| 1 |
#ifndef INC_CharScanner_hpp__ |
| 2 |
#define INC_CharScanner_hpp__ |
| 3 |
|
| 4 |
/* ANTLR Translator Generator |
| 5 |
* Project led by Terence Parr at http://www.jGuru.com |
| 6 |
* Software rights: http://www.antlr.org/license.html |
| 7 |
* |
| 8 |
* $Id$ |
| 9 |
*/ |
| 10 |
|
| 11 |
#include <antlr/config.hpp> |
| 12 |
|
| 13 |
#include <map> |
| 14 |
#include <cstdio> |
| 15 |
|
| 16 |
#ifdef HAS_NOT_CCTYPE_H |
| 17 |
#include <ctype.h> |
| 18 |
#else |
| 19 |
#include <cctype> |
| 20 |
#endif |
| 21 |
|
| 22 |
#if ( _MSC_VER == 1200 ) |
| 23 |
// VC6 seems to need this |
| 24 |
// note that this is not a standard C++ include file. |
| 25 |
# include <stdio.h> |
| 26 |
#endif |
| 27 |
|
| 28 |
#include <antlr/TokenStream.hpp> |
| 29 |
#include <antlr/RecognitionException.hpp> |
| 30 |
#include <antlr/SemanticException.hpp> |
| 31 |
#include <antlr/MismatchedCharException.hpp> |
| 32 |
#include <antlr/InputBuffer.hpp> |
| 33 |
#include <antlr/BitSet.hpp> |
| 34 |
#include <antlr/LexerSharedInputState.hpp> |
| 35 |
|
| 36 |
#ifdef ANTLR_CXX_SUPPORTS_NAMESPACE |
| 37 |
namespace antlr { |
| 38 |
#endif |
| 39 |
|
| 40 |
class ANTLR_API CharScanner; |
| 41 |
|
| 42 |
ANTLR_C_USING(tolower) |
| 43 |
|
| 44 |
#if !defined(HAVE_STRCASECMP) && defined(HAVE_STRICMP) && !defined(stricmp) |
| 45 |
#define strcasecmp stricmp |
| 46 |
#endif |
| 47 |
#if !defined(HAVE_STRNCASECMP) && defined(HAVE_STRNICMP) && !defined(strnicmp) |
| 48 |
#define strncasecmp strnicmp |
| 49 |
#endif |
| 50 |
|
| 51 |
|
| 52 |
#if !defined(HAVE_STRCASECMP) && !defined(HAVE_STRICMP) |
| 53 |
inline int strcasecmp(const char *s1, const char *s2) |
| 54 |
{ |
| 55 |
while (true) |
| 56 |
{ |
| 57 |
char c1 = tolower(*s1++), |
| 58 |
c2 = tolower(*s2++); |
| 59 |
if (c1 < c2) return -1; |
| 60 |
if (c1 > c2) return 1; |
| 61 |
if (c1 == 0) return 0; |
| 62 |
} |
| 63 |
} |
| 64 |
#endif |
| 65 |
|
| 66 |
/** Functor for the literals map |
| 67 |
*/ |
| 68 |
class ANTLR_API CharScannerLiteralsLess : public ANTLR_USE_NAMESPACE(std)binary_function<ANTLR_USE_NAMESPACE(std)string,ANTLR_USE_NAMESPACE(std)string,bool> { |
| 69 |
private: |
| 70 |
const CharScanner* scanner; |
| 71 |
public: |
| 72 |
#ifdef NO_TEMPLATE_PARTS |
| 73 |
CharScannerLiteralsLess() {} // not really used, definition to appease MSVC |
| 74 |
#endif |
| 75 |
CharScannerLiteralsLess(const CharScanner* theScanner) |
| 76 |
: scanner(theScanner) |
| 77 |
{ |
| 78 |
} |
| 79 |
bool operator() (const ANTLR_USE_NAMESPACE(std)string& x,const ANTLR_USE_NAMESPACE(std)string& y) const; |
| 80 |
// defaults are good enough.. |
| 81 |
// CharScannerLiteralsLess(const CharScannerLiteralsLess&); |
| 82 |
// CharScannerLiteralsLess& operator=(const CharScannerLiteralsLess&); |
| 83 |
}; |
| 84 |
|
| 85 |
/** Superclass of generated lexers |
| 86 |
*/ |
| 87 |
class ANTLR_API CharScanner : public TokenStream { |
| 88 |
protected: |
| 89 |
typedef RefToken (*factory_type)(); |
| 90 |
public: |
| 91 |
CharScanner(InputBuffer& cb, bool case_sensitive ); |
| 92 |
CharScanner(InputBuffer* cb, bool case_sensitive ); |
| 93 |
CharScanner(const LexerSharedInputState& state, bool case_sensitive ); |
| 94 |
|
| 95 |
virtual ~CharScanner() |
| 96 |
{ |
| 97 |
} |
| 98 |
|
| 99 |
virtual int LA(unsigned int i); |
| 100 |
|
| 101 |
virtual void append(char c) |
| 102 |
{ |
| 103 |
if (saveConsumedInput) |
| 104 |
{ |
| 105 |
size_t l = text.length(); |
| 106 |
|
| 107 |
if ((l%256) == 0) |
| 108 |
text.reserve(l+256); |
| 109 |
|
| 110 |
text.replace(l,0,&c,1); |
| 111 |
} |
| 112 |
} |
| 113 |
|
| 114 |
virtual void append(const ANTLR_USE_NAMESPACE(std)string& s) |
| 115 |
{ |
| 116 |
if( saveConsumedInput ) |
| 117 |
text += s; |
| 118 |
} |
| 119 |
|
| 120 |
virtual void commit() |
| 121 |
{ |
| 122 |
inputState->getInput().commit(); |
| 123 |
} |
| 124 |
|
| 125 |
virtual void consume() |
| 126 |
{ |
| 127 |
if (inputState->guessing == 0) |
| 128 |
{ |
| 129 |
int c = LA(1); |
| 130 |
if (caseSensitive) |
| 131 |
{ |
| 132 |
append(c); |
| 133 |
} |
| 134 |
else |
| 135 |
{ |
| 136 |
// use input.LA(), not LA(), to get original case |
| 137 |
// CharScanner.LA() would toLower it. |
| 138 |
append(inputState->getInput().LA(1)); |
| 139 |
} |
| 140 |
|
| 141 |
// RK: in a sense I don't like this automatic handling. |
| 142 |
if (c == '\t') |
| 143 |
tab(); |
| 144 |
else |
| 145 |
inputState->column++; |
| 146 |
} |
| 147 |
inputState->getInput().consume(); |
| 148 |
} |
| 149 |
|
| 150 |
/** Consume chars until one matches the given char */ |
| 151 |
virtual void consumeUntil(int c) |
| 152 |
{ |
| 153 |
for(;;) |
| 154 |
{ |
| 155 |
int la_1 = LA(1); |
| 156 |
if( la_1 == EOF_CHAR || la_1 == c ) |
| 157 |
break; |
| 158 |
consume(); |
| 159 |
} |
| 160 |
} |
| 161 |
|
| 162 |
/** Consume chars until one matches the given set */ |
| 163 |
virtual void consumeUntil(const BitSet& set) |
| 164 |
{ |
| 165 |
for(;;) |
| 166 |
{ |
| 167 |
int la_1 = LA(1); |
| 168 |
if( la_1 == EOF_CHAR || set.member(la_1) ) |
| 169 |
break; |
| 170 |
consume(); |
| 171 |
} |
| 172 |
} |
| 173 |
|
| 174 |
/// Mark the current position and return a id for it |
| 175 |
virtual unsigned int mark() |
| 176 |
{ |
| 177 |
return inputState->getInput().mark(); |
| 178 |
} |
| 179 |
/// Rewind the scanner to a previously marked position |
| 180 |
virtual void rewind(unsigned int pos) |
| 181 |
{ |
| 182 |
inputState->getInput().rewind(pos); |
| 183 |
} |
| 184 |
|
| 185 |
/// See if input contains character 'c' throw MismatchedCharException if not |
| 186 |
virtual void match(int c) |
| 187 |
{ |
| 188 |
int la_1 = LA(1); |
| 189 |
if ( la_1 != c ) |
| 190 |
throw MismatchedCharException(la_1, c, false, this); |
| 191 |
consume(); |
| 192 |
} |
| 193 |
|
| 194 |
/** See if input contains element from bitset b |
| 195 |
* throw MismatchedCharException if not |
| 196 |
*/ |
| 197 |
virtual void match(const BitSet& b) |
| 198 |
{ |
| 199 |
int la_1 = LA(1); |
| 200 |
|
| 201 |
if ( !b.member(la_1) ) |
| 202 |
throw MismatchedCharException( la_1, b, false, this ); |
| 203 |
consume(); |
| 204 |
} |
| 205 |
|
| 206 |
/** See if input contains string 's' throw MismatchedCharException if not |
| 207 |
* @note the string cannot match EOF |
| 208 |
*/ |
| 209 |
virtual void match( const char* s ) |
| 210 |
{ |
| 211 |
while( *s != '\0' ) |
| 212 |
{ |
| 213 |
// the & 0xFF is here to prevent sign extension lateron |
| 214 |
int la_1 = LA(1), c = (*s++ & 0xFF); |
| 215 |
|
| 216 |
if ( la_1 != c ) |
| 217 |
throw MismatchedCharException(la_1, c, false, this); |
| 218 |
|
| 219 |
consume(); |
| 220 |
} |
| 221 |
} |
| 222 |
/** See if input contains string 's' throw MismatchedCharException if not |
| 223 |
* @note the string cannot match EOF |
| 224 |
*/ |
| 225 |
virtual void match(const ANTLR_USE_NAMESPACE(std)string& s) |
| 226 |
{ |
| 227 |
size_t len = s.length(); |
| 228 |
|
| 229 |
for (size_t i = 0; i < len; i++) |
| 230 |
{ |
| 231 |
// the & 0xFF is here to prevent sign extension lateron |
| 232 |
int la_1 = LA(1), c = (s[i] & 0xFF); |
| 233 |
|
| 234 |
if ( la_1 != c ) |
| 235 |
throw MismatchedCharException(la_1, c, false, this); |
| 236 |
|
| 237 |
consume(); |
| 238 |
} |
| 239 |
} |
| 240 |
/** See if input does not contain character 'c' |
| 241 |
* throw MismatchedCharException if not |
| 242 |
*/ |
| 243 |
virtual void matchNot(int c) |
| 244 |
{ |
| 245 |
int la_1 = LA(1); |
| 246 |
|
| 247 |
if ( la_1 == c ) |
| 248 |
throw MismatchedCharException(la_1, c, true, this); |
| 249 |
|
| 250 |
consume(); |
| 251 |
} |
| 252 |
/** See if input contains character in range c1-c2 |
| 253 |
* throw MismatchedCharException if not |
| 254 |
*/ |
| 255 |
virtual void matchRange(int c1, int c2) |
| 256 |
{ |
| 257 |
int la_1 = LA(1); |
| 258 |
|
| 259 |
if ( la_1 < c1 || la_1 > c2 ) |
| 260 |
throw MismatchedCharException(la_1, c1, c2, false, this); |
| 261 |
|
| 262 |
consume(); |
| 263 |
} |
| 264 |
|
| 265 |
virtual bool getCaseSensitive() const |
| 266 |
{ |
| 267 |
return caseSensitive; |
| 268 |
} |
| 269 |
|
| 270 |
virtual void setCaseSensitive(bool t) |
| 271 |
{ |
| 272 |
caseSensitive = t; |
| 273 |
} |
| 274 |
|
| 275 |
virtual bool getCaseSensitiveLiterals() const=0; |
| 276 |
|
| 277 |
/// Get the line the scanner currently is in (starts at 1) |
| 278 |
virtual int getLine() const |
| 279 |
{ |
| 280 |
return inputState->line; |
| 281 |
} |
| 282 |
|
| 283 |
/// set the line number |
| 284 |
virtual void setLine(int l) |
| 285 |
{ |
| 286 |
inputState->line = l; |
| 287 |
} |
| 288 |
|
| 289 |
/// Get the column the scanner currently is in (starts at 1) |
| 290 |
virtual int getColumn() const |
| 291 |
{ |
| 292 |
return inputState->column; |
| 293 |
} |
| 294 |
/// set the column number |
| 295 |
virtual void setColumn(int c) |
| 296 |
{ |
| 297 |
inputState->column = c; |
| 298 |
} |
| 299 |
|
| 300 |
/// get the filename for the file currently used |
| 301 |
virtual const ANTLR_USE_NAMESPACE(std)string& getFilename() const |
| 302 |
{ |
| 303 |
return inputState->filename; |
| 304 |
} |
| 305 |
/// Set the filename the scanner is using (used in error messages) |
| 306 |
virtual void setFilename(const ANTLR_USE_NAMESPACE(std)string& f) |
| 307 |
{ |
| 308 |
inputState->filename = f; |
| 309 |
} |
| 310 |
|
| 311 |
virtual bool getCommitToPath() const |
| 312 |
{ |
| 313 |
return commitToPath; |
| 314 |
} |
| 315 |
|
| 316 |
virtual void setCommitToPath(bool commit) |
| 317 |
{ |
| 318 |
commitToPath = commit; |
| 319 |
} |
| 320 |
|
| 321 |
/** return a copy of the current text buffer */ |
| 322 |
virtual const ANTLR_USE_NAMESPACE(std)string& getText() const |
| 323 |
{ |
| 324 |
return text; |
| 325 |
} |
| 326 |
|
| 327 |
virtual void setText(const ANTLR_USE_NAMESPACE(std)string& s) |
| 328 |
{ |
| 329 |
text = s; |
| 330 |
} |
| 331 |
|
| 332 |
virtual void resetText() |
| 333 |
{ |
| 334 |
text = ""; |
| 335 |
inputState->tokenStartColumn = inputState->column; |
| 336 |
inputState->tokenStartLine = inputState->line; |
| 337 |
} |
| 338 |
|
| 339 |
virtual RefToken getTokenObject() const |
| 340 |
{ |
| 341 |
return _returnToken; |
| 342 |
} |
| 343 |
|
| 344 |
/** Used to keep track of line breaks, needs to be called from |
| 345 |
* within generated lexers when a \n \r is encountered. |
| 346 |
*/ |
| 347 |
virtual void newline() |
| 348 |
{ |
| 349 |
++inputState->line; |
| 350 |
inputState->column = 1; |
| 351 |
} |
| 352 |
|
| 353 |
/** Advance the current column number by an appropriate amount according |
| 354 |
* to the tabsize. This method needs to be explicitly called from the |
| 355 |
* lexer rules encountering tabs. |
| 356 |
*/ |
| 357 |
virtual void tab() |
| 358 |
{ |
| 359 |
int c = getColumn(); |
| 360 |
int nc = ( ((c-1)/tabsize) + 1) * tabsize + 1; // calculate tab stop |
| 361 |
setColumn( nc ); |
| 362 |
} |
| 363 |
/// set the tabsize. Returns the old tabsize |
| 364 |
int setTabsize( int size ) |
| 365 |
{ |
| 366 |
int oldsize = tabsize; |
| 367 |
tabsize = size; |
| 368 |
return oldsize; |
| 369 |
} |
| 370 |
/// Return the tabsize used by the scanner |
| 371 |
int getTabSize() const |
| 372 |
{ |
| 373 |
return tabsize; |
| 374 |
} |
| 375 |
|
| 376 |
/** Report exception errors caught in nextToken() */ |
| 377 |
virtual void reportError(const RecognitionException& e); |
| 378 |
|
| 379 |
/** Parser error-reporting function can be overridden in subclass */ |
| 380 |
virtual void reportError(const ANTLR_USE_NAMESPACE(std)string& s); |
| 381 |
|
| 382 |
/** Parser warning-reporting function can be overridden in subclass */ |
| 383 |
virtual void reportWarning(const ANTLR_USE_NAMESPACE(std)string& s); |
| 384 |
|
| 385 |
virtual InputBuffer& getInputBuffer() |
| 386 |
{ |
| 387 |
return inputState->getInput(); |
| 388 |
} |
| 389 |
|
| 390 |
virtual LexerSharedInputState getInputState() |
| 391 |
{ |
| 392 |
return inputState; |
| 393 |
} |
| 394 |
|
| 395 |
/** set the input state for the lexer. |
| 396 |
* @note state is a reference counted object, hence no reference */ |
| 397 |
virtual void setInputState(LexerSharedInputState state) |
| 398 |
{ |
| 399 |
inputState = state; |
| 400 |
} |
| 401 |
|
| 402 |
/// Set the factory for created tokens |
| 403 |
virtual void setTokenObjectFactory(factory_type factory) |
| 404 |
{ |
| 405 |
tokenFactory = factory; |
| 406 |
} |
| 407 |
|
| 408 |
/** Test the token text against the literals table |
| 409 |
* Override this method to perform a different literals test |
| 410 |
*/ |
| 411 |
virtual int testLiteralsTable(int ttype) const |
| 412 |
{ |
| 413 |
ANTLR_USE_NAMESPACE(std)map<ANTLR_USE_NAMESPACE(std)string,int,CharScannerLiteralsLess>::const_iterator i = literals.find(text); |
| 414 |
if (i != literals.end()) |
| 415 |
ttype = (*i).second; |
| 416 |
return ttype; |
| 417 |
} |
| 418 |
|
| 419 |
/** Test the text passed in against the literals table |
| 420 |
* Override this method to perform a different literals test |
| 421 |
* This is used primarily when you want to test a portion of |
| 422 |
* a token |
| 423 |
*/ |
| 424 |
virtual int testLiteralsTable(const ANTLR_USE_NAMESPACE(std)string& txt,int ttype) const |
| 425 |
{ |
| 426 |
ANTLR_USE_NAMESPACE(std)map<ANTLR_USE_NAMESPACE(std)string,int,CharScannerLiteralsLess>::const_iterator i = literals.find(txt); |
| 427 |
if (i != literals.end()) |
| 428 |
ttype = (*i).second; |
| 429 |
return ttype; |
| 430 |
} |
| 431 |
|
| 432 |
/// Override this method to get more specific case handling |
| 433 |
virtual int toLower(int c) const |
| 434 |
{ |
| 435 |
// test on EOF_CHAR for buggy (?) STLPort tolower (or HPUX tolower?) |
| 436 |
// also VC++ 6.0 does this. (see fix 422 (is reverted by this fix) |
| 437 |
// this one is more structural. Maybe make this configurable. |
| 438 |
return (c == EOF_CHAR ? EOF_CHAR : tolower(c)); |
| 439 |
} |
| 440 |
|
| 441 |
/** This method is called by YourLexer::nextToken() when the lexer has |
| 442 |
* hit EOF condition. EOF is NOT a character. |
| 443 |
* This method is not called if EOF is reached during |
| 444 |
* syntactic predicate evaluation or during evaluation |
| 445 |
* of normal lexical rules, which presumably would be |
| 446 |
* an IOException. This traps the "normal" EOF condition. |
| 447 |
* |
| 448 |
* uponEOF() is called after the complete evaluation of |
| 449 |
* the previous token and only if your parser asks |
| 450 |
* for another token beyond that last non-EOF token. |
| 451 |
* |
| 452 |
* You might want to throw token or char stream exceptions |
| 453 |
* like: "Heh, premature eof" or a retry stream exception |
| 454 |
* ("I found the end of this file, go back to referencing file"). |
| 455 |
*/ |
| 456 |
virtual void uponEOF() |
| 457 |
{ |
| 458 |
} |
| 459 |
|
| 460 |
/// Methods used to change tracing behavior |
| 461 |
virtual void traceIndent(); |
| 462 |
virtual void traceIn(const char* rname); |
| 463 |
virtual void traceOut(const char* rname); |
| 464 |
|
| 465 |
#ifndef NO_STATIC_CONSTS |
| 466 |
static const int EOF_CHAR = EOF; |
| 467 |
#else |
| 468 |
enum { |
| 469 |
EOF_CHAR = EOF |
| 470 |
}; |
| 471 |
#endif |
| 472 |
protected: |
| 473 |
ANTLR_USE_NAMESPACE(std)string text; ///< Text of current token |
| 474 |
/// flag indicating wether consume saves characters |
| 475 |
bool saveConsumedInput; |
| 476 |
factory_type tokenFactory; ///< Factory for tokens |
| 477 |
bool caseSensitive; ///< Is this lexer case sensitive |
| 478 |
ANTLR_USE_NAMESPACE(std)map<ANTLR_USE_NAMESPACE(std)string,int,CharScannerLiteralsLess> literals; // set by subclass |
| 479 |
|
| 480 |
RefToken _returnToken; ///< used to return tokens w/o using return val |
| 481 |
|
| 482 |
/// Input state, gives access to input stream, shared among different lexers |
| 483 |
LexerSharedInputState inputState; |
| 484 |
|
| 485 |
/** Used during filter mode to indicate that path is desired. |
| 486 |
* A subsequent scan error will report an error as usual |
| 487 |
* if acceptPath=true; |
| 488 |
*/ |
| 489 |
bool commitToPath; |
| 490 |
|
| 491 |
int tabsize; ///< tab size the scanner uses. |
| 492 |
|
| 493 |
/// Create a new RefToken of type t |
| 494 |
virtual RefToken makeToken(int t) |
| 495 |
{ |
| 496 |
RefToken tok = tokenFactory(); |
| 497 |
tok->setType(t); |
| 498 |
tok->setColumn(inputState->tokenStartColumn); |
| 499 |
tok->setLine(inputState->tokenStartLine); |
| 500 |
return tok; |
| 501 |
} |
| 502 |
|
| 503 |
/** Tracer class, used when -traceLexer is passed to antlr |
| 504 |
*/ |
| 505 |
class Tracer { |
| 506 |
private: |
| 507 |
CharScanner* parser; |
| 508 |
const char* text; |
| 509 |
|
| 510 |
Tracer(const Tracer& other); // undefined |
| 511 |
Tracer& operator=(const Tracer& other); // undefined |
| 512 |
public: |
| 513 |
Tracer( CharScanner* p,const char* t ) |
| 514 |
: parser(p), text(t) |
| 515 |
{ |
| 516 |
parser->traceIn(text); |
| 517 |
} |
| 518 |
~Tracer() |
| 519 |
{ |
| 520 |
parser->traceOut(text); |
| 521 |
} |
| 522 |
}; |
| 523 |
|
| 524 |
int traceDepth; |
| 525 |
private: |
| 526 |
CharScanner( const CharScanner& other ); // undefined |
| 527 |
CharScanner& operator=( const CharScanner& other ); // undefined |
| 528 |
|
| 529 |
#ifndef NO_STATIC_CONSTS |
| 530 |
static const int NO_CHAR = 0; |
| 531 |
#else |
| 532 |
enum { |
| 533 |
NO_CHAR = 0 |
| 534 |
}; |
| 535 |
#endif |
| 536 |
}; |
| 537 |
|
| 538 |
inline int CharScanner::LA(unsigned int i) |
| 539 |
{ |
| 540 |
int c = inputState->getInput().LA(i); |
| 541 |
|
| 542 |
if ( caseSensitive ) |
| 543 |
return c; |
| 544 |
else |
| 545 |
return toLower(c); // VC 6 tolower bug caught in toLower. |
| 546 |
} |
| 547 |
|
| 548 |
inline bool CharScannerLiteralsLess::operator() (const ANTLR_USE_NAMESPACE(std)string& x,const ANTLR_USE_NAMESPACE(std)string& y) const |
| 549 |
{ |
| 550 |
if (scanner->getCaseSensitiveLiterals()) |
| 551 |
return ANTLR_USE_NAMESPACE(std)less<ANTLR_USE_NAMESPACE(std)string>()(x,y); |
| 552 |
else |
| 553 |
{ |
| 554 |
#ifdef NO_STRCASECMP |
| 555 |
return (stricmp(x.c_str(),y.c_str())<0); |
| 556 |
#else |
| 557 |
return (strcasecmp(x.c_str(),y.c_str())<0); |
| 558 |
#endif |
| 559 |
} |
| 560 |
} |
| 561 |
|
| 562 |
#ifdef ANTLR_CXX_SUPPORTS_NAMESPACE |
| 563 |
} |
| 564 |
#endif |
| 565 |
|
| 566 |
#endif //INC_CharScanner_hpp__ |
