* 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.
*/
using support::FileName;
using support::findtexfile;
using support::isFileReadable;
+using support::isValidLaTeXFilename;
using support::latex_path;
using support::ltrim;
using support::makeAbsPath;
{}
-std::auto_ptr<Inset> InsetBibtex::doClone() const
+Inset * InsetBibtex::clone() const
{
- return std::auto_ptr<Inset>(new InsetBibtex(*this));
+ return new InsetBibtex(*this);
}
else
// FIXME UNICODE
return to_utf8(makeRelPath(from_utf8(fname),
- from_utf8(buffer.getMasterBuffer()->filePath())));
+ from_utf8(buffer.getMasterBuffer()->filePath())));
}
}
<< "' 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)
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
ifs.get(ch);
} while (ifs && isSpace(ch));
- if (!ifs)
+ if (!ifs)
return false;
if (ch != ',') {
/// 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
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)) {
}
/// 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)
val.clear();
- if (!ifs)
+ if (!ifs)
return false;
do {
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 '{':
return false;
// continue reading next value on concatenate with '#'
- } while (ch == '#');
+ } while (ch == '#');
ifs.putback(ch);
// 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
// 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 != '{')) {
// 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;
} else if (entryType == from_ascii("preamble")) {
- // preamble definitions are discarded.
+ // preamble definitions are discarded.
// can they be of any use in lyx?
docstring value;
} 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
}
// 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
}
projects / lyx.git / blobdiff