* 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/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] = {
/// 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;
typedef map<char_type, CharInfo> CharInfoMap;
CharInfoMap unicodesymbols;
+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;
}
-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
void Encoding::init() const
{
+ if (complete_)
+ return;
+
start_encodable_ = 0;
// temporarily switch off lyxerr, since we will generate iconv errors
lyxerr.disable();
for (unsigned short j = 0; j < 256; ++j) {
char const c = char(j);
vector<char_type> 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);
- }
+ if (ucs4.size() != 1)
+ continue;
+ char_type const uc = ucs4[0];
+ CharInfoMap::const_iterator const it = unicodesymbols.find(uc);
+ if (it == unicodesymbols.end() || !it->second.force)
+ encodable_.insert(uc);
}
} else {
// We do not know how many code points this encoding has, and
}
-docstring const Encoding::latexChar(char_type c) const
+docstring Encoding::latexChar(char_type c, bool no_commands) const
{
// assure the used encoding is properly initialized
- if (!complete_)
- init();
- BOOST_ASSERT(complete_);
+ 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 (encodable_.find(c) == encodable_.end()) {
- // c cannot be encoded in this encoding
- CharInfoMap::const_iterator const it = unicodesymbols.find(c);
- if (it == unicodesymbols.end())
+ if (no_commands)
+ return docstring();
+
+ // 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 "\\ensuremath{" + it->second.mathcommand + '}';
+ return it->second.textcommand;
+}
+
+
+vector<char_type> Encoding::symbolsList() const
+{
+ // assure the used encoding is properly initialized
+ init();
+
+ // first all encodable characters
+ vector<char_type> 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)
+{
+ 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);
- else
- return it->second.command;
+ 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,
+ int cmdtype)
+{
+ bool const mathmode = cmdtype & MATH_CMD;
+ bool const textmode = cmdtype & TEXT_CMD;
+ 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 = 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
+ 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 docstring(1, c);
+ return symbols;
}
-void Encodings::validate(char_type c, LaTeXFeatures & features)
+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() && !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;
+ return c <= 0x05c2 && c >= 0x05b0 && c != 0x05be && c != 0x05c0;
}
// 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;
}
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;
+}
+
+
+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::getFromLyXName(string const & name) const
+Encoding const * Encodings::fromLyXName(string const & name) const
{
- EncodingList::const_iterator it = encodinglist.find(name);
- if (it != encodinglist.end())
- return &it->second;
- else
- return 0;
+ 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 & 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.
{
// 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)) {
- istringstream is(symbolslex.getString());
- // reading symbol directly does not work if
- // char_type == wchar_t.
- boost::uint32_t tmp;
- if(!(is >> 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.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
+ 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:
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();
fixedwidth = true;
else if (width == "variable")
fixedwidth = false;
- else {
- lex.printError("Encodings::read: "
- "Unknown width: `$$Token'");
- }
-
+ else
+ lex.printError("Unknown width");
+
lex.next();
string const p = lex.getString();
Encoding::Package package = Encoding::none;
if (p == "none")
- package = Encoding::none;
+ package = Encoding::none;
else if (p == "inputenc")
package = Encoding::inputenc;
else if (p == "CJK")
package = Encoding::CJK;
- else {
- lex.printError("Encodings::read: "
- "Unknown package: `$$Token'");
- }
-
+ 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, 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;
}
}