1 // -*- C++ -*- 2 /** 3 * \file Encoding.h 4 * This file is part of LyX, the document processor. 5 * Licence details can be found in the file COPYING. 6 * 7 * \author Lars Gullik Bjønnes 8 * \author Jean-Marc Lasgouttes 9 * 10 * Full author contact details are available in file CREDITS. 11 */ 12 13 #ifndef ENCODING_H 14 #define ENCODING_H 15 16 #include "support/docstring.h" 17 #include "support/trivstring.h" 18 #include "support/types.h" 19 20 #include <map> 21 #include <set> 22 #include <vector> 23 24 namespace lyx { 25 26 namespace support { class FileName; } 27 28 class EncodingException : public std::exception { 29 public: 30 EncodingException(char_type c); ~EncodingException()31 virtual ~EncodingException() throw() {} 32 virtual const char * what() const throw(); 33 34 char_type failed_char; 35 int par_id; 36 pos_type pos; 37 }; 38 39 40 enum CharInfoFlags { 41 /// 42 CharInfoCombining = 1, 43 /// 44 CharInfoTextFeature = 2, 45 /// 46 CharInfoMathFeature = 4, 47 /// 48 CharInfoForce = 8, 49 /// 50 CharInfoTextNoTermination = 16, 51 /// 52 CharInfoMathNoTermination = 32, 53 /// 54 CharInfoForceSelected = 64, 55 /// 56 CharInfoDeprecated = 128 57 }; 58 59 60 /// Information about a single UCS4 character 61 class CharInfo { 62 public: CharInfo()63 CharInfo() : flags_(0) {} 64 CharInfo( 65 docstring const & textcommand, docstring const & mathcommand, 66 std::string const & textpreamble, std::string const & mathpreamble, 67 std::string const & tipashortcut, unsigned int flags); 68 // we assume that at least one command is nonempty when using unicodesymbols isUnicodeSymbol()69 bool isUnicodeSymbol() const { return !textcommand_.empty() || !mathcommand_.empty(); } 70 /// LaTeX command (text mode) for this character textcommand()71 docstring const textcommand() const { return textcommand_; } 72 /// LaTeX command (math mode) for this character mathcommand()73 docstring mathcommand() const { return mathcommand_; } 74 /// Needed LaTeX preamble (or feature) for text mode textpreamble()75 std::string textpreamble() const { return textpreamble_; } 76 /// Needed LaTeX preamble (or feature) for math mode mathpreamble()77 std::string mathpreamble() const { return mathpreamble_; } 78 /// Is this a combining character? combining()79 bool combining() const { return flags_ & CharInfoCombining ? true : false; } 80 /// Is \c textpreamble a feature known by LaTeXFeatures, or a raw LaTeX 81 /// command? textfeature()82 bool textfeature() const { return flags_ & CharInfoTextFeature ? true : false; } 83 /// Is \c mathpreamble a feature known by LaTeXFeatures, or a raw LaTeX 84 /// command? mathfeature()85 bool mathfeature() const { return flags_ & CharInfoMathFeature ? true : false; } 86 /// Always force the LaTeX command, even if the encoding contains 87 /// this character? force()88 bool force() const { return flags_ & CharInfoForce ? true : false; } 89 /// Force the LaTeX command for some encodings? forceselected()90 bool forceselected() const { return flags_ & CharInfoForceSelected ? true : false; } 91 /// Disable LaTeX command => char_type conversion for this deprecated symbol? deprecated()92 bool deprecated() const { return flags_ & CharInfoDeprecated ? true : false; } 93 /// TIPA shortcut tipashortcut()94 std::string const tipashortcut() const { return tipashortcut_; } 95 /// \c textcommand needs no termination (such as {} or space). textnotermination()96 bool textnotermination() const { return flags_ & CharInfoTextNoTermination ? true : false; } 97 /// \c mathcommand needs no termination (such as {} or space). mathnotermination()98 bool mathnotermination() const { return flags_ & CharInfoMathNoTermination ? true : false; } 99 /// 100 private: 101 /// LaTeX command (text mode) for this character 102 trivdocstring textcommand_; 103 /// LaTeX command (math mode) for this character 104 trivdocstring mathcommand_; 105 /// Needed LaTeX preamble (or feature) for text mode 106 trivstring textpreamble_; 107 /// Needed LaTeX preamble (or feature) for math mode 108 trivstring mathpreamble_; 109 /// TIPA shortcut 110 trivstring tipashortcut_; 111 /// feature flags 112 unsigned int flags_; 113 }; 114 115 116 /** 117 * An encoding as defined in lib/encodings. 118 * All const methods are thread-safe, so the caller does not need any locking. 119 * This property must be kept when changing the class. 120 */ 121 class Encoding { 122 public: 123 /// Which LaTeX package handles this encoding? 124 enum Package { 125 none = 1, 126 inputenc = 2, 127 CJK = 4, 128 japanese = 8 129 }; 130 /// Represent any of the above packages 131 static int const any; 132 /// Encoding()133 Encoding() : fixedwidth_(true), unsafe_(false), forced_(0), 134 start_encodable_(0), package_(none), complete_(false) {} 135 /// 136 Encoding(std::string const & n, std::string const & l, 137 std::string const & g, std::string const & i, 138 bool f, bool u, Package p); 139 /// 140 void init() const; 141 /// name()142 std::string const name() const { return name_; } 143 /// latexName()144 std::string const latexName() const { return latexName_; } 145 /// guiName()146 std::string const guiName() const { return guiName_; } 147 /// iconvName()148 std::string const iconvName() const { return iconvName_; } 149 /// hasFixedWidth()150 bool hasFixedWidth() const { return fixedwidth_; } 151 /// unsafe()152 bool unsafe() const { return unsafe_; } 153 /// \p c is representable in this encoding without a LaTeX macro 154 bool encodable(char_type c) const; 155 /** 156 * Convert \p c to something that LaTeX can understand. 157 * This is either the character itself (if it is representable 158 * in this encoding), or a LaTeX macro. 159 * If the character is not representable in this encoding, but no 160 * LaTeX macro is known, a warning is given of lyxerr, and the 161 * character is returned. 162 * \return the converted character and a flag indicating whether 163 * the command needs to be terminated by {} or a space. 164 */ 165 std::pair<docstring, bool> latexChar(char_type c) const; 166 /** 167 * Convert \p input to something that LaTeX can understand. 168 * This is either the string itself (if it is representable 169 * in this encoding), or a LaTeX macro. 170 * If a character is not representable in this encoding, but no 171 * LaTeX macro is known, a warning is given of lyxerr, and the 172 * character is returned in the second string of the pair and 173 * omitted in the first. 174 * \p dryrun specifies whether the string is used within source 175 * preview (which yields a special warning). 176 */ 177 std::pair<docstring, docstring> latexString(docstring const & input, 178 bool dryrun = false) const; 179 /// Which LaTeX package handles this encoding? package()180 Package package() const { return package_; } 181 /// A list of all characters usable in this encoding 182 std::vector<char_type> symbolsList() const; 183 private: 184 /** 185 * Do we have to output this character as LaTeX command in any case? 186 * This is true if the "force" flag is set. 187 * We need this if the inputencoding does not support a certain glyph. 188 */ 189 bool isForced(char_type c) const; 190 /// 191 trivstring name_; 192 /// 193 trivstring latexName_; 194 /// 195 trivstring guiName_; 196 /// 197 trivstring iconvName_; 198 /// Is this a fixed width encoding? 199 bool fixedwidth_; 200 /// Is this encoding TeX unsafe, e.g. control characters like {, } 201 /// and \\ may appear in high bytes? 202 bool unsafe_; 203 /// 204 typedef std::set<char_type> CharSet; 205 /// Set of UCS4 characters that we can encode (for singlebyte 206 /// encodings only) 207 CharSet encodable_; 208 /// Set of UCS4 characters that we can't encode 209 CharSet const * forced_; 210 /// All code points below this are encodable. This helps us to avoid 211 /// lokup of ASCII characters in encodable_ and gives about 1 sec 212 /// speedup on export of the Userguide. 213 char_type start_encodable_; 214 /// Which LaTeX package handles this encoding? 215 Package package_; 216 /** 217 * If this is true the stored information about the encoding covers 218 * all encodable characters. We set this to false initially so that 219 * we only need to query iconv for the actually used encodings. 220 * This is needed especially for the multibyte encodings, if we 221 * complete all encoding info on startup it takes 2-3 minutes. 222 */ 223 bool complete_; 224 }; 225 226 class Encodings { 227 public: 228 /// 229 typedef std::set<char_type> MathCommandSet; 230 /// 231 typedef std::set<char_type> TextCommandSet; 232 /// 233 typedef std::set<char_type> MathSymbolSet; 234 /// 235 typedef std::map<trivstring, Encoding> EncodingList; 236 /// iterator to iterate over all encodings. 237 /// We hide the fact that our encoding list is implemented as a map. 238 class const_iterator : public EncodingList::const_iterator { 239 typedef EncodingList::const_iterator base; 240 public: const_iterator()241 const_iterator() : base() {} const_iterator(base const & b)242 const_iterator(base const & b) : base(b) {} 243 Encoding const & operator*() const { return base::operator*().second; } 244 Encoding const * operator->() const { return &(base::operator*().second); } 245 }; 246 /// 247 Encodings(); 248 /// Read the encodings. 249 /// \param encfile encodings definition file 250 /// \param symbolsfile unicode->LaTeX mapping file 251 void read(support::FileName const & encfile, 252 support::FileName const & symbolsfile); 253 /// Get encoding from LyX name \p name 254 Encoding const * 255 fromLyXName(std::string const & name, bool allowUnsafe = false) const; 256 /// Get encoding from LaTeX name \p name and package \p package 257 Encoding const * fromLaTeXName(std::string const & name, 258 int const & package = Encoding::any, bool allowUnsafe = false) const; 259 /// Get encoding from iconv name \p name and package \p package 260 Encoding const * fromIconvName(std::string const & name, 261 int const & package = Encoding::any, bool allowUnsafe = false) const; 262 263 /// begin()264 const_iterator begin() const { return encodinglist.begin(); } 265 /// end()266 const_iterator end() const { return encodinglist.end(); } 267 268 /// Accessor for the unicode information table. 269 static CharInfo const & unicodeCharInfo(char_type c); 270 /// Is this a combining char? 271 static bool isCombiningChar(char_type c); 272 /// Return the TIPA shortcut 273 static std::string const TIPAShortcut(char_type c); 274 /** 275 * Is this a known char from some language? 276 * If \p preamble is empty and code point \p c is known to belong 277 * to a supported script, true is returned and \p preamble is set 278 * to the corresponding entry in the unicodesymbols file. 279 * If \p preamble is not empty, a check is made whether code point 280 * \p c is a known character matching the preamble entry. 281 */ 282 static bool isKnownScriptChar(char_type const c, std::string & preamble); 283 /** 284 * Do we have to display in italics this character when in mathmode? 285 * This is true if the "mathalpha" flag is set. We use this for 286 * letters and accented characters that are output as math commands. 287 */ 288 static bool isMathAlpha(char_type c); 289 /** 290 * Do we have to wrap in \text this character when in mathmode? 291 * This is true if \p c is not ascii and the "mathalpha" flag is not 292 * set and a mathcommand is not defined in the unicodesymbols file. 293 */ 294 static bool isUnicodeTextOnly(char_type c); 295 /** 296 * Register \p c as a mathmode command. 297 */ addMathCmd(char_type c)298 static void addMathCmd(char_type c) { mathcmd.insert(c); } 299 /** 300 * Register \p c as a textmode command. 301 */ addTextCmd(char_type c)302 static void addTextCmd(char_type c) { textcmd.insert(c); } 303 /** 304 * Register \p c as a mathmode symbol. 305 */ addMathSym(char_type c)306 static void addMathSym(char_type c) { mathsym.insert(c); } 307 /** 308 * Tell whether \p c is registered as a mathmode command. 309 */ isMathCmd(char_type c)310 static bool isMathCmd(char_type c) { return mathcmd.count(c); } 311 /** 312 * Tell whether \p c is registered as a textmode command. 313 */ isTextCmd(char_type c)314 static bool isTextCmd(char_type c) { return textcmd.count(c); } 315 /** 316 * Tell whether \p c is registered as a mathmode symbol. 317 */ isMathSym(char_type c)318 static bool isMathSym(char_type c) { return mathsym.count(c); } 319 /** 320 * If \p c cannot be encoded in the given \p encoding, convert 321 * it to something that LaTeX can understand in mathmode. 322 * \p needsTermination indicates whether the command needs to be 323 * terminated by {} or a space. 324 * \return whether \p command is a mathmode command 325 */ 326 static bool latexMathChar(char_type c, bool mathmode, 327 Encoding const * encoding, docstring & command, 328 bool & needsTermination); 329 /** 330 * Convert the LaTeX command in \p cmd to the corresponding unicode 331 * point and set \p combining to true if it is a combining symbol. 332 * \p needsTermination indicates whether the command needs to be 333 * terminated by {} or a space. 334 */ 335 static char_type fromLaTeXCommand(docstring const & cmd, int cmdtype, 336 bool & combining, bool & needsTermination, 337 std::set<std::string> * req = 0); 338 /// 339 enum LatexCmd { 340 /// 341 MATH_CMD = 1, 342 /// 343 TEXT_CMD = 2 344 }; 345 /** 346 * Convert the LaTeX commands in \p cmd and \return a docstring 347 * of corresponding unicode points. The conversion stops at the 348 * first command which could not be converted, and the remaining 349 * unconverted commands are returned in \p rem. 350 * The \p cmdtype parameter can be used to limit recognized 351 * commands to math or text mode commands only. 352 * \p needsTermination indicates whether the command needs to be 353 * terminated by {} or a space. 354 */ 355 static docstring fromLaTeXCommand(docstring const & cmd, int cmdtype, 356 bool & needsTermination, docstring & rem, 357 std::set<std::string> * req = 0); 358 359 protected: 360 /// 361 EncodingList encodinglist; 362 /// 363 static MathCommandSet mathcmd; 364 /// 365 static TextCommandSet textcmd; 366 /// 367 static MathSymbolSet mathsym; 368 }; 369 370 extern Encodings encodings; 371 372 373 } // namespace lyx 374 375 #endif 376