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 <boost/regex.hpp>
  42
  43 #include <limits>
  44
  45 using namespace std;
  46 using namespace lyx::support;
  47
  48 namespace lyx {
  49
  50 namespace Alert = frontend::Alert;
  51 namespace os = support::os;
  52
  53
  54 InsetBibtex::InsetBibtex(Buffer const & buf, InsetCommandParams const & p)
  55         : InsetCommand(p, "bibtex")
  56 {
  57         Inset::setBuffer(const_cast<Buffer &>(buf));
  58         buffer_->invalidateBibinfoCache();
  59 }
  60
  61
  62 InsetBibtex::~InsetBibtex()
  63 {
  64         if (isBufferValid())
  65                 buffer_->invalidateBibinfoCache();
  66 }
  67
  68
  69 ParamInfo const & InsetBibtex::findInfo(string const & /* cmdName */)
  70 {
  71         static ParamInfo param_info_;
  72         if (param_info_.empty()) {
  73                 param_info_.add("btprint", ParamInfo::LATEX_OPTIONAL);
  74                 param_info_.add("bibfiles", ParamInfo::LATEX_REQUIRED);
  75                 param_info_.add("options", ParamInfo::LYX_INTERNAL);
  76         }
  77         return param_info_;
  78 }
  79
  80
  81 void InsetBibtex::doDispatch(Cursor & cur, FuncRequest & cmd)
  82 {
  83         switch (cmd.action) {
  84
  85         case LFUN_INSET_EDIT:
  86                 editDatabases();
  87                 break;
  88
  89         case LFUN_INSET_MODIFY: {
  90                 InsetCommandParams p(BIBTEX_CODE);
  91                 try {
  92                         if (!InsetCommand::string2params("bibtex",
  93                                         to_utf8(cmd.argument()), p)) {
  94                                 cur.noUpdate();
  95                                 break;
  96                         }
  97                 } catch (ExceptionMessage const & message) {
  98                         if (message.type_ == WarningException) {
  99                                 Alert::warning(message.title_, message.details_);
 100                                 cur.noUpdate();
 101                         } else
 102                                 throw message;
 103                         break;
 104                 }
 105                 //
 106                 setParams(p);
 107                 buffer().updateBibfilesCache();
 108                 break;
 109         }
 110
 111         default:
 112                 InsetCommand::doDispatch(cur, cmd);
 113                 break;
 114         }
 115 }
 116
 117
 118 bool InsetBibtex::getStatus(Cursor & cur, FuncRequest const & cmd,
 119                 FuncStatus & flag) const
 120 {
 121         switch (cmd.action) {
 122         case LFUN_INSET_EDIT:
 123                 flag.setEnabled(true);
 124                 return true;
 125
 126         default:
 127                 return InsetCommand::getStatus(cur, cmd, flag);
 128         }
 129 }
 130
 131
 132 void InsetBibtex::editDatabases() const
 133 {
 134         vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
 135
 136         if (bibfilelist.empty())
 137                 return;
 138
 139         int nr_databases = bibfilelist.size();
 140         if (nr_databases > 1) {
 141                         docstring message = bformat(_("The BibTeX inset includes %1$s databases.\n"
 142                                                        "If you proceed, all of them will be opened."),
 143                                                         convert<docstring>(nr_databases));
 144                         int const ret = Alert::prompt(_("Open Databases?"),
 145                                 message, 0, 1, _("&Cancel"), _("&Proceed"));
 146
 147                         if (ret == 0)
 148                                 return;
 149         }
 150
 151         vector<docstring>::const_iterator it = bibfilelist.begin();
 152         vector<docstring>::const_iterator en = bibfilelist.end();
 153         for (; it != en; ++it) {
 154                 FileName bibfile = getBibTeXPath(*it, buffer());
 155                 formats.edit(buffer(), bibfile,
 156                      formats.getFormatFromFile(bibfile));
 157         }
 158 }
 159
 160
 161 docstring InsetBibtex::screenLabel() const
 162 {
 163         return _("BibTeX Generated Bibliography");
 164 }
 165
 166
 167 docstring InsetBibtex::toolTip(BufferView const & /*bv*/, int /*x*/, int /*y*/) const
 168 {
 169         docstring item = from_ascii("* ");
 170         docstring tip = _("Databases:") + "\n";
 171         vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
 172
 173         if (bibfilelist.empty()) {
 174                 tip += item;
 175                 tip += _("none");
 176         } else {
 177                 vector<docstring>::const_iterator it = bibfilelist.begin();
 178                 vector<docstring>::const_iterator en = bibfilelist.end();
 179                 for (; it != en; ++it) {
 180                         tip += item;
 181                         tip += *it + "\n";
 182                 }
 183         }
 184
 185         // Style-Options
 186         bool toc = false;
 187         docstring style = getParam("options"); // maybe empty! and with bibtotoc
 188         docstring bibtotoc = from_ascii("bibtotoc");
 189         if (prefixIs(style, bibtotoc)) {
 190                 toc = true;
 191                 if (contains(style, char_type(',')))
 192                         style = split(style, bibtotoc, char_type(','));
 193         }
 194
 195         tip += _("Style File:") +"\n";
 196         tip += item;
 197         if (!style.empty())
 198                 tip += style;
 199         else
 200                 tip += _("none");
 201
 202         tip += "\n" + _("Lists:") + " ";
 203         docstring btprint = getParam("btprint");
 204                 if (btprint == "btPrintAll")
 205                         tip += _("all references");
 206                 else if (btprint == "btPrintNotCited")
 207                         tip += _("all uncited references");
 208                 else
 209                         tip += _("all cited references");
 210
 211         if (toc) {
 212                 tip += ", ";
 213                 tip += _("included in TOC");
 214         }
 215
 216         return tip;
 217 }
 218
 219
 220 static string normalizeName(Buffer const & buffer,
 221         OutputParams const & runparams, string const & name, string const & ext)
 222 {
 223         string const fname = makeAbsPath(name, buffer.filePath()).absFilename();
 224         if (FileName(name).isAbsolute() || !FileName(fname + ext).isReadableFile())
 225                 return name;
 226         if (!runparams.nice)
 227                 return fname;
 228
 229         // FIXME UNICODE
 230         return to_utf8(makeRelPath(from_utf8(fname),
 231                                          from_utf8(buffer.masterBuffer()->filePath())));
 232 }
 233
 234
 235 int InsetBibtex::latex(odocstream & os, OutputParams const & runparams) const
 236 {
 237         // the sequence of the commands:
 238         // 1. \bibliographystyle{style}
 239         // 2. \addcontentsline{...} - if option bibtotoc set
 240         // 3. \bibliography{database}
 241         // and with bibtopic:
 242         // 1. \bibliographystyle{style}
 243         // 2. \begin{btSect}{database}
 244         // 3. \btPrint{Cited|NotCited|All}
 245         // 4. \end{btSect}
 246
 247         // Database(s)
 248         // If we are processing the LaTeX file in a temp directory then
 249         // copy the .bib databases to this temp directory, mangling their
 250         // names in the process. Store this mangled name in the list of
 251         // all databases.
 252         // (We need to do all this because BibTeX *really*, *really*
 253         // can't handle "files with spaces" and Windows users tend to
 254         // use such filenames.)
 255         // Otherwise, store the (maybe absolute) path to the original,
 256         // unmangled database name.
 257         vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
 258         vector<docstring>::const_iterator it = bibfilelist.begin();
 259         vector<docstring>::const_iterator en = bibfilelist.end();
 260         odocstringstream dbs;
 261         bool didone = false;
 262
 263         for (; it != en; ++it) {
 264                 string utf8input = to_utf8(*it);
 265                 string database =
 266                         normalizeName(buffer(), runparams, utf8input, ".bib");
 267                 FileName const try_in_file =
 268                         makeAbsPath(database + ".bib", buffer().filePath());
 269                 bool const not_from_texmf = try_in_file.isReadableFile();
 270
 271                 if (!runparams.inComment && !runparams.dryrun && !runparams.nice &&
 272                     not_from_texmf) {
 273                         // mangledFilename() needs the extension
 274                         DocFileName const in_file = DocFileName(try_in_file);
 275                         database = removeExtension(in_file.mangledFilename());
 276                         FileName const out_file = makeAbsPath(database + ".bib",
 277                                         buffer().masterBuffer()->temppath());
 278
 279                         bool const success = in_file.copyTo(out_file);
 280                         if (!success) {
 281                                 lyxerr << "Failed to copy '" << in_file
 282                                        << "' to '" << out_file << "'"
 283                                        << endl;
 284                         }
 285                 } else if (!runparams.inComment && runparams.nice && not_from_texmf &&
 286                            !isValidLaTeXFilename(database)) {
 287                                 frontend::Alert::warning(_("Invalid filename"),
 288                                                          _("The following filename is likely to cause trouble "
 289                                                            "when running the exported file through LaTeX: ") +
 290                                                             from_utf8(database));
 291                 }
 292
 293                 if (didone)
 294                         dbs << ',';
 295                 else
 296                         didone = true;
 297                 // FIXME UNICODE
 298                 dbs << from_utf8(latex_path(database));
 299         }
 300         docstring const db_out = dbs.str();
 301
 302         // Post this warning only once.
 303         static bool warned_about_spaces = false;
 304         if (!warned_about_spaces &&
 305             runparams.nice && db_out.find(' ') != docstring::npos) {
 306                 warned_about_spaces = true;
 307                 Alert::warning(_("Export Warning!"),
 308                                _("There are spaces in the paths to your BibTeX databases.\n"
 309                                               "BibTeX will be unable to find them."));
 310         }
 311         // Style-Options
 312         string style = to_utf8(getParam("options")); // maybe empty! and with bibtotoc
 313         string bibtotoc;
 314         if (prefixIs(style, "bibtotoc")) {
 315                 bibtotoc = "bibtotoc";
 316                 if (contains(style, ','))
 317                         style = split(style, bibtotoc, ',');
 318         }
 319
 320         // line count
 321         int nlines = 0;
 322
 323         if (!style.empty()) {
 324                 string base = normalizeName(buffer(), runparams, style, ".bst");
 325                 FileName const try_in_file =
 326                         makeAbsPath(base + ".bst", buffer().filePath());
 327                 bool const not_from_texmf = try_in_file.isReadableFile();
 328                 // If this style does not come from texmf and we are not
 329                 // exporting to .tex copy it to the tmp directory.
 330                 // This prevents problems with spaces and 8bit charcaters
 331                 // in the file name.
 332                 if (!runparams.inComment && !runparams.dryrun && !runparams.nice &&
 333                     not_from_texmf) {
 334                         // use new style name
 335                         DocFileName const in_file = DocFileName(try_in_file);
 336                         base = removeExtension(in_file.mangledFilename());
 337                         FileName const out_file = makeAbsPath(base + ".bst",
 338                                         buffer().masterBuffer()->temppath());
 339                         bool const success = in_file.copyTo(out_file);
 340                         if (!success) {
 341                                 lyxerr << "Failed to copy '" << in_file
 342                                        << "' to '" << out_file << "'"
 343                                        << endl;
 344                         }
 345                 }
 346                 // FIXME UNICODE
 347                 os << "\\bibliographystyle{"
 348                    << from_utf8(latex_path(normalizeName(buffer(), runparams, base, ".bst")))
 349                    << "}\n";
 350                 nlines += 1;
 351         }
 352
 353         // Post this warning only once.
 354         static bool warned_about_bst_spaces = false;
 355         if (!warned_about_bst_spaces && runparams.nice && contains(style, ' ')) {
 356                 warned_about_bst_spaces = true;
 357                 Alert::warning(_("Export Warning!"),
 358                                _("There are spaces in the path to your BibTeX style file.\n"
 359                                               "BibTeX will be unable to find it."));
 360         }
 361
 362         if (!db_out.empty() && buffer().params().use_bibtopic) {
 363                 os << "\\begin{btSect}{" << db_out << "}\n";
 364                 docstring btprint = getParam("btprint");
 365                 if (btprint.empty())
 366                         // default
 367                         btprint = from_ascii("btPrintCited");
 368                 os << "\\" << btprint << "\n"
 369                    << "\\end{btSect}\n";
 370                 nlines += 3;
 371         }
 372
 373         // bibtotoc-Option
 374         if (!bibtotoc.empty() && !buffer().params().use_bibtopic) {
 375                 if (buffer().params().documentClass().hasLaTeXLayout("chapter")) {
 376                         if (buffer().params().sides == OneSide) {
 377                                 // oneside
 378                                 os << "\\clearpage";
 379                         } else {
 380                                 // twoside
 381                                 os << "\\cleardoublepage";
 382                         }
 383                         os << "\\addcontentsline{toc}{chapter}{\\bibname}";
 384                 } else if (buffer().params().documentClass().hasLaTeXLayout("section"))
 385                         os << "\\addcontentsline{toc}{section}{\\refname}";
 386         }
 387
 388         if (!db_out.empty() && !buffer().params().use_bibtopic) {
 389                 docstring btprint = getParam("btprint");
 390                 if (btprint == "btPrintAll") {
 391                         os << "\\nocite{*}\n";
 392                         nlines += 1;
 393                 }
 394                 os << "\\bibliography{" << db_out << "}\n";
 395                 nlines += 1;
 396         }
 397
 398         return nlines;
 399 }
 400
 401
 402 support::FileNameList InsetBibtex::getBibFiles() const
 403 {
 404         FileName path(buffer().filePath());
 405         support::PathChanger p(path);
 406
 407         support::FileNameList vec;
 408
 409         vector<docstring> bibfilelist = getVectorFromString(getParam("bibfiles"));
 410         vector<docstring>::const_iterator it = bibfilelist.begin();
 411         vector<docstring>::const_iterator en = bibfilelist.end();
 412         for (; it != en; ++it) {
 413                 FileName const file =
 414                         findtexfile(changeExtension(to_utf8(*it), "bib"), "bib");
 415
 416                 if (!file.empty())
 417                         vec.push_back(file);
 418                 else
 419                         LYXERR0("Couldn't find " + to_utf8(*it) + " in InsetBibtex::getBibFiles()!");
 420         }
 421
 422         return vec;
 423
 424 }
 425
 426 namespace {
 427
 428         // methods for parsing bibtex files
 429
 430         typedef map<docstring, docstring> VarMap;
 431
 432         /// remove whitespace characters, optionally a single comma,
 433         /// and further whitespace characters from the stream.
 434         /// @return true if a comma was found, false otherwise
 435         ///
 436         bool removeWSAndComma(ifdocstream & ifs) {
 437                 char_type ch;
 438
 439                 if (!ifs)
 440                         return false;
 441
 442                 // skip whitespace
 443                 do {
 444                         ifs.get(ch);
 445                 } while (ifs && isSpace(ch));
 446
 447                 if (!ifs)
 448                         return false;
 449
 450                 if (ch != ',') {
 451                         ifs.putback(ch);
 452                         return false;
 453                 }
 454
 455                 // skip whitespace
 456                 do {
 457                         ifs.get(ch);
 458                 } while (ifs && isSpace(ch));
 459
 460                 if (ifs) {
 461                         ifs.putback(ch);
 462                 }
 463
 464                 return true;
 465         }
 466
 467
 468         enum charCase {
 469                 makeLowerCase,
 470                 keepCase
 471         };
 472
 473         /// remove whitespace characters, read characer sequence
 474         /// not containing whitespace characters or characters in
 475         /// delimChars, and remove further whitespace characters.
 476         ///
 477         /// @return true if a string of length > 0 could be read.
 478         ///
 479         bool readTypeOrKey(docstring & val, ifdocstream & ifs,
 480                 docstring const & delimChars, docstring const &illegalChars,
 481                 charCase chCase) {
 482
 483                 char_type ch;
 484
 485                 val.clear();
 486
 487                 if (!ifs)
 488                         return false;
 489
 490                 // skip whitespace
 491                 do {
 492                         ifs.get(ch);
 493                 } while (ifs && isSpace(ch));
 494
 495                 if (!ifs)
 496                         return false;
 497
 498                 // read value
 499                 bool legalChar = true;
 500                 while (ifs && !isSpace(ch) &&
 501                                                  delimChars.find(ch) == docstring::npos &&
 502                                                  (legalChar = (illegalChars.find(ch) == docstring::npos))
 503                                         )
 504                 {
 505                         if (chCase == makeLowerCase)
 506                                 val += lowercase(ch);
 507                         else
 508                                 val += ch;
 509                         ifs.get(ch);
 510                 }
 511
 512                 if (!legalChar) {
 513                         ifs.putback(ch);
 514                         return false;
 515                 }
 516
 517                 // skip whitespace
 518                 while (ifs && isSpace(ch)) {
 519                         ifs.get(ch);
 520                 }
 521
 522                 if (ifs) {
 523                         ifs.putback(ch);
 524                 }
 525
 526                 return val.length() > 0;
 527         }
 528
 529         /// read subsequent bibtex values that are delimited with a #-character.
 530         /// Concatenate all parts and replace names with the associated string in
 531         /// the variable strings.
 532         /// @return true if reading was successfull (all single parts were delimited
 533         /// correctly)
 534         bool readValue(docstring & value, ifdocstream & ifs, const VarMap & strings) {
 535
 536                 char_type ch;
 537
 538                 value.clear();
 539
 540                 if (!ifs)
 541                         return false;
 542
 543                 docstring val;
 544                 do {
 545                         // skip whitespace
 546                         do {
 547                                 ifs.get(ch);
 548                         } while (ifs && isSpace(ch));
 549
 550                         if (!ifs)
 551                                 return false;
 552
 553                         // check for field type
 554                         if (isDigit(ch)) {
 555
 556                                 // read integer value
 557                                 do {
 558                                         val += ch;
 559                                         ifs.get(ch);
 560                                 } while (ifs && isDigit(ch));
 561
 562                                 if (!ifs)
 563                                         return false;
 564
 565                         } else if (ch == '"' || ch == '{') {
 566                                 // set end delimiter
 567                                 char_type delim = ch == '"' ? '"': '}';
 568
 569                                 // Skip whitespace
 570                                 do {
 571                                         ifs.get(ch);
 572                                 } while (ifs && isSpace(ch));
 573
 574                                 if (!ifs)
 575                                         return false;
 576
 577                                 // We now have the first non-whitespace character
 578                                 // We'll collapse adjacent whitespace.
 579                                 bool lastWasWhiteSpace = false;
 580
 581                                 // inside this delimited text braces must match.
 582                                 // Thus we can have a closing delimiter only
 583                                 // when nestLevel == 0
 584                                 int nestLevel = 0;
 585
 586                                 while (ifs && (nestLevel > 0 || ch != delim)) {
 587                                         if (isSpace(ch)) {
 588                                                 lastWasWhiteSpace = true;
 589                                                 ifs.get(ch);
 590                                                 continue;
 591                                         }
 592                                         // We output the space only after we stop getting
 593                                         // whitespace so as not to output any whitespace
 594                                         // at the end of the value.
 595                                         if (lastWasWhiteSpace) {
 596                                                 lastWasWhiteSpace = false;
 597                                                 val += ' ';
 598                                         }
 599
 600                                         val += ch;
 601
 602                                         // update nesting level
 603                                         switch (ch) {
 604                                                 case '{':
 605                                                         ++nestLevel;
 606                                                         break;
 607                                                 case '}':
 608                                                         --nestLevel;
 609                                                         if (nestLevel < 0) return false;
 610                                                         break;
 611                                         }
 612
 613                                         ifs.get(ch);
 614                                 }
 615
 616                                 if (!ifs)
 617                                         return false;
 618
 619                                 ifs.get(ch);
 620
 621                                 if (!ifs)
 622                                         return false;
 623
 624                         } else {
 625
 626                                 // reading a string name
 627                                 docstring strName;
 628
 629                                 while (ifs && !isSpace(ch) && ch != '#' && ch != ',' && ch != '}' && ch != ')') {
 630                                         strName += lowercase(ch);
 631                                         ifs.get(ch);
 632                                 }
 633
 634                                 if (!ifs)
 635                                         return false;
 636
 637                                 // replace the string with its assigned value or
 638                                 // discard it if it's not assigned
 639                                 if (strName.length()) {
 640                                         VarMap::const_iterator pos = strings.find(strName);
 641                                         if (pos != strings.end()) {
 642                                                 val += pos->second;
 643                                         }
 644                                 }
 645                         }
 646
 647                         // skip WS
 648                         while (ifs && isSpace(ch)) {
 649                                 ifs.get(ch);
 650                         }
 651
 652                         if (!ifs)
 653                                 return false;
 654
 655                         // continue reading next value on concatenate with '#'
 656                 } while (ch == '#');
 657
 658                 ifs.putback(ch);
 659
 660                 // Ok, we now have the value. Now we are going to go
 661                 // through it and replace e.g. \"a with its unicode value.
 662                 // We'll also strip commands, like \emph, and the like, so
 663                 // it will look nice in the UI.
 664                 bool scanning_cmd = false;
 665                 bool scanning_math = false;
 666                 bool escaped = false; // used to catch \$, etc.
 667                 while (val.size()) {
 668                         char_type const ch = val[0];
 669
 670                         // if we're scanning math, we output everything until we
 671                         // find an unescaped $, at which point we break out.
 672                         if (scanning_math) {
 673                                 if (escaped)
 674                                         escaped = false;
 675                                 else if (ch == '\\')
 676                                         escaped = true;
 677                                 else if (ch == '$')
 678                                         scanning_math = false;
 679                                 value += ch;
 680                                 val = val.substr(1);
 681                                 continue;
 682                         }
 683
 684                         // if we're scanning a command name, then we just
 685                         // discard characters until we hit something that
 686                         // isn't alpha.
 687                         if (scanning_cmd) {
 688                                 if (isAlphaASCII(ch)) {
 689                                         val = val.substr(1);
 690                                         escaped = false;
 691                                         continue;
 692                                 }
 693                                 // so we're done with this command.
 694                                 // now we fall through and check this character.
 695                                 scanning_cmd = false;
 696                         }
 697
 698                         // was the last character a \? If so, then this is something like: \\,
 699                         // or \$, so we'll just output it. That's probably not always right...
 700                         if (escaped) {
 701                                 value += ch;
 702                                 val = val.substr(1);
 703                                 escaped = false;
 704                                 continue;
 705                         }
 706
 707                         if (ch == '$') {
 708                                 value += ch;
 709                                 val = val.substr(1);
 710                                 scanning_math = true;
 711                                 continue;
 712                         }
 713
 714                         // we just ignore braces
 715                         if (ch == '{' || ch == '}') {
 716                                 val = val.substr(1);
 717                                 continue;
 718                         }
 719
 720                         // we're going to check things that look like commands, so if
 721                         // this doesn't, just output it.
 722                         if (ch != '\\') {
 723                                 value += ch;
 724                                 val = val.substr(1);
 725                                 continue;
 726                         }
 727
 728                         // ok, could be a command of some sort
 729                         // let's see if it corresponds to some unicode
 730                         // unicodesymbols has things in the form: \"{u},
 731                         // whereas we may see things like: \"u. So we'll
 732                         // look for that and change it, if necessary.
 733                         static boost::regex const reg("^\\\\\\W\\w");
 734                         if (boost::regex_search(to_utf8(val), reg)) {
 735                                 val.insert(3, from_ascii("}"));
 736                                 val.insert(2, from_ascii("{"));
 737                         }
 738                         docstring rem;
 739                         docstring const cnvtd = Encodings::fromLaTeXCommand(val, rem);
 740                         if (!cnvtd.empty()) {
 741                                 // it did, so we'll take that bit and proceed with what's left
 742                                 value += cnvtd;
 743                                 val = rem;
 744                                 continue;
 745                         }
 746                         // it's a command of some sort
 747                         scanning_cmd = true;
 748                         escaped = true;
 749                         val = val.substr(1);
 750                 }
 751
 752                 return true;
 753         }
 754 }
 755
 756
 757 // This method returns a comma separated list of Bibtex entries
 758 void InsetBibtex::fillWithBibKeys(BiblioInfo & keylist,
 759         InsetIterator const & /*di*/) const
 760 {
 761         // This bibtex parser is a first step to parse bibtex files
 762         // more precisely.
 763         //
 764         // - it reads the whole bibtex entry and does a syntax check
 765         //   (matching delimiters, missing commas,...
 766         // - it recovers from errors starting with the next @-character
 767         // - it reads @string definitions and replaces them in the
 768         //   field values.
 769         // - it accepts more characters in keys or value names than
 770         //   bibtex does.
 771         //
 772         // Officially bibtex does only support ASCII, but in practice
 773         // you can use the encoding of the main document as long as
 774         // some elements like keys and names are pure ASCII. Therefore
 775         // we convert the file from the buffer encoding.
 776         // We don't restrict keys to ASCII in LyX, since our own
 777         // InsetBibitem can generate non-ASCII keys, and nonstandard
 778         // 8bit clean bibtex forks exist.
 779         support::FileNameList const files = getBibFiles();
 780         support::FileNameList::const_iterator it = files.begin();
 781         support::FileNameList::const_iterator en = files.end();
 782         for (; it != en; ++ it) {
 783                 ifdocstream ifs(it->toFilesystemEncoding().c_str(),
 784                         ios_base::in, buffer().params().encoding().iconvName());
 785
 786                 char_type ch;
 787                 VarMap strings;
 788
 789                 while (ifs) {
 790
 791                         ifs.get(ch);
 792                         if (!ifs)
 793                                 break;
 794
 795                         if (ch != '@')
 796                                 continue;
 797
 798                         docstring entryType;
 799
 800                         if (!readTypeOrKey(entryType, ifs, from_ascii("{("), docstring(), makeLowerCase)) {
 801                                 lyxerr << "InsetBibtex::fillWithBibKeys: Error reading entry type." << std::endl;
 802                                 continue;
 803                         }
 804
 805                         if (!ifs) {
 806                                 lyxerr << "InsetBibtex::fillWithBibKeys: Unexpected end of file." << std::endl;
 807                                 continue;
 808                         }
 809
 810                         if (entryType == from_ascii("comment")) {
 811                                 ifs.ignore(numeric_limits<int>::max(), '\n');
 812                                 continue;
 813                         }
 814
 815                         ifs.get(ch);
 816                         if (!ifs) {
 817                                 lyxerr << "InsetBibtex::fillWithBibKeys: Unexpected end of file." << std::endl;
 818                                 break;
 819                         }
 820
 821                         if ((ch != '(') && (ch != '{')) {
 822                                 lyxerr << "InsetBibtex::fillWithBibKeys: Invalid entry delimiter." << std::endl;
 823                                 ifs.putback(ch);
 824                                 continue;
 825                         }
 826
 827                         // process the entry
 828                         if (entryType == from_ascii("string")) {
 829
 830                                 // read string and add it to the strings map
 831                                 // (or replace it's old value)
 832                                 docstring name;
 833                                 docstring value;
 834
 835                                 if (!readTypeOrKey(name, ifs, from_ascii("="), from_ascii("#{}(),"), makeLowerCase)) {
 836                                         lyxerr << "InsetBibtex::fillWithBibKeys: Error reading string name." << std::endl;
 837                                         continue;
 838                                 }
 839
 840                                 if (!ifs) {
 841                                         lyxerr << "InsetBibtex::fillWithBibKeys: Unexpected end of file." << std::endl;
 842                                         continue;
 843                                 }
 844
 845                                 // next char must be an equal sign
 846                                 ifs.get(ch);
 847                                 if (!ifs || ch != '=') {
 848                                         lyxerr << "InsetBibtex::fillWithBibKeys: No `=' after string name: " <<
 849                                                         name << "." << std::endl;
 850                                         continue;
 851                                 }
 852
 853                                 if (!readValue(value, ifs, strings)) {
 854                                         lyxerr << "InsetBibtex::fillWithBibKeys: Unable to read value for string: " <<
 855                                                         name << "." << std::endl;
 856                                         continue;
 857                                 }
 858
 859                                 strings[name] = value;
 860
 861                         } else if (entryType == from_ascii("preamble")) {
 862
 863                                 // preamble definitions are discarded.
 864                                 // can they be of any use in lyx?
 865                                 docstring value;
 866
 867                                 if (!readValue(value, ifs, strings)) {
 868                                         lyxerr << "InsetBibtex::fillWithBibKeys: Unable to read preamble value." << std::endl;
 869                                         continue;
 870                                 }
 871
 872                         } else {
 873
 874                                 // Citation entry. Try to read the key.
 875                                 docstring key;
 876
 877                                 if (!readTypeOrKey(key, ifs, from_ascii(","), from_ascii("}"), keepCase)) {
 878                                         lyxerr << "InsetBibtex::fillWithBibKeys: Unable to read key for entry type:" <<
 879                                                         entryType << "." << std::endl;
 880                                         continue;
 881                                 }
 882
 883                                 if (!ifs) {
 884                                         lyxerr << "InsetBibtex::fillWithBibKeys: Unexpected end of file." << std::endl;
 885                                         continue;
 886                                 }
 887
 888                                 /////////////////////////////////////////////
 889                                 // now we have a key, so we will add an entry
 890                                 // (even if it's empty, as bibtex does)
 891                                 //
 892                                 // we now read the field = value pairs.
 893                                 // all items must be separated by a comma. If
 894                                 // it is missing the scanning of this entry is
 895                                 // stopped and the next is searched.
 896                                 docstring fields;
 897                                 docstring name;
 898                                 docstring value;
 899                                 docstring commaNewline;
 900                                 docstring data;
 901                                 BibTeXInfo keyvalmap(key, entryType);
 902
 903                                 bool readNext = removeWSAndComma(ifs);
 904
 905                                 while (ifs && readNext) {
 906
 907                                         // read field name
 908                                         if (!readTypeOrKey(name, ifs, from_ascii("="),
 909                                                            from_ascii("{}(),"), makeLowerCase) || !ifs)
 910                                                 break;
 911
 912                                         // next char must be an equal sign
 913                                         ifs.get(ch);
 914                                         if (!ifs) {
 915                                                 lyxerr << "InsetBibtex::fillWithBibKeys: Unexpected end of file." << std::endl;
 916                                                 break;
 917                                         }
 918                                         if (ch != '=') {
 919                                                 lyxerr << "InsetBibtex::fillWithBibKeys: Missing `=' after field name: " <<
 920                                                                 name << ", for key: " << key << "." << std::endl;
 921                                                 ifs.putback(ch);
 922                                                 break;
 923                                         }
 924
 925                                         // read field value
 926                                         if (!readValue(value, ifs, strings)) {
 927                                                 lyxerr << "InsetBibtex::fillWithBibKeys: Unable to read value for field: " <<
 928                                                                 name << ", for key: " << key << "." << std::endl;
 929                                                 break;
 930                                         }
 931
 932                                         keyvalmap[name] = value;
 933                                         data += "\n\n" + value;
 934                                         keylist.addFieldName(name);
 935                                         readNext = removeWSAndComma(ifs);
 936                                 }
 937
 938                                 // add the new entry
 939                                 keylist.addEntryType(entryType);
 940                                 keyvalmap.setAllData(data);
 941                                 keylist[key] = keyvalmap;
 942                         } //< else (citation entry)
 943                 } //< searching '@'
 944         } //< for loop over files
 945 }
 946
 947
 948 FileName InsetBibtex::getBibTeXPath(docstring const & filename, Buffer const & buf)
 949 {
 950         string texfile = changeExtension(to_utf8(filename), "bib");
 951         // note that, if the filename can be found directly from the path,
 952         // findtexfile will just return a FileName object for that path.
 953         FileName file(findtexfile(texfile, "bib"));
 954         if (file.empty())
 955                 file = FileName(makeAbsPath(texfile, buf.filePath()));
 956         return file;
 957 }
 958
 959
 960 bool InsetBibtex::addDatabase(docstring const & db)
 961 {
 962         docstring bibfiles = getParam("bibfiles");
 963         if (tokenPos(bibfiles, ',', db) != -1)
 964                 return false;
 965         if (!bibfiles.empty())
 966                 bibfiles += ',';
 967         setParam("bibfiles", bibfiles + db);
 968         return true;
 969 }
 970
 971
 972 bool InsetBibtex::delDatabase(docstring const & db)
 973 {
 974         docstring bibfiles = getParam("bibfiles");
 975         if (contains(bibfiles, db)) {
 976                 int const n = tokenPos(bibfiles, ',', db);
 977                 docstring bd = db;
 978                 if (n > 0) {
 979                         // this is not the first database
 980                         docstring tmp = ',' + bd;
 981                         setParam("bibfiles", subst(bibfiles, tmp, docstring()));
 982                 } else if (n == 0)
 983                         // this is the first (or only) database
 984                         setParam("bibfiles", split(bibfiles, bd, ','));
 985                 else
 986                         return false;
 987         }
 988         return true;
 989 }
 990
 991
 992 void InsetBibtex::validate(LaTeXFeatures & features) const
 993 {
 994         if (features.bufferParams().use_bibtopic)
 995                 features.require("bibtopic");
 996 }
 997
 998
 999 docstring InsetBibtex::contextMenu(BufferView const &, int, int) const
1000 {
1001         return from_ascii("context-bibtex");
1002 }
1003
1004
1005 } // namespace lyx