| OLD | NEW |
| 1 /* | 1 /* |
| 2 ******************************************************************************* | 2 ******************************************************************************* |
| 3 * Copyright (C) 2013, International Business Machines | 3 * Copyright (C) 2013-2014, International Business Machines |
| 4 * Corporation and others. All Rights Reserved. | 4 * Corporation and others. All Rights Reserved. |
| 5 ******************************************************************************* | 5 ******************************************************************************* |
| 6 * file name: uscript_props.cpp | 6 * file name: uscript_props.cpp |
| 7 * encoding: US-ASCII | 7 * encoding: US-ASCII |
| 8 * tab size: 8 (not used) | 8 * tab size: 8 (not used) |
| 9 * indentation:4 | 9 * indentation:4 |
| 10 * | 10 * |
| 11 * created on: 2013feb16 | 11 * created on: 2013feb16 |
| 12 * created by: Markus W. Scherer | 12 * created by: Markus W. Scherer |
| 13 */ | 13 */ |
| 14 | 14 |
| 15 #include "unicode/utypes.h" | 15 #include "unicode/utypes.h" |
| 16 #include "unicode/unistr.h" | 16 #include "unicode/unistr.h" |
| 17 #include "unicode/uscript.h" | 17 #include "unicode/uscript.h" |
| 18 #include "unicode/utf16.h" | 18 #include "unicode/utf16.h" |
| 19 #include "ustr_imp.h" | 19 #include "ustr_imp.h" |
| 20 | 20 #include "cmemory.h" |
| 21 #define LENGTHOF(array) (int32_t)(sizeof(array)/sizeof((array)[0])) | |
| 22 | 21 |
| 23 namespace { | 22 namespace { |
| 24 | 23 |
| 25 // Script metadata (script properties). | 24 // Script metadata (script properties). |
| 26 // See http://unicode.org/cldr/trac/browser/trunk/common/properties/scriptMetada
ta.txt | 25 // See http://unicode.org/cldr/trac/browser/trunk/common/properties/scriptMetada
ta.txt |
| 27 | 26 |
| 28 // 0 = NOT_ENCODED, no sample character, default false script properties. | 27 // 0 = NOT_ENCODED, no sample character, default false script properties. |
| 29 // Bits 20.. 0: sample character | 28 // Bits 20.. 0: sample character |
| 30 | 29 |
| 31 // Bits 23..21: usage | 30 // Bits 23..21: usage |
| 32 const int32_t UNKNOWN = 1 << 21; | 31 const int32_t UNKNOWN = 1 << 21; |
| 33 const int32_t EXCLUSION = 2 << 21; | 32 const int32_t EXCLUSION = 2 << 21; |
| 34 const int32_t LIMITED_USE = 3 << 21; | 33 const int32_t LIMITED_USE = 3 << 21; |
| 35 const int32_t ASPIRATIONAL = 4 << 21; | 34 const int32_t ASPIRATIONAL = 4 << 21; |
| 36 const int32_t RECOMMENDED = 5 << 21; | 35 const int32_t RECOMMENDED = 5 << 21; |
| 37 | 36 |
| 38 // Bits 31..24: Single-bit flags | 37 // Bits 31..24: Single-bit flags |
| 39 const int32_t RTL = 1 << 24; | 38 const int32_t RTL = 1 << 24; |
| 40 const int32_t LB_LETTERS = 1 << 25; | 39 const int32_t LB_LETTERS = 1 << 25; |
| 41 const int32_t CASED = 1 << 26; | 40 const int32_t CASED = 1 << 26; |
| 42 | 41 |
| 43 const int32_t SCRIPT_PROPS[] = { | 42 const int32_t SCRIPT_PROPS[] = { |
| 44 // Begin copy-paste output from | 43 // Begin copy-paste output from |
| 45 // tools/trunk/unicode/py/parsescriptmetadata.py | 44 // tools/trunk/unicode/py/parsescriptmetadata.py |
| 46 0x0040 | UNKNOWN, // Zyyy | 45 0x0040 | RECOMMENDED, // Zyyy |
| 47 0x0308 | UNKNOWN, // Zinh | 46 0x0308 | UNKNOWN, // Zinh |
| 48 0x0628 | RECOMMENDED | RTL, // Arab | 47 0x0628 | RECOMMENDED | RTL, // Arab |
| 49 0x0531 | RECOMMENDED | CASED, // Armn | 48 0x0531 | RECOMMENDED | CASED, // Armn |
| 50 0x0995 | RECOMMENDED, // Beng | 49 0x0995 | RECOMMENDED, // Beng |
| 51 0x3105 | RECOMMENDED | LB_LETTERS, // Bopo | 50 0x3105 | RECOMMENDED | LB_LETTERS, // Bopo |
| 52 0x13C4 | LIMITED_USE, // Cher | 51 0x13C4 | LIMITED_USE, // Cher |
| 53 0x03E2 | EXCLUSION | CASED, // Copt | 52 0x03E2 | EXCLUSION | CASED, // Copt |
| 54 0x042F | RECOMMENDED | CASED, // Cyrl | 53 0x042F | RECOMMENDED | CASED, // Cyrl |
| 55 0x10414 | EXCLUSION | CASED, // Dsrt | 54 0x10414 | EXCLUSION | CASED, // Dsrt |
| 56 0x0905 | RECOMMENDED, // Deva | 55 0x0905 | RECOMMENDED, // Deva |
| (...skipping 41 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
| 98 0x1950 | LIMITED_USE | LB_LETTERS, // Tale | 97 0x1950 | LIMITED_USE | LB_LETTERS, // Tale |
| 99 0x10380 | EXCLUSION, // Ugar | 98 0x10380 | EXCLUSION, // Ugar |
| 100 0, | 99 0, |
| 101 0x1A00 | EXCLUSION, // Bugi | 100 0x1A00 | EXCLUSION, // Bugi |
| 102 0x2C00 | EXCLUSION | CASED, // Glag | 101 0x2C00 | EXCLUSION | CASED, // Glag |
| 103 0x10A00 | EXCLUSION | RTL, // Khar | 102 0x10A00 | EXCLUSION | RTL, // Khar |
| 104 0xA800 | LIMITED_USE, // Sylo | 103 0xA800 | LIMITED_USE, // Sylo |
| 105 0x1980 | LIMITED_USE | LB_LETTERS, // Talu | 104 0x1980 | LIMITED_USE | LB_LETTERS, // Talu |
| 106 0x2D30 | ASPIRATIONAL, // Tfng | 105 0x2D30 | ASPIRATIONAL, // Tfng |
| 107 0x103A0 | EXCLUSION, // Xpeo | 106 0x103A0 | EXCLUSION, // Xpeo |
| 108 0x1B05 | LIMITED_USE | LB_LETTERS, // Bali | 107 0x1B05 | LIMITED_USE, // Bali |
| 109 0x1BC0 | LIMITED_USE, // Batk | 108 0x1BC0 | LIMITED_USE, // Batk |
| 110 0, | 109 0, |
| 111 0x11005 | EXCLUSION, // Brah | 110 0x11005 | EXCLUSION, // Brah |
| 112 0xAA00 | LIMITED_USE, // Cham | 111 0xAA00 | LIMITED_USE, // Cham |
| 113 0, | 112 0, |
| 114 0, | 113 0, |
| 115 0, | 114 0, |
| 116 0, | 115 0, |
| 117 0x13153 | EXCLUSION, // Egyp | 116 0x13153 | EXCLUSION, // Egyp |
| 118 0, | 117 0, |
| 119 0x5B57 | RECOMMENDED | LB_LETTERS, // Hans | 118 0x5B57 | RECOMMENDED | LB_LETTERS, // Hans |
| 120 0x5B57 | RECOMMENDED | LB_LETTERS, // Hant | 119 0x5B57 | RECOMMENDED | LB_LETTERS, // Hant |
| 120 0x16B1C | EXCLUSION, // Hmng |
| 121 0, | 121 0, |
| 122 0, | 122 0, |
| 123 0, | 123 0xA984 | LIMITED_USE, // Java |
| 124 0xA984 | LIMITED_USE | LB_LETTERS, // Java | |
| 125 0xA90A | LIMITED_USE, // Kali | 124 0xA90A | LIMITED_USE, // Kali |
| 126 0, | 125 0, |
| 127 0, | 126 0, |
| 128 0x1C00 | LIMITED_USE, // Lepc | 127 0x1C00 | LIMITED_USE, // Lepc |
| 129 0, | 128 0x10647 | EXCLUSION, // Lina |
| 130 0x0840 | LIMITED_USE | RTL, // Mand | 129 0x0840 | LIMITED_USE | RTL, // Mand |
| 131 0, | 130 0, |
| 132 0x10980 | EXCLUSION | RTL, // Mero | 131 0x10980 | EXCLUSION | RTL, // Mero |
| 133 0x07CA | LIMITED_USE | RTL, // Nkoo | 132 0x07CA | LIMITED_USE | RTL, // Nkoo |
| 134 0x10C00 | EXCLUSION | RTL, // Orkh | 133 0x10C00 | EXCLUSION | RTL, // Orkh |
| 135 0, | 134 0x1036B | EXCLUSION, // Perm |
| 136 0xA840 | EXCLUSION, // Phag | 135 0xA840 | EXCLUSION, // Phag |
| 137 0x10900 | EXCLUSION | RTL, // Phnx | 136 0x10900 | EXCLUSION | RTL, // Phnx |
| 138 0x16F00 | ASPIRATIONAL, // Plrd | 137 0x16F00 | ASPIRATIONAL, // Plrd |
| 139 0, | 138 0, |
| 140 0, | 139 0, |
| 141 0, | 140 0, |
| 142 0, | 141 0, |
| 143 0, | 142 0, |
| 144 0, | 143 0, |
| 145 0xA549 | LIMITED_USE, // Vaii | 144 0xA549 | LIMITED_USE, // Vaii |
| (...skipping 11 matching lines...) Expand all Loading... |
| 157 0xA882 | LIMITED_USE, // Saur | 156 0xA882 | LIMITED_USE, // Saur |
| 158 0, | 157 0, |
| 159 0x1B83 | LIMITED_USE, // Sund | 158 0x1B83 | LIMITED_USE, // Sund |
| 160 0, | 159 0, |
| 161 0xABC0 | LIMITED_USE, // Mtei | 160 0xABC0 | LIMITED_USE, // Mtei |
| 162 0x10840 | EXCLUSION | RTL, // Armi | 161 0x10840 | EXCLUSION | RTL, // Armi |
| 163 0x10B00 | EXCLUSION | RTL, // Avst | 162 0x10B00 | EXCLUSION | RTL, // Avst |
| 164 0x11103 | LIMITED_USE, // Cakm | 163 0x11103 | LIMITED_USE, // Cakm |
| 165 0xAC00 | RECOMMENDED, // Kore | 164 0xAC00 | RECOMMENDED, // Kore |
| 166 0x11083 | EXCLUSION, // Kthi | 165 0x11083 | EXCLUSION, // Kthi |
| 167 0, | 166 0x10AD8 | EXCLUSION | RTL, // Mani |
| 168 0x10B60 | EXCLUSION | RTL, // Phli | 167 0x10B60 | EXCLUSION | RTL, // Phli |
| 169 0, | 168 0x10B8F | EXCLUSION | RTL, // Phlp |
| 170 0, | 169 0, |
| 171 0x10B40 | EXCLUSION | RTL, // Prti | 170 0x10B40 | EXCLUSION | RTL, // Prti |
| 172 0x0800 | EXCLUSION | RTL, // Samr | 171 0x0800 | EXCLUSION | RTL, // Samr |
| 173 0xAA80 | LIMITED_USE | LB_LETTERS, // Tavt | 172 0xAA80 | LIMITED_USE | LB_LETTERS, // Tavt |
| 174 0, | 173 0, |
| 175 0, | 174 0, |
| 176 0xA6A0 | LIMITED_USE, // Bamu | 175 0xA6A0 | LIMITED_USE, // Bamu |
| 177 0xA4D0 | LIMITED_USE, // Lisu | 176 0xA4D0 | LIMITED_USE, // Lisu |
| 178 0, | 177 0, |
| 179 0x10A60 | EXCLUSION | RTL, // Sarb | 178 0x10A60 | EXCLUSION | RTL, // Sarb |
| 179 0x16AE6 | EXCLUSION, // Bass |
| 180 0x1BC20 | EXCLUSION, // Dupl |
| 181 0x10500 | EXCLUSION, // Elba |
| 182 0x11315 | EXCLUSION, // Gran |
| 180 0, | 183 0, |
| 181 0, | 184 0, |
| 185 0x1E802 | EXCLUSION | RTL, // Mend |
| 186 0x109A0 | EXCLUSION | RTL, // Merc |
| 187 0x10A95 | EXCLUSION | RTL, // Narb |
| 188 0x10896 | EXCLUSION | RTL, // Nbat |
| 189 0x10873 | EXCLUSION | RTL, // Palm |
| 190 0x112BE | EXCLUSION, // Sind |
| 191 0x118B4 | EXCLUSION | CASED, // Wara |
| 182 0, | 192 0, |
| 183 0, | 193 0, |
| 184 0, | 194 0x16A4F | EXCLUSION, // Mroo |
| 185 0, | |
| 186 0, | |
| 187 0x109A0 | EXCLUSION | RTL, // Merc | |
| 188 0, | |
| 189 0, | |
| 190 0, | |
| 191 0, | |
| 192 0, | |
| 193 0, | |
| 194 0, | |
| 195 0, | |
| 196 0, | 195 0, |
| 197 0x11183 | EXCLUSION, // Shrd | 196 0x11183 | EXCLUSION, // Shrd |
| 198 0x110D0 | EXCLUSION, // Sora | 197 0x110D0 | EXCLUSION, // Sora |
| 199 0x11680 | EXCLUSION, // Takr | 198 0x11680 | EXCLUSION, // Takr |
| 200 0, | 199 0, |
| 201 0, | 200 0, |
| 202 0, | 201 0, |
| 202 0x11208 | EXCLUSION, // Khoj |
| 203 0x11484 | EXCLUSION, // Tirh |
| 204 0x10537 | EXCLUSION, // Aghb |
| 205 0x11152 | EXCLUSION, // Mahj |
| 203 0, | 206 0, |
| 204 0, | 207 0, |
| 208 0x1160E | EXCLUSION, // Modi |
| 205 0, | 209 0, |
| 206 0, | 210 0x11AC0 | EXCLUSION, // Pauc |
| 211 0x1158E | EXCLUSION, // Sidd |
| 207 // End copy-paste from parsescriptmetadata.py | 212 // End copy-paste from parsescriptmetadata.py |
| 208 }; | 213 }; |
| 209 | 214 |
| 210 int32_t getScriptProps(UScriptCode script) { | 215 int32_t getScriptProps(UScriptCode script) { |
| 211 if (0 <= script && script < LENGTHOF(SCRIPT_PROPS)) { | 216 if (0 <= script && script < UPRV_LENGTHOF(SCRIPT_PROPS)) { |
| 212 return SCRIPT_PROPS[script]; | 217 return SCRIPT_PROPS[script]; |
| 213 } else { | 218 } else { |
| 214 return 0; | 219 return 0; |
| 215 } | 220 } |
| 216 } | 221 } |
| 217 | 222 |
| 218 } // namespace | 223 } // namespace |
| 219 | 224 |
| 220 U_CAPI int32_t U_EXPORT2 | 225 U_CAPI int32_t U_EXPORT2 |
| 221 uscript_getSampleString(UScriptCode script, UChar *dest, int32_t capacity, UErro
rCode *pErrorCode) { | 226 uscript_getSampleString(UScriptCode script, UChar *dest, int32_t capacity, UErro
rCode *pErrorCode) { |
| (...skipping 38 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
| 260 | 265 |
| 261 U_CAPI UBool U_EXPORT2 | 266 U_CAPI UBool U_EXPORT2 |
| 262 uscript_breaksBetweenLetters(UScriptCode script) { | 267 uscript_breaksBetweenLetters(UScriptCode script) { |
| 263 return (getScriptProps(script) & LB_LETTERS) != 0; | 268 return (getScriptProps(script) & LB_LETTERS) != 0; |
| 264 } | 269 } |
| 265 | 270 |
| 266 U_CAPI UBool U_EXPORT2 | 271 U_CAPI UBool U_EXPORT2 |
| 267 uscript_isCased(UScriptCode script) { | 272 uscript_isCased(UScriptCode script) { |
| 268 return (getScriptProps(script) & CASED) != 0; | 273 return (getScriptProps(script) & CASED) != 0; |
| 269 } | 274 } |
| OLD | NEW |