3 * This file is part of LyX, the document processor.
4 * Licence details can be found in the file COPYING.
8 * Full author contact details are available in file CREDITS.
24 CatCode theCatcode[256];
28 static bool init_done = false;
33 fill(theCatcode, theCatcode + 256, catOther);
34 fill(theCatcode + 'a', theCatcode + 'z' + 1, catLetter);
35 fill(theCatcode + 'A', theCatcode + 'Z' + 1, catLetter);
37 theCatcode[int('\\')] = catEscape;
38 theCatcode[int('{')] = catBegin;
39 theCatcode[int('}')] = catEnd;
40 theCatcode[int('$')] = catMath;
41 theCatcode[int('&')] = catAlign;
42 theCatcode[int('\n')] = catNewline;
43 theCatcode[int('#')] = catParameter;
44 theCatcode[int('^')] = catSuper;
45 theCatcode[int('_')] = catSub;
46 theCatcode[0x7f] = catIgnore;
47 theCatcode[int(' ')] = catSpace;
48 theCatcode[int('\t')] = catSpace;
49 theCatcode[int('\r')] = catNewline;
50 theCatcode[int('~')] = catActive;
51 theCatcode[int('%')] = catComment;
54 theCatcode[int('@')] = catLetter;
58 * Translate a line ending to '\n'.
59 * \p c must have catcode catNewline, and it must be the last character read
62 char_type getNewline(idocstream & is, char_type c)
64 // we have to handle 3 different line endings:
71 if (is.get(wc) && wc != '\n') {
81 CatCode catcode(char_type c)
84 return theCatcode[(unsigned char)c];
95 ostream & operator<<(ostream & os, Token const & t)
97 if (t.cat() == catComment)
98 os << '%' << t.cs() << '\n';
99 else if (t.cat() == catSpace)
101 else if (t.cat() == catEscape)
102 os << '\\' << t.cs() << ' ';
103 else if (t.cat() == catLetter)
105 else if (t.cat() == catNewline)
106 os << "[" << t.cs().size() << "\\n," << t.cat() << "]\n";
108 os << '[' << t.cs() << ',' << t.cat() << ']';
113 string Token::asString() const
119 string Token::asInput() const
121 if (cat_ == catComment)
122 return '%' + cs_ + '\n';
123 if (cat_ == catEscape)
134 Parser::Parser(idocstream & is)
135 : lineno_(0), pos_(0), iss_(0), is_(is), encoding_latex_("utf8")
140 Parser::Parser(string const & s)
141 : lineno_(0), pos_(0),
142 iss_(new idocstringstream(from_utf8(s))), is_(*iss_),
143 encoding_latex_("utf8")
154 void Parser::setEncoding(std::string const & e)
156 Encoding const * enc = encodings.fromLaTeXName(e);
157 //cerr << "setting encoding to " << enc->iconvName()<<std::endl;
158 is_ << lyx::setEncoding(enc->iconvName());
163 void Parser::push_back(Token const & t)
165 tokens_.push_back(t);
169 // We return a copy here because the tokens_ vector may get reallocated
170 Token const Parser::prev_token() const
172 static const Token dummy;
173 return pos_ > 1 ? tokens_[pos_ - 2] : dummy;
177 // We return a copy here because the tokens_ vector may get reallocated
178 Token const Parser::curr_token() const
180 static const Token dummy;
181 return pos_ > 0 ? tokens_[pos_ - 1] : dummy;
185 // We return a copy here because the tokens_ vector may get reallocated
186 Token const Parser::next_token()
188 static const Token dummy;
189 return good() ? tokens_[pos_] : dummy;
193 // We return a copy here because the tokens_ vector may get reallocated
194 Token const Parser::get_token()
196 static const Token dummy;
197 //cerr << "looking at token " << tokens_[pos_] << " pos: " << pos_ << '\n';
198 return good() ? tokens_[pos_++] : dummy;
202 bool Parser::isParagraph()
204 // A new paragraph in TeX ist started
205 // - either by a newline, following any amount of whitespace
206 // characters (including zero), and another newline
207 // - or the token \par
208 if (curr_token().cat() == catNewline &&
209 (curr_token().cs().size() > 1 ||
210 (next_token().cat() == catSpace &&
211 pos_ < tokens_.size() - 1 &&
212 tokens_[pos_ + 1].cat() == catNewline)))
214 if (curr_token().cat() == catEscape && curr_token().cs() == "par")
220 void Parser::skip_spaces(bool skip_comments)
222 // We just silently return if we have no more tokens.
223 // skip_spaces() should be callable at any time,
224 // the caller must check p::good() anyway.
231 if ( curr_token().cat() == catSpace ||
232 curr_token().cat() == catNewline ||
233 (curr_token().cat() == catComment && curr_token().cs().empty()))
235 if (skip_comments && curr_token().cat() == catComment)
236 cerr << " Ignoring comment: " << curr_token().asInput();
245 void Parser::unskip_spaces(bool skip_comments)
248 if ( curr_token().cat() == catSpace ||
249 (curr_token().cat() == catNewline && curr_token().cs().size() == 1))
251 else if (skip_comments && curr_token().cat() == catComment) {
252 // TODO: Get rid of this
253 cerr << "Unignoring comment: " << curr_token().asInput();
262 void Parser::putback()
270 if (pos_ < tokens_.size())
273 return pos_ < tokens_.size();
277 char Parser::getChar()
280 error("The input stream is not well...");
281 return get_token().character();
285 Parser::Arg Parser::getFullArg(char left, char right)
289 // This is needed if a partial file ends with a command without arguments,
292 return make_pair(false, string());
299 return make_pair(false, string());
301 while ((c = getChar()) != right && good()) {
303 if (curr_token().cat() == catComment) {
304 if (!curr_token().cs().empty())
305 cerr << "Ignoring comment: " << curr_token().asInput();
308 result += curr_token().asInput();
311 return make_pair(true, result);
315 string Parser::getArg(char left, char right)
317 return getFullArg(left, right).second;
321 string Parser::getFullOpt()
323 Arg arg = getFullArg('[', ']');
325 return '[' + arg.second + ']';
330 string Parser::getOpt()
332 string const res = getArg('[', ']');
333 return res.empty() ? string() : '[' + res + ']';
337 string Parser::getOptContent()
338 // the same as getOpt but without the brackets
340 string const res = getArg('[', ']');
341 return res.empty() ? string() : res;
345 string Parser::getFullParentheseArg()
347 Arg arg = getFullArg('(', ')');
349 return '(' + arg.second + ')';
354 string const Parser::verbatimEnvironment(string const & name)
360 for (Token t = get_token(); good(); t = get_token()) {
361 if (t.cat() == catBegin) {
363 os << '{' << verbatim_item() << '}';
364 } else if (t.asInput() == "\\begin") {
365 string const env = getArg('{', '}');
366 os << "\\begin{" << env << '}'
367 << verbatimEnvironment(env)
368 << "\\end{" << env << '}';
369 } else if (t.asInput() == "\\end") {
370 string const end = getArg('{', '}');
372 cerr << "\\end{" << end
373 << "} does not match \\begin{" << name
379 cerr << "unexpected end of input" << endl;
384 void Parser::tokenize_one()
391 switch (catcode(c)) {
394 while (is_.get(c) && catcode(c) == catSpace)
396 if (catcode(c) != catSpace)
398 push_back(Token(s, catSpace));
404 docstring s(1, getNewline(is_, c));
405 while (is_.get(c) && catcode(c) == catNewline) {
407 s += getNewline(is_, c);
409 if (catcode(c) != catNewline)
411 push_back(Token(s, catNewline));
416 // We don't treat "%\n" combinations here specially because
417 // we want to preserve them in the preamble
419 while (is_.get(c) && catcode(c) != catNewline)
421 // handle possible DOS line ending
422 if (catcode(c) == catNewline)
423 c = getNewline(is_, c);
424 // Note: The '%' at the beginning and the '\n' at the end
425 // of the comment are not stored.
427 push_back(Token(s, catComment));
434 error("unexpected end of input");
437 if (catcode(c) == catLetter) {
439 while (is_.get(c) && catcode(c) == catLetter)
441 if (catcode(c) != catLetter)
444 push_back(Token(s, catEscape));
450 cerr << "ignoring a char: " << c << "\n";
455 push_back(Token(docstring(1, c), catcode(c)));
457 //cerr << tokens_.back();
461 void Parser::dump() const
463 cerr << "\nTokens: ";
464 for (unsigned i = 0; i < tokens_.size(); ++i) {
469 cerr << " pos: " << pos_ << "\n";
473 void Parser::error(string const & msg)
475 cerr << "Line ~" << lineno_ << ": parse error: " << msg << endl;
481 string Parser::verbatimOption()
484 if (next_token().character() == '[') {
485 Token t = get_token();
486 for (t = get_token(); t.character() != ']' && good(); t = get_token()) {
487 if (t.cat() == catBegin) {
489 res += '{' + verbatim_item() + '}';
498 string Parser::verbatim_item()
503 if (next_token().cat() == catBegin) {
504 Token t = get_token(); // skip brace
506 for (Token t = get_token(); t.cat() != catEnd && good(); t = get_token()) {
507 if (t.cat() == catBegin) {
509 res += '{' + verbatim_item() + '}';
516 return get_token().asInput();
526 void Parser::setCatCode(char c, CatCode cat)
528 theCatcode[(unsigned char)c] = cat;
532 CatCode Parser::getCatCode(char c) const
534 return theCatcode[(unsigned char)c];