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