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_UNDERBAR,   0x0320, "underbar",        LFUN_ACCENT_UNDERBAR},
  75         {TEX_CEDILLA,    0x0327, "cedilla",         LFUN_ACCENT_CEDILLA},
  76         {TEX_UNDERDOT,   0x0323, "underdot",        LFUN_ACCENT_UNDERDOT},
  77         {TEX_CIRCUMFLEX, 0x0302, "circumflex",      LFUN_ACCENT_CIRCUMFLEX},
  78         {TEX_CIRCLE,     0x030a, "circle",          LFUN_ACCENT_CIRCLE},
  79         {TEX_TIE,        0x0361, "tie",             LFUN_ACCENT_TIE},
  80         {TEX_BREVE,      0x0306, "breve",           LFUN_ACCENT_BREVE},
  81         {TEX_CARON,      0x030c, "caron",           LFUN_ACCENT_CARON},
  82         // Don't fix this typo for compatibility reasons!
  83         {TEX_HUNGUML,    0x030b, "hugarian_umlaut", LFUN_ACCENT_HUNGARIAN_UMLAUT},
  84         {TEX_UMLAUT,     0x0308, "umlaut",          LFUN_ACCENT_UMLAUT},
  85         {TEX_DOT,        0x0307, "dot",             LFUN_ACCENT_DOT},
  86         {TEX_OGONEK,     0x0328, "ogonek",          LFUN_ACCENT_OGONEK}
  87 };
  88
  89
  90 TeXAccent get_accent(FuncCode action)
  91 {
  92         int i = 0;
  93         while (i <= TEX_MAX_ACCENT) {
  94                 if (lyx_accent_table[i].action == action)
  95                         return lyx_accent_table[i];
  96                 ++i;
  97         }
  98         struct TeXAccent temp = { static_cast<tex_accent>(0), 0,
  99                                           0, static_cast<FuncCode>(0)};
 100         return temp;
 101 }
 102
 103
 104 static docstring const doAccent(docstring const & s, tex_accent accent)
 105 {
 106         if (s.empty())
 107                 return docstring(1, lyx_accent_table[accent].ucs4);
 108
 109         odocstringstream os;
 110         os.put(s[0]);
 111         os.put(lyx_accent_table[accent].ucs4);
 112         if (s.length() > 1) {
 113                 if (accent != TEX_TIE || s.length() > 2)
 114                         lyxerr << "Warning: Too many characters given for accent "
 115                                << lyx_accent_table[accent].name << '.' << endl;
 116                 os << s.substr(1);
 117         }
 118         return normalize_c(os.str());
 119 }
 120
 121
 122 static docstring const doAccent(char_type c, tex_accent accent)
 123 {
 124         return doAccent(docstring(1, c), accent);
 125 }
 126
 127
 128
 129 /////////////////////////////////////////////////////////////////////
 130 //
 131 // Trans
 132 //
 133 /////////////////////////////////////////////////////////////////////
 134
 135
 136 void Trans::insertException(KmodException & exclist, char_type c,
 137         docstring const & data, bool flag, tex_accent accent)
 138 {
 139         Keyexc p;
 140         p.c = c;
 141         p.data = data;
 142         p.combined = flag;
 143         p.accent = accent;
 144         exclist.insert(exclist.begin(), p);
 145         // or just
 146         // exclist.push_back(p);
 147 }
 148
 149
 150 void Trans::freeException(KmodException & exclist)
 151 {
 152         exclist.clear();
 153 }
 154
 155
 156 void Trans::freeKeymap()
 157 {
 158         kmod_list_.clear();
 159         keymap_.clear();
 160 }
 161
 162
 163 bool Trans::isDefined() const
 164 {
 165         return !name_.empty();
 166 }
 167
 168
 169 enum {
 170         KCOMB = 1,
 171         KMOD,
 172         KMAP,
 173         KXMOD,
 174 };
 175
 176
 177 tex_accent getkeymod(string const &);
 178
 179
 180 void Trans::addDeadkey(tex_accent accent, docstring const & keys)
 181 {
 182         KmodInfo tmp;
 183         tmp.data = keys;
 184         tmp.accent = accent;
 185         kmod_list_[accent] = tmp;
 186
 187         for (docstring::size_type i = 0; i < keys.length(); ++i) {
 188                 // FIXME This is a hack.
 189                 // tmp is no valid UCS4 string, but misused to store the
 190                 // accent.
 191                 docstring tmp;
 192                 tmp += char_type(0);
 193                 tmp += char_type(accent);
 194                 keymap_[keys[i]] = tmp;
 195         }
 196 }
 197
 198
 199 int Trans::load(Lexer & lex)
 200 {
 201         bool error = false;
 202
 203         while (lex.isOK() && !error) {
 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                         docstring allowed;
 282                         if (!lex.next())
 283                                 return -1;
 284
 285                         allowed = lex.getDocString();
 286                         LYXERR(Debug::KBMAP, "allowed: " << to_utf8(allowed));
 287
 288                         insertException(kmod_list_[accent_1].exception_list,
 289                                         it->first, allowed, true, accent_2);
 290                 }
 291                 break;
 292                 case KMAP: {
 293                         unsigned char key_from;
 294
 295                         LYXERR(Debug::KBMAP, "KMAP:\t" << lex.getString());
 296
 297                         if (!lex.next(true))
 298                                 return -1;
 299
 300                         key_from = lex.getString()[0];
 301                         LYXERR(Debug::KBMAP, "\t`" << lex.getString() << '\'');
 302
 303                         if (!lex.next(true))
 304                                 return -1;
 305
 306                         docstring const string_to = lex.getDocString();
 307                         keymap_[key_from] = string_to;
 308                         LYXERR(Debug::KBMAP, "\t`" << to_utf8(string_to) << '\'');
 309                         break;
 310                 }
 311                 case KXMOD: {
 312                         tex_accent accent;
 313                         char_type key;
 314                         docstring str;
 315
 316                         LYXERR(Debug::KBMAP, "KXMOD:\t" << lex.getString());
 317
 318                         if (!lex.next(true))
 319                                 return -1;
 320
 321                         LYXERR(Debug::KBMAP, "\t`" << lex.getString() << '\'');
 322                         accent = getkeymod(lex.getString());
 323
 324                         if (!lex.next(true))
 325                                 return -1;
 326
 327                         LYXERR(Debug::KBMAP, "\t`" << lex.getString() << '\'');
 328                         key = lex.getDocString()[0];
 329
 330                         if (!lex.next(true))
 331                                 return -1;
 332
 333                         LYXERR(Debug::KBMAP, "\t`" << lex.getString() << '\'');
 334                         str = lex.getDocString();
 335
 336                         insertException(kmod_list_[accent].exception_list,
 337                                         key, str);
 338                         break;
 339                 }
 340                 case Lexer::LEX_FEOF:
 341                         LYXERR(Debug::PARSER, "End of parsing");
 342                         break;
 343                 default:
 344                         lex.printError("ParseKeymapFile: Unknown tag: `$$Token'");
 345                         return -1;
 346                 }
 347         }
 348         return 0;
 349 }
 350
 351
 352 bool Trans::isAccentDefined(tex_accent accent, KmodInfo & i) const
 353 {
 354         map<tex_accent, KmodInfo>::const_iterator cit = kmod_list_.find(accent);
 355         if (cit == kmod_list_.end())
 356                 return false;
 357         i = cit->second;
 358         return true;
 359 }
 360
 361
 362 docstring const Trans::process(char_type c, TransManager & k)
 363 {
 364         docstring const t = match(c);
 365
 366         if (t.empty() && c != 0)
 367                 return k.normalkey(c);
 368
 369         if (!t.empty() && t[0] != 0)
 370                 return t; //return k.normalkey(c);
 371
 372         return k.deadkey(c, kmod_list_[static_cast<tex_accent>(t[1])]);
 373 }
 374
 375
 376 int Trans::load(string const & language)
 377 {
 378         LexerKeyword kmapTags[] = {
 379                 {"\\kcomb", KCOMB },
 380                 { "\\kmap", KMAP },
 381                 { "\\kmod", KMOD },
 382                 { "\\kxmod", KXMOD }
 383         };
 384
 385         FileName const filename = libFileSearch("kbd", language, "kmap");
 386         if (filename.empty())
 387                 return -1;
 388
 389         freeKeymap();
 390         Lexer lex(kmapTags);
 391         lex.setFile(filename);
 392
 393         int const res = load(lex);
 394
 395         if (res == 0)
 396                 name_ = language;
 397         else
 398                 name_.erase();
 399
 400         return res;
 401 }
 402
 403
 404 tex_accent getkeymod(string const & p)
 405         /* return modifier - decoded from p and update p */
 406 {
 407         for (int i = 1; i <= TEX_MAX_ACCENT; ++i) {
 408                 LYXERR(Debug::KBMAP, "p = " << p
 409                        << ", lyx_accent_table[" << i
 410                        << "].name = `" << lyx_accent_table[i].name << '\'');
 411
 412                 if (lyx_accent_table[i].name
 413                      && contains(p, lyx_accent_table[i].name)) {
 414                         LYXERR(Debug::KBMAP, "Found it!");
 415                         return static_cast<tex_accent>(i);
 416                 }
 417         }
 418         return TEX_NOACCENT;
 419 }
 420
 421
 422 /////////////////////////////////////////////////////////////////////
 423 //
 424 // TransState
 425 //
 426 /////////////////////////////////////////////////////////////////////
 427
 428
 429 // TransFSMData
 430 TransFSMData::TransFSMData()
 431 {
 432         deadkey_ = deadkey2_ = 0;
 433         deadkey_info_.accent = deadkey2_info_.accent = TEX_NOACCENT;
 434 }
 435
 436
 437 // TransState
 438 char_type const TransState::TOKEN_SEP = 4;
 439
 440
 441 // TransInitState
 442 TransInitState::TransInitState()
 443 {
 444         init_state_ = this;
 445 }
 446
 447
 448 docstring const TransInitState::normalkey(char_type c)
 449 {
 450         docstring res;
 451         res = c;
 452         return res;
 453 }
 454
 455
 456 docstring const TransInitState::deadkey(char_type c, KmodInfo d)
 457 {
 458         deadkey_ = c;
 459         deadkey_info_ = d;
 460         currentState = deadkey_state_;
 461         return docstring();
 462 }
 463
 464
 465 // TransDeadkeyState
 466 TransDeadkeyState::TransDeadkeyState()
 467 {
 468         deadkey_state_ = this;
 469 }
 470
 471
 472 docstring const TransDeadkeyState::normalkey(char_type c)
 473 {
 474         docstring res;
 475
 476         KmodException::iterator it = deadkey_info_.exception_list.begin();
 477         KmodException::iterator end = deadkey_info_.exception_list.end();
 478
 479         for (; it != end; ++it) {
 480                 if (it->c == c) {
 481                         res = it->data;
 482                         break;
 483                 }
 484         }
 485         if (it == end) {
 486                 res = doAccent(c, deadkey_info_.accent);
 487         }
 488         currentState = init_state_;
 489         return res;
 490 }
 491
 492
 493 docstring const TransDeadkeyState::deadkey(char_type c, KmodInfo d)
 494 {
 495         docstring res;
 496
 497         // Check if the same deadkey was typed twice
 498         if (deadkey_ == c) {
 499                 res = deadkey_;
 500                 deadkey_ = 0;
 501                 deadkey_info_.accent = TEX_NOACCENT;
 502                 currentState = init_state_;
 503                 return res;
 504         }
 505
 506         // Check if it is a combination or an exception
 507         KmodException::const_iterator cit = deadkey_info_.exception_list.begin();
 508         KmodException::const_iterator end = deadkey_info_.exception_list.end();
 509         for (; cit != end; ++cit) {
 510                 if (cit->combined == true && cit->accent == d.accent) {
 511                         deadkey2_ = c;
 512                         deadkey2_info_ = d;
 513                         comb_info_ = (*cit);
 514                         currentState = combined_state_;
 515                         return docstring();
 516                 }
 517                 if (cit->c == c) {
 518                         res = cit->data;
 519                         deadkey_ = 0;
 520                         deadkey_info_.accent = TEX_NOACCENT;
 521                         currentState = init_state_;
 522                         return res;
 523                 }
 524         }
 525
 526         // Not a combination or an exception.
 527         // Output deadkey1 and keep deadkey2
 528
 529         if (deadkey_!= 0)
 530                 res = deadkey_;
 531         deadkey_ = c;
 532         deadkey_info_ = d;
 533         currentState = deadkey_state_;
 534         return res;
 535 }
 536
 537
 538 TransCombinedState::TransCombinedState()
 539 {
 540         combined_state_ = this;
 541 }
 542
 543
 544 docstring const TransCombinedState::normalkey(char_type c)
 545 {
 546         docstring const temp = doAccent(c, deadkey2_info_.accent);
 547         docstring const res = doAccent(temp, deadkey_info_.accent);
 548         currentState = init_state_;
 549         return res;
 550 }
 551
 552
 553 docstring const TransCombinedState::deadkey(char_type c, KmodInfo d)
 554 {
 555         // Third key in a row. Output the first one and
 556         // reenter with shifted deadkeys
 557         docstring res;
 558         if (deadkey_ != 0)
 559                 res = deadkey_;
 560         res += TOKEN_SEP;
 561         deadkey_ = deadkey2_;
 562         deadkey_info_ = deadkey2_info_;
 563         res += deadkey_state_->deadkey(c, d);
 564         return res;
 565 }
 566
 567
 568 // TransFSM
 569 TransFSM::TransFSM()
 570         : TransFSMData(), TransInitState(), TransDeadkeyState(), TransCombinedState()
 571 {
 572         currentState = init_state_;
 573 }
 574
 575
 576 // TransManager
 577
 578 // Initialize static member.
 579 Trans TransManager::default_;
 580
 581
 582 TransManager::TransManager()
 583         : active_(0)
 584 {}
 585
 586
 587 int TransManager::setPrimary(string const & language)
 588 {
 589         if (t1_.getName() == language)
 590                 return 0;
 591
 592         return t1_.load(language);
 593 }
 594
 595
 596 int TransManager::setSecondary(string const & language)
 597 {
 598         if (t2_.getName() == language)
 599                 return 0;
 600
 601         return t2_.load(language);
 602 }
 603
 604
 605 void TransManager::enablePrimary()
 606 {
 607         if (t1_.isDefined())
 608                 active_ = &t1_;
 609
 610         LYXERR(Debug::KBMAP, "Enabling primary keymap");
 611 }
 612
 613
 614 void TransManager::enableSecondary()
 615 {
 616         if (t2_.isDefined())
 617                 active_ = &t2_;
 618         LYXERR(Debug::KBMAP, "Enabling secondary keymap");
 619 }
 620
 621
 622 void TransManager::disableKeymap()
 623 {
 624         active_ = &default_;
 625         LYXERR(Debug::KBMAP, "Disabling keymap");
 626 }
 627
 628
 629 void  TransManager::translateAndInsert(char_type c, Text * text, Cursor & cur)
 630 {
 631         docstring res = active_->process(c, *this);
 632
 633         // Process with tokens
 634         docstring temp;
 635
 636         while (res.length() > 0) {
 637                 res = split(res, temp, TransState::TOKEN_SEP);
 638                 insert(temp, text, cur);
 639         }
 640 }
 641
 642
 643 void TransManager::insert(docstring const & str, Text * text, Cursor & cur)
 644 {
 645         for (size_t i = 0, n = str.size(); i != n; ++i)
 646                 text->insertChar(cur, str[i]);
 647 }
 648
 649
 650 void TransManager::deadkey(char_type c, tex_accent accent, Text * t, Cursor & cur)
 651 {
 652         if (c == 0 && active_ != &default_) {
 653                 // A deadkey was pressed that cannot be printed
 654                 // or a accent command was typed in the minibuffer
 655                 KmodInfo i;
 656                 if (active_->isAccentDefined(accent, i) == true) {
 657                         docstring const res = trans_fsm_
 658                                 .currentState->deadkey(c, i);
 659                         insert(res, t, cur);
 660                         return;
 661                 }
 662         }
 663
 664         if (active_ == &default_ || c == 0) {
 665                 KmodInfo i;
 666                 i.accent = accent;
 667                 i.data.erase();
 668                 docstring res = trans_fsm_.currentState->deadkey(c, i);
 669                 insert(res, t, cur);
 670         } else {
 671                 // Go through the translation
 672                 translateAndInsert(c, t, cur);
 673         }
 674 }
 675
 676
 677 } // namespace lyx