1 /* 2 * Copyright (C) 2015 The Android Open Source Project 3 * 4 * Licensed under the Apache License, Version 2.0 (the "License"); 5 * you may not use this file except in compliance with the License. 6 * You may obtain a copy of the License at 7 * 8 * http://www.apache.org/licenses/LICENSE-2.0 9 * 10 * Unless required by applicable law or agreed to in writing, software 11 * distributed under the License is distributed on an "AS IS" BASIS, 12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 13 * See the License for the specific language governing permissions and 14 * limitations under the License. 15 */ 16 17 #ifndef MINIKIN_LOCALE_LIST_H 18 #define MINIKIN_LOCALE_LIST_H 19 20 #include <string> 21 #include <vector> 22 23 #include <hb.h> 24 25 #include "StringPiece.h" 26 27 namespace minikin { 28 29 // Due to the limits in font fallback score calculation, we can't use anything more than 12 locales. 30 const size_t FONT_LOCALE_LIMIT = 12; 31 32 // The language or region code is encoded to 15 bits. 33 constexpr uint16_t NO_LANGUAGE = 0x7fff; 34 constexpr uint16_t NO_REGION = 0x7fff; 35 // The script code is encoded to 20 bits. 36 constexpr uint32_t NO_SCRIPT = 0xfffff; 37 38 class LocaleList; 39 40 // Enum for making sub-locale from FontLangauge. 41 enum class SubtagBits : uint8_t { 42 EMPTY = 0b00000000, 43 LANGUAGE = 0b00000001, 44 SCRIPT = 0b00000010, 45 REGION = 0b00000100, 46 VARIANT = 0b00001000, 47 EMOJI = 0b00010000, 48 ALL = 0b00011111, 49 }; 50 51 inline constexpr SubtagBits operator&(SubtagBits l, SubtagBits r) { 52 return static_cast<SubtagBits>(static_cast<uint8_t>(l) & static_cast<uint8_t>(r)); 53 } 54 inline constexpr SubtagBits operator|(SubtagBits l, SubtagBits r) { 55 return static_cast<SubtagBits>(static_cast<uint8_t>(l) | static_cast<uint8_t>(r)); 56 } 57 58 // Enum for emoji style. 59 enum class EmojiStyle : uint8_t { 60 EMPTY = 0, // No emoji style is specified. 61 DEFAULT = 1, // Default emoji style is specified. 62 EMOJI = 2, // Emoji (color) emoji style is specified. 63 TEXT = 3, // Text (black/white) emoji style is specified. 64 }; 65 66 // Enum for line break style. 67 enum class LineBreakStyle : uint8_t { 68 EMPTY = 0, // No line break style is specified. 69 LOOSE = 1, // line break style is loose. 70 NORMAL = 2, // line break style is normal. 71 STRICT = 3, // line break style is strict. 72 }; 73 74 // Locale is a compact representation of a BCP 47 language tag. 75 // It does not capture all possible information, only what directly affects text layout: 76 // font rendering, hyphenation, word breaking, etc. 77 struct Locale { 78 public: 79 enum class Variant : uint16_t { 80 NO_VARIANT = 0x0000, 81 GERMAN_1901_ORTHOGRAPHY = 0x0001, 82 GERMAN_1996_ORTHOGRAPHY = 0x0002, 83 }; 84 85 // Default constructor creates the unsupported locale. LocaleLocale86 Locale() 87 : mScript(NO_SCRIPT), 88 mLanguage(NO_LANGUAGE), 89 mRegion(NO_REGION), 90 mSubScriptBits(0ul), 91 mVariant(Variant::NO_VARIANT), 92 mEmojiStyle(EmojiStyle::EMPTY), 93 mLBStyle(LineBreakStyle::EMPTY) {} 94 95 // Parse from string 96 Locale(const StringPiece& buf); 97 98 // Parse from identifier. See getIdentifier() for the identifier format. LocaleLocale99 explicit Locale(uint64_t identifier) 100 : mScript(extractBits(identifier, 29, 20)), 101 mLanguage(extractBits(identifier, 49, 15)), 102 mRegion(extractBits(identifier, 14, 15)), 103 mSubScriptBits(scriptToSubScriptBits(mScript)), 104 mVariant(static_cast<Variant>(extractBits(identifier, 0, 2))), 105 mEmojiStyle(static_cast<EmojiStyle>(extractBits(identifier, 12, 2))), 106 mLBStyle(static_cast<LineBreakStyle>(extractBits(identifier, 10, 2))) {} 107 108 bool operator==(const Locale& other) const { 109 return !isUnsupported() && isEqualScript(other) && mLanguage == other.mLanguage && 110 mRegion == other.mRegion && mVariant == other.mVariant && 111 mLBStyle == other.mLBStyle && mEmojiStyle == other.mEmojiStyle; 112 } 113 114 bool operator!=(const Locale other) const { return !(*this == other); } 115 hasLanguageLocale116 inline bool hasLanguage() const { return mLanguage != NO_LANGUAGE; } hasScriptLocale117 inline bool hasScript() const { return mScript != NO_SCRIPT; } hasRegionLocale118 inline bool hasRegion() const { return mRegion != NO_REGION; } hasVariantLocale119 inline bool hasVariant() const { return mVariant != Variant::NO_VARIANT; } hasLBStyleLocale120 inline bool hasLBStyle() const { return mLBStyle != LineBreakStyle::EMPTY; } hasEmojiStyleLocale121 inline bool hasEmojiStyle() const { return mEmojiStyle != EmojiStyle::EMPTY; } 122 isSupportedLocale123 inline bool isSupported() const { 124 return hasLanguage() || hasScript() || hasRegion() || hasVariant() || hasLBStyle() || 125 hasEmojiStyle(); 126 } 127 isUnsupportedLocale128 inline bool isUnsupported() const { return !isSupported(); } 129 getEmojiStyleLocale130 EmojiStyle getEmojiStyle() const { return mEmojiStyle; } 131 132 bool isEqualScript(const Locale& other) const; 133 134 // Returns true if this script supports the given script. For example, ja-Jpan supports Hira, 135 // ja-Hira doesn't support Jpan. 136 bool supportsHbScript(hb_script_t script) const; 137 138 std::string getString() const; 139 140 // Calculates a matching score. This score represents how well the input locales cover this 141 // locale. The maximum score in the locale list is returned. 142 // 0 = no match, 1 = script match, 2 = script and primary language match. 143 int calcScoreFor(const LocaleList& supported) const; 144 145 // Identifier pattern: 146 // |-------|-------|-------|-------|-------|-------|-------|-------| 147 // lllllllllllllll Language Code (15 bits) 148 // ssssssssssssssssssss Script Code (20 bits) 149 // rrrrrrrrrrrrrrr Region Code (15 bits) 150 // ee Emoji Style (2 bits) 151 // bb Line Break Style (2 bits) 152 // XXXXXXXX Free (8 bits) 153 // vv German Variant (2 bits) getIdentifierLocale154 uint64_t getIdentifier() const { 155 return ((uint64_t)mLanguage << 49) | ((uint64_t)mScript << 29) | ((uint64_t)mRegion << 14) | 156 ((uint64_t)mEmojiStyle << 12) | ((uint64_t)mLBStyle << 10) | (uint64_t)mVariant; 157 } 158 159 Locale getPartialLocale(SubtagBits bits) const; 160 161 private: 162 friend class LocaleList; // for LocaleList constructor 163 164 // ISO 15924 compliant script code. The 4 chars script code are packed into a 20 bit integer. 165 // If not specified, this is kInvalidScript. 166 uint32_t mScript; 167 168 // ISO 639-1 or ISO 639-2 compliant language code. 169 // The two- or three-letter language code is packed into a 15 bit integer. 170 // mLanguage = 0 means the Locale is unsupported. 171 uint16_t mLanguage; 172 173 // ISO 3166-1 or UN M.49 compliant region code. The two-letter or three-digit region code is 174 // packed into a 15 bit integer. 175 uint16_t mRegion; 176 177 // For faster comparing, use 7 bits for specific scripts. 178 static const uint8_t kBopomofoFlag = 1u; 179 static const uint8_t kHanFlag = 1u << 1; 180 static const uint8_t kHangulFlag = 1u << 2; 181 static const uint8_t kHiraganaFlag = 1u << 3; 182 static const uint8_t kKatakanaFlag = 1u << 4; 183 static const uint8_t kSimplifiedChineseFlag = 1u << 5; 184 static const uint8_t kTraditionalChineseFlag = 1u << 6; 185 uint8_t mSubScriptBits; 186 187 Variant mVariant; 188 189 EmojiStyle mEmojiStyle; 190 LineBreakStyle mLBStyle; 191 192 void resolveUnicodeExtension(const char* buf, size_t length); 193 extractBitsLocale194 inline static uint64_t extractBits(uint64_t value, uint8_t shift, uint8_t nBits) { 195 return (value >> shift) & ((1 << nBits) - 1); 196 } 197 198 static uint8_t scriptToSubScriptBits(uint32_t rawScript); 199 200 static LineBreakStyle resolveLineBreakStyle(const char* buf, size_t length); 201 static EmojiStyle resolveEmojiStyle(const char* buf, size_t length); 202 static EmojiStyle scriptToEmojiStyle(uint32_t script); 203 204 // Returns true if the provide subscript bits has the requested subscript bits. 205 // Note that this function returns false if the requested subscript bits are empty. 206 static bool supportsScript(uint8_t providedBits, uint8_t requestedBits); 207 }; 208 209 // An immutable list of locale. 210 class LocaleList { 211 public: 212 explicit LocaleList(std::vector<Locale>&& locales); LocaleList()213 LocaleList() 214 : mUnionOfSubScriptBits(0), 215 mIsAllTheSameLocale(false), 216 mEmojiStyle(EmojiStyle::EMPTY) {} 217 LocaleList(LocaleList&&) = default; 218 size()219 size_t size() const { return mLocales.size(); } empty()220 bool empty() const { return mLocales.empty(); } 221 const Locale& operator[](size_t n) const { return mLocales[n]; } 222 getHbLanguage(size_t n)223 hb_language_t getHbLanguage(size_t n) const { return mHbLangs[n]; } 224 225 // Returns an effective emoji style of this locale list. 226 // The effective means the first non empty emoji style in the list. getEmojiStyle()227 EmojiStyle getEmojiStyle() const { return mEmojiStyle; } 228 229 private: 230 friend struct Locale; // for calcScoreFor 231 232 std::vector<Locale> mLocales; 233 234 // The languages to be passed to HarfBuzz shaper. 235 std::vector<hb_language_t> mHbLangs; 236 uint8_t mUnionOfSubScriptBits; 237 bool mIsAllTheSameLocale; 238 EmojiStyle mEmojiStyle; 239 getUnionOfSubScriptBits()240 uint8_t getUnionOfSubScriptBits() const { return mUnionOfSubScriptBits; } isAllTheSameLocale()241 bool isAllTheSameLocale() const { return mIsAllTheSameLocale; } 242 243 // Do not copy and assign. 244 LocaleList(const LocaleList&) = delete; 245 void operator=(const LocaleList&) = delete; 246 }; 247 248 } // namespace minikin 249 250 #endif // MINIKIN_LOCALE_LIST_H 251