+ 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()) {
+ // 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::isHebrewComposeChar(char_type c)
+{
+ return c <= 0x05c2 && c >= 0x05b0 && c != 0x05be && c != 0x05c0;