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 "Lexer.h"
  20 #include "LyXRC.h"
  21 #include "Text.h"
  22
  23 #include "support/debug.h"
  24 #include "support/docstream.h"
  25 #include "support/FileName.h"
  26 #include "support/filetools.h"
  27 #include "support/lstrings.h"
  28
  29 using namespace std;
  30 using namespace lyx::support;
  31
  32 namespace lyx {
  33
  34 /////////////////////////////////////////////////////////////////////
  35 //
  36 // TeXAccents
  37 //
  38 /////////////////////////////////////////////////////////////////////
  39
  40 /* the names used by TeX and XWindows for deadkeys/accents are not the same
  41    so here follows a table to clearify the differences. Please correct this
  42    if I got it wrong
  43
  44    |------------------|------------------|------------------|--------------|
  45    |      TeX         |     XWindows     |   \bind/LFUN     | used by intl |
  46    |------------------|------------------|------------------|--------------|
  47    |    grave         |    grave         |LFUN_ACCENT_GRAVE        | grave
  48    |    acute         |    acute         |LFUN_ACCENT_ACUTE        | acute
  49    |    circumflex    |    circumflex    |LFUN_ACCENT_CIRCUMFLEX   | circumflex
  50    | umlaut/dieresis  |    diaeresis     |LFUN_ACCENT_UMLAUT       | umlaut
  51    |    tilde         |    tilde         |LFUN_ACCENT_TILDE        | tilde
  52    |    macron        |    maron         |LFUN_ACCENT_MACRON       | macron
  53    |    dot           |    abovedot      |LFUN_ACCENT_DOT          | dot
  54    |    cedilla       |    cedilla       |LFUN_ACCENT_CEDILLA      | cedilla
  55    |    underdot      |                  |LFUN_ACCENT_UNDERDOT     | underdot
  56    |    underbar      |                  |LFUN_ACCENT_UNDERBAR     | underbar
  57    |    hácek         |    caron         |LFUN_ACCENT_CARON        | caron
  58    |    breve         |    breve         |LFUN_ACCENT_BREVE        | breve
  59    |    tie           |                  |LFUN_ACCENT_TIE          | tie
  60    | Hungarian umlaut |    doubleacute   |LFUN_ACCENT_HUNGARIAN_UMLAUT  | hungarian umlaut
  61    |    circle        |    abovering     |LFUN_ACCENT_CIRCLE       | circle
  62    |                  |    ogonek        |                  |
  63    |                  |    iota          |                  |
  64    |                  |    voiced_sound  |                  |
  65    |                  | semivoiced_sound |                  |
  66    */
  67 static TeXAccent lyx_accent_table[] = {
  68         {TEX_NOACCENT,   0,      "",                LFUN_NOACTION},
  69         {TEX_ACUTE,      0x0301, "acute",           LFUN_ACCENT_ACUTE},
  70         {TEX_GRAVE,      0x0300, "grave",           LFUN_ACCENT_GRAVE},
  71         {TEX_MACRON,     0x0304, "macron",          LFUN_ACCENT_MACRON},
  72         {TEX_TILDE,      0x0303, "tilde",           LFUN_ACCENT_TILDE},
  73         {TEX_PERISPOMENI, 0x0342, "perispomeni",    LFUN_ACCENT_PERISPOMENI},
  74         {TEX_UNDERBAR,   0x0320, "underbar",        LFUN_ACCENT_UNDERBAR}, // COMBINING MINUS SIGN BELOW or 0x0331 COMBINING MACRON BELOW ?
  75
  76         {TEX_CEDILLA,    0x0327, "cedilla",         LFUN_ACCENT_CEDILLA},
  77         {TEX_UNDERDOT,   0x0323, "underdot",        LFUN_ACCENT_UNDERDOT},
  78         {TEX_CIRCUMFLEX, 0x0302, "circumflex",      LFUN_ACCENT_CIRCUMFLEX},
  79         {TEX_CIRCLE,     0x030a, "circle",          LFUN_ACCENT_CIRCLE},
  80         {TEX_TIE,        0x0361, "tie",             LFUN_ACCENT_TIE},
  81         {TEX_BREVE,      0x0306, "breve",           LFUN_ACCENT_BREVE},
  82         {TEX_CARON,      0x030c, "caron",           LFUN_ACCENT_CARON},
  83         // Don't fix this typo for compatibility reasons!
  84         {TEX_HUNGUML,    0x030b, "hugarian_umlaut", LFUN_ACCENT_HUNGARIAN_UMLAUT},
  85         {TEX_UMLAUT,     0x0308, "umlaut",          LFUN_ACCENT_UMLAUT},
  86         {TEX_DOT,        0x0307, "dot",             LFUN_ACCENT_DOT},
  87         {TEX_OGONEK,     0x0328, "ogonek",          LFUN_ACCENT_OGONEK}
  88 };
  89
  90
  91 TeXAccent get_accent(FuncCode action)
  92 {
  93         int i = 0;
  94         while (i <= TEX_MAX_ACCENT) {
  95                 if (lyx_accent_table[i].action == action)
  96                         return lyx_accent_table[i];
  97                 ++i;
  98         }
  99         struct TeXAccent temp = { static_cast<tex_accent>(0), 0,
 100                                           nullptr, static_cast<FuncCode>(0)};
 101         return temp;
 102 }
 103
 104
 105 static docstring const doAccent(docstring const & s, tex_accent accent)
 106 {
 107         if (s.empty())
 108                 return docstring(1, lyx_accent_table[accent].ucs4);
 109
 110         odocstringstream os;
 111         os.put(s[0]);
 112         os.put(lyx_accent_table[accent].ucs4);
 113         if (s.length() > 1) {
 114                 if (accent != TEX_TIE || s.length() > 2)
 115                         lyxerr << "Warning: Too many characters given for accent "
 116                                << lyx_accent_table[accent].name << '.' << endl;
 117                 os << s.substr(1);
 118         }
 119         return normalize_c(os.str());
 120 }
 121
 122
 123 static docstring const doAccent(char_type c, tex_accent accent)
 124 {
 125         return doAccent(docstring(1, c), accent);
 126 }
 127
 128
 129
 130 /////////////////////////////////////////////////////////////////////
 131 //
 132 // Trans
 133 //
 134 /////////////////////////////////////////////////////////////////////
 135
 136
 137 void Trans::insertException(KmodException & exclist, char_type c,
 138         docstring const & data, bool flag, tex_accent accent)
 139 {
 140         Keyexc p;
 141         p.c = c;
 142         p.data = data;
 143         p.combined = flag;
 144         p.accent = accent;
 145         exclist.insert(exclist.begin(), p);
 146         // or just
 147         // exclist.push_back(p);
 148 }
 149
 150
 151 void Trans::freeException(KmodException & exclist)
 152 {
 153         exclist.clear();
 154 }
 155
 156
 157 void Trans::freeKeymap()
 158 {
 159         kmod_list_.clear();
 160         keymap_.clear();
 161 }
 162
 163
 164 bool Trans::isDefined() const
 165 {
 166         return !name_.empty();
 167 }
 168
 169
 170 enum {
 171         KCOMB = 1,
 172         KMOD,
 173         KMAP,
 174         KXMOD
 175 };
 176
 177
 178 tex_accent getkeymod(string const &);
 179
 180
 181 void Trans::addDeadkey(tex_accent accent, docstring const & keys)
 182 {
 183         KmodInfo tmp;
 184         tmp.data = keys;
 185         tmp.accent = accent;
 186         kmod_list_[accent] = tmp;
 187
 188         for (char_type key : keys) {
 189                 // FIXME This is a hack.
 190                 // tmp is no valid UCS4 string, but misused to store the
 191                 // accent.
 192                 docstring tmpd;
 193                 tmpd += char_type(0);
 194                 tmpd += char_type(accent);
 195                 keymap_[key] = tmpd;
 196         }
 197 }
 198
 199
 200 int Trans::load(Lexer & lex)
 201 {
 202         while (lex.isOK()) {
 203                 switch (lex.lex()) {
 204                 case KMOD:
 205                 {
 206                         LYXERR(Debug::KBMAP, "KMOD:\t" << lex.getString());
 207                         if (!lex.next(true))
 208                                 return -1;
 209
 210                         LYXERR(Debug::KBMAP, "key\t`" << lex.getString() << '\'');
 211
 212                         docstring const keys = lex.getDocString();
 213
 214                         if (!lex.next(true))
 215                                 return -1;
 216
 217                         LYXERR(Debug::KBMAP, "accent\t`" << lex.getString() << '\'');
 218
 219                         tex_accent accent = getkeymod(lex.getString());
 220
 221                         if (accent == TEX_NOACCENT)
 222                                 return -1;
 223
 224 #if 1
 225                         // FIXME: This code should be removed...
 226                         // But we need to fix up all the kmap files first
 227                         // so that this field is not present anymore.
 228                         if (!lex.next(true))
 229                                 return -1;
 230
 231                         LYXERR(Debug::KBMAP, "allowed\t`" << lex.getString() << '\'');
 232
 233                         /* string const allowed = lex.getString(); */
 234                         addDeadkey(accent, keys /*, allowed*/);
 235 #else
 236                         addDeadkey(accent, keys);
 237 #endif
 238                         break;
 239                 }
 240                 case KCOMB: {
 241                         string str;
 242
 243                         LYXERR(Debug::KBMAP, "KCOMB:");
 244                         if (!lex.next(true))
 245                                 return -1;
 246
 247                         str = lex.getString();
 248                         LYXERR(Debug::KBMAP, str);
 249
 250                         tex_accent accent_1 = getkeymod(str);
 251                         if (accent_1 == TEX_NOACCENT)
 252                                 return -1;
 253
 254                         if (!lex.next(true))
 255                                 return -1;
 256
 257                         str = lex.getString();
 258                         LYXERR(Debug::KBMAP, str);
 259
 260                         tex_accent accent_2 = getkeymod(str);
 261                         if (accent_2 == TEX_NOACCENT) return -1;
 262
 263                         map<tex_accent, KmodInfo>::iterator it1 =
 264                                 kmod_list_.find(accent_1);
 265                         map<tex_accent, KmodInfo>::iterator it2 =
 266                                 kmod_list_.find(accent_2);
 267                         if (it1 == kmod_list_.end() || it2 == kmod_list_.end())
 268                                 return -1;
 269
 270                         // Find what key accent_2 is on - should
 271                         // check about accent_1 also
 272                         map<char_type, docstring>::iterator it = keymap_.begin();
 273                         map<char_type, docstring>::iterator end = keymap_.end();
 274                         for (; it != end; ++it) {
 275                                 if (!it->second.empty()
 276                                     && it->second[0] == 0
 277                                     && it->second[1] == accent_2)
 278                                         break;
 279                         }
 280
 281                         // could not find accent2 on a key -- this should not happen.
 282                         if (it == end)
 283                                 return -1;
 284
 285                         docstring allowed;
 286                         if (!lex.next())
 287                                 return -1;
 288                         allowed = lex.getDocString();
 289                         LYXERR(Debug::KBMAP, "allowed: " << to_utf8(allowed));
 290
 291                         insertException(kmod_list_[accent_1].exception_list,
 292                                         it->first, allowed, true, accent_2);
 293                 }
 294                 break;
 295                 case KMAP: {
 296                         unsigned char key_from;
 297
 298                         LYXERR(Debug::KBMAP, "KMAP:\t" << lex.getString());
 299
 300                         if (!lex.next(true))
 301                                 return -1;
 302
 303                         key_from = static_cast<unsigned char>(lex.getString()[0]);
 304                         LYXERR(Debug::KBMAP, "\t`" << lex.getString() << '\'');
 305
 306                         if (!lex.next(true))
 307                                 return -1;
 308
 309                         docstring const string_to = lex.getDocString();
 310                         keymap_[key_from] = string_to;
 311                         LYXERR(Debug::KBMAP, "\t`" << to_utf8(string_to) << '\'');
 312                         break;
 313                 }
 314                 case KXMOD: {
 315                         tex_accent accent;
 316                         char_type key;
 317                         docstring str;
 318
 319                         LYXERR(Debug::KBMAP, "KXMOD:\t" << lex.getString());
 320
 321                         if (!lex.next(true))
 322                                 return -1;
 323
 324                         LYXERR(Debug::KBMAP, "\t`" << lex.getString() << '\'');
 325                         accent = getkeymod(lex.getString());
 326
 327                         if (!lex.next(true))
 328                                 return -1;
 329
 330                         LYXERR(Debug::KBMAP, "\t`" << lex.getString() << '\'');
 331                         key = lex.getDocString()[0];
 332
 333                         if (!lex.next(true))
 334                                 return -1;
 335
 336                         LYXERR(Debug::KBMAP, "\t`" << lex.getString() << '\'');
 337                         str = lex.getDocString();
 338
 339                         insertException(kmod_list_[accent].exception_list,
 340                                         key, str);
 341                         break;
 342                 }
 343                 case Lexer::LEX_FEOF:
 344                         LYXERR(Debug::PARSER, "End of parsing");
 345                         break;
 346                 default:
 347                         lex.printError("ParseKeymapFile: Unknown tag: `$$Token'");
 348                         return -1;
 349                 }
 350         }
 351         return 0;
 352 }
 353
 354
 355 bool Trans::isAccentDefined(tex_accent accent, KmodInfo & i) const
 356 {
 357         map<tex_accent, KmodInfo>::const_iterator cit = kmod_list_.find(accent);
 358         if (cit == kmod_list_.end())
 359                 return false;
 360         i = cit->second;
 361         return true;
 362 }
 363
 364
 365 docstring const Trans::process(char_type c, TransManager & k)
 366 {
 367         docstring const t = match(c);
 368
 369         if (t.empty() && c != 0)
 370                 return k.normalkey(c);
 371
 372         if (!t.empty() && t[0] != 0)
 373                 return t; //return k.normalkey(c);
 374
 375         return k.deadkey(c, kmod_list_[static_cast<tex_accent>(t[1])]);
 376 }
 377
 378
 379 int Trans::load(string const & language)
 380 {
 381         LexerKeyword kmapTags[] = {
 382                 {"\\kcomb", KCOMB },
 383                 { "\\kmap", KMAP },
 384                 { "\\kmod", KMOD },
 385                 { "\\kxmod", KXMOD }
 386         };
 387
 388         FileName const filename = libFileSearch("kbd", language, "kmap");
 389         if (filename.empty())
 390                 return -1;
 391
 392         freeKeymap();
 393         Lexer lex(kmapTags);
 394         lex.setFile(filename);
 395
 396         int const res = load(lex);
 397
 398         if (res == 0)
 399                 name_ = language;
 400         else
 401                 name_.erase();
 402
 403         return res;
 404 }
 405
 406
 407 tex_accent getkeymod(string const & p)
 408         /* return modifier - decoded from p and update p */
 409 {
 410         for (int i = 1; i <= TEX_MAX_ACCENT; ++i) {
 411                 LYXERR(Debug::KBMAP, "p = " << p
 412                        << ", lyx_accent_table[" << i
 413                        << "].name = `" << lyx_accent_table[i].name << '\'');
 414
 415                 if (lyx_accent_table[i].name
 416                      && contains(p, lyx_accent_table[i].name)) {
 417                         LYXERR(Debug::KBMAP, "Found it!");
 418                         return static_cast<tex_accent>(i);
 419                 }
 420         }
 421         return TEX_NOACCENT;
 422 }
 423
 424
 425 /////////////////////////////////////////////////////////////////////
 426 //
 427 // TransState
 428 //
 429 /////////////////////////////////////////////////////////////////////
 430
 431
 432 // TransFSMData
 433 TransFSMData::TransFSMData() : deadkey_(0), deadkey2_(0), init_state_(nullptr),
 434         deadkey_state_(nullptr), combined_state_(nullptr), currentState(nullptr)
 435 {
 436 }
 437
 438
 439 // TransState
 440 char_type const TransState::TOKEN_SEP = 4;
 441
 442
 443 // TransInitState
 444 TransInitState::TransInitState()
 445 {
 446         init_state_ = this;
 447 }
 448
 449
 450 docstring const TransInitState::normalkey(char_type c)
 451 {
 452         docstring res;
 453         res = c;
 454         return res;
 455 }
 456
 457
 458 docstring const TransInitState::deadkey(char_type c, KmodInfo d)
 459 {
 460         deadkey_ = c;
 461         deadkey_info_ = d;
 462         currentState = deadkey_state_;
 463         return docstring();
 464 }
 465
 466
 467 // TransDeadkeyState
 468 TransDeadkeyState::TransDeadkeyState()
 469 {
 470         deadkey_state_ = this;
 471 }
 472
 473
 474 docstring const TransDeadkeyState::normalkey(char_type c)
 475 {
 476         docstring res;
 477
 478         KmodException::iterator it = deadkey_info_.exception_list.begin();
 479         KmodException::iterator end = deadkey_info_.exception_list.end();
 480
 481         for (; it != end; ++it) {
 482                 if (it->c == c) {
 483                         res = it->data;
 484                         break;
 485                 }
 486         }
 487         if (it == end) {
 488                 res = doAccent(c, deadkey_info_.accent);
 489         }
 490         currentState = init_state_;
 491         return res;
 492 }
 493
 494
 495 docstring const TransDeadkeyState::deadkey(char_type c, KmodInfo d)
 496 {
 497         docstring res;
 498
 499         // Check if the same deadkey was typed twice
 500         if (deadkey_ == c) {
 501                 res = deadkey_;
 502                 deadkey_ = 0;
 503                 deadkey_info_.accent = TEX_NOACCENT;
 504                 currentState = init_state_;
 505                 return res;
 506         }
 507
 508         // Check if it is a combination or an exception
 509         KmodException::const_iterator cit = deadkey_info_.exception_list.begin();
 510         KmodException::const_iterator end = deadkey_info_.exception_list.end();
 511         for (; cit != end; ++cit) {
 512                 if (cit->combined && cit->accent == d.accent) {
 513                         deadkey2_ = c;
 514                         deadkey2_info_ = d;
 515                         comb_info_ = (*cit);
 516                         currentState = combined_state_;
 517                         return docstring();
 518                 }
 519                 if (cit->c == c) {
 520                         res = cit->data;
 521                         deadkey_ = 0;
 522                         deadkey_info_.accent = TEX_NOACCENT;
 523                         currentState = init_state_;
 524                         return res;
 525                 }
 526         }
 527
 528         // Not a combination or an exception.
 529         // Output deadkey1 and keep deadkey2
 530
 531         if (deadkey_!= 0)
 532                 res = deadkey_;
 533         deadkey_ = c;
 534         deadkey_info_ = d;
 535         currentState = deadkey_state_;
 536         return res;
 537 }
 538
 539
 540 TransCombinedState::TransCombinedState()
 541 {
 542         combined_state_ = this;
 543 }
 544
 545
 546 docstring const TransCombinedState::normalkey(char_type c)
 547 {
 548         docstring const temp = doAccent(c, deadkey2_info_.accent);
 549         docstring const res = doAccent(temp, deadkey_info_.accent);
 550         currentState = init_state_;
 551         return res;
 552 }
 553
 554
 555 docstring const TransCombinedState::deadkey(char_type c, KmodInfo d)
 556 {
 557         // Third key in a row. Output the first one and
 558         // reenter with shifted deadkeys
 559         docstring res;
 560         if (deadkey_ != 0)
 561                 res = deadkey_;
 562         res += TOKEN_SEP;
 563         deadkey_ = deadkey2_;
 564         deadkey_info_ = deadkey2_info_;
 565         res += deadkey_state_->deadkey(c, d);
 566         return res;
 567 }
 568
 569
 570 // TransFSM
 571 TransFSM::TransFSM()
 572         : TransFSMData(), TransInitState(), TransDeadkeyState(), TransCombinedState()
 573 {
 574         currentState = init_state_;
 575 }
 576
 577
 578 // TransManager
 579
 580 // Initialize static member.
 581 Trans TransManager::default_;
 582
 583
 584 TransManager::TransManager()
 585         : active_(&default_)
 586 {}
 587
 588
 589 int TransManager::setPrimary(string const & language)
 590 {
 591         if (t1_.getName() == language)
 592                 return 0;
 593
 594         return t1_.load(language);
 595 }
 596
 597
 598 int TransManager::setSecondary(string const & language)
 599 {
 600         if (t2_.getName() == language)
 601                 return 0;
 602
 603         return t2_.load(language);
 604 }
 605
 606
 607 void TransManager::enablePrimary()
 608 {
 609         if (t1_.isDefined())
 610                 active_ = &t1_;
 611
 612         LYXERR(Debug::KBMAP, "Enabling primary keymap");
 613 }
 614
 615
 616 void TransManager::enableSecondary()
 617 {
 618         if (t2_.isDefined())
 619                 active_ = &t2_;
 620         LYXERR(Debug::KBMAP, "Enabling secondary keymap");
 621 }
 622
 623
 624 void TransManager::disableKeymap()
 625 {
 626         active_ = &default_;
 627         LYXERR(Debug::KBMAP, "Disabling keymap");
 628 }
 629
 630
 631 void  TransManager::translateAndInsert(char_type c, Text * text, Cursor & cur)
 632 {
 633         docstring res = active_->process(c, *this);
 634
 635         // Process with tokens
 636         docstring temp;
 637
 638         while (res.length() > 0) {
 639                 res = split(res, temp, TransState::TOKEN_SEP);
 640                 insert(temp, text, cur);
 641         }
 642 }
 643
 644
 645 void TransManager::insert(docstring const & str, Text * text, Cursor & cur)
 646 {
 647         for (size_t i = 0, n = str.size(); i != n; ++i)
 648                 text->insertChar(cur, str[i]);
 649 }
 650
 651
 652 void TransManager::deadkey(char_type c, tex_accent accent, Text * t, Cursor & cur)
 653 {
 654         if (c == 0 && active_ != &default_) {
 655                 // A deadkey was pressed that cannot be printed
 656                 // or a accent command was typed in the minibuffer
 657                 KmodInfo i;
 658                 if (active_->isAccentDefined(accent, i)) {
 659                         docstring const res = trans_fsm_
 660                                 .currentState->deadkey(c, i);
 661                         insert(res, t, cur);
 662                         return;
 663                 }
 664         }
 665
 666         if (active_ == &default_ || c == 0) {
 667                 KmodInfo i;
 668                 i.accent = accent;
 669                 i.data.erase();
 670                 docstring res = trans_fsm_.currentState->deadkey(c, i);
 671                 insert(res, t, cur);
 672         } else {
 673                 // Go through the translation
 674                 translateAndInsert(c, t, cur);
 675         }
 676 }
 677
 678
 679 } // namespace lyx