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 // Locale is a compact representation of a BCP 47 language tag. 67 // It does not capture all possible information, only what directly affects text layout: 68 // font rendering, hyphenation, word breaking, etc. 69 struct Locale { 70 public: 71 enum class Variant : uint16_t { // Up to 12 bits 72 NO_VARIANT = 0x0000, 73 GERMAN_1901_ORTHOGRAPHY = 0x0001, 74 GERMAN_1996_ORTHOGRAPHY = 0x0002, 75 }; 76 77 // Default constructor creates the unsupported locale. 78 Locale() 79 : mScript(NO_SCRIPT), 80 mLanguage(NO_LANGUAGE), 81 mRegion(NO_REGION), 82 mSubScriptBits(0ul), 83 mVariant(Variant::NO_VARIANT), 84 mEmojiStyle(EmojiStyle::EMPTY) {} 85 86 // Parse from string 87 Locale(const StringPiece& buf); 88 89 bool operator==(const Locale other) const { 90 return !isUnsupported() && isEqualScript(other) && mLanguage == other.mLanguage && 91 mRegion == other.mRegion && mVariant == other.mVariant && 92 mEmojiStyle == other.mEmojiStyle; 93 } 94 95 bool operator!=(const Locale other) const { return !(*this == other); } 96 97 inline bool hasLanguage() const { return mLanguage != NO_LANGUAGE; } 98 inline bool hasScript() const { return mScript != NO_SCRIPT; } 99 inline bool hasRegion() const { return mRegion != NO_REGION; } 100 inline bool hasVariant() const { return mVariant != Variant::NO_VARIANT; } 101 inline bool hasEmojiStyle() const { return mEmojiStyle != EmojiStyle::EMPTY; } 102 103 inline bool isSupported() const { 104 return hasLanguage() || hasScript() || hasRegion() || hasVariant() || hasEmojiStyle(); 105 } 106 107 inline bool isUnsupported() const { return !isSupported(); } 108 109 EmojiStyle getEmojiStyle() const { return mEmojiStyle; } 110 111 bool isEqualScript(const Locale& other) const; 112 113 // Returns true if this script supports the given script. For example, ja-Jpan supports Hira, 114 // ja-Hira doesn't support Jpan. 115 bool supportsHbScript(hb_script_t script) const; 116 117 std::string getString() const; 118 119 // Calculates a matching score. This score represents how well the input locales cover this 120 // locale. The maximum score in the locale list is returned. 121 // 0 = no match, 1 = script match, 2 = script and primary language match. 122 int calcScoreFor(const LocaleList& supported) const; 123 124 uint64_t getIdentifier() const { 125 return ((uint64_t)mLanguage << 49) | ((uint64_t)mScript << 29) | ((uint64_t)mRegion << 14) | 126 ((uint64_t)mEmojiStyle << 12) | (uint64_t)mVariant; 127 } 128 129 Locale getPartialLocale(SubtagBits bits) const; 130 131 private: 132 friend class LocaleList; // for LocaleList constructor 133 134 // ISO 15924 compliant script code. The 4 chars script code are packed into a 20 bit integer. 135 // If not specified, this is kInvalidScript. 136 uint32_t mScript; 137 138 // ISO 639-1 or ISO 639-2 compliant language code. 139 // The two- or three-letter language code is packed into a 15 bit integer. 140 // mLanguage = 0 means the Locale is unsupported. 141 uint16_t mLanguage; 142 143 // ISO 3166-1 or UN M.49 compliant region code. The two-letter or three-digit region code is 144 // packed into a 15 bit integer. 145 uint16_t mRegion; 146 147 // For faster comparing, use 7 bits for specific scripts. 148 static const uint8_t kBopomofoFlag = 1u; 149 static const uint8_t kHanFlag = 1u << 1; 150 static const uint8_t kHangulFlag = 1u << 2; 151 static const uint8_t kHiraganaFlag = 1u << 3; 152 static const uint8_t kKatakanaFlag = 1u << 4; 153 static const uint8_t kSimplifiedChineseFlag = 1u << 5; 154 static const uint8_t kTraditionalChineseFlag = 1u << 6; 155 uint8_t mSubScriptBits; 156 157 Variant mVariant; 158 159 EmojiStyle mEmojiStyle; 160 161 static uint8_t scriptToSubScriptBits(uint32_t rawScript); 162 163 static EmojiStyle resolveEmojiStyle(const char* buf, size_t length); 164 static EmojiStyle scriptToEmojiStyle(uint32_t script); 165 166 // Returns true if the provide subscript bits has the requested subscript bits. 167 // Note that this function returns false if the requested subscript bits are empty. 168 static bool supportsScript(uint8_t providedBits, uint8_t requestedBits); 169 }; 170 171 // An immutable list of locale. 172 class LocaleList { 173 public: 174 explicit LocaleList(std::vector<Locale>&& locales); 175 LocaleList() 176 : mUnionOfSubScriptBits(0), 177 mIsAllTheSameLocale(false), 178 mEmojiStyle(EmojiStyle::EMPTY) {} 179 LocaleList(LocaleList&&) = default; 180 181 size_t size() const { return mLocales.size(); } 182 bool empty() const { return mLocales.empty(); } 183 const Locale& operator[](size_t n) const { return mLocales[n]; } 184 185 hb_language_t getHbLanguage(size_t n) const { return mHbLangs[n]; } 186 187 // Returns an effective emoji style of this locale list. 188 // The effective means the first non empty emoji style in the list. 189 EmojiStyle getEmojiStyle() const { return mEmojiStyle; } 190 191 private: 192 friend struct Locale; // for calcScoreFor 193 194 std::vector<Locale> mLocales; 195 196 // The languages to be passed to HarfBuzz shaper. 197 std::vector<hb_language_t> mHbLangs; 198 uint8_t mUnionOfSubScriptBits; 199 bool mIsAllTheSameLocale; 200 EmojiStyle mEmojiStyle; 201 202 uint8_t getUnionOfSubScriptBits() const { return mUnionOfSubScriptBits; } 203 bool isAllTheSameLocale() const { return mIsAllTheSameLocale; } 204 205 // Do not copy and assign. 206 LocaleList(const LocaleList&) = delete; 207 void operator=(const LocaleList&) = delete; 208 }; 209 210 } // namespace minikin 211 212 #endif // MINIKIN_LOCALE_LIST_H 213
