* 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] = {
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
}
-docstring Encoding::latexChar(char_type c) const
+docstring Encoding::latexChar(char_type c, bool no_commands) 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 (no_commands)
+ return docstring();
// c cannot (or should not) be encoded in this encoding
CharInfoMap::const_iterator const it = unicodesymbols.find(c);
}
-bool Encodings::latexMathChar(char_type c, docstring & command)
+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())
- 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;
+ 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 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()) {
- // 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 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)
- 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 {
+ }
+ if (use_text) {
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
}
}
+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);
}
-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.
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.
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();
package = Encoding::inputenc;
else if (p == "CJK")
package = Encoding::CJK;
+ 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("Missing end");