src/Trans.cpp

   1 /**
   2  * \file Trans.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 Matthias Ettrich
   8  *
   9  * Full author contact details are available in file CREDITS.
  10  */
  11
  12 #include <config.h>
  13
  14 #include "Trans.h"
  15
  16 #include "Buffer.h"
  17 #include "BufferView.h"
  18 #include "Cursor.h"
  19 #include "CutAndPaste.h"
  20 #include "debug.h"
  21 #include "Lexer.h"
  22 #include "LyXRC.h"
  23 #include "Text.h"
  24
  25 #include "support/filetools.h"
  26 #include "support/lstrings.h"
  27 #include "support/convert.h"
  28 #include "support/docstream.h"
  29
  30 namespace lyx {
  31
  32 using support::split;
  33 using support::contains;
  34 using support::libFileSearch;
  35
  36 using std::endl;
  37 using std::string;
  38 using std::pair;
  39 using std::map;
  40
  41
  42 /////////////////////////////////////////////////////////////////////
  43 //
  44 // TeXAccents
  45 //
  46 /////////////////////////////////////////////////////////////////////
  47
  48 /* the names used by TeX and XWindows for deadkeys/accents are not the same
  49    so here follows a table to clearify the differences. Please correct this
  50    if I got it wrong
  51
  52    |------------------|------------------|------------------|--------------|
  53    |      TeX         |     XWindows     |   \bind/LFUN     | used by intl |
  54    |------------------|------------------|------------------|--------------|
  55    |    grave         |    grave         |LFUN_ACCENT_GRAVE        | grave
  56    |    acute         |    acute         |LFUN_ACCENT_ACUTE        | acute
  57    |    circumflex    |    circumflex    |LFUN_ACCENT_CIRCUMFLEX   | circumflex
  58    | umlaut/dieresis  |    diaeresis     |LFUN_ACCENT_UMLAUT       | umlaut
  59    |    tilde         |    tilde         |LFUN_ACCENT_TILDE        | tilde
  60    |    macron        |    maron         |LFUN_ACCENT_MACRON       | macron
  61    |    dot           |    abovedot      |LFUN_ACCENT_DOT          | dot
  62    |    cedilla       |    cedilla       |LFUN_ACCENT_CEDILLA      | cedilla
  63    |    underdot      |                  |LFUN_ACCENT_UNDERDOT     | underdot
  64    |    underbar      |                  |LFUN_ACCENT_UNDERBAR     | underbar
  65    |    hácek         |    caron         |LFUN_ACCENT_CARON        | caron
  66    |    breve         |    breve         |LFUN_ACCENT_BREVE        | breve
  67    |    tie           |                  |LFUN_ACCENT_TIE          | tie
  68    | Hungarian umlaut |    doubleacute   |LFUN_ACCENT_HUNGARIAN_UMLAUT  | hungarian umlaut
  69    |    circle        |    abovering     |LFUN_ACCENT_CIRCLE       | circle
  70    |                  |    ogonek        |                  |
  71    |                  |    iota          |                  |
  72    |                  |    voiced_sound  |                  |
  73    |                  | semivoiced_sound |                  |
  74    |                  |                  |LFUN_ACCENT_SPECIAL_CARON| special caron
  75    */
  76 static tex_accent_struct lyx_accent_table[] = {
  77         {TEX_NOACCENT,   0,      "",                LFUN_NOACTION},
  78         {TEX_ACUTE,      0x0301, "acute",           LFUN_ACCENT_ACUTE},
  79         {TEX_GRAVE,      0x0300, "grave",           LFUN_ACCENT_GRAVE},
  80         {TEX_MACRON,     0x0304, "macron",          LFUN_ACCENT_MACRON},
  81         {TEX_TILDE,      0x0303, "tilde",           LFUN_ACCENT_TILDE},
  82         {TEX_UNDERBAR,   0x0320, "underbar",        LFUN_ACCENT_UNDERBAR},
  83         {TEX_CEDILLA,    0x0327, "cedilla",         LFUN_ACCENT_CEDILLA},
  84         {TEX_UNDERDOT,   0x0323, "underdot",        LFUN_ACCENT_UNDERDOT},
  85         {TEX_CIRCUMFLEX, 0x0302, "circumflex",      LFUN_ACCENT_CIRCUMFLEX},
  86         {TEX_CIRCLE,     0x030a, "circle",          LFUN_ACCENT_CIRCLE},
  87         {TEX_TIE,        0x0361, "tie",             LFUN_ACCENT_TIE},
  88         {TEX_BREVE,      0x0306, "breve",           LFUN_ACCENT_BREVE},
  89         {TEX_CARON,      0x030c, "caron",           LFUN_ACCENT_CARON},
  90 //      {TEX_SPECIAL_CARON, 0x030c, "ooo",          LFUN_ACCENT_SPECIAL_CARON},
  91         // Don't fix this typo for compatibility reasons!
  92         {TEX_HUNGUML,    0x030b, "hugarian_umlaut", LFUN_ACCENT_HUNGARIAN_UMLAUT},
  93         {TEX_UMLAUT,     0x0308, "umlaut",          LFUN_ACCENT_UMLAUT},
  94         {TEX_DOT,        0x0307, "dot",             LFUN_ACCENT_DOT},
  95         {TEX_OGONEK,     0x0328, "ogonek",          LFUN_ACCENT_OGONEK}
  96 };
  97
  98
  99 tex_accent_struct get_accent(kb_action action)
 100 {
 101         int i = 0;
 102         while (i <= TEX_MAX_ACCENT) {
 103                 if (lyx_accent_table[i].action == action)
 104                         return lyx_accent_table[i];
 105                 ++i;
 106         }
 107         struct tex_accent_struct temp = { static_cast<tex_accent>(0), 0,
 108                                           0, static_cast<kb_action>(0)};
 109         return temp;
 110 }
 111
 112
 113 static docstring const doAccent(docstring const & s, tex_accent accent)
 114 {
 115         if (s.empty())
 116                 return docstring(1, lyx_accent_table[accent].ucs4);
 117
 118         odocstringstream os;
 119         os.put(s[0]);
 120         os.put(lyx_accent_table[accent].ucs4);
 121         if (s.length() > 1) {
 122                 if (accent != TEX_TIE || s.length() > 2)
 123                         lyxerr << "Warning: Too many characters given for accent "
 124                                << lyx_accent_table[accent].name << '.' << std::endl;
 125                 os << s.substr(1);
 126         }
 127         return normalize_c(os.str());
 128 }
 129
 130
 131 static docstring const doAccent(char_type c, tex_accent accent)
 132 {
 133         return doAccent(docstring(1, c), accent);
 134 }
 135
 136
 137
 138 /////////////////////////////////////////////////////////////////////
 139 //
 140 // Trans
 141 //
 142 /////////////////////////////////////////////////////////////////////
 143
 144
 145 void Trans::insertException(KmodException & exclist, char_type c,
 146         docstring const & data, bool flag, tex_accent accent)
 147 {
 148         Keyexc p;
 149         p.c = c;
 150         p.data = data;
 151         p.combined = flag;
 152         p.accent = accent;
 153         exclist.insert(exclist.begin(), p);
 154         // or just
 155         // exclist.push_back(p);
 156 }
 157
 158
 159 void Trans::freeException(KmodException & exclist)
 160 {
 161         exclist.clear();
 162 }
 163
 164
 165 void Trans::freeKeymap()
 166 {
 167         kmod_list_.clear();
 168         keymap_.clear();
 169 }
 170
 171
 172 bool Trans::isDefined() const
 173 {
 174         return !name_.empty();
 175 }
 176
 177
 178 enum kmaptags_ {
 179         KCOMB = 1,
 180         KMOD,
 181         KMAP,
 182         KXMOD,
 183         K_LAST
 184 };
 185
 186
 187 struct keyword_item kmapTags[K_LAST - 1] = {
 188         {"\\kcomb", KCOMB },
 189         { "\\kmap", KMAP },
 190         { "\\kmod", KMOD },
 191         { "\\kxmod", KXMOD }
 192 };
 193
 194
 195 tex_accent getkeymod(string const &);
 196
 197
 198 void Trans::addDeadkey(tex_accent accent, docstring const & keys)
 199 {
 200         KmodInfo tmp;
 201         tmp.data = keys;
 202         tmp.accent = accent;
 203         kmod_list_[accent] = tmp;
 204
 205         for (docstring::size_type i = 0; i < keys.length(); ++i) {
 206                 // FIXME This is a hack.
 207                 // tmp is no valid UCS4 string, but misused to store the
 208                 // accent.
 209                 docstring tmp;
 210                 tmp += char_type(0);
 211                 tmp += char_type(accent);
 212                 keymap_[keys[i]] = tmp;
 213         }
 214 }
 215
 216
 217 int Trans::load(Lexer & lex)
 218 {
 219         bool error = false;
 220
 221         while (lex.isOK() && !error) {
 222                 switch (lex.lex()) {
 223                 case KMOD:
 224                 {
 225                         LYXERR(Debug::KBMAP, "KMOD:\t" << lex.getString());
 226                         if (!lex.next(true))
 227                                 return -1;
 228
 229                         LYXERR(Debug::KBMAP, "key\t`" << lex.getString() << '\'');
 230
 231                         docstring const keys = lex.getDocString();
 232
 233                         if (!lex.next(true))
 234                                 return -1;
 235
 236                         LYXERR(Debug::KBMAP, "accent\t`" << lex.getString() << '\'');
 237
 238                         tex_accent accent = getkeymod(lex.getString());
 239
 240                         if (accent == TEX_NOACCENT)
 241                                 return -1;
 242
 243 #if 1
 244                         // FIXME: This code should be removed...
 245                         // But we need to fix up all the kmap files first
 246                         // so that this field is not present anymore.
 247                         if (!lex.next(true))
 248                                 return -1;
 249
 250                         LYXERR(Debug::KBMAP, "allowed\t`" << lex.getString() << '\'');
 251
 252                         /* string const allowed = lex.getString(); */
 253                         addDeadkey(accent, keys /*, allowed*/);
 254 #else
 255                         addDeadkey(accent, keys);
 256 #endif
 257                         break;
 258                 }
 259                 case KCOMB: {
 260                         string str;
 261
 262                         LYXERR(Debug::KBMAP, "KCOMB:");
 263                         if (!lex.next(true))
 264                                 return -1;
 265
 266                         str = lex.getString();
 267                         LYXERR(Debug::KBMAP, str);
 268
 269                         tex_accent accent_1 = getkeymod(str);
 270                         if (accent_1 == TEX_NOACCENT)
 271                                 return -1;
 272
 273                         if (!lex.next(true))
 274                                 return -1;
 275
 276                         str = lex.getString();
 277                         LYXERR(Debug::KBMAP, str);
 278
 279                         tex_accent accent_2 = getkeymod(str);
 280                         if (accent_2 == TEX_NOACCENT) return -1;
 281
 282                         map<tex_accent, KmodInfo>::iterator it1 =
 283                                 kmod_list_.find(accent_1);
 284                         map<tex_accent, KmodInfo>::iterator it2 =
 285                                 kmod_list_.find(accent_2);
 286                         if (it1 == kmod_list_.end() || it2 == kmod_list_.end())
 287                                 return -1;
 288
 289                         // Find what key accent_2 is on - should
 290                         // check about accent_1 also
 291                         map<char_type, docstring>::iterator it = keymap_.begin();
 292                         map<char_type, docstring>::iterator end = keymap_.end();
 293                         for (; it != end; ++it) {
 294                                 if (!it->second.empty()
 295                                     && it->second[0] == 0
 296                                     && it->second[1] == accent_2)
 297                                         break;
 298                         }
 299                         docstring allowed;
 300                         if (!lex.next())
 301                                 return -1;
 302
 303                         allowed = lex.getDocString();
 304                         LYXERR(Debug::KBMAP, "allowed: " << to_utf8(allowed));
 305
 306                         insertException(kmod_list_[accent_1].exception_list,
 307                                         it->first, allowed, true, accent_2);
 308                 }
 309                 break;
 310                 case KMAP: {
 311                         unsigned char key_from;
 312
 313                         LYXERR(Debug::KBMAP, "KMAP:\t" << lex.getString());
 314
 315                         if (!lex.next(true))
 316                                 return -1;
 317
 318                         key_from = lex.getString()[0];
 319                         LYXERR(Debug::KBMAP, "\t`" << lex.getString() << '\'');
 320
 321                         if (!lex.next(true))
 322                                 return -1;
 323
 324                         docstring const string_to = lex.getDocString();
 325                         keymap_[key_from] = string_to;
 326                         LYXERR(Debug::KBMAP, "\t`" << to_utf8(string_to) << '\'');
 327                         break;
 328                 }
 329                 case KXMOD: {
 330                         tex_accent accent;
 331                         char_type key;
 332                         docstring str;
 333
 334                         LYXERR(Debug::KBMAP, "KXMOD:\t" << lex.getString());
 335
 336                         if (!lex.next(true))
 337                                 return -1;
 338
 339                         LYXERR(Debug::KBMAP, "\t`" << lex.getString() << '\'');
 340                         accent = getkeymod(lex.getString());
 341
 342                         if (!lex.next(true))
 343                                 return -1;
 344
 345                         LYXERR(Debug::KBMAP, "\t`" << lex.getString() << '\'');
 346                         key = lex.getDocString()[0];
 347
 348                         if (!lex.next(true))
 349                                 return -1;
 350
 351                         LYXERR(Debug::KBMAP, "\t`" << lex.getString() << '\'');
 352                         str = lex.getDocString();
 353
 354                         insertException(kmod_list_[accent].exception_list,
 355                                         key, str);
 356                         break;
 357                 }
 358                 case Lexer::LEX_FEOF:
 359                         LYXERR(Debug::PARSER, "End of parsing");
 360                         break;
 361                 default:
 362                         lex.printError("ParseKeymapFile: Unknown tag: `$$Token'");
 363                         return -1;
 364                 }
 365         }
 366         return 0;
 367 }
 368
 369
 370 bool Trans::isAccentDefined(tex_accent accent, KmodInfo & i) const
 371 {
 372         map<tex_accent, KmodInfo>::const_iterator cit = kmod_list_.find(accent);
 373         if (cit == kmod_list_.end())
 374                 return false;
 375         i = cit->second;
 376         return true;
 377 }
 378
 379
 380 docstring const Trans::process(char_type c, TransManager & k)
 381 {
 382         docstring const t = match(c);
 383
 384         if (t.empty() && c != 0)
 385                 return k.normalkey(c);
 386
 387         if (!t.empty() && t[0] != 0)
 388                 return t; //return k.normalkey(c);
 389
 390         return k.deadkey(c, kmod_list_[static_cast<tex_accent>(t[1])]);
 391 }
 392
 393
 394 int Trans::load(string const & language)
 395 {
 396         support::FileName const filename = libFileSearch("kbd", language, "kmap");
 397         if (filename.empty())
 398                 return -1;
 399
 400         freeKeymap();
 401         Lexer lex(kmapTags, K_LAST - 1);
 402         lex.setFile(filename);
 403
 404         int const res = load(lex);
 405
 406         if (res == 0)
 407                 name_ = language;
 408         else
 409                 name_.erase();
 410
 411         return res;
 412 }
 413
 414
 415 tex_accent getkeymod(string const & p)
 416         /* return modifier - decoded from p and update p */
 417 {
 418         for (int i = 1; i <= TEX_MAX_ACCENT; ++i) {
 419                 LYXERR(Debug::KBMAP, "p = " << p
 420                        << ", lyx_accent_table[" << i
 421                        << "].name = `" << lyx_accent_table[i].name << '\'');
 422
 423                 if (lyx_accent_table[i].name
 424                      && contains(p, lyx_accent_table[i].name)) {
 425                         LYXERR(Debug::KBMAP, "Found it!");
 426                         return static_cast<tex_accent>(i);
 427                 }
 428         }
 429         return TEX_NOACCENT;
 430 }
 431
 432
 433 /////////////////////////////////////////////////////////////////////
 434 //
 435 // TransState
 436 //
 437 /////////////////////////////////////////////////////////////////////
 438
 439
 440 // TransFSMData
 441 TransFSMData::TransFSMData()
 442 {
 443         deadkey_ = deadkey2_ = 0;
 444         deadkey_info_.accent = deadkey2_info_.accent = TEX_NOACCENT;
 445 }
 446
 447
 448 // TransState
 449 char_type const TransState::TOKEN_SEP = 4;
 450
 451
 452 // TransInitState
 453 TransInitState::TransInitState()
 454 {
 455         init_state_ = this;
 456 }
 457
 458
 459 docstring const TransInitState::normalkey(char_type c)
 460 {
 461         docstring res;
 462         res = c;
 463         return res;
 464 }
 465
 466
 467 docstring const TransInitState::deadkey(char_type c, KmodInfo d)
 468 {
 469         deadkey_ = c;
 470         deadkey_info_ = d;
 471         currentState = deadkey_state_;
 472         return docstring();
 473 }
 474
 475
 476 // TransDeadkeyState
 477 TransDeadkeyState::TransDeadkeyState()
 478 {
 479         deadkey_state_ = this;
 480 }
 481
 482
 483 docstring const TransDeadkeyState::normalkey(char_type c)
 484 {
 485         docstring res;
 486
 487         KmodException::iterator it = deadkey_info_.exception_list.begin();
 488         KmodException::iterator end = deadkey_info_.exception_list.end();
 489
 490         for (; it != end; ++it) {
 491                 if (it->c == c) {
 492                         res = it->data;
 493                         break;
 494                 }
 495         }
 496         if (it == end) {
 497                 res = doAccent(c, deadkey_info_.accent);
 498         }
 499         currentState = init_state_;
 500         return res;
 501 }
 502
 503
 504 docstring const TransDeadkeyState::deadkey(char_type c, KmodInfo d)
 505 {
 506         docstring res;
 507
 508         // Check if the same deadkey was typed twice
 509         if (deadkey_ == c) {
 510                 res = deadkey_;
 511                 deadkey_ = 0;
 512                 deadkey_info_.accent = TEX_NOACCENT;
 513                 currentState = init_state_;
 514                 return res;
 515         }
 516
 517         // Check if it is a combination or an exception
 518         KmodException::const_iterator cit = deadkey_info_.exception_list.begin();
 519         KmodException::const_iterator end = deadkey_info_.exception_list.end();
 520         for (; cit != end; ++cit) {
 521                 if (cit->combined == true && cit->accent == d.accent) {
 522                         deadkey2_ = c;
 523                         deadkey2_info_ = d;
 524                         comb_info_ = (*cit);
 525                         currentState = combined_state_;
 526                         return docstring();
 527                 }
 528                 if (cit->c == c) {
 529                         res = cit->data;
 530                         deadkey_ = 0;
 531                         deadkey_info_.accent = TEX_NOACCENT;
 532                         currentState = init_state_;
 533                         return res;
 534                 }
 535         }
 536
 537         // Not a combination or an exception.
 538         // Output deadkey1 and keep deadkey2
 539
 540         if (deadkey_!= 0)
 541                 res = deadkey_;
 542         deadkey_ = c;
 543         deadkey_info_ = d;
 544         currentState = deadkey_state_;
 545         return res;
 546 }
 547
 548
 549 TransCombinedState::TransCombinedState()
 550 {
 551         combined_state_ = this;
 552 }
 553
 554
 555 docstring const TransCombinedState::normalkey(char_type c)
 556 {
 557         docstring const temp = doAccent(c, deadkey2_info_.accent);
 558         docstring const res = doAccent(temp, deadkey_info_.accent);
 559         currentState = init_state_;
 560         return res;
 561 }
 562
 563
 564 docstring const TransCombinedState::deadkey(char_type c, KmodInfo d)
 565 {
 566         // Third key in a row. Output the first one and
 567         // reenter with shifted deadkeys
 568         docstring res;
 569         if (deadkey_ != 0)
 570                 res = deadkey_;
 571         res += TOKEN_SEP;
 572         deadkey_ = deadkey2_;
 573         deadkey_info_ = deadkey2_info_;
 574         res += deadkey_state_->deadkey(c, d);
 575         return res;
 576 }
 577
 578
 579 // TransFSM
 580 TransFSM::TransFSM()
 581         : TransFSMData(), TransInitState(), TransDeadkeyState(), TransCombinedState()
 582 {
 583         currentState = init_state_;
 584 }
 585
 586
 587 // TransManager
 588
 589 // Initialize static member.
 590 Trans TransManager::default_;
 591
 592
 593 TransManager::TransManager()
 594         : active_(0)
 595 {}
 596
 597
 598 int TransManager::setPrimary(string const & language)
 599 {
 600         if (t1_.getName() == language)
 601                 return 0;
 602
 603         return t1_.load(language);
 604 }
 605
 606
 607 int TransManager::setSecondary(string const & language)
 608 {
 609         if (t2_.getName() == language)
 610                 return 0;
 611
 612         return t2_.load(language);
 613 }
 614
 615
 616 void TransManager::enablePrimary()
 617 {
 618         if (t1_.isDefined())
 619                 active_ = &t1_;
 620
 621         LYXERR(Debug::KBMAP, "Enabling primary keymap");
 622 }
 623
 624
 625 void TransManager::enableSecondary()
 626 {
 627         if (t2_.isDefined())
 628                 active_ = &t2_;
 629         LYXERR(Debug::KBMAP, "Enabling secondary keymap");
 630 }
 631
 632
 633 void TransManager::disableKeymap()
 634 {
 635         active_ = &default_;
 636         LYXERR(Debug::KBMAP, "Disabling keymap");
 637 }
 638
 639
 640 void  TransManager::translateAndInsert(char_type c, Text * text, Cursor & cur)
 641 {
 642         docstring res = active_->process(c, *this);
 643
 644         // Process with tokens
 645         docstring temp;
 646
 647         while (res.length() > 0) {
 648                 res = split(res, temp, TransState::TOKEN_SEP);
 649                 insert(temp, text, cur);
 650         }
 651 }
 652
 653
 654 void TransManager::insert(docstring const & str, Text * text, Cursor & cur)
 655 {
 656         for (size_t i = 0, n = str.size(); i != n; ++i)
 657                 text->insertChar(cur, str[i]);
 658 }
 659
 660
 661 void TransManager::deadkey(char_type c, tex_accent accent, Text * t, Cursor & cur)
 662 {
 663         if (c == 0 && active_ != &default_) {
 664                 // A deadkey was pressed that cannot be printed
 665                 // or a accent command was typed in the minibuffer
 666                 KmodInfo i;
 667                 if (active_->isAccentDefined(accent, i) == true) {
 668                         docstring const res = trans_fsm_
 669                                 .currentState->deadkey(c, i);
 670                         insert(res, t, cur);
 671                         return;
 672                 }
 673         }
 674
 675         if (active_ == &default_ || c == 0) {
 676                 KmodInfo i;
 677                 i.accent = accent;
 678                 i.data.erase();
 679                 docstring res = trans_fsm_.currentState->deadkey(c, i);
 680                 insert(res, t, cur);
 681         } else {
 682                 // Go through the translation
 683                 translateAndInsert(c, t, cur);
 684         }
 685 }
 686
 687
 688 } // namespace lyx