]> git.lyx.org Git - features.git/blobdiff - src/insets/InsetBibtex.cpp
cosmetics. mainly replace #include "dostring.h" by #include "strfwd.h"
[features.git] / src / insets / InsetBibtex.cpp
index 3eb2b99414a0780f52b5c09015b95ac547b68e0f..d5a076671fbcfc876690bc639bd6ed61cf8de43c 100644 (file)
@@ -4,6 +4,7 @@
  * Licence details can be found in the file COPYING.
  *
  * \author Alejandro Aguilar Sierra
+ * \author Richard Heck (BibTeX parser improvements)
  *
  * Full author contact details are available in file CREDITS.
  */
 
 #include "InsetBibtex.h"
 
-#include "buffer.h"
-#include "bufferparams.h"
-#include "dispatchresult.h"
+#include "Buffer.h"
+#include "BufferParams.h"
+#include "DispatchResult.h"
 #include "debug.h"
-#include "encoding.h"
-#include "funcrequest.h"
+#include "Encoding.h"
+#include "FuncRequest.h"
 #include "gettext.h"
 #include "LaTeXFeatures.h"
-#include "metricsinfo.h"
-#include "outputparams.h"
+#include "MetricsInfo.h"
+#include "OutputParams.h"
 
-#include "frontends/Alert.h"
+#include "frontends/alert.h"
 
+#include "support/ExceptionMessage.h"
+#include "support/docstream.h"
 #include "support/filetools.h"
 #include "support/lstrings.h"
 #include "support/lyxlib.h"
 #include "support/os.h"
-#include "support/path.h"
+#include "support/Path.h"
 #include "support/textutils.h"
 
 #include <boost/tokenizer.hpp>
@@ -45,7 +48,7 @@ using support::copy;
 using support::DocFileName;
 using support::FileName;
 using support::findtexfile;
-using support::isFileReadable;
+using support::isValidLaTeXFilename;
 using support::latex_path;
 using support::ltrim;
 using support::makeAbsPath;
@@ -76,24 +79,44 @@ InsetBibtex::InsetBibtex(InsetCommandParams const & p)
 {}
 
 
-std::auto_ptr<InsetBase> InsetBibtex::doClone() const
+CommandInfo const * InsetBibtex::findInfo(std::string const & /* cmdName */)
 {
-       return std::auto_ptr<InsetBase>(new InsetBibtex(*this));
+       static const char * const paramnames[] = 
+               {"options", "btprint", "bibfiles", ""};
+       static const bool isoptional[] = {true, true, false};
+       static const CommandInfo info = {3, paramnames, isoptional};
+       return &info;
 }
 
 
-void InsetBibtex::doDispatch(LCursor & cur, FuncRequest & cmd)
+Inset * InsetBibtex::clone() const
+{
+       return new InsetBibtex(*this);
+}
+
+
+void InsetBibtex::doDispatch(Cursor & cur, FuncRequest & cmd)
 {
        switch (cmd.action) {
 
        case LFUN_INSET_MODIFY: {
-               InsetCommandParams p("bibtex");
-               InsetCommandMailer::string2params("bibtex", to_utf8(cmd.argument()), p);
-               if (!p.getCmdName().empty()) {
-                       setParams(p);
-                       cur.buffer().updateBibfilesCache();
-               } else
-                       cur.noUpdate();
+               InsetCommandParams p(BIBTEX_CODE);
+               try {
+                       if (!InsetCommandMailer::string2params("bibtex", 
+                                       to_utf8(cmd.argument()), p)) {
+                               cur.noUpdate();
+                               break;
+                       }
+               } catch (support::ExceptionMessage const & message) {
+                       if (message.type_ == support::WarningException) {
+                               Alert::warning(message.title_, message.details_);
+                               cur.noUpdate();
+                       } else 
+                               throw message;
+                       break;
+               }
+               setParams(p);
+               cur.buffer().updateBibfilesCache();
                break;
        }
 
@@ -112,18 +135,18 @@ docstring const InsetBibtex::getScreenLabel(Buffer const &) const
 
 namespace {
 
-string normalize_name(Buffer const & buffer, OutputParams const & runparams,
+string normalizeName(Buffer const & buffer, OutputParams const & runparams,
                      string const & name, string const & ext)
 {
        string const fname = makeAbsPath(name, buffer.filePath()).absFilename();
-       if (absolutePath(name) || !isFileReadable(FileName(fname + ext)))
+       if (absolutePath(name) || !FileName(fname + ext).isFileReadable())
                return name;
-       else if (!runparams.nice)
+       if (!runparams.nice)
                return fname;
-       else
-               // FIXME UNICODE
-               return to_utf8(makeRelPath(from_utf8(fname),
-                                          from_utf8(buffer.getMasterBuffer()->filePath())));
+
+       // FIXME UNICODE
+       return to_utf8(makeRelPath(from_utf8(fname),
+                                        from_utf8(buffer.masterBuffer()->filePath())));
 }
 
 }
@@ -167,11 +190,11 @@ int InsetBibtex::latex(Buffer const & buffer, odocstream & os,
        for (Tokenizer::const_iterator it = begin; it != end; ++it) {
                docstring const input = trim(*it);
                // FIXME UNICODE
-               string utf8input(to_utf8(input));
+               string utf8input = to_utf8(input);
                string database =
-                       normalize_name(buffer, runparams, utf8input, ".bib");
+                       normalizeName(buffer, runparams, utf8input, ".bib");
                FileName const try_in_file(makeAbsPath(database + ".bib", buffer.filePath()));
-               bool const not_from_texmf = isFileReadable(try_in_file);
+               bool const not_from_texmf = try_in_file.isFileReadable();
 
                if (!runparams.inComment && !runparams.dryrun && !runparams.nice &&
                    not_from_texmf) {
@@ -179,8 +202,8 @@ int InsetBibtex::latex(Buffer const & buffer, odocstream & os,
                        // mangledFilename() needs the extension
                        DocFileName const in_file = DocFileName(try_in_file);
                        database = removeExtension(in_file.mangledFilename());
-                       FileName const out_file(makeAbsPath(database + ".bib",
-                                       buffer.getMasterBuffer()->temppath()));
+                       FileName const out_file = makeAbsPath(database + ".bib",
+                                       buffer.masterBuffer()->temppath());
 
                        bool const success = copy(in_file, out_file);
                        if (!success) {
@@ -188,6 +211,12 @@ int InsetBibtex::latex(Buffer const & buffer, odocstream & os,
                                       << "' to '" << out_file << "'"
                                       << endl;
                        }
+               } else if (!runparams.inComment && runparams.nice && not_from_texmf &&
+                          !isValidLaTeXFilename(database)) {
+                               frontend::Alert::warning(_("Invalid filename"),
+                                                        _("The following filename is likely to cause trouble "
+                                                          "when running the exported file through LaTeX: ") +
+                                                           from_utf8(database));
                }
 
                if (it != begin)
@@ -206,7 +235,6 @@ int InsetBibtex::latex(Buffer const & buffer, odocstream & os,
                Alert::warning(_("Export Warning!"),
                               _("There are spaces in the paths to your BibTeX databases.\n"
                                              "BibTeX will be unable to find them."));
-
        }
 
        // Style-Options
@@ -214,19 +242,17 @@ int InsetBibtex::latex(Buffer const & buffer, odocstream & os,
        string bibtotoc;
        if (prefixIs(style, "bibtotoc")) {
                bibtotoc = "bibtotoc";
-               if (contains(style, ',')) {
+               if (contains(style, ','))
                        style = split(style, bibtotoc, ',');
-               }
        }
 
        // line count
        int nlines = 0;
 
        if (!style.empty()) {
-               string base =
-                       normalize_name(buffer, runparams, style, ".bst");
+               string base = normalizeName(buffer, runparams, style, ".bst");
                FileName const try_in_file(makeAbsPath(base + ".bst", buffer.filePath()));
-               bool const not_from_texmf = isFileReadable(try_in_file);
+               bool const not_from_texmf = try_in_file.isFileReadable();
                // If this style does not come from texmf and we are not
                // exporting to .tex copy it to the tmp directory.
                // This prevents problems with spaces and 8bit charcaters
@@ -237,7 +263,7 @@ int InsetBibtex::latex(Buffer const & buffer, odocstream & os,
                        DocFileName const in_file = DocFileName(try_in_file);
                        base = removeExtension(in_file.mangledFilename());
                        FileName const out_file(makeAbsPath(base + ".bst",
-                                       buffer.getMasterBuffer()->temppath()));
+                                       buffer.masterBuffer()->temppath()));
                        bool const success = copy(in_file, out_file);
                        if (!success) {
                                lyxerr << "Failed to copy '" << in_file
@@ -247,7 +273,7 @@ int InsetBibtex::latex(Buffer const & buffer, odocstream & os,
                }
                // FIXME UNICODE
                os << "\\bibliographystyle{"
-                  << from_utf8(latex_path(normalize_name(buffer, runparams, base, ".bst")))
+                  << from_utf8(latex_path(normalizeName(buffer, runparams, base, ".bst")))
                   << "}\n";
                nlines += 1;
        }
@@ -278,9 +304,9 @@ int InsetBibtex::latex(Buffer const & buffer, odocstream & os,
                // part of its name, because it's than book.
                // For the "official" lyx-layouts it's no problem to support
                // all well
-               if (!contains(buffer.params().getLyXTextClass().name(),
+               if (!contains(buffer.params().getTextClass().name(),
                              "art")) {
-                       if (buffer.params().sides == LyXTextClass::OneSide) {
+                       if (buffer.params().sides == TextClass::OneSide) {
                                // oneside
                                os << "\\clearpage";
                        } else {
@@ -338,14 +364,14 @@ namespace {
 
        typedef map<docstring, docstring> VarMap;
 
-       /// remove whitespace characters, optionally a single comma, 
+       /// remove whitespace characters, optionally a single comma,
        /// and further whitespace characters from the stream.
        /// @return true if a comma was found, false otherwise
        ///
        bool removeWSAndComma(idocfstream & ifs) {
                char_type ch;
 
-               if (!ifs) 
+               if (!ifs)
                        return false;
 
                // skip whitespace
@@ -353,7 +379,7 @@ namespace {
                        ifs.get(ch);
                } while (ifs && isSpace(ch));
 
-               if (!ifs) 
+               if (!ifs)
                        return false;
 
                if (ch != ',') {
@@ -384,15 +410,16 @@ namespace {
        /// delimChars, and remove further whitespace characters.
        ///
        /// @return true if a string of length > 0 could be read.
-       /// 
-       bool readTypeOrKey(docstring & val, idocfstream & ifs, 
-               docstring const & delimChars, charCase chCase) {
+       ///
+       bool readTypeOrKey(docstring & val, idocfstream & ifs,
+               docstring const & delimChars, docstring const &illegalChars, 
+               charCase chCase) {
 
                char_type ch;
 
                val.clear();
 
-               if (!ifs) 
+               if (!ifs)
                        return false;
 
                // skip whitespace
@@ -400,18 +427,27 @@ namespace {
                        ifs.get(ch);
                } while (ifs && isSpace(ch));
 
-               if (!ifs) 
+               if (!ifs)
                        return false;
 
-               // read value 
-               while (ifs && !isSpace(ch) && delimChars.find(ch) == docstring::npos) {
-                       if (chCase == makeLowerCase) {
+               // read value
+               bool legalChar = true;
+               while (ifs && !isSpace(ch) && 
+                                                delimChars.find(ch) == docstring::npos &&
+                                                (legalChar = (illegalChars.find(ch) == docstring::npos))
+                                       ) 
+               {
+                       if (chCase == makeLowerCase)
                                val += lowercase(ch);
-                       } else {
+                       else
                                val += ch;
-                       }
                        ifs.get(ch);
                }
+               
+               if (!legalChar) {
+                       ifs.putback(ch);
+                       return false;
+               }
 
                // skip whitespace
                while (ifs && isSpace(ch)) {
@@ -426,7 +462,7 @@ namespace {
        }
 
        /// read subsequent bibtex values that are delimited with a #-character.
-       /// Concatenate all parts and replace names with the associated string in 
+       /// Concatenate all parts and replace names with the associated string in
        /// the variable strings.
        /// @return true if reading was successfull (all single parts were delimited
        /// correctly)
@@ -436,7 +472,7 @@ namespace {
 
                val.clear();
 
-               if (!ifs) 
+               if (!ifs)
                        return false;
 
                do {
@@ -461,19 +497,42 @@ namespace {
                                        return false;
 
                        } else if (ch == '"' || ch == '{') {
+                               // set end delimiter
+                               char_type delim = ch == '"' ? '"': '}';
 
-                               // read delimited text - set end delimiter
-                               char_type delim = ch == '"'? '"': '}';
-
-                               // inside this delimited text braces must match.
-                               // Thus we can have a closing delimiter only
-                               // when nestLevel == 0
+                               //Skip whitespace
+                               do {
+                                       ifs.get(ch);
+                               } while (ifs && isSpace(ch));
+                               
+                               if (!ifs)
+                                       return false;
+                               
+                               //We now have the first non-whitespace character
+                               //We'll collapse adjacent whitespace.
+                               bool lastWasWhiteSpace = false;
+                               
+                               // inside this delimited text braces must match.
+                               // Thus we can have a closing delimiter only
+                               // when nestLevel == 0
                                int nestLevel = 0;
-
-                               ifs.get(ch);
                                while (ifs && (nestLevel > 0 || ch != delim)) {
-                                       val += ch;
+                                       if (isSpace(ch)) {
+                                               lastWasWhiteSpace = true;
+                                               ifs.get(ch);
+                                               continue;
+                                       }
+                                       //We output the space only after we stop getting 
+                                       //whitespace so as not to output any whitespace
+                                       //at the end of the value.
+                                       if (lastWasWhiteSpace) {
+                                               lastWasWhiteSpace = false;
+                                               val += ' ';
+                                       }
                                        
+                                       val += ch;
+
                                        // update nesting level
                                        switch (ch) {
                                                case '{':
@@ -528,7 +587,7 @@ namespace {
                                return false;
 
                        // continue reading next value on concatenate with '#'
-               } while (ch == '#');  
+               } while (ch == '#');
 
                ifs.putback(ch);
 
@@ -539,31 +598,22 @@ namespace {
 
 // This method returns a comma separated list of Bibtex entries
 void InsetBibtex::fillWithBibKeys(Buffer const & buffer,
-               std::vector<std::pair<string, docstring> > & keys) const
+               BiblioInfo & keylist, InsetIterator const & /*di*/) const
 {
        vector<FileName> const files = getFiles(buffer);
        for (vector<FileName>::const_iterator it = files.begin();
             it != files.end(); ++ it) {
                // This bibtex parser is a first step to parse bibtex files
-               // more precisely. 
-               // 
+               // more precisely.
+               //
                // - it reads the whole bibtex entry and does a syntax check
                //   (matching delimiters, missing commas,...
                // - it recovers from errors starting with the next @-character
-               // - it reads @string definitions and replaces them in the 
+               // - it reads @string definitions and replaces them in the
                //   field values.
-               // - it accepts more characters in keys or value names than 
+               // - it accepts more characters in keys or value names than
                //   bibtex does.
                //
-               // TODOS:
-               // - the entries are split into name = value pairs by the 
-               //   parser. These have to be merged again because of the 
-               //   way lyx treats the entries ( pair<...>(...) ). The citation
-               //   mechanism in lyx should be changed such that it can use
-               //   the split entries.
-               // - messages on parsing errors can be generated.
-               //
-
                // Officially bibtex does only support ASCII, but in practice
                // you can use the encoding of the main document as long as
                // some elements like keys and names are pure ASCII. Therefore
@@ -571,35 +621,37 @@ void InsetBibtex::fillWithBibKeys(Buffer const & buffer,
                // We don't restrict keys to ASCII in LyX, since our own
                // InsetBibitem can generate non-ASCII keys, and nonstandard
                // 8bit clean bibtex forks exist.
-               idocfstream ifs(it->toFilesystemEncoding().c_str(),
-                               std::ios_base::in,
-                               buffer.params().encoding().iconvName());
                
+               idocfstream ifs(it->toFilesystemEncoding().c_str(),
+                       std::ios_base::in,
+                       buffer.params().encoding().iconvName());
+
                char_type ch;
                VarMap strings;
 
                while (ifs) {
 
                        ifs.get(ch);
-                       if (!ifs) 
+                       if (!ifs)
                                break;
 
-                       if (ch != '@') 
+                       if (ch != '@')
                                continue;
 
                        docstring entryType;
 
-                       if (!readTypeOrKey(entryType, ifs, from_ascii("{("), makeLowerCase) || !ifs)
+                       if (!readTypeOrKey(entryType, ifs, from_ascii("{("), 
+                                          docstring(), makeLowerCase) || !ifs)
                                continue;
 
                        if (entryType == from_ascii("comment")) {
 
                                ifs.ignore(std::numeric_limits<int>::max(), '\n');
                                continue;
-                       } 
+                       }
 
                        ifs.get(ch);
-                       if (!ifs) 
+                       if (!ifs)
                                break;
 
                        if ((ch != '(') && (ch != '{')) {
@@ -611,14 +663,16 @@ void InsetBibtex::fillWithBibKeys(Buffer const & buffer,
                        // process the entry
                        if (entryType == from_ascii("string")) {
 
-                               // read string and add it to the strings map 
+                               // read string and add it to the strings map
                                // (or replace it's old value)
                                docstring name;
                                docstring value;
 
-                               if (!readTypeOrKey(name, ifs, from_ascii("#=}),"), makeLowerCase) || !ifs)
+                               if (!readTypeOrKey(name, ifs, from_ascii("="), 
+                                                  from_ascii("#{}(),"), makeLowerCase) || !ifs)
                                        continue;
 
+                               // next char must be an equal sign
                                ifs.get(ch);
                                if (!ifs || ch != '=')
                                        continue;
@@ -630,7 +684,7 @@ void InsetBibtex::fillWithBibKeys(Buffer const & buffer,
 
                        } else if (entryType == from_ascii("preamble")) {
 
-                               // preamble definitions are discarded. 
+                               // preamble definitions are discarded.
                                // can they be of any use in lyx?
                                docstring value;
 
@@ -639,28 +693,36 @@ void InsetBibtex::fillWithBibKeys(Buffer const & buffer,
 
                        } else {
 
-                               // Citation entry. Read the key and all name = value pairs
+                               // Citation entry. Try to read the key.
                                docstring key;
-                               docstring fields;
-                               docstring name;
-                               docstring value;
-                               docstring commaNewline;
 
-                               if (!readTypeOrKey(key, ifs, from_ascii(",})"), keepCase) || !ifs)
+                               if (!readTypeOrKey(key, ifs, from_ascii(","), 
+                                                  from_ascii("}"), keepCase) || !ifs)
                                        continue;
 
+                               /////////////////////////////////////////////
                                // now we have a key, so we will add an entry 
-                               // (even if it's empty, as bibtex does)
-                               // 
-                               // all items must be separated by a comma. If
-                               // it is missing the scanning of this entry is
-                               // stopped and the next is searched.
+                               // (even if it's empty, as bibtex does)
+                               //
+                               // we now read the field = value pairs.
+                               // all items must be separated by a comma. If
+                               // it is missing the scanning of this entry is
+                               // stopped and the next is searched.
+                               docstring fields;
+                               docstring name;
+                               docstring value;
+                               docstring commaNewline;
+                               docstring data;
+                               BibTeXInfo keyvalmap;
+                               keyvalmap.entryType = entryType;
+                               
                                bool readNext = removeWSAndComma(ifs);
-
                                while (ifs && readNext) {
 
                                        // read field name
-                                       if (!readTypeOrKey(name, ifs, from_ascii("=}),"), makeLowerCase) || !ifs)
+                                       if (!readTypeOrKey(name, ifs, from_ascii("="), 
+                                                          from_ascii("{}(),"), makeLowerCase) || !ifs)
                                                break;
 
                                        // next char must be an equal sign
@@ -673,30 +735,23 @@ void InsetBibtex::fillWithBibKeys(Buffer const & buffer,
                                        }
 
                                        // read field value
-                                       if (!readValue(value, ifs, strings)) 
+                                       if (!readValue(value, ifs, strings))
                                                break;
 
-                                       // append field to the total entry string.
-                                       //
-                                       // TODO: Here is where the fields can be put in 
-                                       //       a more intelligent structure that preserves
-                                       //           the already known parts.
-                                       fields += commaNewline;
-                                       fields += name + from_ascii(" = {") + value + '}';
-
-                                       if (!commaNewline.length()) 
-                                               commaNewline = from_ascii(",\n"); 
-
+                                       keyvalmap[name] = value;
+                                       data += "\n\n" + value;
+                                       keylist.fieldNames.insert(name);
                                        readNext = removeWSAndComma(ifs);
                                }
 
                                // add the new entry
-                               keys.push_back(pair<string, docstring>(
-                               to_utf8(key), fields));
+                               keylist.entryTypes.insert(entryType);
+                               keyvalmap.allData = data;
+                               keyvalmap.isBibTeX = true;
+                               keyvalmap.bibKey = key;
+                               keylist[key] = keyvalmap;
                        }
-
                } //< searching '@'
-
        } //< for loop over files
 }