X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2FEncoding.cpp;h=ff83ab8c139a972f2392b399043e08a717947447;hb=952853eb2358472644e45f3fe3a6f3a68fac3556;hp=ca896d4c433bfc1fc15fc15192add67f29ea0a6d;hpb=f48d7caee3fab40528ceedc7a05f67004ba59d83;p=lyx.git diff --git a/src/Encoding.cpp b/src/Encoding.cpp index ca896d4c43..ff83ab8c13 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 * @@ -14,30 +14,37 @@ #include "Encoding.h" -#include "debug.h" +#include "Buffer.h" +#include "BufferList.h" +#include "InsetIterator.h" #include "LaTeXFeatures.h" #include "Lexer.h" #include "LyXRC.h" +#include "support/debug.h" +#include "support/gettext.h" #include "support/FileName.h" #include "support/lstrings.h" +#include "support/textutils.h" #include "support/unicode.h" -#include +#include -#ifndef CXX_GLOBAL_CSTD -using std::strtol; -#endif -using std::endl; -using std::string; +#include +using namespace std; +using namespace lyx::support; namespace lyx { -using support::FileName; +int const Encoding::any = -1; Encodings encodings; +Encodings::MathCommandSet Encodings::mathcmd; +Encodings::TextCommandSet Encodings::textcmd; +Encodings::MathSymbolSet Encodings::mathsym; + namespace { char_type arabic_table[172][4] = { @@ -138,7 +145,7 @@ char_type arabic_table[172][4] = { {0, 0, 0, 0}, // 0x067b {0, 0, 0, 0}, // 0x067c {0, 0, 0, 0}, // 0x067d - {0xfb56, 0xfb57, 0xfb58, 0xfb59}, // 0x067e = peh + {0xfb56, 0xfb57, 0xfb58, 0xfb59}, // 0x067e = peh {0, 0, 0, 0}, // 0x067f {0, 0, 0, 0}, // 0x0680 {0, 0, 0, 0}, // 0x0681 @@ -146,7 +153,7 @@ char_type arabic_table[172][4] = { {0, 0, 0, 0}, // 0x0683 {0, 0, 0, 0}, // 0x0684 {0, 0, 0, 0}, // 0x0685 - {0xfb7a, 0xfb7b, 0xfb7c, 0xfb7d}, // 0x0686 = tcheh + {0xfb7a, 0xfb7b, 0xfb7c, 0xfb7d}, // 0x0686 = tcheh {0, 0, 0, 0}, // 0x0687 {0, 0, 0, 0}, // 0x0688 {0, 0, 0, 0}, // 0x0689 @@ -181,13 +188,13 @@ char_type arabic_table[172][4] = { {0, 0, 0, 0}, // 0x06a6 {0, 0, 0, 0}, // 0x06a7 {0, 0, 0, 0}, // 0x06a8 - {0xfb8e, 0xfb8f, 0xfb90, 0xfb91}, // 0x06a9 = farsi kaf + {0xfb8e, 0xfb8f, 0xfb90, 0xfb91}, // 0x06a9 = farsi kaf {0, 0, 0, 0}, // 0x06aa {0, 0, 0, 0}, // 0x06ab {0, 0, 0, 0}, // 0x06ac {0, 0, 0, 0}, // 0x06ad {0, 0, 0, 0}, // 0x06ae - {0xfb92, 0xfb93, 0xfb94, 0xfb95}, // 0x06af = gaf + {0xfb92, 0xfb93, 0xfb94, 0xfb95}, // 0x06af = gaf {0, 0, 0, 0}, // 0x06b0 {0, 0, 0, 0}, // 0x06b1 {0, 0, 0, 0}, // 0x06b2 @@ -216,7 +223,7 @@ char_type arabic_table[172][4] = { {0, 0, 0, 0}, // 0x06c9 {0, 0, 0, 0}, // 0x06ca {0, 0, 0, 0}, // 0x06cb - {0xfbfc, 0xfbfd, 0xfbfe, 0xfbff} // 0x06cc = farsi yeh + {0xfbfc, 0xfbfd, 0xfbfe, 0xfbff} // 0x06cc = farsi yeh }; @@ -224,26 +231,68 @@ char_type const arabic_start = 0x0621; char_type const arabic_end = 0x06cc; +enum CharInfoFlags { + /// + CharInfoCombining = 1, + /// + CharInfoTextFeature = 2, + /// + CharInfoMathFeature = 4, + /// + CharInfoForce = 8, + /// + CharInfoTextNoTermination = 16, + /// + CharInfoMathNoTermination = 32, + /// + CharInfoForceSelected = 64, +}; + /// 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 + bool combining() const { return flags & CharInfoCombining ? true : false; } + /// Is \c textpreamble a feature known by LaTeXFeatures, or a raw LaTeX + /// command? + bool textfeature() const { return flags & CharInfoTextFeature ? true : false; } + /// Is \c mathpreamble a feature known by LaTeXFeatures, or a raw LaTeX /// command? - bool feature; + bool mathfeature() const { return flags & CharInfoMathFeature ? true : false; } /// Always force the LaTeX command, even if the encoding contains /// this character? - bool force; + bool force() const { return flags & CharInfoForce ? true : false; } + /// Force the LaTeX command for some encodings? + bool forceselected() const { return flags & CharInfoForceSelected ? true : false; } + /// TIPA shortcut + string tipashortcut; + /// \c textcommand needs no termination (such as {} or space). + bool textnotermination() const { return flags & CharInfoTextNoTermination ? true : false; } + /// \c mathcommand needs no termination (such as {} or space). + bool mathnotermination() const { return flags & CharInfoMathNoTermination ? true : false; } + /// + unsigned int flags; }; -typedef std::map CharInfoMap; +typedef map CharInfoMap; CharInfoMap unicodesymbols; +typedef std::set CharSet; +typedef std::map CharSetMap; +CharSet forced; +CharSetMap forcedselected; + +typedef std::set MathAlphaSet; +MathAlphaSet mathalpha; + /// The highest code point in UCS4 encoding (1<<20 + 1<<16) char_type const max_ucs4 = 0x110000; @@ -251,9 +300,22 @@ char_type const max_ucs4 = 0x110000; } // namespace anon -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) +EncodingException::EncodingException(char_type c) + : failed_char(c), par_id(0), pos(0) +{ +} + + +const char * EncodingException::what() const throw() +{ + return "Could not find LaTeX command for a character"; +} + + +Encoding::Encoding(string const & n, string const & l, string const & g, + string const & i, bool f, bool u, Encoding::Package p) + : name_(n), latexName_(l), guiName_(g), iconvName_(i), fixedwidth_(f), + unsafe_(u), forced_(&forcedselected[n]), package_(p) { if (n == "ascii") { // ASCII can encode 128 code points and nothing else @@ -264,6 +326,7 @@ Encoding::Encoding(string const & n, string const & l, string const & i, start_encodable_ = max_ucs4; complete_ = true; } else { + start_encodable_ = 0; complete_ = false; } } @@ -271,6 +334,9 @@ Encoding::Encoding(string const & n, string const & l, string const & i, void Encoding::init() const { + if (complete_) + return; + start_encodable_ = 0; // temporarily switch off lyxerr, since we will generate iconv errors lyxerr.disable(); @@ -278,13 +344,17 @@ void Encoding::init() const // We do not need to check all UCS4 code points, it is enough // if we check all 256 code points of this encoding. for (unsigned short j = 0; j < 256; ++j) { - char const c = j; - std::vector const ucs4 = eightbit_to_ucs4(&c, 1, iconvName_); - if (ucs4.size() == 1) { - char_type const c = ucs4[0]; - CharInfoMap::const_iterator const it = unicodesymbols.find(c); - if (it == unicodesymbols.end() || !it->second.force) - encodable_.insert(c); + char const c = char(j); + vector const ucs4 = eightbit_to_ucs4(&c, 1, iconvName_); + if (ucs4.size() != 1) + continue; + char_type const uc = ucs4[0]; + CharInfoMap::const_iterator const it = unicodesymbols.find(uc); + if (it == unicodesymbols.end()) + encodable_.insert(uc); + else if (!it->second.force()) { + if (forced_->empty() || forced_->find(uc) == forced_->end()) + encodable_.insert(uc); } } } else { @@ -293,11 +363,15 @@ void Encoding::init() const // therefore we need to check all UCS4 code points. // This is expensive! for (char_type c = 0; c < max_ucs4; ++c) { - std::vector const eightbit = ucs4_to_eightbit(&c, 1, iconvName_); + vector const eightbit = ucs4_to_eightbit(&c, 1, iconvName_); if (!eightbit.empty()) { CharInfoMap::const_iterator const it = unicodesymbols.find(c); - if (it == unicodesymbols.end() || !it->second.force) + if (it == unicodesymbols.end()) encodable_.insert(c); + else if (!it->second.force()) { + if (forced_->empty() || forced_->find(c) == forced_->end()) + encodable_.insert(c); + } } } } @@ -312,46 +386,465 @@ void Encoding::init() const } -docstring const Encoding::latexChar(char_type c) const +bool Encoding::isForced(char_type c) const +{ + if (!forced.empty() && forced.find(c) != forced.end()) + return true; + return !forced_->empty() && forced_->find(c) != forced_->end(); +} + + +bool Encoding::encodable(char_type c) const { // assure the used encoding is properly initialized - if (!complete_) - init(); - BOOST_ASSERT(complete_); - - if (c < start_encodable_) - return docstring(1, c); - if (encodable_.find(c) == encodable_.end()) { - // c cannot be encoded in this encoding - CharInfoMap::const_iterator const it = unicodesymbols.find(c); - if (it == unicodesymbols.end()) - lyxerr << "Could not find LaTeX command for character 0x" - << std::hex << c << std::dec - << ".\nLaTeX export will fail." - << endl; - else - return it->second.command; + init(); + + if (iconvName_ == "UTF-8" && package_ == none) + return true; + if (c < start_encodable_ && !isForced(c)) + return true; + if (encodable_.find(c) != encodable_.end()) + return true; + return false; +} + + +pair Encoding::latexChar(char_type c) const +{ + if (encodable(c)) + return make_pair(docstring(1, c), false); + + // 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); + // at least one of mathcommand and textcommand is nonempty + if (it->second.textcommand.empty()) + return make_pair( + "\\ensuremath{" + it->second.mathcommand + '}', false); + return make_pair(it->second.textcommand, !it->second.textnotermination()); +} + + +pair Encoding::latexString(docstring const input, bool dryrun) const +{ + docstring result; + docstring uncodable; + bool terminate = false; + for (size_t n = 0; n < input.size(); ++n) { + try { + char_type const c = input[n]; + pair latex_char = latexChar(c); + docstring const latex = latex_char.first; + if (terminate && !prefixIs(latex, '\\') + && !prefixIs(latex, '{') + && !prefixIs(latex, '}')) { + // Prevent eating of a following + // space or command corruption by + // following characters + if (latex == " ") + result += "{}"; + else + result += " "; + } + result += latex; + terminate = latex_char.second; + } catch (EncodingException & /* e */) { + LYXERR0("Uncodable character in latexString!"); + if (dryrun) { + result += "<" + _("LyX Warning: ") + + _("uncodable character") + " '"; + result += docstring(1, input[n]); + result += "'>"; + } else + uncodable += input[n]; + } + } + return make_pair(result, uncodable); +} + + +vector Encoding::symbolsList() const +{ + // assure the used encoding is properly initialized + init(); + + // first all encodable characters + vector symbols(encodable_.begin(), encodable_.end()); + // add those below start_encodable_ + for (char_type c = 0; c < start_encodable_; ++c) + symbols.push_back(c); + // now the ones from the unicodesymbols file + CharInfoMap::const_iterator const end = unicodesymbols.end(); + CharInfoMap::const_iterator it = unicodesymbols.begin(); + for (; it != end; ++it) + symbols.push_back(it->first); + return symbols; +} + + +bool Encodings::latexMathChar(char_type c, bool mathmode, + Encoding const * encoding, docstring & command, + bool & needsTermination) +{ + command = empty_docstring(); + if (encoding) + if (encoding->encodable(c)) + command = docstring(1, c); + needsTermination = false; + + 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; + needsTermination = !it->second.mathnotermination(); + addMathCmd(c); + } else { + if (!encoding || command.empty()) { + command = it->second.textcommand; + needsTermination = !it->second.textnotermination(); + addTextCmd(c); + } + if (mathmode) + addMathSym(c); + } + return use_math; +} + + +char_type Encodings::fromLaTeXCommand(docstring const & cmd, int cmdtype, + bool & combining, bool & needsTermination, set * req) +{ + 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 ((cmdtype & MATH_CMD) && math == cmd) { + combining = it->second.combining(); + needsTermination = !it->second.mathnotermination(); + if (req && it->second.mathfeature() && + !it->second.mathpreamble.empty()) + req->insert(it->second.mathpreamble); + return it->first; + } + if ((cmdtype & TEXT_CMD) && text == cmd) { + combining = it->second.combining(); + needsTermination = !it->second.textnotermination(); + if (req && it->second.textfeature() && + !it->second.textpreamble.empty()) + req->insert(it->second.textpreamble); + return it->first; + } + } + needsTermination = false; + return 0; +} + + +docstring Encodings::fromLaTeXCommand(docstring const & cmd, int cmdtype, + bool & needsTermination, docstring & rem, set * req) +{ + needsTermination = false; + rem = empty_docstring(); + bool const mathmode = cmdtype & MATH_CMD; + bool const textmode = cmdtype & TEXT_CMD; + docstring symbols; + size_t const cmdend = cmd.size(); + size_t prefix = 0; + CharInfoMap::const_iterator const uniend = unicodesymbols.end(); + for (size_t i = 0, j = 0; j < cmdend; ++j) { + // Also get the char after a backslash + if (j + 1 < cmdend && cmd[j] == '\\') { + ++j; + prefix = 1; + // Detect things like \=*{e} as well + if (j + 3 < cmdend && cmd[j+1] == '*' && + cmd[j+2] == '{') { + ++j; + prefix = 2; + } + } + // position of the last character before a possible macro + // argument + size_t m = j; + // If a macro argument follows, get it, too + // Do it here only for single character commands. Other + // combining commands need this too, but they are handled in + // the loop below for performance reasons. + 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; + } else if (m + 1 < cmdend && isAlphaASCII(cmd[m])) { + while (m + 2 < cmdend && isAlphaASCII(cmd[m+1])) + m++; + } + // Start with this substring and try augmenting it when it is + // the prefix of some command in the unicodesymbols file + docstring subcmd = cmd.substr(i, j - i + 1); + + CharInfoMap::const_iterator it = unicodesymbols.begin(); + // First part of subcmd which might be a combining character + docstring combcmd = (m == j) ? docstring() : cmd.substr(i, m - i + 1); + // The combining character of combcmd if it exists + CharInfoMap::const_iterator combining = uniend; + size_t unicmd_size = 0; + char_type c = 0; + for (; it != uniend; ++it) { + docstring const math = mathmode ? it->second.mathcommand + : docstring(); + docstring const text = textmode ? it->second.textcommand + : docstring(); + if (!combcmd.empty() && it->second.combining() && + (math == combcmd || text == combcmd)) + combining = it; + 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 (math != tmp && text != tmp) + continue; + // If we found a combining command, we need to append + // the macro argument if this has not been done above. + if (tmp == combcmd && combining != uniend && + k < cmdend && cmd[k] == '{') { + size_t l = k; + int count = 1; + while (l < cmdend && count && l != docstring::npos) { + l = cmd.find_first_of(from_ascii("{}"), l + 1); + if (cmd[l] == '{') + ++count; + else + --count; + } + if (l != docstring::npos) { + j = l; + subcmd = cmd.substr(i, j - i + 1); + } + } + // 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 }. + else if (tmp[0] != '\\' + || (tmp.size() == prefix + 1 && + !isAlphaASCII(tmp[1]) && + (prefix == 1 || !isAlphaASCII(tmp[2]))) + || k == cmdend + || !isAlphaASCII(cmd[k]) + || tmp[tmp.size() - 1] == '}' + ) { + c = it->first; + j = k - 1; + i = j + 1; + unicmd_size = cur_size; + if (math == tmp) + needsTermination = !it->second.mathnotermination(); + else + needsTermination = !it->second.textnotermination(); + if (req) { + if (math == tmp && it->second.mathfeature() && + !it->second.mathpreamble.empty()) + req->insert(it->second.mathpreamble); + if (text == tmp && it->second.textfeature() && + !it->second.textpreamble.empty()) + req->insert(it->second.textpreamble); + } + } + } + if (unicmd_size) + symbols += c; + else if (combining != uniend && + prefixIs(subcmd, combcmd + '{')) { + // We know that subcmd starts with combcmd and + // contains an argument in braces. + docstring const arg = subcmd.substr( + combcmd.length() + 1, + subcmd.length() - combcmd.length() - 2); + // If arg is a single character we can construct a + // combining sequence. + char_type a; + bool argcomb = false; + if (arg.size() == 1 && isAlnumASCII(arg[0])) + a = arg[0]; + else { + // Use the version of fromLaTeXCommand() that + // parses only one command, since we cannot + // use more than one character. + bool dummy = false; + set r; + a = fromLaTeXCommand(arg, cmdtype, argcomb, + dummy, &r); + if (a && req && !argcomb) + req->insert(r.begin(), r.end()); + } + if (a && !argcomb) { + // In unicode the combining character comes + // after its base + symbols += a; + symbols += combining->first; + i = j + 1; + unicmd_size = 2; + } + } + if (j + 1 == cmdend && !unicmd_size) { + // No luck. Return what remains + rem = cmd.substr(i); + if (needsTermination && !rem.empty()) { + if (rem.substr(0, 2) == "{}") { + rem = rem.substr(2); + needsTermination = false; + } else if (rem[0] == ' ') { + needsTermination = false; + // LaTeX would swallow all spaces + rem = ltrim(rem); + } + } + } + } + return symbols; +} + + +void Encodings::initUnicodeMath(Buffer const & buffer, bool for_master) +{ +#ifdef TEX2LYX + // The code below is not needed in tex2lyx and requires additional stuff + (void)buffer; + (void)for_master; +#else + if (for_master) { + mathcmd.clear(); + textcmd.clear(); + mathsym.clear(); } - return docstring(1, c); + + // Check this buffer + Inset & inset = buffer.inset(); + InsetIterator it = inset_iterator_begin(inset); + InsetIterator const end = inset_iterator_end(inset); + for (; it != end; ++it) + it->initUnicodeMath(); + + if (!for_master) + return; + + // Check children + ListOfBuffers blist = buffer.getDescendents(); + ListOfBuffers::const_iterator bit = blist.begin(); + ListOfBuffers::const_iterator const bend = blist.end(); + for (; bit != bend; ++bit) + initUnicodeMath(**bit, false); +#endif } -void Encodings::validate(char_type c, LaTeXFeatures & features) +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 math_mode = for_mathed && isMathCmd(c); + bool const use_math = math_mode || + (!for_mathed && it->second.textcommand.empty()); + bool const use_text = (for_mathed && isTextCmd(c)) || + (!for_mathed && !it->second.textcommand.empty()); + bool const plain_utf8 = (features.runparams().encoding->name() == "utf8-plain"); + bool const unicode_math = (features.isRequired("unicode-math") + && features.isAvailable("unicode-math")); + // with utf8-plain, we only load packages when in mathed (see #7766) + // and if we do not use unicode-math + if ((math_mode && !unicode_math) + || (use_math && !plain_utf8)) { + 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); + } + } + // with utf8-plain, we do not load packages (see #7766) + if (use_text && !plain_utf8) { + 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; + return c <= 0x05c2 && c >= 0x05b0 && c != 0x05be && c != 0x05c0; } @@ -359,36 +852,30 @@ 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 == 0x0627 || c == 0x0629 || - c == 0x062f || c == 0x0648 || - (c >= 0x0630 && c <= 0x0632) || - c == 0x0649 || c == 0x0698; + return (c >= 0x0621 && c <= 0x0625) || (c >= 0x0630 && c <= 0x0632) + || c == 0x0627 || c == 0x0629 || c == 0x062f || c == 0x0648 + || c == 0x0649 || c == 0x0698; } -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]; + 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) { - if (!is_arabic(c)) - return c; - - return arabic_table[c-arabic_start][form]; + return isArabicChar(c) ? arabic_table[c-arabic_start][form] : c; } @@ -396,11 +883,20 @@ bool Encodings::isCombiningChar(char_type c) { CharInfoMap::const_iterator const it = unicodesymbols.find(c); if (it != unicodesymbols.end()) - return it->second.combining; + return it->second.combining(); return false; } +string const Encodings::TIPAShortcut(char_type c) +{ + CharInfoMap::const_iterator const it = unicodesymbols.find(c); + if (it != unicodesymbols.end()) + return it->second.tipashortcut; + return string(); +} + + bool Encodings::isKnownScriptChar(char_type const c, string & preamble) { CharInfoMap::const_iterator const it = unicodesymbols.find(c); @@ -408,124 +904,221 @@ 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::isMathAlpha(char_type c) { - EncodingList::const_iterator it = encodinglist.find(name); - if (it != encodinglist.end()) - return &it->second; - else + return mathalpha.count(c); +} + + +Encoding const * +Encodings::fromLyXName(string const & name, bool allowUnsafe) const +{ + EncodingList::const_iterator const it = encodinglist.find(name); + if (!allowUnsafe && it->second.unsafe()) return 0; + return it != encodinglist.end() ? &it->second : 0; } -Encoding const * Encodings::getFromLaTeXName(string const & name) const +Encoding const * +Encodings::fromLaTeXName(string const & n, int const & p, bool allowUnsafe) const { - // We don't use std::find_if because it makes copies of the pairs in + string name = n; + // FIXME: if we have to test for too many of these synonyms, + // we should instead extend the format of lib/encodings + if (n == "ansinew") + name = "cp1252"; + + // We don't use find_if because it makes copies of the pairs in // the map. // This linear search is OK since we don't have many encodings. // Users could even optimize it by putting the encodings they use // most at the top of lib/encodings. EncodingList::const_iterator const end = encodinglist.end(); for (EncodingList::const_iterator it = encodinglist.begin(); it != end; ++it) - if (it->second.latexName() == name) + if ((it->second.latexName() == name) && (it->second.package() & p) + && (!it->second.unsafe() || allowUnsafe)) return &it->second; return 0; } -Encodings::Encodings() +Encoding const * +Encodings::fromIconvName(string const & n, int const & p, bool allowUnsafe) const { + EncodingList::const_iterator const end = encodinglist.end(); + for (EncodingList::const_iterator it = encodinglist.begin(); it != end; ++it) + if ((it->second.iconvName() == n) && (it->second.package() & p) + && (!it->second.unsafe() || allowUnsafe)) + return &it->second; + return 0; } +Encodings::Encodings() +{} + + 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); + CharSetMap forcednotselected; + Lexer symbolslex; symbolslex.setFile(symbolsfile); + bool getNextToken = true; while (symbolslex.isOK()) { char_type symbol; CharInfo info; string flags; - if (symbolslex.next(true)) { - std::istringstream is(symbolslex.getString()); - // reading symbol directly does not work if - // char_type == std::wchar_t. - boost::uint32_t tmp; - if(!(is >> std::hex >> tmp)) + if (getNextToken) { + if (!symbolslex.next(true)) break; - symbol = tmp; } else + getNextToken = true; + + istringstream is(symbolslex.getString()); + // reading symbol directly does not work if + // char_type == wchar_t. + boost::uint32_t tmp; + if(!(is >> hex >> tmp)) break; - if (symbolslex.next(true)) - info.command = symbolslex.getDocString(); - else + symbol = tmp; + + if (!symbolslex.next(true)) break; - if (symbolslex.next(true)) - info.preamble = symbolslex.getString(); - else + info.textcommand = symbolslex.getDocString(); + if (!symbolslex.next(true)) break; - if (symbolslex.next(true)) - flags = symbolslex.getString(); - else + info.textpreamble = symbolslex.getString(); + if (!symbolslex.next(true)) break; + flags = symbolslex.getString(); - info.combining = false; - info.feature = false; - info.force = false; + info.flags = 0; + if (suffixIs(info.textcommand, '}')) + info.flags |= CharInfoTextNoTermination; + if (suffixIs(info.mathcommand, '}')) + info.flags |= CharInfoMathNoTermination; while (!flags.empty()) { string flag; - flags = support::split(flags, flag, ','); - if (flag == "combining") - info.combining = true; - else if (flag == "force") - info.force = true; - else + flags = split(flags, flag, ','); + if (flag == "combining") { + info.flags |= CharInfoCombining; + } else if (flag == "force") { + info.flags |= CharInfoForce; + forced.insert(symbol); + } else if (prefixIs(flag, "force=")) { + vector encodings = + getVectorFromString(flag.substr(6), ";"); + for (size_t i = 0; i < encodings.size(); ++i) + forcedselected[encodings[i]].insert(symbol); + info.flags |= CharInfoForceSelected; + } else if (prefixIs(flag, "force!=")) { + vector encodings = + getVectorFromString(flag.substr(7), ";"); + for (size_t i = 0; i < encodings.size(); ++i) + forcednotselected[encodings[i]].insert(symbol); + info.flags |= CharInfoForceSelected; + } else if (flag == "mathalpha") { + mathalpha.insert(symbol); + } else if (flag == "notermination=text") { + info.flags |= CharInfoTextNoTermination; + } else if (flag == "notermination=math") { + info.flags |= CharInfoMathNoTermination; + } else if (flag == "notermination=both") { + info.flags |= CharInfoTextNoTermination; + info.flags |= CharInfoMathNoTermination; + } else if (flag == "notermination=none") { + info.flags &= ~CharInfoTextNoTermination; + info.flags &= ~CharInfoMathNoTermination; + } else if (contains(flag, "tipashortcut=")) { + info.tipashortcut = split(flag, '='); + } else { lyxerr << "Ignoring unknown flag `" << flag << "' for symbol `0x" - << std::hex << symbol << std::dec + << 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] != '\\'; - - LYXERR(Debug::INFO) - << "Read unicode symbol " << symbol << " '" - << to_utf8(info.command) << "' '" << info.preamble - << "' " << info.combining << ' ' << info.feature - << endl; - unicodesymbols[symbol] = info; + // backward compatibility + if (info.mathpreamble == "esintoramsmath") + info.mathpreamble = "esint|amsmath"; + + if (!info.textpreamble.empty()) + if (info.textpreamble[0] != '\\') + info.flags |= CharInfoTextFeature; + if (!info.mathpreamble.empty()) + if (info.mathpreamble[0] != '\\') + info.flags |= CharInfoMathFeature; + + LYXERR(Debug::INFO, "Read unicode symbol " << symbol << " '" + << to_utf8(info.textcommand) << "' '" << info.textpreamble + << " '" << info.textfeature() << ' ' << info.textnotermination() + << ' ' << to_utf8(info.mathcommand) << "' '" << info.mathpreamble + << "' " << info.mathfeature() << ' ' << info.mathnotermination() + << ' ' << info.combining() << ' ' << info.force() + << ' ' << info.forceselected()); + + // 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 + et_end }; - 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: @@ -535,53 +1128,66 @@ 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(); bool fixedwidth = false; + bool unsafe = false; if (width == "fixed") fixedwidth = true; else if (width == "variable") fixedwidth = false; - else { - lex.printError("Encodings::read: " - "Unknown width: `$$Token'"); - } - + else if (width == "variableunsafe") { + fixedwidth = false; + unsafe = true; + } + else + lex.printError("Unknown width"); + lex.next(); string const p = lex.getString(); Encoding::Package package = Encoding::none; if (p == "none") - package = Encoding::none; + package = Encoding::none; else if (p == "inputenc") package = Encoding::inputenc; else if (p == "CJK") package = Encoding::CJK; - else { - lex.printError("Encodings::read: " - "Unknown package: `$$Token'"); - } - - LYXERR(Debug::INFO) << "Reading encoding " << name << endl; + else if (p == "japanese") + package = Encoding::japanese; + else + lex.printError("Unknown package"); + + LYXERR(Debug::INFO, "Reading encoding " << name); encodinglist[name] = Encoding(name, latexname, - iconvname, fixedwidth, - package); + guiname, iconvname, fixedwidth, unsafe, + 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; } } + + // Move all information from forcednotselected to forcedselected + for (CharSetMap::const_iterator it1 = forcednotselected.begin(); it1 != forcednotselected.end(); ++it1) { + for (CharSetMap::iterator it2 = forcedselected.begin(); it2 != forcedselected.end(); ++it2) { + if (it2->first != it1->first) + it2->second.insert(it1->second.begin(), it1->second.end()); + } + } + }