* 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
*
#include "Encoding.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 <boost/cstdint.hpp>
Encodings encodings;
+Encodings::MathCommandSet Encodings::mathcmd;
+Encodings::TextCommandSet Encodings::textcmd;
+Encodings::MathSymbolSet Encodings::mathsym;
+
namespace {
char_type arabic_table[172][4] = {
char_type const arabic_end = 0x06cc;
+enum CharInfoFlags {
+ ///
+ CharInfoCombining = 1,
+ ///
+ CharInfoTextFeature = 2,
+ ///
+ CharInfoMathFeature = 4,
+ ///
+ CharInfoForce = 8,
+ ///
+ CharInfoTextNoTermination = 16,
+ ///
+ CharInfoMathNoTermination = 32,
+};
+
/// Information about a single UCS4 character
struct CharInfo {
/// LaTeX command (text mode) for this character
/// Needed LaTeX preamble (or feature) for math mode
string mathpreamble;
/// Is this a combining character?
- bool combining;
+ bool combining() const { return flags & CharInfoCombining ? true : false; }
/// Is \c textpreamble a feature known by LaTeXFeatures, or a raw LaTeX
/// command?
- bool textfeature;
+ bool textfeature() const { return flags & CharInfoTextFeature ? true : false; }
/// Is \c mathpreamble a feature known by LaTeXFeatures, or a raw LaTeX
/// command?
- bool mathfeature;
+ 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; }
+ /// 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::set<char_type> CharSet;
CharSet forced;
+typedef std::set<char_type> MathAlphaSet;
+MathAlphaSet mathalpha;
+
/// The highest code point in UCS4 encoding (1<<20 + 1<<16)
char_type const max_ucs4 = 0x110000;
continue;
char_type const uc = ucs4[0];
CharInfoMap::const_iterator const it = unicodesymbols.find(uc);
- if (it == unicodesymbols.end() || !it->second.force)
+ if (it == unicodesymbols.end() || !it->second.force())
encodable_.insert(uc);
}
} else {
vector<char> 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() || !it->second.force())
encodable_.insert(c);
}
}
}
-docstring Encoding::latexChar(char_type c) const
+bool Encoding::encodable(char_type c) const
{
// assure the used encoding is properly initialized
init();
if (iconvName_ == "UTF-8" && package_ == none)
- return docstring(1, c);
+ return true;
if (c < start_encodable_ && !encodings.isForced(c))
- return docstring(1, c);
+ return true;
if (encodable_.find(c) != encodable_.end())
- return docstring(1, c);
+ return true;
+ return false;
+}
+
+
+pair<docstring, bool> 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);
throw EncodingException(c);
// at least one of mathcommand and textcommand is nonempty
if (it->second.textcommand.empty())
- return "\\ensuremath{" + it->second.mathcommand + '}';
- return it->second.textcommand;
+ return make_pair(
+ "\\ensuremath{" + it->second.mathcommand + '}', false);
+ return make_pair(it->second.textcommand, !it->second.textnotermination());
+}
+
+
+pair<docstring, docstring> 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<docstring, bool> 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);
}
}
-bool Encodings::latexMathChar(char_type c, docstring & command)
+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())
- throw EncodingException(c);
- if (it->second.mathcommand.empty()) {
- if (it->second.textcommand.empty())
+ if (it == unicodesymbols.end()) {
+ if (!encoding || command.empty())
throw EncodingException(c);
- command = it->second.textcommand;
+ if (mathmode)
+ addMathSym(c);
return false;
}
- command = it->second.mathcommand;
- return true;
+ // 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, bool & combining)
+char_type Encodings::fromLaTeXCommand(docstring const & cmd, int cmdtype,
+ bool & combining, bool & needsTermination, set<string> * req)
{
CharInfoMap::const_iterator const end = unicodesymbols.end();
CharInfoMap::const_iterator it = unicodesymbols.begin();
- for (; it != end; ++it) {
+ 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;
+ 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, docstring & rem)
+docstring Encodings::fromLaTeXCommand(docstring const & cmd, int cmdtype,
+ bool & needsTermination, docstring & rem, set<string> * req)
{
+ needsTermination = false;
+ rem = empty_docstring();
+ bool const mathmode = cmdtype & MATH_CMD;
+ bool const textmode = cmdtype & TEXT_CMD;
docstring symbols;
size_t i = 0;
size_t const cmdend = cmd.size();
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;
+ docstring const math = mathmode ? it->second.mathcommand
+ : docstring();
+ docstring const text = textmode ? it->second.textcommand
+ : docstring();
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
tmp.resize(tmp.size() - 1);
// If this is an exact match, we found a (longer)
- // matching command in the unicodesymbols file
- if (math == tmp || text == tmp) {
+ // 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 (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 (j + 1 == cmdend)
+ else if (j + 1 == cmdend) {
// 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();
+ }
+
+ // 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, 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()) {
- // at least one of mathcommand and textcommand is nonempty
- bool const use_math = (for_mathed && !it->second.mathcommand.empty()) ||
+ // 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());
- if (use_math) {
+ 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)
- features.require(it->second.mathpreamble);
- else
+ 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);
}
- } else {
+ }
+ // with utf8-plain, we do not load packages (see #7766)
+ if (use_text && !plain_utf8) {
if (!it->second.textpreamble.empty()) {
- if (it->second.textfeature)
- features.require(it->second.textpreamble);
- else
+ 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) {
- features.require("relsize");
- features.require("lyxmathsym");
- }
}
}
+ if (for_mathed && isMathSym(c)) {
+ features.require("amstext");
+ features.require("lyxmathsym");
+ }
+#endif
}
{
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);
}
+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);
}
-Encoding const * Encodings::fromLaTeXName(string const & name) const
+Encoding const * Encodings::fromLaTeXName(string const & n) const
{
+ 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.
Encodings::Encodings()
-{
-}
+{}
void Encodings::read(FileName const & encfile, FileName const & symbolsfile)
break;
flags = symbolslex.getString();
- info.combining = false;
- info.textfeature = 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 = split(flags, flag, ',');
- if (flag == "combining")
- info.combining = true;
- else if (flag == "force") {
- info.force = true;
+ if (flag == "combining") {
+ info.flags |= CharInfoCombining;
+ } else if (flag == "force") {
+ info.flags |= CharInfoForce;
forced.insert(symbol);
- } else
+ } 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"
<< hex << symbol << dec
<< "'." << endl;
+ }
}
// mathcommand and mathpreamble have been added for 1.6.0.
// make them optional so that old files still work.
breakout = true;
}
+ // backward compatibility
+ if (info.mathpreamble == "esintoramsmath")
+ info.mathpreamble = "esint|amsmath";
+
if (!info.textpreamble.empty())
- info.textfeature = info.textpreamble[0] != '\\';
+ if (info.textpreamble[0] != '\\')
+ info.flags |= CharInfoTextFeature;
if (!info.mathpreamble.empty())
- info.mathfeature = info.mathpreamble[0] != '\\';
+ if (info.mathpreamble[0] != '\\')
+ info.flags |= CharInfoMathFeature;
LYXERR(Debug::INFO, "Read unicode symbol " << symbol << " '"
<< to_utf8(info.textcommand) << "' '" << info.textpreamble
- << "' " << info.combining << ' ' << info.textfeature
- << " '" << to_utf8(info.mathcommand) << "' '"
- << info.mathpreamble << "' " << info.mathfeature);
+ << " '" << info.textfeature() << ' ' << info.textnotermination()
+ << ' ' << to_utf8(info.mathcommand) << "' '" << info.mathpreamble
+ << "' " << info.mathfeature() << ' ' << info.mathnotermination()
+ << ' ' << info.combining() << ' ' << info.force());
// we assume that at least one command is nonempty when using unicodesymbols
if (!info.textcommand.empty() || !info.mathcommand.empty())
// Now read the encodings
enum {
et_encoding = 1,
- et_end,
+ et_end
};
LexerKeyword encodingtags[] = {
package = Encoding::inputenc;
else if (p == "CJK")
package = Encoding::CJK;
+ else if (p == "japanese")
+ package = Encoding::japanese;
else
lex.printError("Unknown package");