}
-bool Counters::read(Lexer & lex, docstring const & name)
+bool Counters::read(Lexer & lex, docstring const & name, bool makenew)
{
if (hasCounter(name)) {
LYXERR(Debug::TCLASS, "Reading existing counter " << to_utf8(name));
return counterList_[name].read(lex);
}
+
LYXERR(Debug::TCLASS, "Reading new counter " << to_utf8(name));
Counter cnt;
bool success = cnt.read(lex);
- if (success)
+ // if makenew is false, we will just discard what we read
+ if (success && makenew)
counterList_[name] = cnt;
- else
+ else if (!success)
LYXERR0("Error reading counter `" << name << "'!");
return success;
}
// FIXME: Using regexps would be better, but we compile boost without
// wide regexps currently.
+ docstring const the = from_ascii("\\the");
while (true) {
- //lyxerr << "label=" << to_utf8(label) << endl;
- size_t const i = label.find(from_ascii("\\the"), 0);
+ //lyxerr << "label=" << label << endl;
+ size_t const i = label.find(the, 0);
if (i == docstring::npos)
break;
size_t const j = i + 4;
while (k < label.size() && lowercase(label[k]) >= 'a'
&& lowercase(label[k]) <= 'z')
++k;
- docstring const newc = label.substr(j, k - j);
- docstring const repl = theCounter(newc, lang);
- label.replace(i, k - j + 4, repl);
+ docstring const newc(label, j, k - j);
+ label.replace(i, k - i, theCounter(newc, lang));
}
while (true) {
- //lyxerr << "label=" << to_utf8(label) << endl;
+ //lyxerr << "label=" << label << endl;
size_t const i = label.find('\\', 0);
if (i == docstring::npos)
break;
docstring const numbertype(label, i + 1, j - i - 1);
docstring const counter(label, j + 1, k - j - 1);
- docstring const rep = labelItem(counter, numbertype);
- label = docstring(label, 0, i) + rep
- + docstring(label, k + 1, docstring::npos);
+ label.replace(i, k + 1 - i, labelItem(counter, numbertype));
}
- //lyxerr << "DONE! label=" << to_utf8(label) << endl;
+ //lyxerr << "DONE! label=" << label << endl;
return label;
}