X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2FBiblioInfo.cpp;h=70127798522effae211a946b48d67658c7600892;hb=3176ac365ce5bdb85c00d1afc36d7feac2438847;hp=93094f8bfa0e704e10b584fcfb1dd69d3b0c9af4;hpb=43f6b1672b5ef442b4e9b7e4f7ea76dc14ac4fc8;p=lyx.git

diff --git a/src/BiblioInfo.cpp b/src/BiblioInfo.cpp
index 93094f8bfa..7012779852 100644
--- a/src/BiblioInfo.cpp
+++ b/src/BiblioInfo.cpp
@@ -6,6 +6,7 @@
  * \author Angus Leeming
  * \author Herbert Voß
  * \author Richard Heck
+ * \author Julien Rioux
  *
  * Full author contact details are available in file CREDITS.
  */
@@ -19,6 +20,7 @@
 #include "Encoding.h"
 #include "InsetIterator.h"
 #include "Language.h"
+#include "output_xhtml.h"
 #include "Paragraph.h"
 #include "TextClass.h"
 #include "TocBackend.h"
@@ -89,6 +91,7 @@ docstring familyName(docstring const & name)
 	return retval;
 }
 
+
 // converts a string containing LaTeX commands into unicode
 // for display.
 docstring convertLaTeXCommands(docstring const & str)
@@ -194,6 +197,49 @@ docstring convertLaTeXCommands(docstring const & str)
 	return ret;
 }
 
+
+// Escape '<' and '>' and remove richtext markers (e.g. {!this is richtext!}) from a string.
+docstring processRichtext(docstring const & str, bool richtext)
+{
+	docstring val = str;
+	docstring ret;
+
+	bool scanning_rich = false;
+	while (!val.empty()) {
+		char_type const ch = val[0];
+		if (ch == '{' && val.size() > 1 && val[1] == '!') {
+			// beginning of rich text
+			scanning_rich = true;
+			val = val.substr(2);
+			continue;
+		}
+		if (scanning_rich && ch == '!' && val.size() > 1 && val[1] == '}') {
+			// end of rich text
+			scanning_rich = false;
+			val = val.substr(2);
+			continue;
+		}
+		if (richtext) {
+			if (scanning_rich)
+				ret += ch;
+			else {
+				// we need to escape '<' and '>'
+				if (ch == '<')
+					ret += "&lt;";
+				else if (ch == '>')
+					ret += "&gt;";
+				else
+					ret += ch;
+			}
+		} else if (!scanning_rich /* && !richtext */)
+			ret += ch;
+		// else the character is discarded, which will happen only if
+		// richtext == false and we are scanning rich text
+		val = val.substr(1);
+	}
+	return ret;
+}
+
 } // anon namespace
 
 
@@ -225,9 +271,9 @@ docstring const BibTeXInfo::getAbbreviatedAuthor(bool jurabib_style) const
 		return authors;
 	}
 
-	docstring author = convertLaTeXCommands(operator[]("author"));
+	docstring author = operator[]("author");
 	if (author.empty()) {
-		author = convertLaTeXCommands(operator[]("editor"));
+		author = operator[]("editor");
 		if (author.empty())
 			return author;
 	}
@@ -245,17 +291,35 @@ docstring const BibTeXInfo::getAbbreviatedAuthor(bool jurabib_style) const
 			+ "/" + familyName(authors[1]);
 		if (authors.size() == 3)
 			shortauthor += "/" + familyName(authors[2]);
-		return shortauthor;
+		return convertLaTeXCommands(shortauthor);
 	}
 
-	if (authors.size() == 2)
-		return bformat(_("%1$s and %2$s"),
+	docstring retval = familyName(authors[0]);
+
+	if (authors.size() == 2 && authors[1] != "others")
+		retval = bformat(from_ascii("%1$s and %2$s"),
 			familyName(authors[0]), familyName(authors[1]));
 
-	if (authors.size() > 2)
-		return bformat(_("%1$s et al."), familyName(authors[0]));
+	if (authors.size() >= 2)
+		retval = bformat(from_ascii("%1$s et al."),
+			familyName(authors[0]));
+
+	return convertLaTeXCommands(retval);
+}
+
 
-	return familyName(authors[0]);
+docstring const BibTeXInfo::getAbbreviatedAuthor(Buffer const & buf, bool jurabib_style) const
+{
+	docstring const author = getAbbreviatedAuthor(jurabib_style);
+	if (!is_bibtex_)
+		return author;
+	vector<docstring> const authors = getVectorFromString(author, from_ascii(" and "));
+	if (authors.size() == 2)
+		return bformat(buf.B_("%1$s and %2$s"), authors[0], authors[1]);
+	docstring::size_type const idx = author.rfind(from_ascii(" et al."));
+	if (idx != docstring::npos)
+		return bformat(buf.B_("%1$s et al."), author.substr(0, idx));
+	return author;
 }
 
 
@@ -288,135 +352,151 @@ docstring const BibTeXInfo::getXRef() 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;
+
+docstring parseOptions(docstring const & format, string & optkey,
+		    docstring & ifpart, docstring & 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.
+docstring parseEmbeddedOption(docstring const & format, docstring & ifelsepart)
+{
+	LASSERT(format[0] == '{' && format[1] == '%', return format);
+	string optkey;
+	docstring ifpart;
+	docstring elsepart;
+	docstring 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 = docstring(); return format; });
+	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.
+docstring getClause(docstring const & format, docstring & clause)
+{
+	docstring fmt = format;
+	// remove '[['
+	fmt = fmt.substr(2);
+	// we'll remove characters from the front of fmt as we
+	// deal with them
+	while (!fmt.empty()) {
+		if (fmt[0] == ']' && fmt.size() > 1 && fmt[1] == ']') {
+			// that's the end
+			fmt = fmt.substr(2);
+			break;
 		}
-		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.empty()) {
-			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] == '%') {
+			docstring part;
+			docstring const rest = parseEmbeddedOption(fmt, part);
+			if (fmt == rest) {
+				LYXERR0("ERROR! Couldn't parse embedded option in `" << format <<"'.");
+				return format;
 			}
-			// 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 += 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;
 		}
+	}
+	return fmt;
+}
 
-		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);
+// 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.
+docstring parseOptions(docstring const & format, string & optkey,
+		    docstring & ifpart, docstring & elsepart)
+{
+	LASSERT(format[0] == '{' && format[1] == '%', return format);
+	// strip '{%'
+	docstring 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 = to_utf8(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;
+	}
 
-		// else part should follow
-		if (fmt[0] != '[' || fmt[1] != '[') {
-			LYXERR0("Error parsing  `" << format <<"'. Can't find else clause.");
-			return format;
-		}
+	docstring curfmt = fmt;
+	fmt = getClause(curfmt, ifpart);
+	if (fmt == curfmt) {
+		LYXERR0("Error parsing  `" << format <<"'. Couldn't get if 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;
-		}
+	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
 
+} // anon namespace
 
-docstring BibTeXInfo::expandFormat(string const & format,
+/* FIXME
+Bug #9131 revealed an oddity in how we are generating citation information
+when more than one key is given. We end up building a longer and longer format 
+string as we go, which we then have to re-parse, over and over and over again,
+rather than generating the information for the individual keys and then putting
+all of that together. We do that to deal with the way separators work, from what
+I can tell, but it still feels like a hack. Fixing this would require quite a
+bit of work, however.
+*/
+docstring BibTeXInfo::expandFormat(docstring const & format,
 		BibTeXInfo const * const xref, int & counter, Buffer const & buf,
-		bool richtext, docstring before, docstring after, docstring dialog, bool next) const
+		docstring before, docstring after, docstring dialog, bool next) const
 {
 	// incorrect use of macros could put us in an infinite loop
-	static int max_passes = 5000;
-	docstring ret; // return value
+	static int const max_passes = 5000;
+	// the use of overly large keys can lead to performance problems, due
+	// to eventual attempts to convert LaTeX macros to unicode. See bug
+	// #8944. This is perhaps not the best solution, but it will have to
+	// do for now.
+	static size_t const max_keysize = 128;
+	odocstringstream ret; // return value
 	string key;
 	bool scanning_key = false;
 	bool scanning_rich = false;
 
 	CiteEngineType const engine_type = buf.params().citeEngineType();
-	string fmt = format;
+	docstring fmt = format;
 	// we'll remove characters from the front of fmt as we
 	// deal with them
 	while (!fmt.empty()) {
-		if (counter++ > max_passes) {
+		if (counter > max_passes) {
 			LYXERR0("Recursion limit reached while parsing `"
 			        << format << "'.");
 			return _("ERROR!");
@@ -431,11 +511,10 @@ docstring BibTeXInfo::expandFormat(string const & format,
 				// so we replace the key with its value, which may be empty
 				if (key[0] == '!') {
 					// macro
-					// FIXME: instead of passing the buf, just past the macros
-					// FIXME: and the language code
 					string const val =
 						buf.params().documentClass().getCiteMacro(engine_type, key);
-					fmt = val + fmt.substr(1);
+					fmt = from_utf8(val) + fmt.substr(1);
+					counter += 1;
 					continue;
 				} else if (key[0] == '_') {
 					// a translatable bit
@@ -443,10 +522,15 @@ docstring BibTeXInfo::expandFormat(string const & format,
 						buf.params().documentClass().getCiteMacro(engine_type, key);
 					docstring const trans =
 						translateIfPossible(from_utf8(val), buf.params().language->code());
-					ret += trans;
+					ret << trans;
 				} else {
-					docstring const val = getValueForKey(key, before, after, dialog, xref);
-					ret += val;
+					docstring const val =
+						getValueForKey(key, buf, before, after, dialog, xref, max_keysize);
+					if (!scanning_rich)
+						ret << from_ascii("{!<span class=\"bib-" + key + "\">!}");
+					ret << val;
+					if (!scanning_rich)
+						ret << from_ascii("{!</span>!}");
 				}
 			} else {
 				// beginning of key
@@ -464,22 +548,26 @@ docstring BibTeXInfo::expandFormat(string const & format,
 				if (fmt[1] == '%') {
 					// it is the beginning of an optional format
 					string optkey;
-					string ifpart;
-					string elsepart;
-					string const newfmt =
+					docstring ifpart;
+					docstring elsepart;
+					docstring const newfmt =
 						parseOptions(fmt, optkey, ifpart, elsepart);
 					if (newfmt == fmt) // parse error
 						return _("ERROR!");
 					fmt = newfmt;
-					docstring const val = getValueForKey(optkey, before, after, dialog, xref);
+					docstring const val =
+						getValueForKey(optkey, buf, before, after, dialog, xref);
 					if (optkey == "next" && next)
-						ret += from_utf8(ifpart); // without expansion
-					else if (!val.empty())
-						ret += expandFormat(ifpart, xref, counter, buf,
-							richtext, before, after, dialog, next);
-					else if (!elsepart.empty())
-						ret += expandFormat(elsepart, xref, counter, buf,
-							richtext, before, after, dialog, next);
+						ret << ifpart; // without expansion
+					else if (!val.empty()) {
+						int newcounter = 0;
+						ret << expandFormat(ifpart, xref, newcounter, buf,
+							before, after, dialog, next);
+					} else if (!elsepart.empty()) {
+						int newcounter = 0;
+						ret << expandFormat(elsepart, xref, newcounter, buf,
+							before, after, dialog, next);
+					}
 					// fmt will have been shortened for us already
 					continue;
 				}
@@ -487,38 +575,31 @@ docstring BibTeXInfo::expandFormat(string const & format,
 					// beginning of rich text
 					scanning_rich = true;
 					fmt = fmt.substr(2);
+					ret << from_ascii("{!");
 					continue;
 				}
 			}
 			// we are here if '{' was not followed by % or !.
 			// So it's just a character.
-			ret += thischar;
+			ret << thischar;
 		}
 		else if (scanning_rich && thischar == '!'
 		         && fmt.size() > 1 && fmt[1] == '}') {
 			// end of rich text
 			scanning_rich = false;
 			fmt = fmt.substr(2);
+			ret << from_ascii("!}");
 			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 {
+			try {
+				ret.put(thischar);
+			} catch (EncodingException & /* e */) {
+				LYXERR0("Uncodable character '" << docstring(1, thischar) << " in citation label!");
 			}
-		}	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) {
@@ -529,15 +610,17 @@ docstring BibTeXInfo::expandFormat(string const & format,
 		LYXERR0("Never found end of rich text in `" << format << "'!");
 		return _("ERROR!");
 	}
-	return ret;
+	return ret.str();
 }
 
 
 docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref,
 	Buffer const & buf, bool richtext) const
 {
-	if (!info_.empty())
+	if (!richtext && !info_.empty())
 		return info_;
+	if (richtext && !info_richtext_.empty())
+		return info_richtext_;
 
 	if (!is_bibtex_) {
 		BibTeXInfo::const_iterator it = find(from_ascii("ref"));
@@ -547,36 +630,43 @@ docstring const & BibTeXInfo::getInfo(BibTeXInfo const * const xref,
 
 	CiteEngineType const engine_type = buf.params().citeEngineType();
 	DocumentClass const & dc = buf.params().documentClass();
-	string const & format = dc.getCiteFormat(engine_type, to_utf8(entry_type_));
+	docstring const & format =
+		from_utf8(dc.getCiteFormat(engine_type, to_utf8(entry_type_)));
 	int counter = 0;
-	info_ = expandFormat(format, xref, counter, buf, richtext);
+	info_ = expandFormat(format, xref, counter, buf,
+		docstring(), docstring(), docstring(), false);
+
+	if (info_.empty()) {
+		// this probably shouldn't happen
+		return info_;
+	}
 
-	if (!info_.empty())
-		info_ = convertLaTeXCommands(info_);
+	if (richtext) {
+		info_richtext_ = convertLaTeXCommands(processRichtext(info_, true));
+		return info_richtext_;
+	}
+
+	info_ = convertLaTeXCommands(processRichtext(info_, false));
 	return info_;
 }
 
 
 docstring const BibTeXInfo::getLabel(BibTeXInfo const * const xref,
-	Buffer const & buf, string const & format, bool richtext,
-	docstring before, docstring after, docstring dialog, bool next) const
+	Buffer const & buf, docstring const & format, bool richtext,
+	docstring const & before, docstring const & after, 
+	docstring const & dialog, bool next) const
 {
 	docstring loclabel;
 
-	/*
-	if (!is_bibtex_) {
-		BibTeXInfo::const_iterator it = find(from_ascii("ref"));
-		label_ = it->second;
-		return label_;
-	}
-	*/
-
 	int counter = 0;
-	loclabel = expandFormat(format, xref, counter, buf, richtext,
+	loclabel = expandFormat(format, xref, counter, buf,
 		before, after, dialog, next);
 
-	if (!loclabel.empty())
+	if (!loclabel.empty() && !next) {
+		loclabel = processRichtext(loclabel, richtext);
 		loclabel = convertLaTeXCommands(loclabel);
+	}
+
 	return loclabel;
 }
 
@@ -597,48 +687,77 @@ docstring const & BibTeXInfo::operator[](string const & field) const
 }
 
 
-docstring BibTeXInfo::getValueForKey(string const & key,
+docstring BibTeXInfo::getValueForKey(string const & oldkey, Buffer const & buf,
 	docstring const & before, docstring const & after, docstring const & dialog,
-	BibTeXInfo const * const xref) const
+	BibTeXInfo const * const xref, size_t maxsize) const
 {
+	// anything less is pointless
+	LASSERT(maxsize >= 16, maxsize = 16);
+	string key = oldkey;
+	bool cleanit = false;
+	if (prefixIs(oldkey, "clean:")) {
+		key = oldkey.substr(6);
+		cleanit = true;
+	}
+
 	docstring ret = operator[](key);
 	if (ret.empty() && xref)
 		ret = (*xref)[key];
-	if (!ret.empty())
-		return ret;
-	// some special keys
-	// FIXME: dialog, textbefore and textafter have nothing to do with this
-	if (key == "dialog")
-		return dialog;
-	else if (key == "entrytype")
-		return entry_type_;
-	else if (key == "key")
-		return bib_key_;
-	else if (key == "label")
-		return label_;
-	else if (key == "abbrvauthor")
-		// Special key to provide abbreviated author names.
-		return getAbbreviatedAuthor();
-	else if (key == "shortauthor")
-		// When shortauthor is not defined, jurabib automatically
-		// provides jurabib-style abbreviated author names. We do
-		// this as well.
-		return getAbbreviatedAuthor(true);
-	else if (key == "shorttitle") {
-		// When shorttitle is not defined, jurabib uses for `article'
-		// and `periodical' entries the form `journal volume [year]'
-		// and for other types of entries it uses the `title' field.
-		if (entry_type_ == "article" || entry_type_ == "periodical")
-			return operator[]("journal") + " " + operator[]("volume")
-				+ " [" + operator[]("year") + "]";
-		else
-			return operator[]("title");
-	} else if (key == "textbefore")
-		return before;
-	else if (key == "textafter")
-		return after;
-	else if (key == "year")
-		return getYear();
+	if (ret.empty()) {
+		// some special keys
+		// FIXME: dialog, textbefore and textafter have nothing to do with this
+		if (key == "dialog")
+			ret = dialog;
+		else if (key == "entrytype")
+			ret = entry_type_;
+		else if (key == "key")
+			ret = bib_key_;
+		else if (key == "label")
+			ret = label_;
+		else if (key == "modifier" && modifier_ != 0)
+			ret = modifier_;
+		else if (key == "numericallabel")
+			ret = cite_number_;
+		else if (key == "abbrvauthor")
+			// Special key to provide abbreviated author names.
+			ret = getAbbreviatedAuthor(buf, false);
+		else if (key == "shortauthor")
+			// When shortauthor is not defined, jurabib automatically
+			// provides jurabib-style abbreviated author names. We do
+			// this as well.
+			ret = getAbbreviatedAuthor(buf, true);
+		else if (key == "shorttitle") {
+			// When shorttitle is not defined, jurabib uses for `article'
+			// and `periodical' entries the form `journal volume [year]'
+			// and for other types of entries it uses the `title' field.
+			if (entry_type_ == "article" || entry_type_ == "periodical")
+				ret = operator[]("journal") + " " + operator[]("volume")
+					+ " [" + operator[]("year") + "]";
+			else
+				ret = operator[]("title");
+		} else if (key == "bibentry") {
+			// Special key to provide the full bibliography entry: see getInfo()
+			CiteEngineType const engine_type = buf.params().citeEngineType();
+			DocumentClass const & dc = buf.params().documentClass();
+			docstring const & format =
+				from_utf8(dc.getCiteFormat(engine_type, to_utf8(entry_type_)));
+			int counter = 0;
+			ret = expandFormat(format, xref, counter, buf,
+				docstring(), docstring(), docstring(), false);
+		} else if (key == "textbefore")
+			ret = before;
+		else if (key == "textafter")
+			ret = after;
+		else if (key == "year")
+			ret = getYear();
+	}
+
+	if (cleanit)
+		ret = html::cleanAttr(ret);
+
+	// make sure it is not too big
+	if (ret.size() > maxsize)
+		ret = ret.substr(0, maxsize - 3) + from_ascii("...");
 	return ret;
 }
 
@@ -650,14 +769,16 @@ docstring BibTeXInfo::getValueForKey(string const & key,
 //////////////////////////////////////////////////////////////////////
 
 namespace {
+
 // A functor for use with sort, leading to case insensitive sorting
-	class compareNoCase: public binary_function<docstring, docstring, bool>
-	{
-		public:
-			bool operator()(docstring const & s1, docstring const & s2) const {
-				return compare_no_case(s1, s2) < 0;
-			}
-	};
+class compareNoCase: public binary_function<docstring, docstring, bool>
+{
+public:
+	bool operator()(docstring const & s1, docstring const & s2) const {
+		return compare_no_case(s1, s2) < 0;
+	}
+};
+
 } // namespace anon
 
 
@@ -696,13 +817,13 @@ vector<docstring> const BiblioInfo::getEntries() const
 }
 
 
-docstring const BiblioInfo::getAbbreviatedAuthor(docstring const & key) const
+docstring const BiblioInfo::getAbbreviatedAuthor(docstring const & key, Buffer const & buf) const
 {
 	BiblioInfo::const_iterator it = find(key);
 	if (it == end())
 		return docstring();
 	BibTeXInfo const & data = it->second;
-	return data.getAbbreviatedAuthor();
+	return data.getAbbreviatedAuthor(buf, false);
 }
 
 
@@ -727,10 +848,12 @@ docstring const BiblioInfo::getYear(docstring const & key, bool use_modifier) co
 		// let's try the crossref
 		docstring const xref = data.getXRef();
 		if (xref.empty())
-			return _("No year"); // no luck
+			// no luck
+			return docstring();
 		BiblioInfo::const_iterator const xrefit = find(xref);
 		if (xrefit == end())
-			return _("No year"); // no luck again
+			// no luck again
+			return docstring();
 		BibTeXInfo const & xref_data = xrefit->second;
 		year = xref_data.getYear();
 	}
@@ -740,6 +863,15 @@ docstring const BiblioInfo::getYear(docstring const & key, bool use_modifier) co
 }
 
 
+docstring const BiblioInfo::getYear(docstring const & key, Buffer const & buf, bool use_modifier) const
+{
+	docstring const year = getYear(key, use_modifier);
+	if (year.empty())
+		return buf.B_("No year");
+	return year;
+}
+
+
 docstring const BiblioInfo::getInfo(docstring const & key,
 	Buffer const & buf, bool richtext) const
 {
@@ -758,14 +890,23 @@ docstring const BiblioInfo::getInfo(docstring const & key,
 }
 
 
-docstring const BiblioInfo::getLabel(vector<docstring> const & keys,
-	Buffer const & buf, string const & style, bool richtext,
-	docstring const & before, docstring const & after, docstring const & dialog) const
+docstring const BiblioInfo::getLabel(vector<docstring> keys,
+	Buffer const & buf, string const & style, bool for_xhtml,
+	size_t max_size, docstring const & before, docstring const & after,
+	docstring const & dialog) const
 {
+	// shorter makes no sense
+	LASSERT(max_size >= 16, max_size = 16);
+
+	// we can't display more than 10 of these, anyway
+	bool const too_many_keys = keys.size() > 10;
+	if (too_many_keys)
+		keys.resize(10);
+
 	CiteEngineType const engine_type = buf.params().citeEngineType();
 	DocumentClass const & dc = buf.params().documentClass();
-	string const & format = dc.getCiteFormat(engine_type, style, "cite");
-	docstring ret = from_utf8(format);
+	docstring const & format = from_utf8(dc.getCiteFormat(engine_type, style, "cite"));
+	docstring ret = format;
 	vector<docstring>::const_iterator key = keys.begin();
 	vector<docstring>::const_iterator ken = keys.end();
 	for (; key != ken; ++key) {
@@ -783,8 +924,17 @@ docstring const BiblioInfo::getLabel(vector<docstring> const & keys,
 					xrefptr = &(xrefit->second);
 			}
 		}
-		ret = data.getLabel(xrefptr, buf, to_utf8(ret), richtext,
-			before, after, dialog, key+1 != ken);
+		ret = data.getLabel(xrefptr, buf, ret, for_xhtml,
+			before, after, dialog, key + 1 != ken);
+	}
+
+	if (ret.size() > max_size) {
+		ret.resize(max_size - 3);
+		ret += "...";
+	} else if (too_many_keys) {
+		if (ret.size() > max_size - 3)
+			ret.resize(max_size - 3);
+		ret += "...";
 	}
 	return ret;
 }
@@ -792,7 +942,9 @@ docstring const BiblioInfo::getLabel(vector<docstring> const & keys,
 
 bool BiblioInfo::isBibtex(docstring const & key) const
 {
-	BiblioInfo::const_iterator it = find(key);
+	docstring key1;
+	split(key, key1, ',');
+	BiblioInfo::const_iterator it = find(key1);
 	if (it == end())
 		return false;
 	return it->second.isBibTeX();
@@ -801,8 +953,8 @@ bool BiblioInfo::isBibtex(docstring const & key) const
 
 vector<docstring> const BiblioInfo::getCiteStrings(
 	vector<docstring> const & keys, vector<CitationStyle> const & styles,
-	Buffer const & buf, bool richtext, docstring const & before,
-	docstring const & after, docstring const & dialog) const
+	Buffer const & buf, docstring const & before,
+	docstring const & after, docstring const & dialog, size_t max_size) const
 {
 	if (empty())
 		return vector<docstring>();
@@ -811,7 +963,7 @@ vector<docstring> const BiblioInfo::getCiteStrings(
 	vector<docstring> vec(styles.size());
 	for (size_t i = 0; i != vec.size(); ++i) {
 		style = styles[i].cmd;
-		vec[i] = getLabel(keys, buf, style, richtext, before, after, dialog);
+		vec[i] = getLabel(keys, buf, style, false, max_size, before, after, dialog);
 	}
 
 	return vec;
@@ -827,19 +979,21 @@ void BiblioInfo::mergeBiblioInfo(BiblioInfo const & info)
 
 
 namespace {
-	// used in xhtml to sort a list of BibTeXInfo objects
-	bool lSorter(BibTeXInfo const * lhs, BibTeXInfo const * rhs)
-	{
-		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);
-	}
+
+// used in xhtml to sort a list of BibTeXInfo objects
+bool lSorter(BibTeXInfo const * lhs, BibTeXInfo const * rhs)
+{
+	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);
+}
+
 }
 
 
@@ -890,7 +1044,7 @@ void BiblioInfo::makeCitationLabels(Buffer const & buf)
 {
 	collectCitedEntries(buf);
 	CiteEngineType const engine_type = buf.params().citeEngineType();
-	bool const numbers = (engine_type == ENGINE_TYPE_NUMERICAL);
+	bool const numbers = (engine_type & ENGINE_TYPE_NUMERICAL);
 
 	int keynumber = 0;
 	char modifier = 0;
@@ -933,6 +1087,28 @@ void BiblioInfo::makeCitationLabels(Buffer const & buf)
 			last = biit;
 		}
 	}
+	// Set the labels
+	it = cited_entries_.begin();
+	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;
+		if (numbers) {
+			entry.label(entry.citeNumber());
+		} else {
+			docstring const auth = entry.getAbbreviatedAuthor(buf, false);
+			// we do it this way so as to access the xref, if necessary
+			// note that this also gives us the modifier
+			docstring const year = getYear(*it, buf, true);
+			if (!auth.empty() && !year.empty())
+				entry.label(auth + ' ' + year);
+			else
+				entry.label(entry.key());
+		}
+	}
 }
 
 
@@ -977,4 +1153,3 @@ string citationStyleToString(const CitationStyle & cs)
 }
 
 } // namespace lyx
-