{}
-docstring const BibTeXInfo::getAbbreviatedAuthor(bool jurabib_style) const
+docstring const BibTeXInfo::getAbbreviatedAuthor(
+ Buffer const * buf, bool jurabib_style) const
{
if (!is_bibtex_) {
docstring const opt = label();
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;
}
+ "/" + familyName(authors[1]);
if (authors.size() == 3)
shortauthor += "/" + familyName(authors[2]);
- return shortauthor;
+ return convertLaTeXCommands(shortauthor);
}
- if (authors.size() == 2 && authors[1] != "others")
- return bformat(from_ascii("%1$s and %2$s"),
- familyName(authors[0]), familyName(authors[1]));
-
- if (authors.size() >= 2)
- return bformat(from_ascii("%1$s et al."),
- familyName(authors[0]));
-
- return familyName(authors[0]);
-}
-
+ docstring retval = familyName(authors[0]);
+
+ if (authors.size() == 2 && authors[1] != "others") {
+ docstring const dformat = buf ?
+ buf->B_("%1$s and %2$s") : from_ascii("%1$s and %2$s");
+ retval = bformat(dformat, familyName(authors[0]), familyName(authors[1]));
+ } else if (authors.size() >= 2) {
+ // we get here either if the author list is longer than two names
+ // or if the second 'name' is "others". we do the same thing either
+ // way.
+ docstring const dformat = buf ?
+ buf->B_("%1$s et al.") : from_ascii("%1$s et al.");
+ retval = bformat(dformat, familyName(authors[0]));
+ }
-docstring const BibTeXInfo::getAbbreviatedAuthor(Buffer const & buf, bool jurabib_style) const
-{
- docstring const author = getAbbreviatedAuthor(jurabib_style);
- if (!is_bibtex_)
- return author;
- vector<docstring> const authors = getVectorFromString(author, from_ascii(" and "));
- if (authors.size() == 2)
- return bformat(buf.B_("%1$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;
+ return convertLaTeXCommands(retval);
}
docstring const BibTeXInfo::getYear() const
{
- if (is_bibtex_)
- return operator[]("year");
+ if (is_bibtex_) {
+ // first try legacy year field
+ docstring year = operator[]("year");
+ if (!year.empty())
+ return year;
+ // now try biblatex's date field
+ year = operator[]("date");
+ // Format is [-]YYYY-MM-DD*/[-]YYYY-MM-DD*
+ // We only want the years.
+ static regex const yreg("[-]?([\\d]{4}).*");
+ static regex const ereg(".*/([\\d]{4}).*");
+ smatch sm;
+ string const date = to_utf8(year);
+ regex_match(date, sm, yreg);
+ year = from_ascii(sm[1]);
+ // check for an endyear
+ if (regex_match(date, sm, ereg))
+ year += char_type(0x2013) + from_ascii(sm[1]);
+ return year;
+ }
docstring const opt = label();
if (opt.empty())
namespace {
-string parseOptions(string const & format, string & optkey,
- string & ifpart, string & elsepart);
+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.
-string parseEmbeddedOption(string const & format, string & ifelsepart)
+docstring parseEmbeddedOption(docstring const & format, docstring & ifelsepart)
{
LASSERT(format[0] == '{' && format[1] == '%', return format);
string optkey;
- string ifpart;
- string elsepart;
- string const rest = parseOptions(format, optkey, ifpart, elsepart);
+ 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(), /* */);
+ LASSERT(rest.size() <= format.size(),
+ { ifelsepart = docstring(); return format; });
ifelsepart = format.substr(0, format.size() - rest.size());
- return rest;
+ 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)
+docstring getClause(docstring const & format, docstring & clause)
{
- string fmt = format;
+ docstring fmt = format;
// remove '[['
fmt = fmt.substr(2);
// we'll remove characters from the front of fmt as we
}
// check for an embedded option
if (fmt[0] == '{' && fmt.size() > 1 && fmt[1] == '%') {
- string part;
- string const rest = parseEmbeddedOption(fmt, part);
+ docstring part;
+ docstring const rest = parseEmbeddedOption(fmt, part);
if (fmt == rest) {
LYXERR0("ERROR! Couldn't parse embedded option in `" << format <<"'.");
return format;
// 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)
+docstring parseOptions(docstring const & format, string & optkey,
+ docstring & ifpart, docstring & elsepart)
{
LASSERT(format[0] == '{' && format[1] == '%', return format);
// strip '{%'
- string fmt = format.substr(2);
+ 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 = fmt.substr(0,pos);
+ optkey = to_utf8(fmt.substr(0, pos));
fmt = fmt.substr(pos + 1);
// [[format]] should be next
if (fmt[0] != '[' || fmt[1] != '[') {
return format;
}
- string curfmt = fmt;
+ docstring curfmt = fmt;
fmt = getClause(curfmt, ifpart);
if (fmt == curfmt) {
LYXERR0("Error parsing `" << format <<"'. Couldn't get if clause.");
} // 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,
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!");
// macro
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
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, buf, before, after, dialog, xref);
+ getValueForKey(key, buf, before, after, dialog, xref, max_keysize);
if (!scanning_rich)
- ret += from_ascii("{!<span class=\"bib-" + key + "\">!}");
- ret += val;
+ ret << from_ascii("{!<span class=\"bib-" + key + "\">!}");
+ ret << val;
if (!scanning_rich)
- ret += from_ascii("{!</span>!}");
+ ret << from_ascii("{!</span>!}");
}
} else {
// beginning of key
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!");
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,
+ 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())
- ret += expandFormat(elsepart, xref, counter, buf,
+ } 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;
}
// beginning of rich text
scanning_rich = true;
fmt = fmt.substr(2);
- ret += from_ascii("{!");
+ 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("!}");
+ ret << from_ascii("!}");
continue;
}
else if (scanning_key)
key += char(thischar);
- else
- ret += thischar;
+ else {
+ try {
+ ret.put(thischar);
+ } catch (EncodingException & /* e */) {
+ LYXERR0("Uncodable character '" << docstring(1, thischar) << " in citation label!");
+ }
+ }
fmt = fmt.substr(1);
} // for loop
if (scanning_key) {
LYXERR0("Never found end of rich text in `" << format << "'!");
return _("ERROR!");
}
- return ret;
+ return ret.str();
}
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,
docstring(), docstring(), docstring(), false);
- if (!info_.empty()) {
+ if (info_.empty()) {
+ // this probably shouldn't happen
+ return info_;
+ }
+
+ if (richtext) {
info_richtext_ = convertLaTeXCommands(processRichtext(info_, true));
- info_ = convertLaTeXCommands(processRichtext(info_, false));
- if (richtext)
- return info_richtext_;
+ 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;
loclabel = processRichtext(loclabel, richtext);
loclabel = convertLaTeXCommands(loclabel);
}
+
return loclabel;
}
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];
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);
+ 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);
+ 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]'
// Special key to provide the full bibliography entry: see getInfo()
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;
ret = expandFormat(format, xref, counter, buf,
docstring(), docstring(), docstring(), false);
else if (key == "year")
ret = getYear();
}
+
if (cleanit)
- return html::cleanAttr(ret);
-
+ ret = html::cleanAttr(ret);
+
+ // make sure it is not too big
+ support::truncateWithEllipsis(ret, maxsize);
return ret;
}
if (it == end())
return docstring();
BibTeXInfo const & data = it->second;
- return data.getAbbreviatedAuthor(buf, false);
+ return data.getAbbreviatedAuthor(&buf, false);
}
}
-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) {
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 (too_many_keys)
+ ret.push_back(0x2026);//HORIZONTAL ELLIPSIS
+ support::truncateWithEllipsis(ret, max_size);
return ret;
}
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();
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>();
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;
// FIXME We may want to collect these differently, in the first case,
// so that we might have them in order of appearance.
set<docstring> citekeys;
- Toc const & toc = buf.tocBackend().toc("citation");
- Toc::const_iterator it = toc.begin();
- Toc::const_iterator const en = toc.end();
+ shared_ptr<Toc const> toc = buf.tocBackend().toc("citation");
+ Toc::const_iterator it = toc->begin();
+ Toc::const_iterator const en = toc->end();
for (; it != en; ++it) {
if (it->str().empty())
continue;
{
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;
docstring const num = convert<docstring>(++keynumber);
entry.setCiteNumber(num);
} else {
+ // coverity complains about our derefercing the iterator last,
+ // which was not initialized above. but it does get initialized
+ // after the first time through the loop, which is the point of
+ // the first test.
+ // coverity[FORWARD_NULL]
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 (numbers) {
entry.label(entry.citeNumber());
} else {
- docstring const auth = entry.getAbbreviatedAuthor(buf, false);
+ 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);
{
string cmd = cs.cmd;
if (cs.forceUpperCase)
- cmd[0] = 'C';
+ cmd[0] = uppercase(cmd[0]);
if (cs.fullAuthorList)
cmd += '*';
return cmd;
projects / lyx.git / blobdiff