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::get_token()
226 static const Token dummy;
227 //cerr << "looking at token " << tokens_[pos_] << " pos: " << pos_ << '\n';
228 return good() ? tokens_[pos_++] : dummy;
232 bool Parser::isParagraph()
234 // A new paragraph in TeX ist started
235 // - either by a newline, following any amount of whitespace
236 // characters (including zero), and another newline
237 // - or the token \par
238 if (curr_token().cat() == catNewline &&
239 (curr_token().cs().size() > 1 ||
240 (next_token().cat() == catSpace &&
241 pos_ < tokens_.size() - 1 &&
242 tokens_[pos_ + 1].cat() == catNewline)))
244 if (curr_token().cat() == catEscape && curr_token().cs() == "par")
250 bool Parser::skip_spaces(bool skip_comments)
252 // We just silently return if we have no more tokens.
253 // skip_spaces() should be callable at any time,
254 // the caller must check p::good() anyway.
255 bool skipped = false;
262 if (curr_token().cat() == catSpace ||
263 curr_token().cat() == catNewline) {
267 if ((curr_token().cat() == catComment && curr_token().cs().empty()))
269 if (skip_comments && curr_token().cat() == catComment) {
270 // If positions_ is not empty we are doing some kind
272 if (!positions_.empty())
273 cerr << " Ignoring comment: "
274 << curr_token().asInput();
284 void Parser::unskip_spaces(bool skip_comments)
287 if ( curr_token().cat() == catSpace ||
288 (curr_token().cat() == catNewline && curr_token().cs().size() == 1))
290 else if (skip_comments && curr_token().cat() == catComment) {
291 // TODO: Get rid of this
292 // If positions_ is not empty we are doing some kind
294 if (!positions_.empty())
295 cerr << "Unignoring comment: "
296 << curr_token().asInput();
305 void Parser::putback()
311 void Parser::pushPosition()
313 positions_.push_back(pos_);
317 void Parser::popPosition()
319 pos_ = positions_.back();
320 positions_.pop_back();
326 if (pos_ < tokens_.size())
329 return pos_ < tokens_.size();
333 char Parser::getChar()
336 error("The input stream is not well...");
337 return get_token().character();
341 bool Parser::hasOpt()
343 // An optional argument can occur in any of the following forms:
351 // remember current position
352 unsigned int oldpos = pos_;
353 // skip spaces and comments
360 if (curr_token().cat() == catSpace ||
361 curr_token().cat() == catNewline ||
362 curr_token().cat() == catComment)
367 bool const retval = (next_token().asInput() == "[");
373 Parser::Arg Parser::getFullArg(char left, char right)
377 // This is needed if a partial file ends with a command without arguments,
380 return make_pair(false, string());
387 return make_pair(false, string());
389 while ((c = getChar()) != right && good()) {
391 if (curr_token().cat() == catComment) {
392 if (!curr_token().cs().empty())
393 cerr << "Ignoring comment: " << curr_token().asInput();
396 result += curr_token().asInput();
399 return make_pair(true, result);
403 string Parser::getArg(char left, char right)
405 return getFullArg(left, right).second;
409 string Parser::getFullOpt()
411 Arg arg = getFullArg('[', ']');
413 return '[' + arg.second + ']';
418 string Parser::getOpt(bool keepws)
420 string const res = getArg('[', ']');
426 return '[' + res + ']';
430 string Parser::getOptContent()
431 // the same as getOpt but without the brackets
433 string const res = getArg('[', ']');
434 return res.empty() ? string() : res;
438 string Parser::getFullParentheseArg()
440 Arg arg = getFullArg('(', ')');
442 return '(' + arg.second + ')';
447 string const Parser::verbatimEnvironment(string const & name)
453 for (Token t = get_token(); good(); t = get_token()) {
454 if (t.cat() == catBegin) {
456 os << '{' << verbatim_item() << '}';
457 } else if (t.asInput() == "\\begin") {
458 string const env = getArg('{', '}');
459 os << "\\begin{" << env << '}'
460 << verbatimEnvironment(env)
461 << "\\end{" << env << '}';
462 } else if (t.asInput() == "\\end") {
463 string const end = getArg('{', '}');
465 cerr << "\\end{" << end
466 << "} does not match \\begin{" << name
472 cerr << "unexpected end of input" << endl;
477 void Parser::tokenize_one()
484 switch (catcode(c)) {
487 while (is_.get(c) && catcode(c) == catSpace)
489 if (catcode(c) != catSpace)
491 push_back(Token(s, catSpace));
497 docstring s(1, getNewline(is_, c));
498 while (is_.get(c) && catcode(c) == catNewline) {
500 s += getNewline(is_, c);
502 if (catcode(c) != catNewline)
504 push_back(Token(s, catNewline));
509 // We don't treat "%\n" combinations here specially because
510 // we want to preserve them in the preamble
512 while (is_.get(c) && catcode(c) != catNewline)
514 // handle possible DOS line ending
515 if (catcode(c) == catNewline)
516 c = getNewline(is_, c);
517 // Note: The '%' at the beginning and the '\n' at the end
518 // of the comment are not stored.
520 push_back(Token(s, catComment));
527 error("unexpected end of input");
530 if (catcode(c) == catLetter) {
532 while (is_.get(c) && catcode(c) == catLetter)
534 if (catcode(c) != catLetter)
537 push_back(Token(s, catEscape));
543 cerr << "ignoring a char: " << c << "\n";
548 push_back(Token(docstring(1, c), catcode(c)));
550 //cerr << tokens_.back();
554 void Parser::dump() const
556 cerr << "\nTokens: ";
557 for (unsigned i = 0; i < tokens_.size(); ++i) {
562 cerr << " pos: " << pos_ << "\n";
566 void Parser::error(string const & msg)
568 cerr << "Line ~" << lineno_ << ": parse error: " << msg << endl;
574 string Parser::verbatimOption()
577 if (next_token().character() == '[') {
578 Token t = get_token();
579 for (t = get_token(); t.character() != ']' && good(); t = get_token()) {
580 if (t.cat() == catBegin) {
582 res += '{' + verbatim_item() + '}';
591 string Parser::verbatim_item()
596 if (next_token().cat() == catBegin) {
597 Token t = get_token(); // skip brace
599 for (Token t = get_token(); t.cat() != catEnd && good(); t = get_token()) {
600 if (t.cat() == catBegin) {
602 res += '{' + verbatim_item() + '}';
609 return get_token().asInput();
619 void Parser::setCatCode(char c, CatCode cat)
621 theCatcode[(unsigned char)c] = cat;
625 CatCode Parser::getCatCode(char c) const
627 return theCatcode[(unsigned char)c];