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.
15 #include "support/textutils.h"
25 CatCode theCatcode[256];
29 static bool init_done = false;
34 fill(theCatcode, theCatcode + 256, catOther);
35 fill(theCatcode + 'a', theCatcode + 'z' + 1, catLetter);
36 fill(theCatcode + 'A', theCatcode + 'Z' + 1, catLetter);
38 theCatcode[int('\\')] = catEscape;
39 theCatcode[int('{')] = catBegin;
40 theCatcode[int('}')] = catEnd;
41 theCatcode[int('$')] = catMath;
42 theCatcode[int('&')] = catAlign;
43 theCatcode[int('\n')] = catNewline;
44 theCatcode[int('#')] = catParameter;
45 theCatcode[int('^')] = catSuper;
46 theCatcode[int('_')] = catSub;
47 theCatcode[0x7f] = catIgnore;
48 theCatcode[int(' ')] = catSpace;
49 theCatcode[int('\t')] = catSpace;
50 theCatcode[int('\r')] = catNewline;
51 theCatcode[int('~')] = catActive;
52 theCatcode[int('%')] = catComment;
55 theCatcode[int('@')] = catLetter;
59 * Translate a line ending to '\n'.
60 * \p c must have catcode catNewline, and it must be the last character read
63 char_type getNewline(idocstream & is, char_type c)
65 // we have to handle 3 different line endings:
72 if (is.get(wc) && wc != '\n') {
82 CatCode catcode(char_type c)
85 return theCatcode[(unsigned char)c];
96 ostream & operator<<(ostream & os, Token const & t)
98 if (t.cat() == catComment)
99 os << '%' << t.cs() << '\n';
100 else if (t.cat() == catSpace)
102 else if (t.cat() == catEscape)
103 os << '\\' << t.cs() << ' ';
104 else if (t.cat() == catLetter)
106 else if (t.cat() == catNewline)
107 os << "[" << t.cs().size() << "\\n," << t.cat() << "]\n";
109 os << '[' << t.cs() << ',' << t.cat() << ']';
114 string Token::asInput() const
116 if (cat_ == catComment)
117 return '%' + cs_ + '\n';
118 if (cat_ == catEscape)
124 bool Token::isAlnumASCII() const
126 return cat_ == catLetter ||
127 (cat_ == catOther && cs_.length() == 1 && isDigitASCII(cs_[0]));
132 void debugToken(std::ostream & os, Token const & t, unsigned int flags)
135 os << "t: " << t << " flags: " << flags;
136 if (flags & FLAG_BRACE_LAST) { os << sep << "BRACE_LAST"; sep = '|'; }
137 if (flags & FLAG_RIGHT ) { os << sep << "RIGHT" ; sep = '|'; }
138 if (flags & FLAG_END ) { os << sep << "END" ; sep = '|'; }
139 if (flags & FLAG_BRACK_LAST) { os << sep << "BRACK_LAST"; sep = '|'; }
140 if (flags & FLAG_TEXTMODE ) { os << sep << "TEXTMODE" ; sep = '|'; }
141 if (flags & FLAG_ITEM ) { os << sep << "ITEM" ; sep = '|'; }
142 if (flags & FLAG_LEAVE ) { os << sep << "LEAVE" ; sep = '|'; }
143 if (flags & FLAG_SIMPLE ) { os << sep << "SIMPLE" ; sep = '|'; }
144 if (flags & FLAG_EQUATION ) { os << sep << "EQUATION" ; sep = '|'; }
145 if (flags & FLAG_SIMPLE2 ) { os << sep << "SIMPLE2" ; sep = '|'; }
146 if (flags & FLAG_OPTION ) { os << sep << "OPTION" ; sep = '|'; }
147 if (flags & FLAG_BRACED ) { os << sep << "BRACED" ; sep = '|'; }
148 if (flags & FLAG_CELL ) { os << sep << "CELL" ; sep = '|'; }
149 if (flags & FLAG_TABBING ) { os << sep << "TABBING" ; sep = '|'; }
160 Parser::Parser(idocstream & is)
161 : lineno_(0), pos_(0), iss_(0), is_(is), encoding_latex_("utf8")
166 Parser::Parser(string const & s)
167 : lineno_(0), pos_(0),
168 iss_(new idocstringstream(from_utf8(s))), is_(*iss_),
169 encoding_latex_("utf8")
180 void Parser::setEncoding(std::string const & e)
182 Encoding const * enc = encodings.fromLaTeXName(e);
184 cerr << "Unknown encoding " << e << ". Ignoring." << std::endl;
187 //cerr << "setting encoding to " << enc->iconvName() << std::endl;
188 is_ << lyx::setEncoding(enc->iconvName());
193 void Parser::push_back(Token const & t)
195 tokens_.push_back(t);
199 // We return a copy here because the tokens_ vector may get reallocated
200 Token const Parser::prev_token() const
202 static const Token dummy;
203 return pos_ > 1 ? tokens_[pos_ - 2] : dummy;
207 // We return a copy here because the tokens_ vector may get reallocated
208 Token const Parser::curr_token() const
210 static const Token dummy;
211 return pos_ > 0 ? tokens_[pos_ - 1] : dummy;
215 // We return a copy here because the tokens_ vector may get reallocated
216 Token const Parser::next_token()
218 static const Token dummy;
219 return good() ? tokens_[pos_] : dummy;
223 // We return a copy here because the tokens_ vector may get reallocated
224 Token const Parser::next_next_token()
226 static const Token dummy;
227 // If good() has not been called after the last get_token() we need
228 // to tokenize two more tokens.
229 if (pos_ + 1 >= tokens_.size()) {
233 return pos_ + 1 < tokens_.size() ? tokens_[pos_ + 1] : dummy;
237 // We return a copy here because the tokens_ vector may get reallocated
238 Token const Parser::get_token()
240 static const Token dummy;
241 //cerr << "looking at token " << tokens_[pos_] << " pos: " << pos_ << '\n';
242 return good() ? tokens_[pos_++] : dummy;
246 bool Parser::isParagraph()
248 // A new paragraph in TeX ist started
249 // - either by a newline, following any amount of whitespace
250 // characters (including zero), and another newline
251 // - or the token \par
252 if (curr_token().cat() == catNewline &&
253 (curr_token().cs().size() > 1 ||
254 (next_token().cat() == catSpace &&
255 next_next_token().cat() == catNewline)))
257 if (curr_token().cat() == catEscape && curr_token().cs() == "par")
263 bool Parser::skip_spaces(bool skip_comments)
265 // We just silently return if we have no more tokens.
266 // skip_spaces() should be callable at any time,
267 // the caller must check p::good() anyway.
268 bool skipped = false;
275 if (curr_token().cat() == catSpace ||
276 curr_token().cat() == catNewline) {
280 if ((curr_token().cat() == catComment && curr_token().cs().empty()))
282 if (skip_comments && curr_token().cat() == catComment) {
283 // If positions_ is not empty we are doing some kind
285 if (!positions_.empty())
286 cerr << " Ignoring comment: "
287 << curr_token().asInput();
297 void Parser::unskip_spaces(bool skip_comments)
300 if ( curr_token().cat() == catSpace ||
301 (curr_token().cat() == catNewline && curr_token().cs().size() == 1))
303 else if (skip_comments && curr_token().cat() == catComment) {
304 // TODO: Get rid of this
305 // If positions_ is not empty we are doing some kind
307 if (!positions_.empty())
308 cerr << "Unignoring comment: "
309 << curr_token().asInput();
318 void Parser::putback()
324 void Parser::pushPosition()
326 positions_.push_back(pos_);
330 void Parser::popPosition()
332 pos_ = positions_.back();
333 positions_.pop_back();
339 if (pos_ < tokens_.size())
342 return pos_ < tokens_.size();
346 char Parser::getChar()
349 error("The input stream is not well...");
350 return get_token().character();
354 bool Parser::hasOpt()
356 // An optional argument can occur in any of the following forms:
364 // remember current position
365 unsigned int oldpos = pos_;
366 // skip spaces and comments
373 if (curr_token().cat() == catSpace ||
374 curr_token().cat() == catNewline ||
375 curr_token().cat() == catComment)
380 bool const retval = (next_token().asInput() == "[");
386 Parser::Arg Parser::getFullArg(char left, char right, bool allow_escaping)
390 // This is needed if a partial file ends with a command without arguments,
393 return make_pair(false, string());
396 Token t = get_token();
398 if (t.cat() == catComment || t.cat() == catEscape ||
399 t.character() != left) {
401 return make_pair(false, string());
403 for (t = get_token(); good(); t = get_token()) {
405 if (t.cat() == catComment) {
407 cerr << "Ignoring comment: " << t.asInput();
410 if (allow_escaping) {
411 if (t.cat() != catEscape && t.character() == right)
414 if (t.character() == right) {
415 if (t.cat() == catEscape)
420 result += t.asInput();
423 return make_pair(true, result);
427 string Parser::getArg(char left, char right, bool allow_escaping)
429 return getFullArg(left, right, allow_escaping).second;
433 string Parser::getFullOpt(bool keepws)
435 Arg arg = getFullArg('[', ']');
437 return '[' + arg.second + ']';
444 string Parser::getOpt(bool keepws)
446 string const res = getArg('[', ']');
452 return '[' + res + ']';
456 string Parser::getFullParentheseArg()
458 Arg arg = getFullArg('(', ')');
460 return '(' + arg.second + ')';
465 string const Parser::verbatimEnvironment(string const & name)
471 for (Token t = get_token(); good(); t = get_token()) {
472 if (t.cat() == catBegin) {
474 os << '{' << verbatim_item() << '}';
475 } else if (t.asInput() == "\\begin") {
476 string const env = getArg('{', '}');
477 os << "\\begin{" << env << '}'
478 << verbatimEnvironment(env)
479 << "\\end{" << env << '}';
480 } else if (t.asInput() == "\\end") {
481 string const end = getArg('{', '}');
483 cerr << "\\end{" << end
484 << "} does not match \\begin{" << name
490 cerr << "unexpected end of input" << endl;
495 string const Parser::plainEnvironment(string const & name)
501 for (Token t = get_token(); good(); t = get_token()) {
502 if (t.asInput() == "\\end") {
503 string const end = getArg('{', '}');
507 os << "\\end{" << end << '}';
511 cerr << "unexpected end of input" << endl;
516 string const Parser::plainCommand(char left, char right, string const & name)
520 // check if first token is really the start character
521 Token tok = get_token();
522 if (tok.character() != left) {
523 cerr << "first character does not match start character of command \\" << name << endl;
527 for (Token t = get_token(); good(); t = get_token()) {
528 if (t.character() == right) {
533 cerr << "unexpected end of input" << endl;
538 void Parser::tokenize_one()
545 switch (catcode(c)) {
548 while (is_.get(c) && catcode(c) == catSpace)
550 if (catcode(c) != catSpace)
552 push_back(Token(s, catSpace));
558 docstring s(1, getNewline(is_, c));
559 while (is_.get(c) && catcode(c) == catNewline) {
561 s += getNewline(is_, c);
563 if (catcode(c) != catNewline)
565 push_back(Token(s, catNewline));
570 // We don't treat "%\n" combinations here specially because
571 // we want to preserve them in the preamble
573 while (is_.get(c) && catcode(c) != catNewline)
575 // handle possible DOS line ending
576 if (catcode(c) == catNewline)
577 c = getNewline(is_, c);
578 // Note: The '%' at the beginning and the '\n' at the end
579 // of the comment are not stored.
581 push_back(Token(s, catComment));
588 error("unexpected end of input");
591 if (catcode(c) == catLetter) {
593 while (is_.get(c) && catcode(c) == catLetter)
595 if (catcode(c) != catLetter)
598 push_back(Token(s, catEscape));
604 cerr << "ignoring a char: " << c << "\n";
609 push_back(Token(docstring(1, c), catcode(c)));
611 //cerr << tokens_.back();
615 void Parser::dump() const
617 cerr << "\nTokens: ";
618 for (unsigned i = 0; i < tokens_.size(); ++i) {
623 cerr << " pos: " << pos_ << "\n";
627 void Parser::error(string const & msg)
629 cerr << "Line ~" << lineno_ << ": parse error: " << msg << endl;
635 string Parser::verbatimOption()
638 if (next_token().character() == '[') {
639 Token t = get_token();
640 for (t = get_token(); t.character() != ']' && good(); t = get_token()) {
641 if (t.cat() == catBegin) {
643 res += '{' + verbatim_item() + '}';
652 string Parser::verbatim_item()
657 if (next_token().cat() == catBegin) {
658 Token t = get_token(); // skip brace
660 for (Token t = get_token(); t.cat() != catEnd && good(); t = get_token()) {
661 if (t.cat() == catBegin) {
663 res += '{' + verbatim_item() + '}';
670 return get_token().asInput();
680 void Parser::setCatCode(char c, CatCode cat)
682 theCatcode[(unsigned char)c] = cat;
686 CatCode Parser::getCatCode(char c) const
688 return theCatcode[(unsigned char)c];