src/Language.cpp

   1 /**
   2  * \file Language.cpp
   3  * This file is part of LyX, the document processor.
   4  * Licence details can be found in the file COPYING.
   5  *
   6  * \author Lars Gullik Bjønnes
   7  * \author Jean-Marc Lasgouttes
   8  * \author Jürgen Spitzmüller
   9  * \author Dekel Tsur
  10  *
  11  * Full author contact details are available in file CREDITS.
  12  */
  13
  14 #include <config.h>
  15
  16 #include "Language.h"
  17
  18 #include "Encoding.h"
  19 #include "LaTeXFonts.h"
  20 #include "Lexer.h"
  21 #include "LyXRC.h"
  22
  23 #include "support/debug.h"
  24 #include "support/FileName.h"
  25 #include "support/filetools.h"
  26 #include "support/lassert.h"
  27 #include "support/lstrings.h"
  28 #include "support/Messages.h"
  29
  30 using namespace std;
  31 using namespace lyx::support;
  32
  33 namespace lyx {
  34
  35 Languages languages;
  36 Language const * ignore_language = 0;
  37 Language const * default_language = 0;
  38 Language const * latex_language = 0;
  39 Language const * reset_language = 0;
  40
  41
  42 bool Language::isPolyglossiaExclusive() const
  43 {
  44         return babel().empty() && !polyglossia().empty() && requires().empty();
  45 }
  46
  47
  48 bool Language::isBabelExclusive() const
  49 {
  50         return !babel().empty() && polyglossia().empty() && requires().empty();
  51 }
  52
  53
  54 docstring const Language::translateLayout(string const & m) const
  55 {
  56         if (m.empty())
  57                 return docstring();
  58
  59         if (!isAscii(m)) {
  60                 lyxerr << "Warning: not translating `" << m
  61                        << "' because it is not pure ASCII.\n";
  62                 return from_utf8(m);
  63         }
  64
  65         TranslationMap::const_iterator it = layoutTranslations_.find(m);
  66         if (it != layoutTranslations_.end())
  67                 return it->second;
  68
  69         docstring t = from_ascii(m);
  70         cleanTranslation(t);
  71         return t;
  72 }
  73
  74
  75 string Language::fontenc(BufferParams const & params) const
  76 {
  77         // Don't use LaTeX fonts, so just return the language's preferred
  78         // (although this is not used with nonTeXFonts anyway).
  79         if (params.useNonTeXFonts)
  80                 return fontenc_.front() == "ASCII" ? "T1" : fontenc_.front();
  81
  82         // Determine optimal font encoding
  83         // We check whether the used rm font supports an encoding our language supports
  84         LaTeXFont const & lf =
  85                 theLaTeXFonts().getLaTeXFont(from_ascii(params.fontsRoman()));
  86         vector<string> const lfe = lf.fontencs();
  87         for (auto & fe : fontenc_) {
  88                 // ASCII means: support all T* encodings plus OT1
  89                 if (fe == "ASCII") {
  90                         for (auto & afe : lfe) {
  91                                 if (afe == "OT1" || prefixIs(afe, "T"))
  92                                         // we found a suitable one; return that.
  93                                         return afe;
  94                         }
  95                 }
  96                 // For other encodings, just check whether the font supports it
  97                 if (lf.hasFontenc(fe))
  98                         return fe;
  99         }
 100         // We did not find a suitable one; just take the first in the list,
 101         // the priorized one (which is "T1" for ASCII).
 102         return fontenc_.front() == "ASCII" ? "T1" : fontenc_.front();
 103 }
 104
 105
 106 string Language::dateFormat(size_t i) const
 107 {
 108         if (i > dateformats_.size())
 109                 return string();
 110         return dateformats_.at(i);
 111 }
 112
 113
 114 bool Language::readLanguage(Lexer & lex)
 115 {
 116         enum LanguageTags {
 117                 LA_BABELNAME = 1,
 118                 LA_DATEFORMATS,
 119                 LA_ENCODING,
 120                 LA_END,
 121                 LA_FONTENC,
 122                 LA_GUINAME,
 123                 LA_HAS_GUI_SUPPORT,
 124                 LA_INTERNAL_ENC,
 125                 LA_LANG_CODE,
 126                 LA_LANG_VARIETY,
 127                 LA_POLYGLOSSIANAME,
 128                 LA_POLYGLOSSIAOPTS,
 129                 LA_POSTBABELPREAMBLE,
 130                 LA_PREBABELPREAMBLE,
 131                 LA_PROVIDES,
 132                 LA_REQUIRES,
 133                 LA_QUOTESTYLE,
 134                 LA_RTL,
 135                 LA_WORDWRAP
 136         };
 137
 138         // Keep these sorted alphabetically!
 139         LexerKeyword languageTags[] = {
 140                 { "babelname",            LA_BABELNAME },
 141                 { "dateformats",          LA_DATEFORMATS },
 142                 { "encoding",             LA_ENCODING },
 143                 { "end",                  LA_END },
 144                 { "fontencoding",         LA_FONTENC },
 145                 { "guiname",              LA_GUINAME },
 146                 { "hasguisupport",        LA_HAS_GUI_SUPPORT },
 147                 { "internalencoding",     LA_INTERNAL_ENC },
 148                 { "langcode",             LA_LANG_CODE },
 149                 { "langvariety",          LA_LANG_VARIETY },
 150                 { "polyglossianame",      LA_POLYGLOSSIANAME },
 151                 { "polyglossiaopts",      LA_POLYGLOSSIAOPTS },
 152                 { "postbabelpreamble",    LA_POSTBABELPREAMBLE },
 153                 { "prebabelpreamble",     LA_PREBABELPREAMBLE },
 154                 { "provides",             LA_PROVIDES },
 155                 { "quotestyle",           LA_QUOTESTYLE },
 156                 { "requires",             LA_REQUIRES },
 157                 { "rtl",                  LA_RTL },
 158                 {"wordwrap",              LA_WORDWRAP }
 159         };
 160
 161         bool error = false;
 162         bool finished = false;
 163         lex.pushTable(languageTags);
 164         // parse style section
 165         while (!finished && lex.isOK() && !error) {
 166                 int le = lex.lex();
 167                 // See comment in LyXRC.cpp.
 168                 switch (le) {
 169                 case Lexer::LEX_FEOF:
 170                         continue;
 171
 172                 case Lexer::LEX_UNDEF: // parse error
 173                         lex.printError("Unknown language tag `$$Token'");
 174                         error = true;
 175                         continue;
 176
 177                 default:
 178                         break;
 179                 }
 180                 switch (static_cast<LanguageTags>(le)) {
 181                 case LA_END: // end of structure
 182                         finished = true;
 183                         break;
 184                 case LA_BABELNAME:
 185                         lex >> babel_;
 186                         break;
 187                 case LA_POLYGLOSSIANAME:
 188                         lex >> polyglossia_name_;
 189                         break;
 190                 case LA_POLYGLOSSIAOPTS:
 191                         lex >> polyglossia_opts_;
 192                         break;
 193                 case LA_QUOTESTYLE:
 194                         lex >> quote_style_;
 195                         break;
 196                 case LA_ENCODING:
 197                         lex >> encodingStr_;
 198                         break;
 199                 case LA_FONTENC: {
 200                         lex.eatLine();
 201                         vector<string> const fe =
 202                                 getVectorFromString(lex.getString(true), "|");
 203                         fontenc_.insert(fontenc_.end(), fe.begin(), fe.end());
 204                         break;
 205                 }
 206                 case LA_DATEFORMATS: {
 207                         lex.eatLine();
 208                         vector<string> const df =
 209                                 getVectorFromString(trim(lex.getString(true), "\""), "|");
 210                         dateformats_.insert(dateformats_.end(), df.begin(), df.end());
 211                         break;
 212                 }
 213                 case LA_GUINAME:
 214                         lex >> display_;
 215                         break;
 216                 case LA_HAS_GUI_SUPPORT:
 217                         lex >> has_gui_support_;
 218                         break;
 219                 case LA_INTERNAL_ENC:
 220                         lex >> internal_enc_;
 221                         break;
 222                 case LA_LANG_CODE:
 223                         lex >> code_;
 224                         break;
 225                 case LA_LANG_VARIETY:
 226                         lex >> variety_;
 227                         break;
 228                 case LA_POSTBABELPREAMBLE:
 229                         babel_postsettings_ =
 230                                 lex.getLongString(from_ascii("EndPostBabelPreamble"));
 231                         break;
 232                 case LA_PREBABELPREAMBLE:
 233                         babel_presettings_ =
 234                                 lex.getLongString(from_ascii("EndPreBabelPreamble"));
 235                         break;
 236                 case LA_REQUIRES:
 237                         lex >> requires_;
 238                         break;
 239                 case LA_PROVIDES:
 240                         lex >> provides_;
 241                         break;
 242                 case LA_RTL:
 243                         lex >> rightToLeft_;
 244                         break;
 245                 case LA_WORDWRAP:
 246                         lex >> word_wrap_;
 247                         break;
 248                 }
 249         }
 250         lex.popTable();
 251         return finished && !error;
 252 }
 253
 254
 255 bool Language::read(Lexer & lex)
 256 {
 257         encoding_ = 0;
 258         internal_enc_ = 0;
 259         rightToLeft_ = 0;
 260
 261         if (!lex.next()) {
 262                 lex.printError("No name given for language: `$$Token'.");
 263                 return false;
 264         }
 265
 266         lang_ = lex.getString();
 267         LYXERR(Debug::INFO, "Reading language " << lang_);
 268         if (!readLanguage(lex)) {
 269                 LYXERR0("Error parsing language `" << lang_ << '\'');
 270                 return false;
 271         }
 272
 273         encoding_ = encodings.fromLyXName(encodingStr_);
 274         if (!encoding_ && !encodingStr_.empty()) {
 275                 encoding_ = encodings.fromLyXName("iso8859-1");
 276                 LYXERR0("Unknown encoding " << encodingStr_);
 277         }
 278         if (fontenc_.empty())
 279                 fontenc_.push_back("ASCII");
 280         if (dateformats_.empty()) {
 281                 dateformats_.push_back("MMMM dd, yyyy");
 282                 dateformats_.push_back("MMM dd, yyyy");
 283                 dateformats_.push_back("M/d/yyyy");
 284         }
 285         return true;
 286 }
 287
 288
 289 void Language::readLayoutTranslations(Language::TranslationMap const & trans, bool replace)
 290 {
 291         for (auto const & t : trans) {
 292                 if (replace
 293                     || layoutTranslations_.find(t.first) == layoutTranslations_.end())
 294                         layoutTranslations_[t.first] = t.second;
 295         }
 296 }
 297
 298
 299 void Languages::read(FileName const & filename)
 300 {
 301         Lexer lex;
 302         lex.setFile(filename);
 303         lex.setContext("Languages::read");
 304         while (lex.isOK()) {
 305                 int le = lex.lex();
 306                 switch (le) {
 307                 case Lexer::LEX_FEOF:
 308                         continue;
 309
 310                 default:
 311                         break;
 312                 }
 313                 if (lex.getString() != "Language") {
 314                         lex.printError("Unknown Language tag `$$Token'");
 315                         continue;
 316                 }
 317                 Language l;
 318                 l.read(lex);
 319                 if (!lex)
 320                         break;
 321                 if (l.lang() == "latex") {
 322                         // Check if latex language was not already defined.
 323                         LASSERT(latex_language == 0, continue);
 324                         static const Language latex_lang = l;
 325                         latex_language = &latex_lang;
 326                 } else if (l.lang() == "ignore") {
 327                         // Check if ignore language was not already defined.
 328                         LASSERT(ignore_language == 0, continue);
 329                         static const Language ignore_lang = l;
 330                         ignore_language = &ignore_lang;
 331                 } else
 332                         languagelist_[l.lang()] = l;
 333         }
 334
 335         default_language = getLanguage("english");
 336         if (!default_language) {
 337                 LYXERR0("Default language \"english\" not found!");
 338                 default_language = &(*languagelist_.begin()).second;
 339                 LYXERR0("Using \"" << default_language->lang() << "\" instead!");
 340         }
 341
 342         // Read layout translations
 343         FileName const path = libFileSearch(string(), "layouttranslations");
 344         readLayoutTranslations(path);
 345 }
 346
 347
 348 namespace {
 349
 350 bool readTranslations(Lexer & lex, Language::TranslationMap & trans)
 351 {
 352         while (lex.isOK()) {
 353                 if (lex.checkFor("End"))
 354                         break;
 355                 if (!lex.next(true))
 356                         return false;
 357                 string const key = lex.getString();
 358                 if (!lex.next(true))
 359                         return false;
 360                 docstring const val = lex.getDocString();
 361                 trans[key] = val;
 362         }
 363         return true;
 364 }
 365
 366
 367 enum Match {
 368         NoMatch,
 369         ApproximateMatch,
 370         VeryApproximateMatch,
 371         ExactMatch
 372 };
 373
 374
 375 Match match(string const & code, Language const & lang)
 376 {
 377         // we need to mimic gettext: code can be a two-letter code, which
 378         // should match all variants, e.g. "de" should match "de_DE",
 379         // "de_AT" etc.
 380         // special case for chinese:
 381         // simplified  => code == "zh_CN", langcode == "zh_CN"
 382         // traditional => code == "zh_TW", langcode == "zh_CN"
 383         string const variety = lang.variety();
 384         string const langcode = variety.empty() ?
 385                                 lang.code() : lang.code() + '_' + variety;
 386         string const name = lang.lang();
 387         if ((code == langcode && name != "chinese-traditional")
 388                 || (code == "zh_TW"  && name == "chinese-traditional"))
 389                 return ExactMatch;
 390         if ((code.size() == 2) && (langcode.size() > 2)
 391                 && (code + '_' == langcode.substr(0, 3)))
 392                 return ApproximateMatch;
 393         if (code.substr(0,2) == langcode.substr(0,2))
 394                 return VeryApproximateMatch;
 395         return NoMatch;
 396 }
 397
 398 } // namespace
 399
 400
 401
 402 Language const * Languages::getFromCode(string const & code) const
 403 {
 404         // 1/ exact match with any known language
 405         for (auto const & l : languagelist_) {
 406                 if (match(code, l.second) == ExactMatch)
 407                         return &l.second;
 408         }
 409
 410         // 2/ approximate with any known language
 411         for (auto const & l : languagelist_) {
 412                 if (match(code, l.second) == ApproximateMatch)
 413                         return &l.second;
 414         }
 415         return 0;
 416 }
 417
 418
 419 Language const * Languages::getFromCode(string const & code,
 420                         set<Language const *> const & tryfirst) const
 421 {
 422         // 1/ exact match with tryfirst list
 423         for (auto const * lptr : tryfirst) {
 424                 if (match(code, *lptr) == ExactMatch)
 425                         return lptr;
 426         }
 427
 428         // 2/ approximate match with tryfirst list
 429         for (auto const * lptr : tryfirst) {
 430                 Match const m = match(code, *lptr);
 431                 if (m == ApproximateMatch || m == VeryApproximateMatch)
 432                         return lptr;
 433         }
 434
 435         // 3/ stricter match in all languages
 436         return getFromCode(code);
 437
 438         LYXERR0("Unknown language `" << code << "'");
 439         return 0;
 440 }
 441
 442
 443 void Languages::readLayoutTranslations(support::FileName const & filename)
 444 {
 445         Lexer lex;
 446         lex.setFile(filename);
 447         lex.setContext("Languages::read");
 448
 449         // 1) read all translations (exact and approximate matches) into trans
 450         std::map<string, Language::TranslationMap> trans;
 451         while (lex.isOK()) {
 452                 if (!lex.checkFor("Translation")) {
 453                         if (lex.isOK())
 454                                 lex.printError("Unknown layout translation tag `$$Token'");
 455                         break;
 456                 }
 457                 if (!lex.next(true))
 458                         break;
 459                 string const code = lex.getString();
 460                 bool found = getFromCode(code);
 461                 if (!found) {
 462                         lex.printError("Unknown language `" + code + "'");
 463                         break;
 464                 }
 465                 if (!readTranslations(lex, trans[code])) {
 466                         lex.printError("Could not read layout translations for language `"
 467                                 + code + "'");
 468                         break;
 469                 }
 470         }
 471
 472         // 2) merge all translations into the languages
 473         // exact translations overwrite approximate ones
 474         for (auto & tr : trans) {
 475                 for (auto & lang : languagelist_) {
 476                         Match const m = match(tr.first, lang.second);
 477                         if (m == NoMatch)
 478                                 continue;
 479                         lang.second.readLayoutTranslations(tr.second, m == ExactMatch);
 480                 }
 481         }
 482
 483 }
 484
 485
 486 Language const * Languages::getLanguage(string const & language) const
 487 {
 488         if (language == "reset")
 489                 return reset_language;
 490         if (language == "ignore")
 491                 return ignore_language;
 492         const_iterator it = languagelist_.find(language);
 493         return it == languagelist_.end() ? reset_language : &it->second;
 494 }
 495
 496
 497 } // namespace lyx