X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2FEncoding.cpp;h=9283e55fc8e8a1b9d6ea7637604a39376fe82638;hb=d1e3d75da226311cb290fc4be3686d6deef7b967;hp=3dcd05c5ab7b0ce8db99db3159dbca090b0e6e09;hpb=f8afb491657fed0851133490d74117dad448848a;p=lyx.git diff --git a/src/Encoding.cpp b/src/Encoding.cpp index 3dcd05c5ab..9283e55fc8 100644 --- a/src/Encoding.cpp +++ b/src/Encoding.cpp @@ -3,7 +3,7 @@ * This file is part of LyX, the document processor. * Licence details can be found in the file COPYING. * - * \author Lars Gullik Bjønnes + * \author Lars Gullik Bjønnes * \author Jean-Marc Lasgouttes * \author Dekel Tsur * @@ -14,6 +14,9 @@ #include "Encoding.h" +#include "Buffer.h" +#include "BufferList.h" +#include "InsetIterator.h" #include "LaTeXFeatures.h" #include "Lexer.h" #include "LyXRC.h" @@ -21,6 +24,7 @@ #include "support/debug.h" #include "support/FileName.h" #include "support/lstrings.h" +#include "support/textutils.h" #include "support/unicode.h" #include @@ -34,6 +38,10 @@ namespace lyx { Encodings encodings; +Encodings::MathCommandSet Encodings::mathcmd; +Encodings::TextCommandSet Encodings::textcmd; +Encodings::MathSymbolSet Encodings::mathsym; + namespace { char_type arabic_table[172][4] = { @@ -222,15 +230,22 @@ char_type const arabic_end = 0x06cc; /// Information about a single UCS4 character struct CharInfo { - /// LaTeX command for this character - docstring command; - /// Needed LaTeX preamble (or feature) - string preamble; + /// LaTeX command (text mode) for this character + docstring textcommand; + /// LaTeX command (math mode) for this character + docstring mathcommand; + /// Needed LaTeX preamble (or feature) for text mode + string textpreamble; + /// Needed LaTeX preamble (or feature) for math mode + string mathpreamble; /// Is this a combining character? bool combining; - /// Is \c preamble a feature known by LaTeXFeatures, or a raw LaTeX + /// Is \c textpreamble a feature known by LaTeXFeatures, or a raw LaTeX + /// command? + bool textfeature; + /// Is \c mathpreamble a feature known by LaTeXFeatures, or a raw LaTeX /// command? - bool feature; + bool mathfeature; /// Always force the LaTeX command, even if the encoding contains /// this character? bool force; @@ -240,6 +255,12 @@ struct CharInfo { typedef map CharInfoMap; CharInfoMap unicodesymbols; +typedef std::set CharSet; +CharSet forced; + +typedef std::set MathAlphaSet; +MathAlphaSet mathalpha; + /// The highest code point in UCS4 encoding (1<<20 + 1<<16) char_type const max_ucs4 = 0x110000; @@ -259,9 +280,9 @@ const char * EncodingException::what() const throw() } -Encoding::Encoding(string const & n, string const & l, string const & i, - bool f, Encoding::Package p) - : Name_(n), LatexName_(l), iconvName_(i), fixedwidth_(f), package_(p) +Encoding::Encoding(string const & n, string const & l, string const & g, + string const & i, bool f, Encoding::Package p) + : name_(n), latexName_(l), guiName_(g), iconvName_(i), fixedwidth_(f), package_(p) { if (n == "ascii") { // ASCII can encode 128 code points and nothing else @@ -323,21 +344,28 @@ void Encoding::init() const } -docstring Encoding::latexChar(char_type c) const +docstring Encoding::latexChar(char_type c, bool for_mathed) const { // assure the used encoding is properly initialized init(); - if (c < start_encodable_) + if (iconvName_ == "UTF-8" && package_ == none) + return docstring(1, c); + if (c < start_encodable_ && !encodings.isForced(c)) return docstring(1, c); if (encodable_.find(c) != encodable_.end()) return docstring(1, c); + if (for_mathed) + return docstring(); - // c cannot be encoded in this encoding + // c cannot (or should not) be encoded in this encoding CharInfoMap::const_iterator const it = unicodesymbols.find(c); if (it == unicodesymbols.end()) throw EncodingException(c); - return it->second.command; + // at least one of mathcommand and textcommand is nonempty + if (it->second.textcommand.empty()) + return "\\ensuremath{" + it->second.mathcommand + '}'; + return it->second.textcommand; } @@ -360,19 +388,227 @@ vector Encoding::symbolsList() const } -void Encodings::validate(char_type c, LaTeXFeatures & features) +bool Encodings::latexMathChar(char_type c, bool mathmode, + Encoding const * encoding, docstring & command) +{ + if (encoding) + command = encoding->latexChar(c, true); + + CharInfoMap::const_iterator const it = unicodesymbols.find(c); + if (it == unicodesymbols.end()) { + if (!encoding || command.empty()) + throw EncodingException(c); + if (mathmode) + addMathSym(c); + return false; + } + // at least one of mathcommand and textcommand is nonempty + bool use_math = (mathmode && !it->second.mathcommand.empty()) || + (!mathmode && it->second.textcommand.empty()); + if (use_math) { + command = it->second.mathcommand; + addMathCmd(c); + } else { + if (!encoding || command.empty()) { + command = it->second.textcommand; + addTextCmd(c); + } + if (mathmode) + addMathSym(c); + } + return use_math; +} + + +char_type Encodings::fromLaTeXCommand(docstring const & cmd, bool & combining) +{ + CharInfoMap::const_iterator const end = unicodesymbols.end(); + CharInfoMap::const_iterator it = unicodesymbols.begin(); + for (combining = false; it != end; ++it) { + docstring const math = it->second.mathcommand; + docstring const text = it->second.textcommand; + if (math == cmd || text == cmd) { + combining = it->second.combining; + return it->first; + } + } + return 0; +} + + +docstring Encodings::fromLaTeXCommand(docstring const & cmd, docstring & rem) +{ + docstring symbols; + size_t i = 0; + size_t const cmdend = cmd.size(); + CharInfoMap::const_iterator const uniend = unicodesymbols.end(); + for (size_t j = 0; j < cmdend; ++j) { + // Also get the char after a backslash + if (j + 1 < cmdend && cmd[j] == '\\') + ++j; + // If a macro argument follows, get it, too + if (j + 1 < cmdend && cmd[j + 1] == '{') { + size_t k = j + 1; + int count = 1; + while (k < cmdend && count && k != docstring::npos) { + k = cmd.find_first_of(from_ascii("{}"), k + 1); + if (cmd[k] == '{') + ++count; + else + --count; + } + if (k != docstring::npos) + j = k; + } + // Start with this substring and try augmenting it when it is + // the prefix of some command in the unicodesymbols file + docstring const subcmd = cmd.substr(i, j - i + 1); + + CharInfoMap::const_iterator it = unicodesymbols.begin(); + size_t unicmd_size = 0; + char_type c = 0; + for (; it != uniend; ++it) { + docstring const math = it->second.mathcommand; + docstring const text = it->second.textcommand; + size_t cur_size = max(math.size(), text.size()); + // The current math or text unicode command cannot + // match, or we already matched a longer one + if (cur_size < subcmd.size() || cur_size <= unicmd_size) + continue; + + docstring tmp = subcmd; + size_t k = j; + while (prefixIs(math, tmp) || prefixIs(text, tmp)) { + ++k; + if (k >= cmdend || cur_size <= tmp.size()) + break; + tmp += cmd[k]; + } + // No match + if (k == j) + continue; + + // The last added char caused a mismatch, because + // we didn't exhaust the chars in cmd and didn't + // exceed the maximum size of the current unicmd + if (k < cmdend && cur_size > tmp.size()) + tmp.resize(tmp.size() - 1); + + // If this is an exact match, we found a (longer) + // matching entry in the unicodesymbols file. + // If the entry doesn't start with '\', we take note + // of the match and continue (this is not a ultimate + // acceptance, as some other entry may match a longer + // portion of the cmd string). However, if the entry + // does start with '\', we accept the match only if + // this is a valid macro, i.e., either it is a single + // (nonletter) char macro, or nothing else follows, + // or what follows is a nonletter char, or the last + // character is a }. + if ((math == tmp || text == tmp) + && (tmp[0] != '\\' + || (tmp.size() == 2 && !isAlphaASCII(tmp[1])) + || k == cmdend + || !isAlphaASCII(cmd[k]) + || tmp[tmp.size() - 1] == '}') + ) { + c = it->first; + j = k - 1; + i = j + 1; + unicmd_size = cur_size; + } + } + if (unicmd_size) + symbols += c; + else if (j + 1 == cmdend) + // No luck. Return what remains + rem = cmd.substr(i); + } + return symbols; +} + + +void Encodings::initUnicodeMath(Buffer const & buffer, bool clear_sets) +{ +#ifdef TEX2LYX + // The code below is not needed in tex2lyx and requires additional stuff + (void)buffer; + (void)clear_sets; +#else + if (clear_sets) { + mathcmd.clear(); + textcmd.clear(); + mathsym.clear(); + } + + // Check master + Inset & inset = buffer.inset(); + InsetIterator it = inset_iterator_begin(inset); + InsetIterator const end = inset_iterator_end(inset); + for (; it != end; ++it) + it->initUnicodeMath(); + + // Check children + BufferList::iterator bit = theBufferList().begin(); + BufferList::iterator const bend = theBufferList().end(); + for (; bit != bend; ++bit) + if (buffer.isChild(*bit)) + initUnicodeMath(**bit, false); +#endif +} + + +void Encodings::validate(char_type c, LaTeXFeatures & features, bool for_mathed) { +#ifdef TEX2LYX + // The code below is not needed in tex2lyx and requires additional stuff + (void)c; + (void)features; + (void)for_mathed; +#else CharInfoMap::const_iterator const it = unicodesymbols.find(c); - if (it != unicodesymbols.end() && !it->second.preamble.empty()) { - if (it->second.feature) - features.require(it->second.preamble); - else - features.addPreambleSnippet(it->second.preamble); + if (it != unicodesymbols.end()) { + // In mathed, c could be used both in textmode and mathmode + bool const use_math = (for_mathed && isMathCmd(c)) || + (!for_mathed && it->second.textcommand.empty()); + bool const use_text = (for_mathed && isTextCmd(c)) || + (!for_mathed && !it->second.textcommand.empty()); + if (use_math) { + if (!it->second.mathpreamble.empty()) { + if (it->second.mathfeature) { + string feats = it->second.mathpreamble; + while (!feats.empty()) { + string feat; + feats = split(feats, feat, ','); + features.require(feat); + } + } else + features.addPreambleSnippet(it->second.mathpreamble); + } + } + if (use_text) { + if (!it->second.textpreamble.empty()) { + if (it->second.textfeature) { + string feats = it->second.textpreamble; + while (!feats.empty()) { + string feat; + feats = split(feats, feat, ','); + features.require(feat); + } + } else + features.addPreambleSnippet(it->second.textpreamble); + } + } } + if (for_mathed && isMathSym(c)) { + features.require("amstext"); + features.require("lyxmathsym"); + } +#endif } -bool Encodings::isComposeChar_hebrew(char_type c) +bool Encodings::isHebrewComposeChar(char_type c) { return c <= 0x05c2 && c >= 0x05b0 && c != 0x05be && c != 0x05c0; } @@ -382,7 +618,7 @@ bool Encodings::isComposeChar_hebrew(char_type c) // they are hamza, alef_madda, alef_hamza, waw_hamza, alef_hamza_under, // alef, tah_marbota, dal, thal, rah, zai, wow, alef_maksoura -bool Encodings::is_arabic_special(char_type c) +bool Encodings::isArabicSpecialChar(char_type c) { return (c >= 0x0621 && c <= 0x0625) || (c >= 0x0630 && c <= 0x0632) || c == 0x0627 || c == 0x0629 || c == 0x062f || c == 0x0648 @@ -390,22 +626,22 @@ bool Encodings::is_arabic_special(char_type c) } -bool Encodings::isComposeChar_arabic(char_type c) +bool Encodings::isArabicComposeChar(char_type c) { return c >= 0x064b && c <= 0x0652; } -bool Encodings::is_arabic(char_type c) +bool Encodings::isArabicChar(char_type c) { return c >= arabic_start && c <= arabic_end && arabic_table[c-arabic_start][0]; } -char_type Encodings::transformChar(char_type c, Encodings::Letter_Form form) +char_type Encodings::transformChar(char_type c, Encodings::LetterForm form) { - return is_arabic(c) ? arabic_table[c-arabic_start][form] : c; + return isArabicChar(c) ? arabic_table[c-arabic_start][form] : c; } @@ -425,25 +661,37 @@ bool Encodings::isKnownScriptChar(char_type const c, string & preamble) if (it == unicodesymbols.end()) return false; - if (it->second.preamble != "textgreek" && it->second.preamble != "textcyr") + if (it->second.textpreamble != "textgreek" && it->second.textpreamble != "textcyr") return false; if (preamble.empty()) { - preamble = it->second.preamble; + preamble = it->second.textpreamble; return true; } - return it->second.preamble == preamble; + return it->second.textpreamble == preamble; } -Encoding const * Encodings::getFromLyXName(string const & name) const +bool Encodings::isForced(char_type c) +{ + return (!forced.empty() && forced.find(c) != forced.end()); +} + + +bool Encodings::isMathAlpha(char_type c) +{ + return mathalpha.count(c); +} + + +Encoding const * Encodings::fromLyXName(string const & name) const { EncodingList::const_iterator const it = encodinglist.find(name); return it != encodinglist.end() ? &it->second : 0; } -Encoding const * Encodings::getFromLaTeXName(string const & name) const +Encoding const * Encodings::fromLaTeXName(string const & name) const { // We don't use find_if because it makes copies of the pairs in // the map. @@ -467,15 +715,19 @@ void Encodings::read(FileName const & encfile, FileName const & symbolsfile) { // We must read the symbolsfile first, because the Encoding // constructor depends on it. - Lexer symbolslex(0, 0); + Lexer symbolslex; symbolslex.setFile(symbolsfile); + bool getNextToken = true; while (symbolslex.isOK()) { char_type symbol; CharInfo info; string flags; - if (!symbolslex.next(true)) - break; + if (getNextToken) { + if (!symbolslex.next(true)) + break; + } else + getNextToken = true; istringstream is(symbolslex.getString()); // reading symbol directly does not work if @@ -487,54 +739,92 @@ void Encodings::read(FileName const & encfile, FileName const & symbolsfile) if (!symbolslex.next(true)) break; - info.command = symbolslex.getDocString(); + info.textcommand = symbolslex.getDocString(); if (!symbolslex.next(true)) break; - info.preamble = symbolslex.getString(); + info.textpreamble = symbolslex.getString(); if (!symbolslex.next(true)) break; flags = symbolslex.getString(); info.combining = false; - info.feature = false; + info.textfeature = false; info.force = false; while (!flags.empty()) { string flag; flags = split(flags, flag, ','); - if (flag == "combining") + if (flag == "combining") { info.combining = true; - else if (flag == "force") + } else if (flag == "force") { info.force = true; - else + forced.insert(symbol); + } else if (flag == "mathalpha") { + mathalpha.insert(symbol); + } else { lyxerr << "Ignoring unknown flag `" << flag << "' for symbol `0x" << hex << symbol << dec << "'." << endl; + } + } + // mathcommand and mathpreamble have been added for 1.6.0. + // make them optional so that old files still work. + int const lineno = symbolslex.lineNumber(); + bool breakout = false; + if (symbolslex.next(true)) { + if (symbolslex.lineNumber() != lineno) { + // line in old format without mathcommand and mathpreamble + getNextToken = false; + } else { + info.mathcommand = symbolslex.getDocString(); + if (symbolslex.next(true)) { + if (symbolslex.lineNumber() != lineno) { + // line in new format with mathcommand only + getNextToken = false; + } else { + // line in new format with mathcommand and mathpreamble + info.mathpreamble = symbolslex.getString(); + } + } else + breakout = true; + } + } else { + breakout = true; } - if (!info.preamble.empty()) - info.feature = info.preamble[0] != '\\'; + if (!info.textpreamble.empty()) + info.textfeature = info.textpreamble[0] != '\\'; + if (!info.mathpreamble.empty()) + info.mathfeature = info.mathpreamble[0] != '\\'; LYXERR(Debug::INFO, "Read unicode symbol " << symbol << " '" - << to_utf8(info.command) << "' '" << info.preamble - << "' " << info.combining << ' ' << info.feature); - unicodesymbols[symbol] = info; + << to_utf8(info.textcommand) << "' '" << info.textpreamble + << "' " << info.combining << ' ' << info.textfeature + << " '" << to_utf8(info.mathcommand) << "' '" + << info.mathpreamble << "' " << info.mathfeature); + + // we assume that at least one command is nonempty when using unicodesymbols + if (!info.textcommand.empty() || !info.mathcommand.empty()) + unicodesymbols[symbol] = info; + + if (breakout) + break; } // Now read the encodings - enum Encodingtags { + enum { et_encoding = 1, et_end, - et_last }; - struct keyword_item encodingtags[et_last - 1] = { + LexerKeyword encodingtags[] = { { "encoding", et_encoding }, { "end", et_end } }; - Lexer lex(encodingtags, et_last - 1); + Lexer lex(encodingtags); lex.setFile(encfile); + lex.setContext("Encodings::read"); while (lex.isOK()) { switch (lex.lex()) { case et_encoding: @@ -544,6 +834,8 @@ void Encodings::read(FileName const & encfile, FileName const & symbolsfile) lex.next(); string const latexname = lex.getString(); lex.next(); + string const guiname = lex.getString(); + lex.next(); string const iconvname = lex.getString(); lex.next(); string const width = lex.getString(); @@ -553,39 +845,37 @@ void Encodings::read(FileName const & encfile, FileName const & symbolsfile) else if (width == "variable") fixedwidth = false; else - lex.printError("Encodings::read: " - "Unknown width: `$$Token'"); + lex.printError("Unknown width"); lex.next(); string const p = lex.getString(); Encoding::Package package = Encoding::none; if (p == "none") - package = Encoding::none; + package = Encoding::none; else if (p == "inputenc") package = Encoding::inputenc; else if (p == "CJK") package = Encoding::CJK; + else if (p == "japanese") + package = Encoding::japanese; else - lex.printError("Encodings::read: " - "Unknown package: `$$Token'"); + lex.printError("Unknown package"); LYXERR(Debug::INFO, "Reading encoding " << name); encodinglist[name] = Encoding(name, latexname, - iconvname, fixedwidth, package); + guiname, iconvname, fixedwidth, package); if (lex.lex() != et_end) - lex.printError("Encodings::read: " - "missing end"); + lex.printError("Missing end"); break; } case et_end: - lex.printError("Encodings::read: Misplaced end"); + lex.printError("Misplaced end"); break; case Lexer::LEX_FEOF: break; default: - lex.printError("Encodings::read: " - "Unknown tag: `$$Token'"); + lex.printError("Unknown tag"); break; } }