1 /* 2 * Copyright 2020 Google LLC 3 * 4 * Use of this source code is governed by a BSD-style license that can be 5 * found in the LICENSE file. 6 */ 7 #ifndef SkUnicode_DEFINED 8 #define SkUnicode_DEFINED 9 #include "include/core/SkRefCnt.h" 10 #include "include/core/SkSpan.h" 11 #include "include/core/SkString.h" 12 #include "include/core/SkTypes.h" 13 #include "include/private/base/SkTArray.h" 14 #include "include/private/base/SkTo.h" 15 #include "src/base/SkUTF.h" 16 #include <cstddef> 17 #include <cstdint> 18 #include <memory> 19 #include <string> 20 #include <vector> 21 namespace sknonstd { template <typename T> struct is_bitmask_enum; } 22 23 #if !defined(SKUNICODE_IMPLEMENTATION) 24 #define SKUNICODE_IMPLEMENTATION 0 25 #endif 26 27 #if !defined(SKUNICODE_API) 28 #if defined(SKUNICODE_DLL) 29 #if defined(_MSC_VER) 30 #if SKUNICODE_IMPLEMENTATION 31 #define SKUNICODE_API __declspec(dllexport) 32 #else 33 #define SKUNICODE_API __declspec(dllimport) 34 #endif 35 #else 36 #define SKUNICODE_API __attribute__((visibility("default"))) 37 #endif 38 #else 39 #define SKUNICODE_API 40 #endif 41 #endif 42 43 class SKUNICODE_API SkBidiIterator { 44 public: 45 typedef int32_t Position; 46 typedef uint8_t Level; 47 struct Region { RegionRegion48 Region(Position start, Position end, Level level) 49 : start(start), end(end), level(level) { } 50 Position start; 51 Position end; 52 Level level; 53 }; 54 enum Direction { 55 kLTR, 56 kRTL, 57 }; 58 virtual ~SkBidiIterator() = default; 59 virtual Position getLength() = 0; 60 virtual Level getLevelAt(Position) = 0; 61 }; 62 63 class SKUNICODE_API SkBreakIterator { 64 public: 65 typedef int32_t Position; 66 typedef int32_t Status; 67 virtual ~SkBreakIterator() = default; 68 virtual Position first() = 0; 69 virtual Position current() = 0; 70 virtual Position next() = 0; 71 virtual Status status() = 0; 72 virtual bool isDone() = 0; 73 virtual bool setText(const char utftext8[], int utf8Units) = 0; 74 virtual bool setText(const char16_t utftext16[], int utf16Units) = 0; 75 }; 76 77 class SKUNICODE_API SkUnicode : public SkRefCnt { 78 public: 79 enum CodeUnitFlags { 80 kNoCodeUnitFlag = 0x00, 81 kPartOfWhiteSpaceBreak = 0x01, 82 kGraphemeStart = 0x02, 83 kSoftLineBreakBefore = 0x04, 84 kHardLineBreakBefore = 0x08, 85 kPartOfIntraWordBreak = 0x10, 86 kControl = 0x20, 87 kTabulation = 0x40, 88 kGlyphClusterStart = 0x80, 89 kIdeographic = 0x100, 90 kEmoji = 0x200, 91 kWordBreak = 0x400, 92 kSentenceBreak = 0x800, 93 }; 94 enum class TextDirection { 95 kLTR, 96 kRTL, 97 }; 98 typedef size_t Position; 99 typedef uint8_t BidiLevel; 100 struct BidiRegion { BidiRegionBidiRegion101 BidiRegion(Position start, Position end, BidiLevel level) 102 : start(start), end(end), level(level) { } 103 Position start; 104 Position end; 105 BidiLevel level; 106 }; 107 enum class LineBreakType { 108 kSoftLineBreak = 0, 109 kHardLineBreak = 100, 110 }; 111 112 enum class BreakType { kWords, kGraphemes, kLines, kSentences }; 113 struct LineBreakBefore { LineBreakBeforeLineBreakBefore114 LineBreakBefore(Position pos, LineBreakType breakType) 115 : pos(pos), breakType(breakType) { } 116 Position pos; 117 LineBreakType breakType; 118 }; 119 120 ~SkUnicode() override = default; 121 122 // deprecated 123 virtual SkString toUpper(const SkString&) = 0; 124 virtual SkString toUpper(const SkString&, const char* locale) = 0; 125 126 virtual bool isControl(SkUnichar utf8) = 0; 127 virtual bool isWhitespace(SkUnichar utf8) = 0; 128 virtual bool isSpace(SkUnichar utf8) = 0; 129 virtual bool isTabulation(SkUnichar utf8) = 0; 130 virtual bool isHardBreak(SkUnichar utf8) = 0; 131 /** 132 * Returns if a code point may start an emoji sequence. 133 * Returns true for '#', '*', and '0'-'9' since they may start an emoji sequence. 134 * To determine if a list of code points begins with an emoji sequence, use 135 * getEmojiSequence. 136 **/ 137 virtual bool isEmoji(SkUnichar utf8) = 0; 138 virtual bool isEmojiComponent(SkUnichar utf8) = 0; 139 virtual bool isEmojiModifierBase(SkUnichar utf8) = 0; 140 virtual bool isEmojiModifier(SkUnichar utf8) = 0; 141 virtual bool isRegionalIndicator(SkUnichar utf8) = 0; 142 virtual bool isIdeographic(SkUnichar utf8) = 0; 143 144 // Methods used in SkShaper and SkText 145 virtual std::unique_ptr<SkBidiIterator> makeBidiIterator 146 (const uint16_t text[], int count, SkBidiIterator::Direction) = 0; 147 virtual std::unique_ptr<SkBidiIterator> makeBidiIterator 148 (const char text[], int count, SkBidiIterator::Direction) = 0; 149 virtual std::unique_ptr<SkBreakIterator> makeBreakIterator 150 (const char locale[], BreakType breakType) = 0; 151 virtual std::unique_ptr<SkBreakIterator> makeBreakIterator(BreakType type) = 0; 152 153 // Methods used in SkParagraph 154 static bool hasTabulationFlag(SkUnicode::CodeUnitFlags flags); 155 static bool hasHardLineBreakFlag(SkUnicode::CodeUnitFlags flags); 156 static bool hasSoftLineBreakFlag(SkUnicode::CodeUnitFlags flags); 157 static bool hasGraphemeStartFlag(SkUnicode::CodeUnitFlags flags); 158 static bool hasControlFlag(SkUnicode::CodeUnitFlags flags); 159 static bool hasPartOfWhiteSpaceBreakFlag(SkUnicode::CodeUnitFlags flags); 160 161 static bool extractBidi(const char utf8[], 162 int utf8Units, 163 TextDirection dir, 164 std::vector<BidiRegion>* bidiRegions); 165 virtual bool getBidiRegions(const char utf8[], 166 int utf8Units, 167 TextDirection dir, 168 std::vector<BidiRegion>* results) = 0; 169 // Returns results in utf16 170 virtual bool getWords(const char utf8[], int utf8Units, const char* locale, 171 std::vector<Position>* results) = 0; 172 virtual bool getUtf8Words(const char utf8[], 173 int utf8Units, 174 const char* locale, 175 std::vector<Position>* results) = 0; 176 virtual bool getSentences(const char utf8[], 177 int utf8Units, 178 const char* locale, 179 std::vector<Position>* results) = 0; 180 virtual bool computeCodeUnitFlags( 181 char utf8[], int utf8Units, bool replaceTabs, 182 skia_private::TArray<SkUnicode::CodeUnitFlags, true>* results) = 0; 183 virtual bool computeCodeUnitFlags( 184 char16_t utf16[], int utf16Units, bool replaceTabs, 185 skia_private::TArray<SkUnicode::CodeUnitFlags, true>* results) = 0; 186 187 static SkString convertUtf16ToUtf8(const char16_t * utf16, int utf16Units); 188 static SkString convertUtf16ToUtf8(const std::u16string& utf16); 189 static std::u16string convertUtf8ToUtf16(const char* utf8, int utf8Units); 190 static std::u16string convertUtf8ToUtf16(const SkString& utf8); 191 192 template <typename Appender8, typename Appender16> extractUtfConversionMapping(SkSpan<const char> utf8,Appender8 && appender8,Appender16 && appender16)193 static bool extractUtfConversionMapping(SkSpan<const char> utf8, Appender8&& appender8, Appender16&& appender16) { 194 size_t size8 = 0; 195 size_t size16 = 0; 196 auto ptr = utf8.begin(); 197 auto end = utf8.end(); 198 while (ptr < end) { 199 200 size_t index = SkToSizeT(ptr - utf8.begin()); 201 SkUnichar u = SkUTF::NextUTF8(&ptr, end); 202 203 // All UTF8 code units refer to the same codepoint 204 size_t next = SkToSizeT(ptr - utf8.begin()); 205 for (auto i = index; i < next; ++i) { 206 //fUTF16IndexForUTF8Index.emplace_back(fUTF8IndexForUTF16Index.size()); 207 appender16(size8); 208 ++size16; 209 } 210 //SkASSERT(fUTF16IndexForUTF8Index.size() == next); 211 SkASSERT(size16 == next); 212 if (size16 != next) { 213 return false; 214 } 215 216 // One or two UTF16 code units refer to the same codepoint 217 uint16_t buffer[2]; 218 size_t count = SkUTF::ToUTF16(u, buffer); 219 //fUTF8IndexForUTF16Index.emplace_back(index); 220 appender8(index); 221 ++size8; 222 if (count > 1) { 223 //fUTF8IndexForUTF16Index.emplace_back(index); 224 appender8(index); 225 ++size8; 226 } 227 } 228 //fUTF16IndexForUTF8Index.emplace_back(fUTF8IndexForUTF16Index.size()); 229 appender16(size8); 230 ++size16; 231 //fUTF8IndexForUTF16Index.emplace_back(fText.size()); 232 appender8(utf8.size()); 233 ++size8; 234 235 return true; 236 } 237 238 template <typename Callback> forEachCodepoint(const char * utf8,int32_t utf8Units,Callback && callback)239 void forEachCodepoint(const char* utf8, int32_t utf8Units, Callback&& callback) { 240 const char* current = utf8; 241 const char* end = utf8 + utf8Units; 242 while (current < end) { 243 auto before = current - utf8; 244 SkUnichar unichar = SkUTF::NextUTF8(¤t, end); 245 if (unichar < 0) unichar = 0xFFFD; 246 auto after = current - utf8; 247 uint16_t buffer[2]; 248 size_t count = SkUTF::ToUTF16(unichar, buffer); 249 callback(unichar, before, after, count); 250 } 251 } 252 253 template <typename Callback> forEachCodepoint(const char16_t * utf16,int32_t utf16Units,Callback && callback)254 void forEachCodepoint(const char16_t* utf16, int32_t utf16Units, Callback&& callback) { 255 const char16_t* current = utf16; 256 const char16_t* end = utf16 + utf16Units; 257 while (current < end) { 258 auto before = current - utf16; 259 SkUnichar unichar = SkUTF::NextUTF16((const uint16_t**)¤t, (const uint16_t*)end); 260 auto after = current - utf16; 261 callback(unichar, before, after); 262 } 263 } 264 265 template <typename Callback> forEachBidiRegion(const uint16_t utf16[],int utf16Units,SkBidiIterator::Direction dir,Callback && callback)266 void forEachBidiRegion(const uint16_t utf16[], int utf16Units, SkBidiIterator::Direction dir, Callback&& callback) { 267 auto iter = makeBidiIterator(utf16, utf16Units, dir); 268 const uint16_t* start16 = utf16; 269 const uint16_t* end16 = utf16 + utf16Units; 270 SkBidiIterator::Level currentLevel = 0; 271 272 SkBidiIterator::Position pos16 = 0; 273 while (pos16 <= iter->getLength()) { 274 auto level = iter->getLevelAt(pos16); 275 if (pos16 == 0) { 276 currentLevel = level; 277 } else if (level != currentLevel) { 278 callback(pos16, start16 - utf16, currentLevel); 279 currentLevel = level; 280 } 281 if (start16 == end16) { 282 break; 283 } 284 SkUnichar u = SkUTF::NextUTF16(&start16, end16); 285 pos16 += SkUTF::ToUTF16(u); 286 } 287 } 288 289 template <typename Callback> forEachBreak(const char16_t utf16[],int utf16Units,SkUnicode::BreakType type,Callback && callback)290 void forEachBreak(const char16_t utf16[], int utf16Units, SkUnicode::BreakType type, Callback&& callback) { 291 auto iter = makeBreakIterator(type); 292 iter->setText(utf16, utf16Units); 293 auto pos = iter->first(); 294 do { 295 callback(pos, iter->status()); 296 pos = iter->next(); 297 } while (!iter->isDone()); 298 } 299 300 virtual void reorderVisual(const BidiLevel runLevels[], int levelsCount, int32_t logicalFromVisual[]) = 0; 301 }; 302 303 namespace sknonstd { 304 template <> struct is_bitmask_enum<SkUnicode::CodeUnitFlags> : std::true_type {}; 305 } // namespace sknonstd 306 307 #endif // SkUnicode_DEFINED 308