X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2FBiblioInfo.cpp;h=f2bf332a2655204b8dbdce3b630bbc4c5fbd0c20;hb=f16b2cab75752e7b0fbec5520e231b068eb29072;hp=4ad119cefd0db38e3e8ff89f6c4a312a16c0efa4;hpb=f6cd5ab863cc9a460efb7a1dc7eb48b87d867655;p=lyx.git diff --git a/src/BiblioInfo.cpp b/src/BiblioInfo.cpp index 4ad119cefd..f2bf332a26 100644 --- a/src/BiblioInfo.cpp +++ b/src/BiblioInfo.cpp @@ -20,6 +20,7 @@ #include "Encoding.h" #include "InsetIterator.h" #include "Language.h" +#include "output_xhtml.h" #include "Paragraph.h" #include "TextClass.h" #include "TocBackend.h" @@ -254,7 +255,7 @@ BibTeXInfo::BibTeXInfo(docstring const & key, docstring const & type) {} -docstring const BibTeXInfo::getAbbreviatedAuthor(bool jurabib_style, string lang) const +docstring const BibTeXInfo::getAbbreviatedAuthor(bool jurabib_style) const { if (!is_bibtex_) { docstring const opt = label(); @@ -270,9 +271,9 @@ docstring const BibTeXInfo::getAbbreviatedAuthor(bool jurabib_style, string lang 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; } @@ -290,18 +291,35 @@ docstring const BibTeXInfo::getAbbreviatedAuthor(bool jurabib_style, string lang + "/" + familyName(authors[1]); if (authors.size() == 3) shortauthor += "/" + familyName(authors[2]); - return shortauthor; + return convertLaTeXCommands(shortauthor); } - if (authors.size() == 2) - return bformat(translateIfPossible(from_ascii("%1$s and %2$s"), lang), + 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(translateIfPossible(from_ascii("%1$s et al."), lang), + if (authors.size() >= 2) + retval = bformat(from_ascii("%1$s et al."), familyName(authors[0])); - return familyName(authors[0]); + return convertLaTeXCommands(retval); +} + + +docstring const BibTeXInfo::getAbbreviatedAuthor(Buffer const & buf, bool jurabib_style) const +{ + docstring const author = getAbbreviatedAuthor(jurabib_style); + if (!is_bibtex_) + return author; + vector 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; } @@ -335,37 +353,38 @@ docstring const BibTeXInfo::getXRef() const 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 @@ -378,8 +397,8 @@ string getClause(string const & format, string & clause) } // 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; @@ -399,18 +418,18 @@ string getClause(string const & format, string & clause) // 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] != '[') { @@ -418,7 +437,7 @@ string parseOptions(string const & format, string & optkey, return format; } - string curfmt = fmt; + docstring curfmt = fmt; fmt = getClause(curfmt, ifpart); if (fmt == curfmt) { LYXERR0("Error parsing `" << format <<"'. Couldn't get if clause."); @@ -447,25 +466,37 @@ string parseOptions(string const & format, string & optkey, } // 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; - string lang = buf.params().language->code(); 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!"); @@ -480,25 +511,26 @@ 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 string const val = buf.params().documentClass().getCiteMacro(engine_type, key); docstring const trans = - translateIfPossible(from_utf8(val), lang); - ret += trans; + translateIfPossible(from_utf8(val), buf.params().language->code()); + ret << trans; } else { docstring const val = - getValueForKey(key, before, after, dialog, xref, lang); - ret += from_ascii("{!!}"); - ret += val; - ret += from_ascii("{!!}"); + getValueForKey(key, buf, before, after, dialog, xref, max_keysize); + if (!scanning_rich) + ret << from_ascii("{!!}"); + ret << val; + if (!scanning_rich) + ret << from_ascii("{!!}"); } } else { // beginning of key @@ -516,23 +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, lang); + 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; } @@ -540,26 +575,31 @@ docstring BibTeXInfo::expandFormat(string const & format, // 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) { @@ -570,15 +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")); @@ -588,31 +630,34 @@ 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, docstring(), docstring(), docstring(), false); - if (!info_.empty()) - info_ = convertLaTeXCommands(info_); + if (info_.empty()) { + // this probably shouldn't happen + return 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, before, after, dialog, next); @@ -621,6 +666,7 @@ docstring const BibTeXInfo::getLabel(BibTeXInfo const * const xref, loclabel = processRichtext(loclabel, richtext); loclabel = convertLaTeXCommands(loclabel); } + return loclabel; } @@ -641,48 +687,76 @@ 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, string lang) 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(false, lang); - 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, lang); - 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 + support::truncateWithEllipsis(ret, maxsize); return ret; } @@ -742,13 +816,13 @@ vector const BiblioInfo::getEntries() const } -docstring const BiblioInfo::getAbbreviatedAuthor(docstring const & key, string lang) 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(false, lang); + return data.getAbbreviatedAuthor(buf, false); } @@ -762,7 +836,7 @@ docstring const BiblioInfo::getCiteNumber(docstring const & key) const } -docstring const BiblioInfo::getYear(docstring const & key, bool use_modifier, string lang) const +docstring const BiblioInfo::getYear(docstring const & key, bool use_modifier) const { BiblioInfo::const_iterator it = find(key); if (it == end()) @@ -774,11 +848,11 @@ docstring const BiblioInfo::getYear(docstring const & key, bool use_modifier, st docstring const xref = data.getXRef(); if (xref.empty()) // no luck - return translateIfPossible(from_ascii("No year"), lang); + return docstring(); BiblioInfo::const_iterator const xrefit = find(xref); if (xrefit == end()) // no luck again - return translateIfPossible(from_ascii("No year"), lang); + return docstring(); BibTeXInfo const & xref_data = xrefit->second; year = xref_data.getYear(); } @@ -788,6 +862,15 @@ docstring const BiblioInfo::getYear(docstring const & key, bool use_modifier, st } +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 { @@ -806,14 +889,23 @@ docstring const BiblioInfo::getInfo(docstring const & key, } -docstring const BiblioInfo::getLabel(vector 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 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::const_iterator key = keys.begin(); vector::const_iterator ken = keys.end(); for (; key != ken; ++key) { @@ -831,16 +923,22 @@ docstring const BiblioInfo::getLabel(vector 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 (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(); @@ -849,8 +947,8 @@ bool BiblioInfo::isBibtex(docstring const & key) const vector const BiblioInfo::getCiteStrings( vector const & keys, vector 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(); @@ -859,7 +957,7 @@ vector const BiblioInfo::getCiteStrings( vector 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; @@ -901,9 +999,9 @@ void BiblioInfo::collectCitedEntries(Buffer const & buf) // FIXME We may want to collect these differently, in the first case, // so that we might have them in order of appearance. set citekeys; - Toc const & toc = buf.tocBackend().toc("citation"); - Toc::const_iterator it = toc.begin(); - Toc::const_iterator const en = toc.end(); + shared_ptr 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; @@ -940,7 +1038,7 @@ 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; @@ -983,6 +1081,28 @@ void BiblioInfo::makeCitationLabels(Buffer const & buf) last = biit; } } + // Set the labels + it = cited_entries_.begin(); + for (; it != en; ++it) { + map::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()); + } + } }