src/insets/InsetBibtex.cpp

   1 /**
   2  * \file InsetBibtex.cpp
   3  * This file is part of LyX, the document processor.
   4  * Licence details can be found in the file COPYING.
   5  *
   6  * \author Alejandro Aguilar Sierra
   7  * \author Richard Heck (BibTeX parser improvements)
   8  *
   9  * Full author contact details are available in file CREDITS.
  10  */
  11
  12 #include <config.h>
  13
  14 #include "InsetBibtex.h"
  15
  16 #include "Buffer.h"
  17 #include "BufferParams.h"
  18 #include "DispatchResult.h"
  19 #include "Encoding.h"
  20 #include "Format.h"
  21 #include "FuncRequest.h"
  22 #include "FuncStatus.h"
  23 #include "LaTeXFeatures.h"
  24 #include "MetricsInfo.h"
  25 #include "OutputParams.h"
  26 #include "TextClass.h"
  27
  28 #include "frontends/alert.h"
  29
  30 #include "support/convert.h"
  31 #include "support/debug.h"
  32 #include "support/docstream.h"
  33 #include "support/ExceptionMessage.h"
  34 #include "support/filetools.h"
  35 #include "support/gettext.h"
  36 #include "support/lstrings.h"
  37 #include "support/os.h"
  38 #include "support/Path.h"
  39 #include "support/textutils.h"
  40
  41 #include <limits>
  42
  43 using namespace std;
  44 using namespace lyx::support;
  45
  46 namespace lyx {
  47
  48 namespace Alert = frontend::Alert;
  49 namespace os = support::os;
  50
  51
  52 InsetBibtex::InsetBibtex(Buffer const & buf, InsetCommandParams const & p)
  53         : InsetCommand(p, "bibtex")
  54 {
  55         Inset::setBuffer(const_cast<Buffer &>(buf));
  56         buffer_->invalidateBibinfoCache();
  57 }
  58
  59
  60 InsetBibtex::~InsetBibtex()
  61 {
  62         if (isBufferValid())
  63                 buffer_->invalidateBibinfoCache();
  64 }
  65
  66
  67 ParamInfo const & InsetBibtex::findInfo(string const & /* cmdName */)
  68 {
  69         static ParamInfo param_info_;
  70         if (param_info_.empty()) {
  71                 param_info_.add("btprint", ParamInfo::LATEX_OPTIONAL);
  72                 param_info_.add("bibfiles", ParamInfo::LATEX_REQUIRED);
  73                 param_info_.add("options", ParamInfo::LYX_INTERNAL);
  74         }
  75         return param_info_;
  76 }
  77
  78
  79 void InsetBibtex::doDispatch(Cursor & cur, FuncRequest & cmd)
  80 {
  81         switch (cmd.action) {
  82
  83         case LFUN_INSET_EDIT:
  84                 editDatabases();
  85                 break;
  86
  87         case LFUN_INSET_MODIFY: {
  88                 InsetCommandParams p(BIBTEX_CODE);
  89                 try {
  90                         if (!InsetCommand::string2params("bibtex",
  91                                         to_utf8(cmd.argument()), p)) {
  92                                 cur.noUpdate();
  93                                 break;
  94                         }
  95                 } catch (ExceptionMessage const & message) {
  96                         if (message.type_ == WarningException) {
  97                                 Alert::warning(message.title_, message.details_);
  98                                 cur.noUpdate();
  99                         } else
 100                                 throw message;
 101                         break;
 102                 }
 103                 //
 104                 setParams(p);
 105                 buffer().updateBibfilesCache();
 106                 break;
 107         }
 108
 109         default:
 110                 InsetCommand::doDispatch(cur, cmd);
 111                 break;
 112         }
 113 }
 114
 115
 116 bool InsetBibtex::getStatus(Cursor & cur, FuncRequest const & cmd,
 117                 FuncStatus & flag) const
 118 {
 119         switch (cmd.action) {
 120         case LFUN_INSET_EDIT:
 121                 flag.setEnabled(true);
 122                 return true;
 123
 124         default:
 125                 return InsetCommand::getStatus(cur, cmd, flag);
 126         }
 127 }
 128
 129
 130 void InsetBibtex::editDatabases() const
 131 {
 132         vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
 133
 134         if (bibfilelist.empty())
 135                 return;
 136
 137         int nr_databases = bibfilelist.size();
 138         if (nr_databases > 1) {
 139                         docstring message = bformat(_("The BibTeX inset includes %1$s databases.\n"
 140                                                        "If you proceed, all of them will be opened."),
 141                                                         convert<docstring>(nr_databases));
 142                         int const ret = Alert::prompt(_("Open Databases?"),
 143                                 message, 0, 1, _("&Cancel"), _("&Proceed"));
 144
 145                         if (ret == 0)
 146                                 return;
 147         }
 148
 149         vector<docstring>::const_iterator it = bibfilelist.begin();
 150         vector<docstring>::const_iterator en = bibfilelist.end();
 151         for (; it != en; ++it) {
 152                 FileName bibfile = getBibTeXPath(*it, buffer());
 153                 formats.edit(buffer(), bibfile,
 154                      formats.getFormatFromFile(bibfile));
 155         }
 156 }
 157
 158
 159 docstring InsetBibtex::screenLabel() const
 160 {
 161         return _("BibTeX Generated Bibliography");
 162 }
 163
 164
 165 docstring InsetBibtex::toolTip(BufferView const & /*bv*/, int /*x*/, int /*y*/) const
 166 {
 167         docstring item = from_ascii("* ");
 168         docstring tip = _("Databases:") + "\n";
 169         vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
 170
 171         if (bibfilelist.empty()) {
 172                 tip += item;
 173                 tip += _("none");
 174         } else {
 175                 vector<docstring>::const_iterator it = bibfilelist.begin();
 176                 vector<docstring>::const_iterator en = bibfilelist.end();
 177                 for (; it != en; ++it) {
 178                         tip += item;
 179                         tip += *it + "\n";
 180                 }
 181         }
 182
 183         // Style-Options
 184         bool toc = false;
 185         docstring style = getParam("options"); // maybe empty! and with bibtotoc
 186         docstring bibtotoc = from_ascii("bibtotoc");
 187         if (prefixIs(style, bibtotoc)) {
 188                 toc = true;
 189                 if (contains(style, char_type(',')))
 190                         style = split(style, bibtotoc, char_type(','));
 191         }
 192
 193         tip += _("Style File:") +"\n";
 194         tip += item;
 195         if (!style.empty())
 196                 tip += style;
 197         else
 198                 tip += _("none");
 199
 200         tip += "\n" + _("Lists:") + " ";
 201         docstring btprint = getParam("btprint");
 202                 if (btprint == "btPrintAll")
 203                         tip += _("all references");
 204                 else if (btprint == "btPrintNotCited")
 205                         tip += _("all uncited references");
 206                 else
 207                         tip += _("all cited references");
 208
 209         if (toc) {
 210                 tip += ", ";
 211                 tip += _("included in TOC");
 212         }
 213
 214         return tip;
 215 }
 216
 217
 218 static string normalizeName(Buffer const & buffer,
 219         OutputParams const & runparams, string const & name, string const & ext)
 220 {
 221         string const fname = makeAbsPath(name, buffer.filePath()).absFilename();
 222         if (FileName(name).isAbsolute() || !FileName(fname + ext).isReadableFile())
 223                 return name;
 224         if (!runparams.nice)
 225                 return fname;
 226
 227         // FIXME UNICODE
 228         return to_utf8(makeRelPath(from_utf8(fname),
 229                                          from_utf8(buffer.masterBuffer()->filePath())));
 230 }
 231
 232
 233 int InsetBibtex::latex(odocstream & os, OutputParams const & runparams) const
 234 {
 235         // the sequence of the commands:
 236         // 1. \bibliographystyle{style}
 237         // 2. \addcontentsline{...} - if option bibtotoc set
 238         // 3. \bibliography{database}
 239         // and with bibtopic:
 240         // 1. \bibliographystyle{style}
 241         // 2. \begin{btSect}{database}
 242         // 3. \btPrint{Cited|NotCited|All}
 243         // 4. \end{btSect}
 244
 245         // Database(s)
 246         // If we are processing the LaTeX file in a temp directory then
 247         // copy the .bib databases to this temp directory, mangling their
 248         // names in the process. Store this mangled name in the list of
 249         // all databases.
 250         // (We need to do all this because BibTeX *really*, *really*
 251         // can't handle "files with spaces" and Windows users tend to
 252         // use such filenames.)
 253         // Otherwise, store the (maybe absolute) path to the original,
 254         // unmangled database name.
 255         vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
 256         vector<docstring>::const_iterator it = bibfilelist.begin();
 257         vector<docstring>::const_iterator en = bibfilelist.end();
 258         odocstringstream dbs;
 259         bool didone = false;
 260
 261         for (; it != en; ++it) {
 262                 string utf8input = to_utf8(*it);
 263                 string database =
 264                         normalizeName(buffer(), runparams, utf8input, ".bib");
 265                 FileName const try_in_file =
 266                         makeAbsPath(database + ".bib", buffer().filePath());
 267                 bool const not_from_texmf = try_in_file.isReadableFile();
 268
 269                 if (!runparams.inComment && !runparams.dryrun && !runparams.nice &&
 270                     not_from_texmf) {
 271                         // mangledFilename() needs the extension
 272                         DocFileName const in_file = DocFileName(try_in_file);
 273                         database = removeExtension(in_file.mangledFilename());
 274                         FileName const out_file = makeAbsPath(database + ".bib",
 275                                         buffer().masterBuffer()->temppath());
 276
 277                         bool const success = in_file.copyTo(out_file);
 278                         if (!success) {
 279                                 lyxerr << "Failed to copy '" << in_file
 280                                        << "' to '" << out_file << "'"
 281                                        << endl;
 282                         }
 283                 } else if (!runparams.inComment && runparams.nice && not_from_texmf &&
 284                            !isValidLaTeXFilename(database)) {
 285                                 frontend::Alert::warning(_("Invalid filename"),
 286                                                          _("The following filename is likely to cause trouble "
 287                                                            "when running the exported file through LaTeX: ") +
 288                                                             from_utf8(database));
 289                 }
 290
 291                 if (didone)
 292                         dbs << ',';
 293                 else
 294                         didone = true;
 295                 // FIXME UNICODE
 296                 dbs << from_utf8(latex_path(database));
 297         }
 298         docstring const db_out = dbs.str();
 299
 300         // Post this warning only once.
 301         static bool warned_about_spaces = false;
 302         if (!warned_about_spaces &&
 303             runparams.nice && db_out.find(' ') != docstring::npos) {
 304                 warned_about_spaces = true;
 305                 Alert::warning(_("Export Warning!"),
 306                                _("There are spaces in the paths to your BibTeX databases.\n"
 307                                               "BibTeX will be unable to find them."));
 308         }
 309         // Style-Options
 310         string style = to_utf8(getParam("options")); // maybe empty! and with bibtotoc
 311         string bibtotoc;
 312         if (prefixIs(style, "bibtotoc")) {
 313                 bibtotoc = "bibtotoc";
 314                 if (contains(style, ','))
 315                         style = split(style, bibtotoc, ',');
 316         }
 317
 318         // line count
 319         int nlines = 0;
 320
 321         if (!style.empty()) {
 322                 string base = normalizeName(buffer(), runparams, style, ".bst");
 323                 FileName const try_in_file =
 324                         makeAbsPath(base + ".bst", buffer().filePath());
 325                 bool const not_from_texmf = try_in_file.isReadableFile();
 326                 // If this style does not come from texmf and we are not
 327                 // exporting to .tex copy it to the tmp directory.
 328                 // This prevents problems with spaces and 8bit charcaters
 329                 // in the file name.
 330                 if (!runparams.inComment && !runparams.dryrun && !runparams.nice &&
 331                     not_from_texmf) {
 332                         // use new style name
 333                         DocFileName const in_file = DocFileName(try_in_file);
 334                         base = removeExtension(in_file.mangledFilename());
 335                         FileName const out_file = makeAbsPath(base + ".bst",
 336                                         buffer().masterBuffer()->temppath());
 337                         bool const success = in_file.copyTo(out_file);
 338                         if (!success) {
 339                                 lyxerr << "Failed to copy '" << in_file
 340                                        << "' to '" << out_file << "'"
 341                                        << endl;
 342                         }
 343                 }
 344                 // FIXME UNICODE
 345                 os << "\\bibliographystyle{"
 346                    << from_utf8(latex_path(normalizeName(buffer(), runparams, base, ".bst")))
 347                    << "}\n";
 348                 nlines += 1;
 349         }
 350
 351         // Post this warning only once.
 352         static bool warned_about_bst_spaces = false;
 353         if (!warned_about_bst_spaces && runparams.nice && contains(style, ' ')) {
 354                 warned_about_bst_spaces = true;
 355                 Alert::warning(_("Export Warning!"),
 356                                _("There are spaces in the path to your BibTeX style file.\n"
 357                                               "BibTeX will be unable to find it."));
 358         }
 359
 360         if (!db_out.empty() && buffer().params().use_bibtopic) {
 361                 os << "\\begin{btSect}{" << db_out << "}\n";
 362                 docstring btprint = getParam("btprint");
 363                 if (btprint.empty())
 364                         // default
 365                         btprint = from_ascii("btPrintCited");
 366                 os << "\\" << btprint << "\n"
 367                    << "\\end{btSect}\n";
 368                 nlines += 3;
 369         }
 370
 371         // bibtotoc-Option
 372         if (!bibtotoc.empty() && !buffer().params().use_bibtopic) {
 373                 if (buffer().params().documentClass().hasLaTeXLayout("chapter")) {
 374                         if (buffer().params().sides == OneSide) {
 375                                 // oneside
 376                                 os << "\\clearpage";
 377                         } else {
 378                                 // twoside
 379                                 os << "\\cleardoublepage";
 380                         }
 381                         os << "\\addcontentsline{toc}{chapter}{\\bibname}";
 382                 } else if (buffer().params().documentClass().hasLaTeXLayout("section"))
 383                         os << "\\addcontentsline{toc}{section}{\\refname}";
 384         }
 385
 386         if (!db_out.empty() && !buffer().params().use_bibtopic) {
 387                 docstring btprint = getParam("btprint");
 388                 if (btprint == "btPrintAll") {
 389                         os << "\\nocite{*}\n";
 390                         nlines += 1;
 391                 }
 392                 os << "\\bibliography{" << db_out << "}\n";
 393                 nlines += 1;
 394         }
 395
 396         return nlines;
 397 }
 398
 399
 400 support::FileNameList InsetBibtex::getBibFiles() const
 401 {
 402         FileName path(buffer().filePath());
 403         support::PathChanger p(path);
 404
 405         support::FileNameList vec;
 406
 407         vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
 408         vector<docstring>::const_iterator it = bibfilelist.begin();
 409         vector<docstring>::const_iterator en = bibfilelist.end();
 410         for (; it != en; ++it) {
 411                 FileName const file =
 412                         findtexfile(changeExtension(to_utf8(*it), "bib"), "bib");
 413
 414                 if (!file.empty())
 415                         vec.push_back(file);
 416                 else
 417                         LYXERR0("Couldn't find " + to_utf8(*it) + " in InsetBibtex::getBibFiles()!");
 418         }
 419
 420         return vec;
 421
 422 }
 423
 424 namespace {
 425
 426         // methods for parsing bibtex files
 427
 428         typedef map<docstring, docstring> VarMap;
 429
 430         /// remove whitespace characters, optionally a single comma,
 431         /// and further whitespace characters from the stream.
 432         /// @return true if a comma was found, false otherwise
 433         ///
 434         bool removeWSAndComma(ifdocstream & ifs) {
 435                 char_type ch;
 436
 437                 if (!ifs)
 438                         return false;
 439
 440                 // skip whitespace
 441                 do {
 442                         ifs.get(ch);
 443                 } while (ifs && isSpace(ch));
 444
 445                 if (!ifs)
 446                         return false;
 447
 448                 if (ch != ',') {
 449                         ifs.putback(ch);
 450                         return false;
 451                 }
 452
 453                 // skip whitespace
 454                 do {
 455                         ifs.get(ch);
 456                 } while (ifs && isSpace(ch));
 457
 458                 if (ifs) {
 459                         ifs.putback(ch);
 460                 }
 461
 462                 return true;
 463         }
 464
 465
 466         enum charCase {
 467                 makeLowerCase,
 468                 keepCase
 469         };
 470
 471         /// remove whitespace characters, read characer sequence
 472         /// not containing whitespace characters or characters in
 473         /// delimChars, and remove further whitespace characters.
 474         ///
 475         /// @return true if a string of length > 0 could be read.
 476         ///
 477         bool readTypeOrKey(docstring & val, ifdocstream & ifs,
 478                 docstring const & delimChars, docstring const &illegalChars,
 479                 charCase chCase) {
 480
 481                 char_type ch;
 482
 483                 val.clear();
 484
 485                 if (!ifs)
 486                         return false;
 487
 488                 // skip whitespace
 489                 do {
 490                         ifs.get(ch);
 491                 } while (ifs && isSpace(ch));
 492
 493                 if (!ifs)
 494                         return false;
 495
 496                 // read value
 497                 bool legalChar = true;
 498                 while (ifs && !isSpace(ch) &&
 499                                                  delimChars.find(ch) == docstring::npos &&
 500                                                  (legalChar = (illegalChars.find(ch) == docstring::npos))
 501                                         )
 502                 {
 503                         if (chCase == makeLowerCase)
 504                                 val += lowercase(ch);
 505                         else
 506                                 val += ch;
 507                         ifs.get(ch);
 508                 }
 509
 510                 if (!legalChar) {
 511                         ifs.putback(ch);
 512                         return false;
 513                 }
 514
 515                 // skip whitespace
 516                 while (ifs && isSpace(ch)) {
 517                         ifs.get(ch);
 518                 }
 519
 520                 if (ifs) {
 521                         ifs.putback(ch);
 522                 }
 523
 524                 return val.length() > 0;
 525         }
 526
 527         /// read subsequent bibtex values that are delimited with a #-character.
 528         /// Concatenate all parts and replace names with the associated string in
 529         /// the variable strings.
 530         /// @return true if reading was successfull (all single parts were delimited
 531         /// correctly)
 532         bool readValue(docstring & value, ifdocstream & ifs, const VarMap & strings) {
 533
 534                 char_type ch;
 535
 536                 value.clear();
 537
 538                 if (!ifs)
 539                         return false;
 540
 541                 docstring val;
 542                 do {
 543                         // skip whitespace
 544                         do {
 545                                 ifs.get(ch);
 546                         } while (ifs && isSpace(ch));
 547
 548                         if (!ifs)
 549                                 return false;
 550
 551                         // check for field type
 552                         if (isDigit(ch)) {
 553
 554                                 // read integer value
 555                                 do {
 556                                         val += ch;
 557                                         ifs.get(ch);
 558                                 } while (ifs && isDigit(ch));
 559
 560                                 if (!ifs)
 561                                         return false;
 562
 563                         } else if (ch == '"' || ch == '{') {
 564                                 // set end delimiter
 565                                 char_type delim = ch == '"' ? '"': '}';
 566
 567                                 // Skip whitespace
 568                                 do {
 569                                         ifs.get(ch);
 570                                 } while (ifs && isSpace(ch));
 571
 572                                 if (!ifs)
 573                                         return false;
 574
 575                                 // We now have the first non-whitespace character
 576                                 // We'll collapse adjacent whitespace.
 577                                 bool lastWasWhiteSpace = false;
 578
 579                                 // inside this delimited text braces must match.
 580                                 // Thus we can have a closing delimiter only
 581                                 // when nestLevel == 0
 582                                 int nestLevel = 0;
 583
 584                                 while (ifs && (nestLevel > 0 || ch != delim)) {
 585                                         if (isSpace(ch)) {
 586                                                 lastWasWhiteSpace = true;
 587                                                 ifs.get(ch);
 588                                                 continue;
 589                                         }
 590                                         // We output the space only after we stop getting
 591                                         // whitespace so as not to output any whitespace
 592                                         // at the end of the value.
 593                                         if (lastWasWhiteSpace) {
 594                                                 lastWasWhiteSpace = false;
 595                                                 val += ' ';
 596                                         }
 597
 598                                         val += ch;
 599
 600                                         // update nesting level
 601                                         switch (ch) {
 602                                                 case '{':
 603                                                         ++nestLevel;
 604                                                         break;
 605                                                 case '}':
 606                                                         --nestLevel;
 607                                                         if (nestLevel < 0) return false;
 608                                                         break;
 609                                         }
 610
 611                                         ifs.get(ch);
 612                                 }
 613
 614                                 if (!ifs)
 615                                         return false;
 616
 617                                 ifs.get(ch);
 618
 619                                 if (!ifs)
 620                                         return false;
 621
 622                         } else {
 623
 624                                 // reading a string name
 625                                 docstring strName;
 626
 627                                 while (ifs && !isSpace(ch) && ch != '#' && ch != ',' && ch != '}' && ch != ')') {
 628                                         strName += lowercase(ch);
 629                                         ifs.get(ch);
 630                                 }
 631
 632                                 if (!ifs)
 633                                         return false;
 634
 635                                 // replace the string with its assigned value or
 636                                 // discard it if it's not assigned
 637                                 if (strName.length()) {
 638                                         VarMap::const_iterator pos = strings.find(strName);
 639                                         if (pos != strings.end()) {
 640                                                 val += pos->second;
 641                                         }
 642                                 }
 643                         }
 644
 645                         // skip WS
 646                         while (ifs && isSpace(ch)) {
 647                                 ifs.get(ch);
 648                         }
 649
 650                         if (!ifs)
 651                                 return false;
 652
 653                         // continue reading next value on concatenate with '#'
 654                 } while (ch == '#');
 655
 656                 ifs.putback(ch);
 657
 658                 // Ok, we now have the value. Now we are going to go
 659                 // through it and replace e.g. \"a with its unicode value.
 660                 // We'll also strip commands, like \emph, and the like, so
 661                 // it will look nice in the UI.
 662                 bool scanning_cmd = false;
 663                 bool scanning_math = false;
 664                 bool escaped = false; // used to catch \$, etc.
 665                 while (val.size()) {
 666                         char_type const ch = val[0];
 667
 668                         // if we're scanning math, we output everything until we
 669                         // find an unescaped $, at which point we break out.
 670                         if (scanning_math) {
 671                                 if (escaped)
 672                                         escaped = false;
 673                                 else if (ch == '\\')
 674                                         escaped = true;
 675                                 else if (ch == '$')
 676                                         scanning_math = false;
 677                                 value += ch;
 678                                 val = val.substr(1);
 679                                 continue;
 680                         }
 681
 682                         // if we're scanning a command name, then we just
 683                         // discard characters until we hit something that
 684                         // isn't alpha.
 685                         if (scanning_cmd) {
 686                                 if (isAlphaASCII(ch)) {
 687                                         val = val.substr(1);
 688                                         escaped = false;
 689                                         continue;
 690                                 }
 691                                 // so we're done with this command.
 692                                 // now we fall through and check this character.
 693                                 scanning_cmd = false;
 694                         }
 695
 696                         // was the last character a \? If so, then this is something like: \\,
 697                         // or \$, so we'll just output it. That's probably not always right...
 698                         if (escaped) {
 699                                 value += ch;
 700                                 val = val.substr(1);
 701                                 escaped = false;
 702                                 continue;
 703                         }
 704
 705                         if (ch == '$') {
 706                                 value += ch;
 707                                 val = val.substr(1);
 708                                 scanning_math = true;
 709                                 continue;
 710                         }
 711
 712                         // we just ignore braces
 713                         if (ch == '{' || ch == '}') {
 714                                 val = val.substr(1);
 715                                 continue;
 716                         }
 717
 718                         // we're going to check things that look like commands, so if
 719                         // this doesn't, just output it.
 720                         if (ch != '\\') {
 721                                 value += ch;
 722                                 val = val.substr(1);
 723                                 continue;
 724                         }
 725
 726                         // ok, could be a command of some sort
 727                         // let's see if it corresponds to some unicode
 728                         docstring rem;
 729                         docstring const cnvtd = Encodings::fromLaTeXCommand(val, rem);
 730                         if (!cnvtd.empty()) {
 731                                 // it did, so we'll take that bit and proceed with what's left
 732                                 value += cnvtd;
 733                                 val = rem;
 734                                 continue;
 735                         }
 736                         // it's a command of some sort
 737                         scanning_cmd = true;
 738                         escaped = true;
 739                         val = val.substr(1);
 740                 }
 741
 742                 return true;
 743         }
 744 }
 745
 746
 747 // This method returns a comma separated list of Bibtex entries
 748 void InsetBibtex::fillWithBibKeys(BiblioInfo & keylist,
 749         InsetIterator const & /*di*/) const
 750 {
 751         // This bibtex parser is a first step to parse bibtex files
 752         // more precisely.
 753         //
 754         // - it reads the whole bibtex entry and does a syntax check
 755         //   (matching delimiters, missing commas,...
 756         // - it recovers from errors starting with the next @-character
 757         // - it reads @string definitions and replaces them in the
 758         //   field values.
 759         // - it accepts more characters in keys or value names than
 760         //   bibtex does.
 761         //
 762         // Officially bibtex does only support ASCII, but in practice
 763         // you can use the encoding of the main document as long as
 764         // some elements like keys and names are pure ASCII. Therefore
 765         // we convert the file from the buffer encoding.
 766         // We don't restrict keys to ASCII in LyX, since our own
 767         // InsetBibitem can generate non-ASCII keys, and nonstandard
 768         // 8bit clean bibtex forks exist.
 769         support::FileNameList const files = getBibFiles();
 770         support::FileNameList::const_iterator it = files.begin();
 771         support::FileNameList::const_iterator en = files.end();
 772         for (; it != en; ++ it) {
 773                 ifdocstream ifs(it->toFilesystemEncoding().c_str(),
 774                         ios_base::in, buffer().params().encoding().iconvName());
 775
 776                 char_type ch;
 777                 VarMap strings;
 778
 779                 while (ifs) {
 780
 781                         ifs.get(ch);
 782                         if (!ifs)
 783                                 break;
 784
 785                         if (ch != '@')
 786                                 continue;
 787
 788                         docstring entryType;
 789
 790                         if (!readTypeOrKey(entryType, ifs, from_ascii("{("), docstring(), makeLowerCase)) {
 791                                 lyxerr << "InsetBibtex::fillWithBibKeys: Error reading entry type." << std::endl;
 792                                 continue;
 793                         }
 794
 795                         if (!ifs) {
 796                                 lyxerr << "InsetBibtex::fillWithBibKeys: Unexpected end of file." << std::endl;
 797                                 continue;
 798                         }
 799
 800                         if (entryType == from_ascii("comment")) {
 801                                 ifs.ignore(numeric_limits<int>::max(), '\n');
 802                                 continue;
 803                         }
 804
 805                         ifs.get(ch);
 806                         if (!ifs) {
 807                                 lyxerr << "InsetBibtex::fillWithBibKeys: Unexpected end of file." << std::endl;
 808                                 break;
 809                         }
 810
 811                         if ((ch != '(') && (ch != '{')) {
 812                                 lyxerr << "InsetBibtex::fillWithBibKeys: Invalid entry delimiter." << std::endl;
 813                                 ifs.putback(ch);
 814                                 continue;
 815                         }
 816
 817                         // process the entry
 818                         if (entryType == from_ascii("string")) {
 819
 820                                 // read string and add it to the strings map
 821                                 // (or replace it's old value)
 822                                 docstring name;
 823                                 docstring value;
 824
 825                                 if (!readTypeOrKey(name, ifs, from_ascii("="), from_ascii("#{}(),"), makeLowerCase)) {
 826                                         lyxerr << "InsetBibtex::fillWithBibKeys: Error reading string name." << std::endl;
 827                                         continue;
 828                                 }
 829
 830                                 if (!ifs) {
 831                                         lyxerr << "InsetBibtex::fillWithBibKeys: Unexpected end of file." << std::endl;
 832                                         continue;
 833                                 }
 834
 835                                 // next char must be an equal sign
 836                                 ifs.get(ch);
 837                                 if (!ifs || ch != '=') {
 838                                         lyxerr << "InsetBibtex::fillWithBibKeys: No `=' after string name: " <<
 839                                                         name << "." << std::endl;
 840                                         continue;
 841                                 }
 842
 843                                 if (!readValue(value, ifs, strings)) {
 844                                         lyxerr << "InsetBibtex::fillWithBibKeys: Unable to read value for string: " <<
 845                                                         name << "." << std::endl;
 846                                         continue;
 847                                 }
 848
 849                                 strings[name] = value;
 850
 851                         } else if (entryType == from_ascii("preamble")) {
 852
 853                                 // preamble definitions are discarded.
 854                                 // can they be of any use in lyx?
 855                                 docstring value;
 856
 857                                 if (!readValue(value, ifs, strings)) {
 858                                         lyxerr << "InsetBibtex::fillWithBibKeys: Unable to read preamble value." << std::endl;
 859                                         continue;
 860                                 }
 861
 862                         } else {
 863
 864                                 // Citation entry. Try to read the key.
 865                                 docstring key;
 866
 867                                 if (!readTypeOrKey(key, ifs, from_ascii(","), from_ascii("}"), keepCase)) {
 868                                         lyxerr << "InsetBibtex::fillWithBibKeys: Unable to read key for entry type:" <<
 869                                                         entryType << "." << std::endl;
 870                                         continue;
 871                                 }
 872
 873                                 if (!ifs) {
 874                                         lyxerr << "InsetBibtex::fillWithBibKeys: Unexpected end of file." << std::endl;
 875                                         continue;
 876                                 }
 877
 878                                 /////////////////////////////////////////////
 879                                 // now we have a key, so we will add an entry
 880                                 // (even if it's empty, as bibtex does)
 881                                 //
 882                                 // we now read the field = value pairs.
 883                                 // all items must be separated by a comma. If
 884                                 // it is missing the scanning of this entry is
 885                                 // stopped and the next is searched.
 886                                 docstring fields;
 887                                 docstring name;
 888                                 docstring value;
 889                                 docstring commaNewline;
 890                                 docstring data;
 891                                 BibTeXInfo keyvalmap(key, entryType);
 892
 893                                 bool readNext = removeWSAndComma(ifs);
 894
 895                                 while (ifs && readNext) {
 896
 897                                         // read field name
 898                                         if (!readTypeOrKey(name, ifs, from_ascii("="),
 899                                                            from_ascii("{}(),"), makeLowerCase) || !ifs)
 900                                                 break;
 901
 902                                         // next char must be an equal sign
 903                                         ifs.get(ch);
 904                                         if (!ifs) {
 905                                                 lyxerr << "InsetBibtex::fillWithBibKeys: Unexpected end of file." << std::endl;
 906                                                 break;
 907                                         }
 908                                         if (ch != '=') {
 909                                                 lyxerr << "InsetBibtex::fillWithBibKeys: Missing `=' after field name: " <<
 910                                                                 name << ", for key: " << key << "." << std::endl;
 911                                                 ifs.putback(ch);
 912                                                 break;
 913                                         }
 914
 915                                         // read field value
 916                                         if (!readValue(value, ifs, strings)) {
 917                                                 lyxerr << "InsetBibtex::fillWithBibKeys: Unable to read value for field: " <<
 918                                                                 name << ", for key: " << key << "." << std::endl;
 919                                                 break;
 920                                         }
 921
 922                                         keyvalmap[name] = value;
 923                                         data += "\n\n" + value;
 924                                         keylist.addFieldName(name);
 925                                         readNext = removeWSAndComma(ifs);
 926                                 }
 927
 928                                 // add the new entry
 929                                 keylist.addEntryType(entryType);
 930                                 keyvalmap.setAllData(data);
 931                                 keylist[key] = keyvalmap;
 932                         } //< else (citation entry)
 933                 } //< searching '@'
 934         } //< for loop over files
 935 }
 936
 937
 938 FileName InsetBibtex::getBibTeXPath(docstring const & filename, Buffer const & buf)
 939 {
 940         string texfile = changeExtension(to_utf8(filename), "bib");
 941         // note that, if the filename can be found directly from the path,
 942         // findtexfile will just return a FileName object for that path.
 943         FileName file(findtexfile(texfile, "bib"));
 944         if (file.empty())
 945                 file = FileName(makeAbsPath(texfile, buf.filePath()));
 946         return file;
 947 }
 948
 949
 950 bool InsetBibtex::addDatabase(docstring const & db)
 951 {
 952         docstring bibfiles = getParam("bibfiles");
 953         if (tokenPos(bibfiles, ',', db) != -1)
 954                 return false;
 955         if (!bibfiles.empty())
 956                 bibfiles += ',';
 957         setParam("bibfiles", bibfiles + db);
 958         return true;
 959 }
 960
 961
 962 bool InsetBibtex::delDatabase(docstring const & db)
 963 {
 964         docstring bibfiles = getParam("bibfiles");
 965         if (contains(bibfiles, db)) {
 966                 int const n = tokenPos(bibfiles, ',', db);
 967                 docstring bd = db;
 968                 if (n > 0) {
 969                         // this is not the first database
 970                         docstring tmp = ',' + bd;
 971                         setParam("bibfiles", subst(bibfiles, tmp, docstring()));
 972                 } else if (n == 0)
 973                         // this is the first (or only) database
 974                         setParam("bibfiles", split(bibfiles, bd, ','));
 975                 else
 976                         return false;
 977         }
 978         return true;
 979 }
 980
 981
 982 void InsetBibtex::validate(LaTeXFeatures & features) const
 983 {
 984         if (features.bufferParams().use_bibtopic)
 985                 features.require("bibtopic");
 986 }
 987
 988
 989 docstring InsetBibtex::contextMenu(BufferView const &, int, int) const
 990 {
 991         return from_ascii("context-bibtex");
 992 }
 993
 994
 995 } // namespace lyx