X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2FBiblioInfo.cpp;h=c2761067d33e2975603791703286bc5b83dc303b;hb=7aee04217d33604f6bfcd2c8e7422dc7520ed193;hp=5fee10ccc2743fac622048457f7417d1d9e3ed97;hpb=833e7085dde3a361b92caeb36beb79232f748e70;p=lyx.git

diff --git a/src/BiblioInfo.cpp b/src/BiblioInfo.cpp
index 5fee10ccc2..c2761067d3 100644
--- a/src/BiblioInfo.cpp
+++ b/src/BiblioInfo.cpp
@@ -18,23 +18,20 @@
 #include "buffer_funcs.h"
 #include "Encoding.h"
 #include "InsetIterator.h"
+#include "Language.h"
 #include "Paragraph.h"
+#include "TextClass.h"
 #include "TocBackend.h"
 
-#include "insets/Inset.h"
-#include "insets/InsetBibitem.h"
-#include "insets/InsetBibtex.h"
-#include "insets/InsetInclude.h"
-
 #include "support/convert.h"
+#include "support/debug.h"
 #include "support/docstream.h"
 #include "support/gettext.h"
 #include "support/lassert.h"
 #include "support/lstrings.h"
+#include "support/regex.h"
 #include "support/textutils.h"
 
-#include "boost/regex.hpp"
-
 #include <set>
 
 using namespace std;
@@ -85,7 +82,7 @@ docstring familyName(docstring const & name)
 	for (; it != en; ++it) {
 		if (!first)
 			retval += " ";
-		else 
+		else
 			first = false;
 		retval += *it;
 	}
@@ -112,7 +109,7 @@ docstring convertLaTeXCommands(docstring const & str)
 				escaped = false;
 			else if (ch == '\\')
 				escaped = true;
-			else if (ch == '$') 
+			else if (ch == '$')
 				scanning_math = false;
 			ret += ch;
 			val = val.substr(1);
@@ -172,8 +169,8 @@ docstring convertLaTeXCommands(docstring const & str)
 		// unicodesymbols has things in the form: \"{u},
 		// whereas we may see things like: \"u. So we'll
 		// look for that and change it, if necessary.
-		static boost::regex const reg("^\\\\\\W\\w");
-		if (boost::regex_search(to_utf8(val), reg)) {
+		static lyx::regex const reg("^\\\\\\W\\w");
+		if (lyx::regex_search(to_utf8(val), reg)) {
 			val.insert(3, from_ascii("}"));
 			val.insert(2, from_ascii("{"));
 		}
@@ -204,7 +201,8 @@ docstring convertLaTeXCommands(docstring const & str)
 //////////////////////////////////////////////////////////////////////
 
 BibTeXInfo::BibTeXInfo(docstring const & key, docstring const & type)
-	: is_bibtex_(true), bib_key_(key), entry_type_(type), info_()
+	: is_bibtex_(true), bib_key_(key), entry_type_(type), info_(),
+	  modifier_(0)
 {}
 
 
@@ -216,7 +214,11 @@ docstring const BibTeXInfo::getAbbreviatedAuthor() const
 			return docstring();
 
 		docstring authors;
-		split(opt, authors, '(');
+		docstring const remainder = trim(split(opt, authors, '('));
+		if (remainder.empty())
+			// in this case, we didn't find a "(",
+			// so we don't have author (year)
+			return docstring();
 		return authors;
 	}
 
@@ -227,6 +229,9 @@ docstring const BibTeXInfo::getAbbreviatedAuthor() const
 			return bib_key_;
 	}
 
+	// FIXME Move this to a separate routine that can
+	// be called from elsewhere.
+	//
 	// OK, we've got some names. Let's format them.
 	// Try to split the author list on " and "
 	vector<docstring> const authors =
@@ -245,7 +250,7 @@ docstring const BibTeXInfo::getAbbreviatedAuthor() const
 
 docstring const BibTeXInfo::getYear() const
 {
-	if (is_bibtex_) 
+	if (is_bibtex_)
 		return operator[]("year");
 
 	docstring const opt = label();
@@ -253,9 +258,12 @@ docstring const BibTeXInfo::getYear() const
 		return docstring();
 
 	docstring authors;
-	docstring const tmp = split(opt, authors, '(');
+	docstring tmp = split(opt, authors, '(');
+	if (tmp.empty())
+		// we don't have author (year)
+		return docstring();
 	docstring year;
-	split(tmp, year, ')');
+	tmp = split(tmp, year, ')');
 	return year;
 }
 
@@ -268,7 +276,247 @@ docstring const BibTeXInfo::getXRef() const
 }
 
 
-docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref) const
+namespace {
+	string parseOptions(string const & format, string & optkey,
+			string & ifpart, string & elsepart);
+
+	// Calls parseOptions to deal with an embedded option, such as:
+	//   {%number%[[, no.~%number%]]}
+	// which must appear at the start of format. ifelsepart gets the
+	// whole of the option, and we return what's left after the option.
+	// we return format if there is an error.
+	string parseEmbeddedOption(string const & format, string & ifelsepart)
+	{
+		LASSERT(format[0] == '{' && format[1] == '%', return format);
+		string optkey;
+		string ifpart;
+		string elsepart;
+		string const rest = parseOptions(format, optkey, ifpart, elsepart);
+		if (format == rest) { // parse error
+			LYXERR0("ERROR! Couldn't parse `" << format <<"'.");
+			return format;
+		}
+		LASSERT(rest.size() <= format.size(), /* */);
+		ifelsepart = format.substr(0, format.size() - rest.size());
+		return rest;
+	}
+
+
+	// Gets a "clause" from a format string, where the clause is
+	// delimited by '[[' and ']]'. Returns what is left after the
+	// clause is removed, and returns format if there is an error.
+	string getClause(string const & format, string & clause)
+	{
+		string fmt = format;
+		// remove '[['
+		fmt = fmt.substr(2);
+		// we'll remove characters from the front of fmt as we
+		// deal with them
+		while (fmt.size()) {
+			if (fmt[0] == ']' && fmt.size() > 1 && fmt[1] == ']') {
+				// that's the end
+				fmt = fmt.substr(2);
+				break;
+			}
+			// check for an embedded option
+			if (fmt[0] == '{' && fmt.size() > 1 && fmt[1] == '%') {
+				string part;
+				string const rest = parseEmbeddedOption(fmt, part);
+				if (fmt == rest) {
+					LYXERR0("ERROR! Couldn't parse embedded option in `" << format <<"'.");
+					return format;
+				}
+				clause += part;
+				fmt = rest;
+			} else { // it's just a normal character
+				clause += fmt[0];
+				fmt = fmt.substr(1);
+			}
+		}
+		return fmt;
+	}
+
+
+	// parse an options string, which must appear at the start of the
+	// format parameter. puts the parsed bits in optkey, ifpart, and
+	// elsepart and returns what's left after the option is removed.
+	// if there's an error, it returns format itself.
+	string parseOptions(string const & format, string & optkey,
+			string & ifpart, string & elsepart)
+	{
+		LASSERT(format[0] == '{' && format[1] == '%', return format);
+		// strip '{%'
+		string fmt = format.substr(2);
+		size_t pos = fmt.find('%'); // end of key
+		if (pos == string::npos) {
+			LYXERR0("Error parsing  `" << format <<"'. Can't find end of key.");
+			return format;
+		}
+		optkey = fmt.substr(0,pos);
+		fmt = fmt.substr(pos + 1);
+		// [[format]] should be next
+		if (fmt[0] != '[' || fmt[1] != '[') {
+			LYXERR0("Error parsing  `" << format <<"'. Can't find '[[' after key.");
+			return format;
+		}
+
+		string curfmt = fmt;
+		fmt = getClause(curfmt, ifpart);
+		if (fmt == curfmt) {
+			LYXERR0("Error parsing  `" << format <<"'. Couldn't get if clause.");
+			return format;
+		}
+
+		if (fmt[0] == '}') // we're done, no else clause
+			return fmt.substr(1);
+
+		// else part should follow
+		if (fmt[0] != '[' || fmt[1] != '[') {
+			LYXERR0("Error parsing  `" << format <<"'. Can't find else clause.");
+			return format;
+		}
+
+		curfmt = fmt;
+		fmt = getClause(curfmt, elsepart);
+		// we should be done
+		if (fmt == curfmt || fmt[0] != '}') {
+			LYXERR0("Error parsing  `" << format <<"'. Can't find end of option.");
+			return format;
+		}
+		return fmt.substr(1);
+}
+
+} // anon namespace
+
+
+docstring BibTeXInfo::expandFormat(string const & format,
+		BibTeXInfo const * const xref, int & counter, Buffer const & buf,
+		bool richtext) const
+{
+	// incorrect use of macros could put us in an infinite loop
+	static int max_passes = 5000;
+	docstring ret; // return value
+	string key;
+	bool scanning_key = false;
+	bool scanning_rich = false;
+
+	string fmt = format;
+	// we'll remove characters from the front of fmt as we
+	// deal with them
+	while (fmt.size()) {
+		if (counter++ > max_passes) {
+			LYXERR0("Recursion limit reached while parsing `"
+			        << format << "'.");
+			return _("ERROR!");
+		}
+
+		char_type thischar = fmt[0];
+		if (thischar == '%') {
+			// beginning or end of key
+			if (scanning_key) {
+				// end of key
+				scanning_key = false;
+				// so we replace the key with its value, which may be empty
+				if (key[0] == '!') {
+					// macro
+					string const val =
+						buf.params().documentClass().getCiteMacro(key);
+					fmt = val + fmt.substr(1);
+					continue;
+				} else if (key[0] == '_') {
+					// a translatable bit
+					string const val =
+						buf.params().documentClass().getCiteMacro(key);
+					docstring const trans =
+						translateIfPossible(from_utf8(val), buf.params().language->code());
+					ret += trans;
+				} else {
+					docstring const val = getValueForKey(key, xref);
+					ret += val;
+				}
+			} else {
+				// beginning of key
+				key.clear();
+				scanning_key = true;
+			}
+		}
+		else if (thischar == '{') {
+			// beginning of option?
+			if (scanning_key) {
+				LYXERR0("ERROR: Found `{' when scanning key in `" << format << "'.");
+				return _("ERROR!");
+			}
+			if (fmt.size() > 1) {
+				if (fmt[1] == '%') {
+					// it is the beginning of an optional format
+					string optkey;
+					string ifpart;
+					string elsepart;
+					string const newfmt =
+						parseOptions(fmt, optkey, ifpart, elsepart);
+					if (newfmt == fmt) // parse error
+						return _("ERROR!");
+					fmt = newfmt;
+					docstring const val = getValueForKey(optkey, xref);
+					if (!val.empty())
+						ret += expandFormat(ifpart, xref, counter, buf, richtext);
+					else if (!elsepart.empty())
+						ret += expandFormat(elsepart, xref, counter, buf, richtext);
+					// fmt will have been shortened for us already
+					continue;
+				}
+				if (fmt[1] == '!') {
+					// beginning of rich text
+					scanning_rich = true;
+					fmt = fmt.substr(2);
+					continue;
+				}
+			}
+			// we are here if '{' was not followed by % or !.
+			// So it's just a character.
+			ret += thischar;
+		}
+		else if (scanning_rich && thischar == '!'
+		         && fmt.size() > 1 && fmt[1] == '}') {
+			// end of rich text
+			scanning_rich = false;
+			fmt = fmt.substr(2);
+			continue;
+		}
+		else if (scanning_key)
+			key += char(thischar);
+		else if (richtext) {
+			if (scanning_rich)
+				ret += thischar;
+			else {
+				// we need to escape '<' and '>'
+				if (thischar == '<')
+					ret += "&lt;";
+				else if (thischar == '>')
+					ret += "&gt;";
+				else
+					ret += thischar;
+			}
+		}	else if (!scanning_rich /* && !richtext */)
+			ret += thischar;
+		// else the character is discarded, which will happen only if
+		// richtext == false and we are scanning rich text
+		fmt = fmt.substr(1);
+	} // for loop
+	if (scanning_key) {
+		LYXERR0("Never found end of key in `" << format << "'!");
+		return _("ERROR!");
+	}
+	if (scanning_rich) {
+		LYXERR0("Never found end of rich text in `" << format << "'!");
+		return _("ERROR!");
+	}
+	return ret;
+}
+
+
+docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref,
+	Buffer const & buf, bool richtext) const
 {
 	if (!info_.empty())
 		return info_;
@@ -278,59 +526,15 @@ docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref) const
 		info_ = it->second;
 		return info_;
 	}
- 
-	// FIXME
-	// This could be made a lot better using the entry_type_
-	// field to customize the output based upon entry type.
-	
-	// Search for all possible "required" fields
-	docstring author = getValueForKey("author", xref);
-	if (author.empty())
-		author = getValueForKey("editor", xref);
- 
-	docstring year   = getValueForKey("year", xref);
-	docstring title  = getValueForKey("title", xref);
-	docstring docLoc = getValueForKey("pages", xref);
-	if (docLoc.empty()) {
-		docLoc = getValueForKey("chapter", xref);
-		if (!docLoc.empty())
-			docLoc = _("Ch. ") + docLoc;
-	}	else {
-		docLoc = _("pp. ") + docLoc;
-	}
 
-	docstring media = getValueForKey("journal", xref);
-	if (media.empty()) {
-		media = getValueForKey("publisher", xref);
-		if (media.empty()) {
-			media = getValueForKey("school", xref);
-			if (media.empty())
-				media = getValueForKey("institution");
-		}
-	}
-	docstring volume = getValueForKey("volume", xref);
-
-	odocstringstream result;
-	if (!author.empty())
-		result << author << ", ";
-	if (!title.empty())
-		result << title;
-	if (!media.empty())
-		result << ", " << media;
-	if (!year.empty())
-		result << " (" << year << ")";
-	if (!docLoc.empty())
-		result << ", " << docLoc;
-
-	docstring const result_str = rtrim(result.str());
-	if (!result_str.empty()) {
-		info_ = convertLaTeXCommands(result_str);
-		return info_;
-	}
+	DocumentClass const & dc = buf.params().documentClass();
+	string const & format = dc.getCiteFormat(to_utf8(entry_type_));
+	int counter = 0;
+	info_ = expandFormat(format, xref, counter, buf, richtext);
 
-	// This should never happen (or at least be very unusual!)
-	static docstring e = docstring();
-	return e;
+	if (!info_.empty())
+		info_ = convertLaTeXCommands(info_);
+	return info_;
 }
 
 
@@ -342,15 +546,15 @@ docstring const & BibTeXInfo::operator[](docstring const & field) const
 	static docstring const empty_value = docstring();
 	return empty_value;
 }
-	
-	
+
+
 docstring const & BibTeXInfo::operator[](string const & field) const
 {
 	return operator[](from_ascii(field));
 }
 
 
-docstring BibTeXInfo::getValueForKey(string const & key, 
+docstring BibTeXInfo::getValueForKey(string const & key,
 		BibTeXInfo const * const xref) const
 {
 	docstring const ret = operator[](key);
@@ -423,34 +627,47 @@ docstring const BiblioInfo::getAbbreviatedAuthor(docstring const & key) const
 }
 
 
-docstring const BiblioInfo::getYear(docstring const & key) const
+docstring const BiblioInfo::getCiteNumber(docstring const & key) const
 {
 	BiblioInfo::const_iterator it = find(key);
 	if (it == end())
 		return docstring();
 	BibTeXInfo const & data = it->second;
-	docstring year = data.getYear();
-	if (!year.empty())
-		return year;
-	// let's try the crossref
-	docstring const xref = data.getXRef();
-	if (xref.empty())
-		return _("No year"); // no luck
-	BiblioInfo::const_iterator const xrefit = find(xref);
-	if (xrefit == end())
-		return _("No year"); // no luck again
-	BibTeXInfo const & xref_data = xrefit->second;
-	return xref_data.getYear();
-	return data.getYear();
+	return data.citeNumber();
 }
 
 
-docstring const BiblioInfo::getInfo(docstring const & key) const
+docstring const BiblioInfo::getYear(docstring const & key, bool use_modifier) const
 {
 	BiblioInfo::const_iterator it = find(key);
 	if (it == end())
 		return docstring();
 	BibTeXInfo const & data = it->second;
+	docstring year = data.getYear();
+	if (year.empty()) {
+		// let's try the crossref
+		docstring const xref = data.getXRef();
+		if (xref.empty())
+			return _("No year"); // no luck
+		BiblioInfo::const_iterator const xrefit = find(xref);
+		if (xrefit == end())
+			return _("No year"); // no luck again
+		BibTeXInfo const & xref_data = xrefit->second;
+		year = xref_data.getYear();
+	}
+	if (use_modifier && data.modifier() != 0)
+		year += data.modifier();
+	return year;
+}
+
+
+docstring const BiblioInfo::getInfo(docstring const & key,
+	Buffer const & buf, bool richtext) const
+{
+	BiblioInfo::const_iterator it = find(key);
+	if (it == end())
+		return docstring(_("BibTeX entry not found!"));
+	BibTeXInfo const & data = it->second;
 	BibTeXInfo const * xrefptr = 0;
 	docstring const xref = data.getXRef();
 	if (!xref.empty()) {
@@ -458,10 +675,20 @@ docstring const BiblioInfo::getInfo(docstring const & key) const
 		if (xrefit != end())
 			xrefptr = &(xrefit->second);
 	}
-	return data.getInfo(xrefptr);
+	return data.getInfo(xrefptr, buf, richtext);
+}
+
+
+bool BiblioInfo::isBibtex(docstring const & key) const
+{
+	BiblioInfo::const_iterator it = find(key);
+	if (it == end())
+		return false;
+	return it->second.isBibTeX();
 }
 
 
+
 vector<docstring> const BiblioInfo::getCiteStrings(
 	docstring const & key, Buffer const & buf) const
 {
@@ -485,7 +712,7 @@ vector<docstring> const BiblioInfo::getNumericalStrings(
 		return vector<docstring>();
 
 	vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
-	
+
 	vector<docstring> vec(styles.size());
 	for (size_t i = 0; i != vec.size(); ++i) {
 		docstring str;
@@ -544,7 +771,7 @@ vector<docstring> const BiblioInfo::getAuthorYearStrings(
 		return vector<docstring>();
 
 	vector<CiteStyle> const & styles = citeStyles(buf.params().citeEngine());
-	
+
 	vector<docstring> vec(styles.size());
 	for (size_t i = 0; i != vec.size(); ++i) {
 		docstring str;
@@ -597,6 +824,8 @@ vector<docstring> const BiblioInfo::getAuthorYearStrings(
 void BiblioInfo::mergeBiblioInfo(BiblioInfo const & info)
 {
 	bimap_.insert(info.begin(), info.end());
+	field_names_.insert(info.field_names_.begin(), info.field_names_.end());
+	entry_types_.insert(info.entry_types_.begin(), info.entry_types_.end());
 }
 
 
@@ -604,7 +833,15 @@ namespace {
 	// used in xhtml to sort a list of BibTeXInfo objects
 	bool lSorter(BibTeXInfo const * lhs, BibTeXInfo const * rhs)
 	{
-		return lhs->getAbbreviatedAuthor() < rhs->getAbbreviatedAuthor();
+		docstring const lauth = lhs->getAbbreviatedAuthor();
+		docstring const rauth = rhs->getAbbreviatedAuthor();
+		docstring const lyear = lhs->getYear();
+		docstring const ryear = rhs->getYear();
+		docstring const ltitl = lhs->operator[]("title");
+		docstring const rtitl = rhs->operator[]("title");
+		return  (lauth < rauth)
+				|| (lauth == rauth && lyear < ryear)
+				|| (lauth == rauth && lyear == ryear && ltitl < rtitl);
 	}
 }
 
@@ -628,9 +865,9 @@ void BiblioInfo::collectCitedEntries(Buffer const & buf)
 	}
 	if (citekeys.empty())
 		return;
-	
+
 	// We have a set of the keys used in this document.
-	// We will now convert it to a list of the BibTeXInfo objects used in 
+	// We will now convert it to a list of the BibTeXInfo objects used in
 	// this document...
 	vector<BibTeXInfo const *> bi;
 	set<docstring>::const_iterator cit = citekeys.begin();
@@ -643,7 +880,7 @@ void BiblioInfo::collectCitedEntries(Buffer const & buf)
 	}
 	// ...and sort it.
 	sort(bi.begin(), bi.end(), lSorter);
-	
+
 	// Now we can write the sorted keys
 	vector<BibTeXInfo const *>::const_iterator bit = bi.begin();
 	vector<BibTeXInfo const *>::const_iterator ben = bi.end();
@@ -655,19 +892,50 @@ void BiblioInfo::collectCitedEntries(Buffer const & buf)
 void BiblioInfo::makeCitationLabels(Buffer const & buf)
 {
 	collectCitedEntries(buf);
-	// FIXME It'd be nice to do author-year as well as numerical
-	// and maybe even some other sorts of labels.
+	CiteEngine const engine = buf.params().citeEngine();
+	bool const numbers =
+		(engine == ENGINE_BASIC || engine == ENGINE_NATBIB_NUMERICAL);
+
+	int keynumber = 0;
+	char modifier = 0;
+	// used to remember the last one we saw
+	// we'll be comparing entries to see if we need to add
+	// modifiers, like "1984a"
+	map<docstring, BibTeXInfo>::iterator last;
+
 	vector<docstring>::const_iterator it = cited_entries_.begin();
 	vector<docstring>::const_iterator const en = cited_entries_.end();
-	int keynumber = 0;
 	for (; it != en; ++it) {
 		map<docstring, BibTeXInfo>::iterator const biit = bimap_.find(*it);
 		// this shouldn't happen, but...
 		if (biit == bimap_.end())
+			// ...fail gracefully, anyway.
 			continue;
 		BibTeXInfo & entry = biit->second;
-		docstring const key = convert<docstring>(++keynumber);
-		entry.setCiteKey(key);
+		if (numbers) {
+			docstring const num = convert<docstring>(++keynumber);
+			entry.setCiteNumber(num);
+		} else {
+			if (it != cited_entries_.begin()
+			    && entry.getAbbreviatedAuthor() == last->second.getAbbreviatedAuthor()
+			    // we access the year via getYear() so as to get it from the xref,
+			    // if we need to do so
+			    && getYear(entry.key()) == getYear(last->second.key())) {
+				if (modifier == 0) {
+					// so the last one should have been 'a'
+					last->second.setModifier('a');
+					modifier = 'b';
+				} else if (modifier == 'z')
+					modifier = 'A';
+				else
+					modifier++;
+			} else {
+				modifier = 0;
+			}
+			entry.setModifier(modifier);
+			// remember the last one
+			last = biit;
+		}
 	}
 }
 
@@ -759,30 +1027,35 @@ string citationStyleToString(const CitationStyle & s)
 
 vector<CiteStyle> citeStyles(CiteEngine engine)
 {
-	unsigned int nStyles = 0;
-	unsigned int start = 0;
+	vector<CiteStyle> styles(0);
 
 	switch (engine) {
 		case ENGINE_BASIC:
-			nStyles = 2;
-			start = 0;
+			styles.push_back(CITE);
 			break;
+		case ENGINE_JURABIB:
+			styles.push_back(CITE);
 		case ENGINE_NATBIB_AUTHORYEAR:
-		case ENGINE_NATBIB_NUMERICAL:
-			nStyles = nCiteStyles - 1;
-			start = 1;
+			styles.push_back(CITET);
+			styles.push_back(CITEP);
+			styles.push_back(CITEALT);
+			styles.push_back(CITEALP);
+			styles.push_back(CITEAUTHOR);
+			styles.push_back(CITEYEAR);
+			styles.push_back(CITEYEARPAR);
 			break;
-		case ENGINE_JURABIB:
-			nStyles = nCiteStyles;
-			start = 0;
+		case ENGINE_NATBIB_NUMERICAL:
+			styles.push_back(CITET);
+			styles.push_back(CITEALT);
+			styles.push_back(CITEAUTHOR);
+			styles.push_back(CITEP);
+			styles.push_back(CITEALP);
+			styles.push_back(CITEYEAR);
+			styles.push_back(CITEYEARPAR);
 			break;
 	}
 
-	vector<CiteStyle> styles(nStyles);
-	size_t i = 0;
-	int j = start;
-	for (; i != styles.size(); ++i, ++j)
-		styles[i] = citeStylesArray[j];
+	styles.push_back(NOCITE);
 
 	return styles;
 }