1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/common/uscript_props.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,269 @@
1.4 +/*
1.5 +*******************************************************************************
1.6 +* Copyright (C) 2013, International Business Machines
1.7 +* Corporation and others. All Rights Reserved.
1.8 +*******************************************************************************
1.9 +* file name: uscript_props.cpp
1.10 +* encoding: US-ASCII
1.11 +* tab size: 8 (not used)
1.12 +* indentation:4
1.13 +*
1.14 +* created on: 2013feb16
1.15 +* created by: Markus W. Scherer
1.16 +*/
1.17 +
1.18 +#include "unicode/utypes.h"
1.19 +#include "unicode/unistr.h"
1.20 +#include "unicode/uscript.h"
1.21 +#include "unicode/utf16.h"
1.22 +#include "ustr_imp.h"
1.23 +
1.24 +#define LENGTHOF(array) (int32_t)(sizeof(array)/sizeof((array)[0]))
1.25 +
1.26 +namespace {
1.27 +
1.28 +// Script metadata (script properties).
1.29 +// See http://unicode.org/cldr/trac/browser/trunk/common/properties/scriptMetadata.txt
1.30 +
1.31 +// 0 = NOT_ENCODED, no sample character, default false script properties.
1.32 +// Bits 20.. 0: sample character
1.33 +
1.34 +// Bits 23..21: usage
1.35 +const int32_t UNKNOWN = 1 << 21;
1.36 +const int32_t EXCLUSION = 2 << 21;
1.37 +const int32_t LIMITED_USE = 3 << 21;
1.38 +const int32_t ASPIRATIONAL = 4 << 21;
1.39 +const int32_t RECOMMENDED = 5 << 21;
1.40 +
1.41 +// Bits 31..24: Single-bit flags
1.42 +const int32_t RTL = 1 << 24;
1.43 +const int32_t LB_LETTERS = 1 << 25;
1.44 +const int32_t CASED = 1 << 26;
1.45 +
1.46 +const int32_t SCRIPT_PROPS[] = {
1.47 + // Begin copy-paste output from
1.48 + // tools/trunk/unicode/py/parsescriptmetadata.py
1.49 + 0x0040 | UNKNOWN, // Zyyy
1.50 + 0x0308 | UNKNOWN, // Zinh
1.51 + 0x0628 | RECOMMENDED | RTL, // Arab
1.52 + 0x0531 | RECOMMENDED | CASED, // Armn
1.53 + 0x0995 | RECOMMENDED, // Beng
1.54 + 0x3105 | RECOMMENDED | LB_LETTERS, // Bopo
1.55 + 0x13C4 | LIMITED_USE, // Cher
1.56 + 0x03E2 | EXCLUSION | CASED, // Copt
1.57 + 0x042F | RECOMMENDED | CASED, // Cyrl
1.58 + 0x10414 | EXCLUSION | CASED, // Dsrt
1.59 + 0x0905 | RECOMMENDED, // Deva
1.60 + 0x12A0 | RECOMMENDED, // Ethi
1.61 + 0x10D3 | RECOMMENDED, // Geor
1.62 + 0x10330 | EXCLUSION, // Goth
1.63 + 0x03A9 | RECOMMENDED | CASED, // Grek
1.64 + 0x0A95 | RECOMMENDED, // Gujr
1.65 + 0x0A15 | RECOMMENDED, // Guru
1.66 + 0x5B57 | RECOMMENDED | LB_LETTERS, // Hani
1.67 + 0xAC00 | RECOMMENDED, // Hang
1.68 + 0x05D0 | RECOMMENDED | RTL, // Hebr
1.69 + 0x304B | RECOMMENDED | LB_LETTERS, // Hira
1.70 + 0x0C95 | RECOMMENDED, // Knda
1.71 + 0x30AB | RECOMMENDED | LB_LETTERS, // Kana
1.72 + 0x1780 | RECOMMENDED | LB_LETTERS, // Khmr
1.73 + 0x0EA5 | RECOMMENDED | LB_LETTERS, // Laoo
1.74 + 0x004C | RECOMMENDED | CASED, // Latn
1.75 + 0x0D15 | RECOMMENDED, // Mlym
1.76 + 0x1826 | ASPIRATIONAL, // Mong
1.77 + 0x1000 | RECOMMENDED | LB_LETTERS, // Mymr
1.78 + 0x168F | EXCLUSION, // Ogam
1.79 + 0x10300 | EXCLUSION, // Ital
1.80 + 0x0B15 | RECOMMENDED, // Orya
1.81 + 0x16A0 | EXCLUSION, // Runr
1.82 + 0x0D85 | RECOMMENDED, // Sinh
1.83 + 0x0710 | LIMITED_USE | RTL, // Syrc
1.84 + 0x0B95 | RECOMMENDED, // Taml
1.85 + 0x0C15 | RECOMMENDED, // Telu
1.86 + 0x078C | RECOMMENDED | RTL, // Thaa
1.87 + 0x0E17 | RECOMMENDED | LB_LETTERS, // Thai
1.88 + 0x0F40 | RECOMMENDED, // Tibt
1.89 + 0x14C0 | ASPIRATIONAL, // Cans
1.90 + 0xA288 | ASPIRATIONAL | LB_LETTERS, // Yiii
1.91 + 0x1703 | EXCLUSION, // Tglg
1.92 + 0x1723 | EXCLUSION, // Hano
1.93 + 0x1743 | EXCLUSION, // Buhd
1.94 + 0x1763 | EXCLUSION, // Tagb
1.95 + 0x2800 | UNKNOWN, // Brai
1.96 + 0x10800 | EXCLUSION | RTL, // Cprt
1.97 + 0x1900 | LIMITED_USE, // Limb
1.98 + 0x10000 | EXCLUSION, // Linb
1.99 + 0x10480 | EXCLUSION, // Osma
1.100 + 0x10450 | EXCLUSION, // Shaw
1.101 + 0x1950 | LIMITED_USE | LB_LETTERS, // Tale
1.102 + 0x10380 | EXCLUSION, // Ugar
1.103 + 0,
1.104 + 0x1A00 | EXCLUSION, // Bugi
1.105 + 0x2C00 | EXCLUSION | CASED, // Glag
1.106 + 0x10A00 | EXCLUSION | RTL, // Khar
1.107 + 0xA800 | LIMITED_USE, // Sylo
1.108 + 0x1980 | LIMITED_USE | LB_LETTERS, // Talu
1.109 + 0x2D30 | ASPIRATIONAL, // Tfng
1.110 + 0x103A0 | EXCLUSION, // Xpeo
1.111 + 0x1B05 | LIMITED_USE | LB_LETTERS, // Bali
1.112 + 0x1BC0 | LIMITED_USE, // Batk
1.113 + 0,
1.114 + 0x11005 | EXCLUSION, // Brah
1.115 + 0xAA00 | LIMITED_USE, // Cham
1.116 + 0,
1.117 + 0,
1.118 + 0,
1.119 + 0,
1.120 + 0x13153 | EXCLUSION, // Egyp
1.121 + 0,
1.122 + 0x5B57 | RECOMMENDED | LB_LETTERS, // Hans
1.123 + 0x5B57 | RECOMMENDED | LB_LETTERS, // Hant
1.124 + 0,
1.125 + 0,
1.126 + 0,
1.127 + 0xA984 | LIMITED_USE | LB_LETTERS, // Java
1.128 + 0xA90A | LIMITED_USE, // Kali
1.129 + 0,
1.130 + 0,
1.131 + 0x1C00 | LIMITED_USE, // Lepc
1.132 + 0,
1.133 + 0x0840 | LIMITED_USE | RTL, // Mand
1.134 + 0,
1.135 + 0x10980 | EXCLUSION | RTL, // Mero
1.136 + 0x07CA | LIMITED_USE | RTL, // Nkoo
1.137 + 0x10C00 | EXCLUSION | RTL, // Orkh
1.138 + 0,
1.139 + 0xA840 | EXCLUSION, // Phag
1.140 + 0x10900 | EXCLUSION | RTL, // Phnx
1.141 + 0x16F00 | ASPIRATIONAL, // Plrd
1.142 + 0,
1.143 + 0,
1.144 + 0,
1.145 + 0,
1.146 + 0,
1.147 + 0,
1.148 + 0xA549 | LIMITED_USE, // Vaii
1.149 + 0,
1.150 + 0x12000 | EXCLUSION, // Xsux
1.151 + 0,
1.152 + 0xFDD0 | UNKNOWN, // Zzzz
1.153 + 0x102A0 | EXCLUSION, // Cari
1.154 + 0x304B | RECOMMENDED | LB_LETTERS, // Jpan
1.155 + 0x1A20 | LIMITED_USE | LB_LETTERS, // Lana
1.156 + 0x10280 | EXCLUSION, // Lyci
1.157 + 0x10920 | EXCLUSION | RTL, // Lydi
1.158 + 0x1C5A | LIMITED_USE, // Olck
1.159 + 0xA930 | EXCLUSION, // Rjng
1.160 + 0xA882 | LIMITED_USE, // Saur
1.161 + 0,
1.162 + 0x1B83 | LIMITED_USE, // Sund
1.163 + 0,
1.164 + 0xABC0 | LIMITED_USE, // Mtei
1.165 + 0x10840 | EXCLUSION | RTL, // Armi
1.166 + 0x10B00 | EXCLUSION | RTL, // Avst
1.167 + 0x11103 | LIMITED_USE, // Cakm
1.168 + 0xAC00 | RECOMMENDED, // Kore
1.169 + 0x11083 | EXCLUSION, // Kthi
1.170 + 0,
1.171 + 0x10B60 | EXCLUSION | RTL, // Phli
1.172 + 0,
1.173 + 0,
1.174 + 0x10B40 | EXCLUSION | RTL, // Prti
1.175 + 0x0800 | EXCLUSION | RTL, // Samr
1.176 + 0xAA80 | LIMITED_USE | LB_LETTERS, // Tavt
1.177 + 0,
1.178 + 0,
1.179 + 0xA6A0 | LIMITED_USE, // Bamu
1.180 + 0xA4D0 | LIMITED_USE, // Lisu
1.181 + 0,
1.182 + 0x10A60 | EXCLUSION | RTL, // Sarb
1.183 + 0,
1.184 + 0,
1.185 + 0,
1.186 + 0,
1.187 + 0,
1.188 + 0,
1.189 + 0,
1.190 + 0x109A0 | EXCLUSION | RTL, // Merc
1.191 + 0,
1.192 + 0,
1.193 + 0,
1.194 + 0,
1.195 + 0,
1.196 + 0,
1.197 + 0,
1.198 + 0,
1.199 + 0,
1.200 + 0x11183 | EXCLUSION, // Shrd
1.201 + 0x110D0 | EXCLUSION, // Sora
1.202 + 0x11680 | EXCLUSION, // Takr
1.203 + 0,
1.204 + 0,
1.205 + 0,
1.206 + 0,
1.207 + 0,
1.208 + 0,
1.209 + 0,
1.210 + // End copy-paste from parsescriptmetadata.py
1.211 +};
1.212 +
1.213 +int32_t getScriptProps(UScriptCode script) {
1.214 + if (0 <= script && script < LENGTHOF(SCRIPT_PROPS)) {
1.215 + return SCRIPT_PROPS[script];
1.216 + } else {
1.217 + return 0;
1.218 + }
1.219 +}
1.220 +
1.221 +} // namespace
1.222 +
1.223 +U_CAPI int32_t U_EXPORT2
1.224 +uscript_getSampleString(UScriptCode script, UChar *dest, int32_t capacity, UErrorCode *pErrorCode) {
1.225 + if(U_FAILURE(*pErrorCode)) { return 0; }
1.226 + if(capacity < 0 || (capacity > 0 && dest == NULL)) {
1.227 + *pErrorCode = U_ILLEGAL_ARGUMENT_ERROR;
1.228 + return 0;
1.229 + }
1.230 + int32_t sampleChar = getScriptProps(script) & 0x1fffff;
1.231 + int32_t length;
1.232 + if(sampleChar == 0) {
1.233 + length = 0;
1.234 + } else {
1.235 + length = U16_LENGTH(sampleChar);
1.236 + if(length <= capacity) {
1.237 + int32_t i = 0;
1.238 + U16_APPEND_UNSAFE(dest, i, sampleChar);
1.239 + }
1.240 + }
1.241 + return u_terminateUChars(dest, capacity, length, pErrorCode);
1.242 +}
1.243 +
1.244 +U_COMMON_API icu::UnicodeString U_EXPORT2
1.245 +uscript_getSampleUnicodeString(UScriptCode script) {
1.246 + icu::UnicodeString sample;
1.247 + int32_t sampleChar = getScriptProps(script) & 0x1fffff;
1.248 + if(sampleChar != 0) {
1.249 + sample.append(sampleChar);
1.250 + }
1.251 + return sample;
1.252 +}
1.253 +
1.254 +U_CAPI UScriptUsage U_EXPORT2
1.255 +uscript_getUsage(UScriptCode script) {
1.256 + return (UScriptUsage)((getScriptProps(script) >> 21) & 7);
1.257 +}
1.258 +
1.259 +U_CAPI UBool U_EXPORT2
1.260 +uscript_isRightToLeft(UScriptCode script) {
1.261 + return (getScriptProps(script) & RTL) != 0;
1.262 +}
1.263 +
1.264 +U_CAPI UBool U_EXPORT2
1.265 +uscript_breaksBetweenLetters(UScriptCode script) {
1.266 + return (getScriptProps(script) & LB_LETTERS) != 0;
1.267 +}
1.268 +
1.269 +U_CAPI UBool U_EXPORT2
1.270 +uscript_isCased(UScriptCode script) {
1.271 + return (getScriptProps(script) & CASED) != 0;
1.272 +}
