and continue parsing with the next byte. Popular replacements are:

  • The replacement character "�" (U+FFFD)
  • The invalid Unicode code points U+DC80..U+DCFF where the low 8 bits are the byte's value.
  • Interpret the bytes according to ISO-8859-1 or CP1252.

Replacing errors is "lossy": more than one UTF-8 string converts to the same Unicode result. Therefore the original UTF-8 should be stored, and translation should only be used when displaying the text to the user.

[edit] Invalid code points

According to the UTF-8 definition (RFC 3629) the high and low surrogate halves used by UTF-16 (U+D800 through U+DFFF) are not legal Unicode values, and the UTF-8 encoding of them is an invalid byte sequence and thus should be treated as described above.

Whether an actual application should do this with surrogate halves is debatable.[who?] Allowing them allows lossless storage of invalid UTF-16, and allows CESU encoding (described below) to be decoded. There are other code points that are far more important to detect and reject, such as the reversed-BOM U+FFFE, or the C1 controls, caused by improper conversion of CP1252 text or double-encoding of UTF-8. These are invalid in HTML.

[edit] Official name and variants

The official name is "UTF-8". All letters are upper-case, and the name is hyphenated. This spelling is used in all the documents relating to the encoding.

Alternatively, the name "utf-8" may be used by all standards conforming to the Internet Assigned Numbers Authority (IANA) list (which include CSS, HTML, XML, and HTTP headers),[16] as the declaration is case insensitive.[17]

Other descriptions that omit the hyphen or replace it with a space, such as "utf8" or "UTF 8", are not accepted as correct by the governing standards.[18] Despite this, most agents such as browsers can understand them, and so standards intended to describe existing practice (such as HTML5) may effectively require their recognition.

MySQL omits the hyphen in the following query:

SET NAMES 'utf8'

[edit] Derivatives

The following implementations show slight differences from the UTF-8 specification. They are incompatible with the UTF-8 specification.

[edit] CESU-8

Many programs added UTF-8 conversions for UCS-2 data and did not alter this UTF-8 conversion when UCS-2 was replaced with the surrogate-pair using UTF-16. In such programs each half of a UTF-16 surrogate pair is encoded as its own 3-byte UTF-8 encoding, resulting in 6-byte sequences rather than 4 bytes for characters outside the Basic Multilingual Plane. Oracle and MySQL databases use this, as well as Java and Tcl as described below, and probably many Windows programs where the programmers were unaware of the complexities of UTF-16. Although this non-optimal encoding is generally not deliberate, a supposed benefit is that it preserves UTF-16 binary sorting order when CESU-8 is binary sorted.

[edit] Modified UTF-8

In Modified UTF-8,[19] the null character (U+0000) is encoded as 0xC0,0x80; this is not valid UTF-8[20] because it is not the shortest possible representation. Modified UTF-8 strings never contain any actual null bytes but can contain all Unicode code points including U+0000,[21] which allows such strings (with a null byte appended) to be processed by traditional null-terminated string functions.

All known Modified UTF-8 implementations also treat the surrogate pairs as in CESU-8.

In normal usage, the Java programming language supports standard UTF-8 when reading and writing strings through InputStreamReader and OutputStreamWriter. However it uses Modified UTF-8 for object serialization,[22] for the Java Native Interface,[23] and for embedding constant strings in class files.[24] Tcl also uses the same modified UTF-8[25] as Java for internal representation of Unicode data, but uses strict CESU-8 for external data.

[edit] Extending from 31 bit to 36 bit range

Extending the accepted input pattern from 6 bytes to 7 bytes would allow over 70 billion code points to be encoded;[26] however, this would require an initial byte value of 0xFE to be accepted as a 7-byte sequence indicator (see under Advantages in section "Compared to single-byte encodings").

[edit] Byte order mark

Many Windows programs (including Windows Notepad) add the bytes 0xEF, 0xBB, 0xBF at the start of any document saved as UTF-8. This is the UTF-8 encoding of the Unicode byte order mark (BOM), and is commonly referred to as a UTF-8 BOM, even though it is not relevant to byte order. The BOM can also appear if another encoding with a BOM is translated to UTF-8 without stripping it. Older text editors may display the BOM as "" at the start of the document.

The Unicode Standard neither requires nor recommends the use of the BOM for UTF-8.[27] The presence of the UTF-8 BOM may cause interoperability problems with existing software that could otherwise handle UTF-8; for example:

  • Programming language parsers not explicitly designed for UTF-8 can often handle UTF-8 in string constants and comments, but cannot parse the BOM at the start of the file.
  • Programs that identify file types by leading characters may fail to identify the file if a BOM is present even if the user of the file could skip the BOM. An example is the Unix shebang syntax. Another example is Internet Explorer which will render pages in standards mode only when it starts with a document type declaration.

If compatibility with existing programs is not important, the BOM could be used to identify UTF-8 encoding. Because checking if text is valid UTF-8 is very reliable (the majority of random byte sequences are not valid UTF-8) such use should not be necessary. Programs that insert information at the start of a file will break this identification (one example is offline browsers that add the originating URL to the start of the file).

Unofficially, UTF-8-BOM or UTF-8-NOBOM are sometimes used to refer to text files which contain or lack a BOM. In Japan especially, "UTF-8 encoding without BOM" is sometimes called "UTF-8N".[citation needed]

[edit] Advantages and disadvantages

[edit] General

[edit] Advantages

  • The ASCII characters are represented by themselves as single bytes that do not appear anywhere else, which makes UTF-8 work with the majority of existing APIs that take bytes strings but only treat a small number of ASCII codes specially. This removes the need to write a new Unicode version of every API, and makes it much easier to convert existing systems to UTF-8 than any other Unicode encoding.
  • UTF-8 is the only encoding for XML entities that does not require a BOM or an indication of the encoding.[28]
  • UTF-8 and UTF-16 are the standard encodings for Unicode text in HTML documents, with UTF-8 as the preferred and most used encoding.
  • UTF-8 strings can be fairly reliably recognized as such by a simple heuristic algorithm.[29] The probability of a random string of bytes which is not pure ASCII being valid UTF-8 is 3.9% for a two-byte sequence,[30] and decreases exponentially for longer sequences. ISO/IEC 8859-1 is even less likely to be mis-recognized as UTF-8: the only non-ASCII characters in it would have to be in sequences starting with either an accented letter or the multiplication symbol and ending with a symbol. This is an advantage that most other encodings do not have, causing errors (mojibake) if the receiving application isn't told and can't guess the correct encoding. Even word-based UTF-16 can be mistaken for byte encodings (like in the "bush hid the facts" bug).
  • Sorting of UTF-8 strings as arrays of unsigned bytes will produce the same results as sorting them based on Unicode code points.
  • Other byte-based encodings can pass through the same API. This means, however, that the encoding must be identified. Because the other encodings are unlikely to be valid UTF-8, a reliable way to implement this is to assume UTF-8 and switch to a legacy encoding only if several invalid UTF-8 byte sequences are encountered.

[edit] Disadvantages

  • A UTF-8 parser that is not compliant with current versions of the standard might accept a number of different pseudo-UTF-8 representations and convert them to the same Unicode output. This provides a way for information to leak past validation routines designed to process data in its eight-bit representation.[31]

[edit] Compared to single-byte encodings

[edit] Advantages

  • UTF-8 can encode any Unicode character, avoiding the need to figure out and set a "code page" or otherwise indicate what character set is in use, and allowing output in multiple scripts at the same time. For many scripts there have been more than one single-byte encoding in usage, so even knowing the script was insufficient information to display it correctly.
  • The bytes 0xFE and 0xFF do not appear, so a valid UTF-8 stream never matches the UTF-16 byte order mark and thus cannot be confused with it. The absence of 0xFF (0377) also eliminates the need to escape this byte in Telnet (and FTP control connection).

[edit] Disadvantages

  • UTF-8 encoded text is larger than the appropriate single-byte encoding except for plain ASCII characters. In the case of scripts which used 8-bit character sets with non-Latin scripts encoded in the upper half (such as most Cyrillic and Greek alphabet code pages), characters in UTF-8 will be double the size. For some scripts such as Thai and Hindi's Devanagari, characters will be triple the size (this has caused objections in India and other countries).
  • It is possible in UTF-8 (or any other multi-byte encoding) to split or truncate a string in the middle of a character, which may result in an invalid string. This will not happen in correct handling of UTF-8.
  • If the code points are all the same size, measurements of a fixed number of them is easy. Due to ASCII-era documentation where "character" is used as a synonym for "byte" this is often considered important. However, by measuring string positions using bytes instead of "characters" most algorithms can be easily and efficiently adapted for UTF-8[citation needed].
  • Some software such as text editors will refuse to correctly display or interpret UTF-8 unless it starts with a Byte Order Mark. This has the effect of making it impossible to use UTF-8 with any older software than can handle ASCII-like encodings but cannot handle the byte order mark. This is considered an incorrect implementation of the text editor, not the older software.

[edit] Compared to other multi-byte encodings

[edit] Advantages

  • UTF-8 uses the codes 0–127 only for the ASCII characters. This means that UTF-8 is an ASCII extension and can with limited change be supported by software that supports an ASCII extension and handles non-ASCII characters as free text.
  • UTF-8 can encode any Unicode character. Files in different scripts can be displayed correctly without having to choose the correct code page or font. For instance Chinese and Arabic can be supported (in the same text) without special codes inserted or manual settings to switch the encoding.
  • UTF-8 is "self-synchronizing": character boundaries are easily found when searching either forwards or backwards. If bytes are lost due to error or corruption, one can always locate the beginning of the next character and thus limit the damage. Many multi-byte encodings are much harder to resynchronize.
  • Any byte oriented string searching algorithm can be used with UTF-8 data, since the sequence of bytes for a character cannot occur anywhere else. Some older variable-length encodings (such as Shift JIS) did not have this property and thus made string-matching algorithms rather complicated. In Shift JIS the end byte of a character and the first byte of the next character could look like another legal character, something that can't happen in UTF-8.
  • Efficient to encode using simple bit operations. UTF-8 does not require slower mathematical operations such as multiplication or division (unlike the obsolete UTF-1 encoding).

[edit] Disadvantages

  • For certain scripts UTF-8 will take more space than an older multi-byte encoding. East Asian scripts generally have two bytes per character in their multi-byte encodings yet take three bytes per character in UTF-8.


[edit] Compared to UTF-16

[edit] Advantages

  • A text byte stream cannot be losslessly converted to UTF-16, due to the possible presence of errors in the byte stream encoding. This causes unexpected and often severe problems attempting to use existing data in a system that uses UTF-16 as an internal encoding. Results are security bugs, DoS if bad encoding throws an exception, and data loss when different byte streams convert to the same UTF-16. Due to the ASCII compatibility and high degree of pattern recognition in UTF-8, random byte streams can be passed losslessly through a system using it, as interpretation can be deferred until display.
  • Converting to UTF-16 while maintaining compatibility with existing programs (such as was done with Windows) requires every API and data structure that takes a string to be duplicated. Invalid encodings make the duplicated APIs not exactly map to each other, often making it impossible to do some action with one of them.
  • Characters outside the basic multilingual plane are not a special case. UTF-16 is often mistaken to be the obsolete constant-length UCS-2 encoding, leading to code that works for most text but suddenly fails for non-BMP characters.[32]
  • Text encoded in UTF-8 is often smaller than (or the same size as) the same text encoded in UTF-16.
    • This is always true for text using only code points below U+0800 (which includes all modern European scripts), as each code point's UTF-8 encoding is one or two bytes then.
    • Even if text contains code points between U+0800 and U+FFFF, it might contain so many code points below U+0080 (which UTF-8 encodes in one byte) that the UTF-8 encoding is still smaller. As HTML markup and line terminators are code points below U+0080, most HTML source is smaller if encoded in UTF-8 even for Asian scripts.
    • Non-BMP characters (U+10000 and above) are encoded in UTF-8 in four bytes, the same size as in UTF-16.
  • Most communication and storage was designed for a stream of bytes. A UTF-16 string must use a pair of bytes for each code unit:
    • The order of those two bytes becomes an issue and must be specified in the UTF-16 protocol, such as with a byte order mark.
    • If an odd number of bytes is missing from UTF-16, the whole rest of the string will be meaningless text. Any bytes missing from UTF-8 will still allow the text to be recovered accurately starting with the next character after the missing bytes. If any partial character is removed the corruption is always recognizable.

[edit] Disadvantages

  • Characters U+0800 through U+FFFF use three bytes in UTF-8, but only two in UTF-16. As a result, text in (for example) Chinese, Japanese or Hindi could take more space in UTF-8 if there are more of these characters than there are ASCII characters. This happens for pure text,[33] but rarely for HTML documents. For example, both the Japanese UTF-8 and the Hindi Unicode articles on Wikipedia take more space in UTF-16 than in UTF-8 .[34]

[edit] See also

[edit] References

  1. ^ [|The Unicode Consortium]. "Chapter 2. General Structure". The Unicode Standard (6.0 ed.). Mountain View, California, USA: The Unicode Consortium. ISBN 978-1-936213-01-6. http://www.unicode.org/versions/Unicode6.0.0/. . RFC 3629 also refers to UTF-8 as "UCS transformation format". Also commonly known as "Unicode Transformation Format".
  2. ^ Mark Davis (28 January 2010). "Unicode nearing 50% of the web". Official Google Blog. Google. http://googleblog.blogspot.com/2010/01/unicode-nearing-50-of-web.html. Retrieved 5 December 2010. 
  3. ^ "UTF-8 Usage Statistics". BuiltWith. http://trends.builtwith.com/encoding/UTF-8. Retrieved 2011-03-28. 
  4. ^ "Usage of character encodings for websites". W3Techs. http://w3techs.com/technologies/overview/character_encoding/all. Retrieved 2010-03-30. 
  5. ^ Alvestrand, H. (1998). "IETF Policy on Character Sets and Languages". RFC 2277. Internet Engineering Task Force. 
  6. ^ "Using International Characters in Internet Mail". Internet Mail Consortium. August 1, 1998. http://www.imc.org/mail-i18n.html. Retrieved 2007-11-08. 
  7. ^ Not all of the 1,112,064 possible code points have been assigned characters; many are reserved for future use, and some are reserved for private use, while still others are specified as permanently undefined.
  8. ^ More precisely, the number of bytes used to encode a character at a given code point is a monotonically increasing function of the numerical value of the code point.
  9. ^ Internet Assigned Numbers Authority (4 November 2010). "CHARACTER SETS". IANA. http://www.iana.org/assignments/character-sets. Retrieved 5 December 2010. 
  10. ^ Pike, Rob (2003-04-03). "UTF-8 history". http://www.cl.cam.ac.uk/~mgk25/ucs/utf-8-history.txt. 
  11. ^ Marin, Marvin (2000-10-17). "Web Server Folder Traversal MS00-078". http://www.sans.org/resources/malwarefaq/wnt-unicode.php. 
  12. ^ "National Vulnerability Database - Summary for CVE-2008-2938". http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2008-2938. 
  13. ^ Yergeau, F. (2003). "UTF-8, a transformation format of ISO 10646". RFC 3629. Internet Engineering Task Force 
  14. ^ Examples: UTF8 (Java Class Library API) or java.nio.charset.CharsetDecoder.decode
  15. ^ "Non-decodable Bytes in System Character Interfaces". http://www.python.org/dev/peps/pep-0383/. 
  16. ^ W3C: Setting the HTTP charset parameter notes that the IANA list is used for HTTP
  17. ^ Internet Assigned Numbers Authority Character Sets
  18. ^ RFC 3629 UTF-8 see chapter 8. MIME registration, first paragraph
  19. ^ "Java SE 6 documentation for Interface java.io.DataInput, subsection on Modified UTF-8". Sun Microsystems. 2008. http://java.sun.com/javase/6/docs/api/java/io/DataInput.html#modified-utf-8. Retrieved 2009-05-22. 
  20. ^ "[...] the overlong UTF-8 sequence C0 80 [...]", "[...] the illegal two-octet sequence C0 80 [...]""Request for Comments 3629: "UTF-8, a transformation format of ISO 10646"". 2003. http://www.apps.ietf.org/rfc/rfc3629.html#page-5. Retrieved 2009-05-22. 
  21. ^ "[...] Java virtual machine UTF-8 strings never have embedded nulls.""The Java Virtual Machine Specification, 2nd Edition, section 4.4.7: "The CONSTANT_Utf8_info Structure"". Sun Microsystems. 1999. http://java.sun.com/docs/books/jvms/second_edition/html/ClassFile.doc.html#7963. Retrieved 2009-05-24. 
  22. ^ "[...] encoded in modified UTF-8.""Java Object Serialization Specification, chapter 6: Object Serialization Stream Protocol, section 2: Stream Elements". Sun Microsystems. 2005. http://java.sun.com/javase/6/docs/platform/serialization/spec/protocol.html#8299. Retrieved 2009-05-22. 
  23. ^ "The JNI uses modified UTF-8 strings to represent various string types.""Java Native Interface Specification, chapter 3: JNI Types and Data Structures, section: Modified UTF-8 Strings". Sun Microsystems. 2003. http://java.sun.com/j2se/1.5.0/docs/guide/jni/spec/types.html#wp16542. Retrieved 2009-05-22. 
  24. ^ "[...] differences between this format and the "standard" UTF-8 format.""The Java Virtual Machine Specification, 2nd Edition, section 4.4.7: "The CONSTANT_Utf8_info Structure"". Sun Microsystems. 1999. http://java.sun.com/docs/books/jvms/second_edition/html/ClassFile.doc.html#7963. Retrieved 2009-05-23. 
  25. ^ "In orthodox UTF-8, a NUL byte(\x00) is represented by a NUL byte. [...] But [...] we [...] want NUL bytes inside [...] strings [...]""Tcler's Wiki: UTF-8 bit by bit (Revision 6)". 2009-04-25. http://wiki.tcl.tk/_/revision?N=1211&V=6. Retrieved 2009-05-22. 
  26. ^ Computational Methods in Linguistic Research, 2004
  27. ^ The Unicode Standard - Chapter 2, see chapter 2.6 page 30 bottom.
  28. ^ W3.org
  29. ^ W3 FAQ: Multilingual Forms: a Perl regular expression to validate a UTF-8 string)
  30. ^ There are 256 × 256 − 128 × 128 not-pure-ASCII two-byte sequences, and of those, only 1920 encode valid UTF-8 characters (the range U+0080 to U+07FF), so the proportion of valid not-pure-ASCII two-byte sequences is 3.9%. Note that this assumes that control characters pass as ASCII; without the control characters, the percentage proportions drop somewhat).
  31. ^ Tools.ietf.org
  32. ^ "Should UTF-16 be considered harmful?". Stackoverflow.com. http://stackoverflow.com/questions/1049947/should-utf-16-be-considered-harmful. Retrieved 2010-09-13. 
  33. ^ Although the difference may not be great: the 2010-11-22 version of hi:यूनिकोड (Unicode in Hindi), when the pure text was pasted to Notepad, generated 19 KB when saved as UTF-16 and 22 KB when saved as UTF-8.
  34. ^ The 2010-10-27 version of ja:UTF-8 generated 169 KB when converted with Notepad to UTF-16, and only 101 KB when converted back to UTF-8. The 2010-11-22 version of hi:यूनिकोड (Unicode in Hindi) required 119 KB in UTF-16 and 76 KB in UTF-8.

[edit] External links

There are several current definitions of UTF-8 in various standards documents:

  • RFC 3629 / STD 63 (2003), which establishes UTF-8 as a standard Internet protocol element
  • The Unicode Standard, Version 6.0, §3.9 D92, §3.10 D95 (2011)
  • ISO/IEC 10646:2003 Annex D (2003)

They supersede the definitions given in the following obsolete works:

  • ISO/IEC 10646-1:1993 Amendment 2 / Annex R (1996)
  • The Unicode Standard, Version 5.0, §3.9 D92, §3.10 D95 (2007)
  • The Unicode Standard, Version 4.0, §3.9–§3.10 (2003)
  • The Unicode Standard, Version 2.0, Appendix A (1996)
  • RFC 2044 (1996)
  • RFC 2279 (1998)
  • The Unicode Standard, Version 3.0, §2.3 (2000) plus Corrigendum #1 : UTF-8 Shortest Form (2000)
  • Unicode Standard Annex #27: Unicode 3.1 (2001)

They are all the same in their general mechanics, with the main differences being on issues such as allowed range of code point values and safe handling of invalid input.

Personal tools
Namespaces

Variants
Actions
Navigation
Interaction
Toolbox
Print/export
Languages
================================================ FILE: unicode.go ================================================ package chardet import ( "bytes" ) var ( utf16beBom = []byte{0xFE, 0xFF} utf16leBom = []byte{0xFF, 0xFE} utf32beBom = []byte{0x00, 0x00, 0xFE, 0xFF} utf32leBom = []byte{0xFF, 0xFE, 0x00, 0x00} ) type recognizerUtf16be struct { } func newRecognizer_utf16be() *recognizerUtf16be { return &recognizerUtf16be{} } func (*recognizerUtf16be) Match(input *recognizerInput) (output recognizerOutput) { output = recognizerOutput{ Charset: "UTF-16BE", } if bytes.HasPrefix(input.raw, utf16beBom) { output.Confidence = 100 } return } type recognizerUtf16le struct { } func newRecognizer_utf16le() *recognizerUtf16le { return &recognizerUtf16le{} } func (*recognizerUtf16le) Match(input *recognizerInput) (output recognizerOutput) { output = recognizerOutput{ Charset: "UTF-16LE", } if bytes.HasPrefix(input.raw, utf16leBom) && !bytes.HasPrefix(input.raw, utf32leBom) { output.Confidence = 100 } return } type recognizerUtf32 struct { name string bom []byte decodeChar func(input []byte) uint32 } func decodeUtf32be(input []byte) uint32 { return uint32(input[0])<<24 | uint32(input[1])<<16 | uint32(input[2])<<8 | uint32(input[3]) } func decodeUtf32le(input []byte) uint32 { return uint32(input[3])<<24 | uint32(input[2])<<16 | uint32(input[1])<<8 | uint32(input[0]) } func newRecognizer_utf32be() *recognizerUtf32 { return &recognizerUtf32{ "UTF-32BE", utf32beBom, decodeUtf32be, } } func newRecognizer_utf32le() *recognizerUtf32 { return &recognizerUtf32{ "UTF-32LE", utf32leBom, decodeUtf32le, } } func (r *recognizerUtf32) Match(input *recognizerInput) (output recognizerOutput) { output = recognizerOutput{ Charset: r.name, } hasBom := bytes.HasPrefix(input.raw, r.bom) var numValid, numInvalid uint32 for b := input.raw; len(b) >= 4; b = b[4:] { if c := r.decodeChar(b); c >= 0x10FFFF || (c >= 0xD800 && c <= 0xDFFF) { numInvalid++ } else { numValid++ } } if hasBom && numInvalid == 0 { output.Confidence = 100 } else if hasBom && numValid > numInvalid*10 { output.Confidence = 80 } else if numValid > 3 && numInvalid == 0 { output.Confidence = 100 } else if numValid > 0 && numInvalid == 0 { output.Confidence = 80 } else if numValid > numInvalid*10 { output.Confidence = 25 } return } ================================================ FILE: utf8.go ================================================ package chardet import ( "bytes" ) var utf8Bom = []byte{0xEF, 0xBB, 0xBF} type recognizerUtf8 struct { } func newRecognizer_utf8() *recognizerUtf8 { return &recognizerUtf8{} } func (*recognizerUtf8) Match(input *recognizerInput) (output recognizerOutput) { output = recognizerOutput{ Charset: "UTF-8", } hasBom := bytes.HasPrefix(input.raw, utf8Bom) inputLen := len(input.raw) var numValid, numInvalid uint32 var trailBytes uint8 for i := 0; i < inputLen; i++ { c := input.raw[i] if c&0x80 == 0 { continue } if c&0xE0 == 0xC0 { trailBytes = 1 } else if c&0xF0 == 0xE0 { trailBytes = 2 } else if c&0xF8 == 0xF0 { trailBytes = 3 } else { numInvalid++ if numInvalid > 5 { break } trailBytes = 0 } for i++; i < inputLen; i++ { c = input.raw[i] if c&0xC0 != 0x80 { numInvalid++ break } if trailBytes--; trailBytes == 0 { numValid++ break } } } if hasBom && numInvalid == 0 { output.Confidence = 100 } else if hasBom && numValid > numInvalid*10 { output.Confidence = 80 } else if numValid > 3 && numInvalid == 0 { output.Confidence = 100 } else if numValid > 0 && numInvalid == 0 { output.Confidence = 80 } else if numValid == 0 && numInvalid == 0 { // Plain ASCII output.Confidence = 10 } else if numValid > numInvalid*10 { output.Confidence = 25 } return }