Mercurial > projects > ldc
annotate dmd/lexer.c @ 341:1bb99290e03a trunk
[svn r362] Started merging the old 'test' dir as well as the newer 'tangotests' dir into 'tests/mini' and 'tests/minicomplex'.
| author | lindquist |
|---|---|
| date | Sun, 13 Jul 2008 02:51:19 +0200 |
| parents | aaade6ded589 |
| children | 1d3026702f65 |
| rev | line source |
|---|---|
| 159 | 1 |
| 2 // Compiler implementation of the D programming language | |
| 3 // Copyright (c) 1999-2008 by Digital Mars | |
| 4 // All Rights Reserved | |
| 5 // written by Walter Bright | |
| 6 // http://www.digitalmars.com | |
| 7 // License for redistribution is by either the Artistic License | |
| 8 // in artistic.txt, or the GNU General Public License in gnu.txt. | |
| 9 // See the included readme.txt for details. | |
| 10 | |
| 11 /* Lexical Analyzer */ | |
| 12 | |
| 13 #include <stdio.h> | |
| 14 #include <string.h> | |
| 15 #include <ctype.h> | |
| 16 #include <stdarg.h> | |
| 17 #include <errno.h> | |
| 18 #include <wchar.h> | |
| 19 #include <stdlib.h> | |
| 20 #include <assert.h> | |
| 21 #include <sys/time.h> | |
| 22 | |
| 23 #ifdef IN_GCC | |
| 24 | |
| 25 #include <time.h> | |
| 26 #include "mem.h" | |
| 27 | |
| 28 #else | |
| 29 | |
| 30 #if __GNUC__ | |
| 31 #include <time.h> | |
| 32 #endif | |
| 33 | |
| 34 #if IN_LLVM | |
| 35 #include "mem.h" | |
| 36 #elif _WIN32 | |
| 37 #include "..\root\mem.h" | |
| 38 #else | |
| 39 #include "../root/mem.h" | |
| 40 #endif | |
| 41 #endif | |
| 42 | |
| 43 #include "stringtable.h" | |
| 44 | |
| 45 #include "lexer.h" | |
| 46 #include "utf.h" | |
| 47 #include "identifier.h" | |
| 48 #include "id.h" | |
| 49 #include "module.h" | |
| 50 | |
| 51 #if _WIN32 && __DMC__ | |
| 52 // from \dm\src\include\setlocal.h | |
| 53 extern "C" char * __cdecl __locale_decpoint; | |
| 54 #endif | |
| 55 | |
| 56 extern int HtmlNamedEntity(unsigned char *p, int length); | |
| 57 | |
| 58 #define LS 0x2028 // UTF line separator | |
| 59 #define PS 0x2029 // UTF paragraph separator | |
| 60 | |
| 61 /******************************************** | |
| 62 * Do our own char maps | |
| 63 */ | |
| 64 | |
| 65 static unsigned char cmtable[256]; | |
| 66 | |
| 67 const int CMoctal = 0x1; | |
| 68 const int CMhex = 0x2; | |
| 69 const int CMidchar = 0x4; | |
| 70 | |
| 71 inline unsigned char isoctal (unsigned char c) { return cmtable[c] & CMoctal; } | |
| 72 inline unsigned char ishex (unsigned char c) { return cmtable[c] & CMhex; } | |
| 73 inline unsigned char isidchar(unsigned char c) { return cmtable[c] & CMidchar; } | |
| 74 | |
| 75 static void cmtable_init() | |
| 76 { | |
| 77 for (unsigned c = 0; c < sizeof(cmtable) / sizeof(cmtable[0]); c++) | |
| 78 { | |
| 79 if ('0' <= c && c <= '7') | |
| 80 cmtable[c] |= CMoctal; | |
| 81 if (isdigit(c) || ('a' <= c && c <= 'f') || ('A' <= c && c <= 'F')) | |
| 82 cmtable[c] |= CMhex; | |
| 83 if (isalnum(c) || c == '_') | |
| 84 cmtable[c] |= CMidchar; | |
| 85 } | |
| 86 } | |
| 87 | |
| 88 | |
| 89 /************************* Token **********************************************/ | |
| 90 | |
| 91 char *Token::tochars[TOKMAX]; | |
| 92 | |
| 93 void *Token::operator new(size_t size) | |
| 94 { Token *t; | |
| 95 | |
| 96 if (Lexer::freelist) | |
| 97 { | |
| 98 t = Lexer::freelist; | |
| 99 Lexer::freelist = t->next; | |
| 100 return t; | |
| 101 } | |
| 102 | |
| 103 return ::operator new(size); | |
| 104 } | |
| 105 | |
| 106 #ifdef DEBUG | |
| 107 void Token::print() | |
| 108 { | |
| 109 fprintf(stdmsg, "%s\n", toChars()); | |
| 110 } | |
| 111 #endif | |
| 112 | |
| 113 char *Token::toChars() | |
| 114 { char *p; | |
| 115 static char buffer[3 + 3 * sizeof(value) + 1]; | |
| 116 | |
| 117 p = buffer; | |
| 118 switch (value) | |
| 119 { | |
| 120 case TOKint32v: | |
| 121 #if IN_GCC | |
| 122 sprintf(buffer,"%d",(d_int32)int64value); | |
| 123 #else | |
| 124 sprintf(buffer,"%d",int32value); | |
| 125 #endif | |
| 126 break; | |
| 127 | |
| 128 case TOKuns32v: | |
| 129 case TOKcharv: | |
| 130 case TOKwcharv: | |
| 131 case TOKdcharv: | |
| 132 #if IN_GCC | |
| 133 sprintf(buffer,"%uU",(d_uns32)uns64value); | |
| 134 #else | |
| 135 sprintf(buffer,"%uU",uns32value); | |
| 136 #endif | |
| 137 break; | |
| 138 | |
| 139 case TOKint64v: | |
|
305
2b72433d5c8c
[svn r326] Fixed a bunch of issues with printf's that MinGW32 did not support.
lindquist
parents:
159
diff
changeset
|
140 sprintf(buffer,"%lldL",int64value); |
| 159 | 141 break; |
| 142 | |
| 143 case TOKuns64v: | |
|
305
2b72433d5c8c
[svn r326] Fixed a bunch of issues with printf's that MinGW32 did not support.
lindquist
parents:
159
diff
changeset
|
144 sprintf(buffer,"%lluUL",uns64value); |
| 159 | 145 break; |
| 146 | |
| 147 #if IN_GCC | |
| 148 case TOKfloat32v: | |
| 149 case TOKfloat64v: | |
| 150 case TOKfloat80v: | |
| 151 float80value.format(buffer, sizeof(buffer)); | |
| 152 break; | |
| 153 case TOKimaginary32v: | |
| 154 case TOKimaginary64v: | |
| 155 case TOKimaginary80v: | |
| 156 float80value.format(buffer, sizeof(buffer)); | |
| 157 // %% buffer | |
| 158 strcat(buffer, "i"); | |
| 159 break; | |
| 160 #else | |
| 161 case TOKfloat32v: | |
| 162 sprintf(buffer,"%Lgf", float80value); | |
| 163 break; | |
| 164 | |
| 165 case TOKfloat64v: | |
| 166 sprintf(buffer,"%Lg", float80value); | |
| 167 break; | |
| 168 | |
| 169 case TOKfloat80v: | |
| 170 sprintf(buffer,"%LgL", float80value); | |
| 171 break; | |
| 172 | |
| 173 case TOKimaginary32v: | |
| 174 sprintf(buffer,"%Lgfi", float80value); | |
| 175 break; | |
| 176 | |
| 177 case TOKimaginary64v: | |
| 178 sprintf(buffer,"%Lgi", float80value); | |
| 179 break; | |
| 180 | |
| 181 case TOKimaginary80v: | |
| 182 sprintf(buffer,"%LgLi", float80value); | |
| 183 break; | |
| 184 #endif | |
| 185 | |
| 186 case TOKstring: | |
| 187 #if CSTRINGS | |
| 188 p = string; | |
| 189 #else | |
| 190 { OutBuffer buf; | |
| 191 | |
| 192 buf.writeByte('"'); | |
| 193 for (size_t i = 0; i < len; ) | |
| 194 { unsigned c; | |
| 195 | |
| 196 utf_decodeChar((unsigned char *)ustring, len, &i, &c); | |
| 197 switch (c) | |
| 198 { | |
| 199 case 0: | |
| 200 break; | |
| 201 | |
| 202 case '"': | |
| 203 case '\\': | |
| 204 buf.writeByte('\\'); | |
| 205 default: | |
| 206 if (isprint(c)) | |
| 207 buf.writeByte(c); | |
| 208 else if (c <= 0x7F) | |
| 209 buf.printf("\\x%02x", c); | |
| 210 else if (c <= 0xFFFF) | |
| 211 buf.printf("\\u%04x", c); | |
| 212 else | |
| 213 buf.printf("\\U%08x", c); | |
| 214 continue; | |
| 215 } | |
| 216 break; | |
| 217 } | |
| 218 buf.writeByte('"'); | |
| 219 if (postfix) | |
| 220 buf.writeByte('"'); | |
| 221 buf.writeByte(0); | |
| 222 p = (char *)buf.extractData(); | |
| 223 } | |
| 224 #endif | |
| 225 break; | |
| 226 | |
| 227 case TOKidentifier: | |
| 228 case TOKenum: | |
| 229 case TOKstruct: | |
| 230 case TOKimport: | |
| 231 CASE_BASIC_TYPES: | |
| 232 p = ident->toChars(); | |
| 233 break; | |
| 234 | |
| 235 default: | |
| 236 p = toChars(value); | |
| 237 break; | |
| 238 } | |
| 239 return p; | |
| 240 } | |
| 241 | |
| 242 char *Token::toChars(enum TOK value) | |
| 243 { char *p; | |
| 244 static char buffer[3 + 3 * sizeof(value) + 1]; | |
| 245 | |
| 246 p = tochars[value]; | |
| 247 if (!p) | |
| 248 { sprintf(buffer,"TOK%d",value); | |
| 249 p = buffer; | |
| 250 } | |
| 251 return p; | |
| 252 } | |
| 253 | |
| 254 /*************************** Lexer ********************************************/ | |
| 255 | |
| 256 Token *Lexer::freelist = NULL; | |
| 257 StringTable Lexer::stringtable; | |
| 258 OutBuffer Lexer::stringbuffer; | |
| 259 | |
| 260 Lexer::Lexer(Module *mod, | |
| 261 unsigned char *base, unsigned begoffset, unsigned endoffset, | |
| 262 int doDocComment, int commentToken) | |
| 263 : loc(mod, 1) | |
| 264 { | |
| 265 //printf("Lexer::Lexer(%p,%d)\n",base,length); | |
| 266 //printf("lexer.mod = %p, %p\n", mod, this->loc.mod); | |
| 267 memset(&token,0,sizeof(token)); | |
| 268 this->base = base; | |
| 269 this->end = base + endoffset; | |
| 270 p = base + begoffset; | |
| 271 this->mod = mod; | |
| 272 this->doDocComment = doDocComment; | |
| 273 this->anyToken = 0; | |
| 274 this->commentToken = commentToken; | |
| 275 //initKeywords(); | |
| 276 | |
| 277 /* If first line starts with '#!', ignore the line | |
| 278 */ | |
| 279 | |
| 280 if (p[0] == '#' && p[1] =='!') | |
| 281 { | |
| 282 p += 2; | |
| 283 while (1) | |
| 284 { unsigned char c = *p; | |
| 285 switch (c) | |
| 286 { | |
| 287 case '\n': | |
| 288 p++; | |
| 289 break; | |
| 290 | |
| 291 case '\r': | |
| 292 p++; | |
| 293 if (*p == '\n') | |
| 294 p++; | |
| 295 break; | |
| 296 | |
| 297 case 0: | |
| 298 case 0x1A: | |
| 299 break; | |
| 300 | |
| 301 default: | |
| 302 if (c & 0x80) | |
| 303 { unsigned u = decodeUTF(); | |
| 304 if (u == PS || u == LS) | |
| 305 break; | |
| 306 } | |
| 307 p++; | |
| 308 continue; | |
| 309 } | |
| 310 break; | |
| 311 } | |
| 312 loc.linnum = 2; | |
| 313 } | |
| 314 } | |
| 315 | |
| 316 | |
| 317 void Lexer::error(const char *format, ...) | |
| 318 { | |
| 319 if (mod && !global.gag) | |
| 320 { | |
| 321 char *p = loc.toChars(); | |
| 322 if (*p) | |
| 323 fprintf(stdmsg, "%s: ", p); | |
| 324 mem.free(p); | |
| 325 | |
| 326 va_list ap; | |
| 327 va_start(ap, format); | |
| 328 vfprintf(stdmsg, format, ap); | |
| 329 va_end(ap); | |
| 330 | |
| 331 fprintf(stdmsg, "\n"); | |
| 332 fflush(stdmsg); | |
| 333 | |
| 334 if (global.errors >= 20) // moderate blizzard of cascading messages | |
| 335 fatal(); | |
| 336 } | |
| 337 global.errors++; | |
| 338 } | |
| 339 | |
| 340 void Lexer::error(Loc loc, const char *format, ...) | |
| 341 { | |
| 342 if (mod && !global.gag) | |
| 343 { | |
| 344 char *p = loc.toChars(); | |
| 345 if (*p) | |
| 346 fprintf(stdmsg, "%s: ", p); | |
| 347 mem.free(p); | |
| 348 | |
| 349 va_list ap; | |
| 350 va_start(ap, format); | |
| 351 vfprintf(stdmsg, format, ap); | |
| 352 va_end(ap); | |
| 353 | |
| 354 fprintf(stdmsg, "\n"); | |
| 355 fflush(stdmsg); | |
| 356 | |
| 357 if (global.errors >= 20) // moderate blizzard of cascading messages | |
| 358 fatal(); | |
| 359 } | |
| 360 global.errors++; | |
| 361 } | |
| 362 | |
| 363 TOK Lexer::nextToken() | |
| 364 { Token *t; | |
| 365 | |
| 366 if (token.next) | |
| 367 { | |
| 368 t = token.next; | |
| 369 memcpy(&token,t,sizeof(Token)); | |
| 370 t->next = freelist; | |
| 371 freelist = t; | |
| 372 } | |
| 373 else | |
| 374 { | |
| 375 scan(&token); | |
| 376 } | |
| 377 //token.print(); | |
| 378 return token.value; | |
| 379 } | |
| 380 | |
| 381 Token *Lexer::peek(Token *ct) | |
| 382 { Token *t; | |
| 383 | |
| 384 if (ct->next) | |
| 385 t = ct->next; | |
| 386 else | |
| 387 { | |
| 388 t = new Token(); | |
| 389 scan(t); | |
| 390 t->next = NULL; | |
| 391 ct->next = t; | |
| 392 } | |
| 393 return t; | |
| 394 } | |
| 395 | |
| 396 /********************************* | |
| 397 * tk is on the opening (. | |
| 398 * Look ahead and return token that is past the closing ). | |
| 399 */ | |
| 400 | |
| 401 Token *Lexer::peekPastParen(Token *tk) | |
| 402 { | |
| 403 //printf("peekPastParen()\n"); | |
| 404 int parens = 1; | |
| 405 int curlynest = 0; | |
| 406 while (1) | |
| 407 { | |
| 408 tk = peek(tk); | |
| 409 //tk->print(); | |
| 410 switch (tk->value) | |
| 411 { | |
| 412 case TOKlparen: | |
| 413 parens++; | |
| 414 continue; | |
| 415 | |
| 416 case TOKrparen: | |
| 417 --parens; | |
| 418 if (parens) | |
| 419 continue; | |
| 420 tk = peek(tk); | |
| 421 break; | |
| 422 | |
| 423 case TOKlcurly: | |
| 424 curlynest++; | |
| 425 continue; | |
| 426 | |
| 427 case TOKrcurly: | |
| 428 if (--curlynest >= 0) | |
| 429 continue; | |
| 430 break; | |
| 431 | |
| 432 case TOKsemicolon: | |
| 433 if (curlynest) | |
| 434 continue; | |
| 435 break; | |
| 436 | |
| 437 case TOKeof: | |
| 438 break; | |
| 439 | |
| 440 default: | |
| 441 continue; | |
| 442 } | |
| 443 return tk; | |
| 444 } | |
| 445 } | |
| 446 | |
| 447 /********************************** | |
| 448 * Determine if string is a valid Identifier. | |
| 449 * Placed here because of commonality with Lexer functionality. | |
| 450 * Returns: | |
| 451 * 0 invalid | |
| 452 */ | |
| 453 | |
| 454 int Lexer::isValidIdentifier(char *p) | |
| 455 { | |
| 456 size_t len; | |
| 457 size_t idx; | |
| 458 | |
| 459 if (!p || !*p) | |
| 460 goto Linvalid; | |
| 461 | |
| 462 if (*p >= '0' && *p <= '9') // beware of isdigit() on signed chars | |
| 463 goto Linvalid; | |
| 464 | |
| 465 len = strlen(p); | |
| 466 idx = 0; | |
| 467 while (p[idx]) | |
| 468 { dchar_t dc; | |
| 469 | |
| 470 char *q = utf_decodeChar((unsigned char *)p, len, &idx, &dc); | |
| 471 if (q) | |
| 472 goto Linvalid; | |
| 473 | |
| 474 if (!((dc >= 0x80 && isUniAlpha(dc)) || isalnum(dc) || dc == '_')) | |
| 475 goto Linvalid; | |
| 476 } | |
| 477 return 1; | |
| 478 | |
| 479 Linvalid: | |
| 480 return 0; | |
| 481 } | |
| 482 | |
| 483 /**************************** | |
| 484 * Turn next token in buffer into a token. | |
| 485 */ | |
| 486 | |
| 487 void Lexer::scan(Token *t) | |
| 488 { | |
| 489 unsigned lastLine = loc.linnum; | |
| 490 unsigned linnum; | |
| 491 | |
| 492 t->blockComment = NULL; | |
| 493 t->lineComment = NULL; | |
| 494 while (1) | |
| 495 { | |
| 496 t->ptr = p; | |
| 497 //printf("p = %p, *p = '%c'\n",p,*p); | |
| 498 switch (*p) | |
| 499 { | |
| 500 case 0: | |
| 501 case 0x1A: | |
| 502 t->value = TOKeof; // end of file | |
| 503 return; | |
| 504 | |
| 505 case ' ': | |
| 506 case '\t': | |
| 507 case '\v': | |
| 508 case '\f': | |
| 509 p++; | |
| 510 continue; // skip white space | |
| 511 | |
| 512 case '\r': | |
| 513 p++; | |
| 514 if (*p != '\n') // if CR stands by itself | |
| 515 loc.linnum++; | |
| 516 continue; // skip white space | |
| 517 | |
| 518 case '\n': | |
| 519 p++; | |
| 520 loc.linnum++; | |
| 521 continue; // skip white space | |
| 522 | |
| 523 case '0': case '1': case '2': case '3': case '4': | |
| 524 case '5': case '6': case '7': case '8': case '9': | |
| 525 t->value = number(t); | |
| 526 return; | |
| 527 | |
| 528 #if CSTRINGS | |
| 529 case '\'': | |
| 530 t->value = charConstant(t, 0); | |
| 531 return; | |
| 532 | |
| 533 case '"': | |
| 534 t->value = stringConstant(t,0); | |
| 535 return; | |
| 536 | |
| 537 case 'l': | |
| 538 case 'L': | |
| 539 if (p[1] == '\'') | |
| 540 { | |
| 541 p++; | |
| 542 t->value = charConstant(t, 1); | |
| 543 return; | |
| 544 } | |
| 545 else if (p[1] == '"') | |
| 546 { | |
| 547 p++; | |
| 548 t->value = stringConstant(t, 1); | |
| 549 return; | |
| 550 } | |
| 551 #else | |
| 552 case '\'': | |
| 553 t->value = charConstant(t,0); | |
| 554 return; | |
| 555 | |
| 556 case 'r': | |
| 557 if (p[1] != '"') | |
| 558 goto case_ident; | |
| 559 p++; | |
| 560 case '`': | |
| 561 t->value = wysiwygStringConstant(t, *p); | |
| 562 return; | |
| 563 | |
| 564 case 'x': | |
| 565 if (p[1] != '"') | |
| 566 goto case_ident; | |
| 567 p++; | |
| 568 t->value = hexStringConstant(t); | |
| 569 return; | |
| 570 | |
| 336 | 571 #if DMDV2 |
| 159 | 572 case 'q': |
| 573 if (p[1] == '"') | |
| 574 { | |
| 575 p++; | |
| 576 t->value = delimitedStringConstant(t); | |
| 577 return; | |
| 578 } | |
| 579 else if (p[1] == '{') | |
| 580 { | |
| 581 p++; | |
| 582 t->value = tokenStringConstant(t); | |
| 583 return; | |
| 584 } | |
| 585 else | |
| 586 goto case_ident; | |
| 587 #endif | |
| 588 | |
| 589 case '"': | |
| 590 t->value = escapeStringConstant(t,0); | |
| 591 return; | |
| 592 | |
| 593 case '\\': // escaped string literal | |
| 594 { unsigned c; | |
| 595 | |
| 596 stringbuffer.reset(); | |
| 597 do | |
| 598 { | |
| 599 p++; | |
| 600 switch (*p) | |
| 601 { | |
| 602 case 'u': | |
| 603 case 'U': | |
| 604 case '&': | |
| 605 c = escapeSequence(); | |
| 606 stringbuffer.writeUTF8(c); | |
| 607 break; | |
| 608 | |
| 609 default: | |
| 610 c = escapeSequence(); | |
| 611 stringbuffer.writeByte(c); | |
| 612 break; | |
| 613 } | |
| 614 } while (*p == '\\'); | |
| 615 t->len = stringbuffer.offset; | |
| 616 stringbuffer.writeByte(0); | |
| 617 t->ustring = (unsigned char *)mem.malloc(stringbuffer.offset); | |
| 618 memcpy(t->ustring, stringbuffer.data, stringbuffer.offset); | |
| 619 t->postfix = 0; | |
| 620 t->value = TOKstring; | |
| 621 return; | |
| 622 } | |
| 623 | |
| 624 case 'l': | |
| 625 case 'L': | |
| 626 #endif | |
| 627 case 'a': case 'b': case 'c': case 'd': case 'e': | |
| 628 case 'f': case 'g': case 'h': case 'i': case 'j': | |
| 629 case 'k': case 'm': case 'n': case 'o': | |
| 336 | 630 #if DMDV2 |
| 159 | 631 case 'p': /*case 'q': case 'r':*/ case 's': case 't': |
| 632 #else | |
| 633 case 'p': case 'q': /*case 'r':*/ case 's': case 't': | |
| 634 #endif | |
| 635 case 'u': case 'v': case 'w': /*case 'x':*/ case 'y': | |
| 636 case 'z': | |
| 637 case 'A': case 'B': case 'C': case 'D': case 'E': | |
| 638 case 'F': case 'G': case 'H': case 'I': case 'J': | |
| 639 case 'K': case 'M': case 'N': case 'O': | |
| 640 case 'P': case 'Q': case 'R': case 'S': case 'T': | |
| 641 case 'U': case 'V': case 'W': case 'X': case 'Y': | |
| 642 case 'Z': | |
| 643 case '_': | |
| 644 case_ident: | |
| 645 { unsigned char c; | |
| 646 StringValue *sv; | |
| 647 Identifier *id; | |
| 648 | |
| 649 do | |
| 650 { | |
| 651 c = *++p; | |
| 652 } while (isidchar(c) || (c & 0x80 && isUniAlpha(decodeUTF()))); | |
| 653 sv = stringtable.update((char *)t->ptr, p - t->ptr); | |
| 654 id = (Identifier *) sv->ptrvalue; | |
| 655 if (!id) | |
| 656 { id = new Identifier(sv->lstring.string,TOKidentifier); | |
| 657 sv->ptrvalue = id; | |
| 658 } | |
| 659 t->ident = id; | |
| 660 t->value = (enum TOK) id->value; | |
| 661 anyToken = 1; | |
| 662 if (*t->ptr == '_') // if special identifier token | |
| 663 { | |
| 664 static char date[11+1]; | |
| 665 static char time[8+1]; | |
| 666 static char timestamp[24+1]; | |
| 667 | |
| 668 if (!date[0]) // lazy evaluation | |
| 669 { time_t t; | |
| 670 char *p; | |
| 671 | |
| 672 ::time(&t); | |
| 673 p = ctime(&t); | |
| 674 assert(p); | |
| 675 sprintf(date, "%.6s %.4s", p + 4, p + 20); | |
| 676 sprintf(time, "%.8s", p + 11); | |
| 677 sprintf(timestamp, "%.24s", p); | |
| 678 } | |
| 679 | |
| 336 | 680 #if DMDV1 |
| 159 | 681 if (mod && id == Id::FILE) |
| 682 { | |
| 683 t->ustring = (unsigned char *)(loc.filename ? loc.filename : mod->ident->toChars()); | |
| 684 goto Lstring; | |
| 685 } | |
| 686 else if (mod && id == Id::LINE) | |
| 687 { | |
| 688 t->value = TOKint64v; | |
| 689 t->uns64value = loc.linnum; | |
| 690 } | |
| 336 | 691 else |
| 692 #endif | |
| 693 if (id == Id::DATE) | |
| 159 | 694 { |
| 695 t->ustring = (unsigned char *)date; | |
| 696 goto Lstring; | |
| 697 } | |
| 698 else if (id == Id::TIME) | |
| 699 { | |
| 700 t->ustring = (unsigned char *)time; | |
| 701 goto Lstring; | |
| 702 } | |
| 703 else if (id == Id::VENDOR) | |
| 704 { | |
| 705 t->ustring = (unsigned char *)"Digital Mars D"; | |
| 706 goto Lstring; | |
| 707 } | |
| 708 else if (id == Id::TIMESTAMP) | |
| 709 { | |
| 710 t->ustring = (unsigned char *)timestamp; | |
| 711 Lstring: | |
| 712 t->value = TOKstring; | |
| 713 Llen: | |
| 714 t->postfix = 0; | |
| 715 t->len = strlen((char *)t->ustring); | |
| 716 } | |
| 717 else if (id == Id::VERSIONX) | |
| 718 { unsigned major = 0; | |
| 719 unsigned minor = 0; | |
| 720 | |
| 721 for (char *p = global.version + 1; 1; p++) | |
| 722 { | |
| 723 char c = *p; | |
| 724 if (isdigit(c)) | |
| 725 minor = minor * 10 + c - '0'; | |
| 726 else if (c == '.') | |
| 727 { major = minor; | |
| 728 minor = 0; | |
| 729 } | |
| 730 else | |
| 731 break; | |
| 732 } | |
| 733 t->value = TOKint64v; | |
| 734 t->uns64value = major * 1000 + minor; | |
| 735 } | |
| 336 | 736 #if DMDV2 |
| 159 | 737 else if (id == Id::EOFX) |
| 738 { | |
| 739 t->value = TOKeof; | |
| 740 // Advance scanner to end of file | |
| 741 while (!(*p == 0 || *p == 0x1A)) | |
| 742 p++; | |
| 743 } | |
| 744 #endif | |
| 745 } | |
| 746 //printf("t->value = %d\n",t->value); | |
| 747 return; | |
| 748 } | |
| 749 | |
| 750 case '/': | |
| 751 p++; | |
| 752 switch (*p) | |
| 753 { | |
| 754 case '=': | |
| 755 p++; | |
| 756 t->value = TOKdivass; | |
| 757 return; | |
| 758 | |
| 759 case '*': | |
| 760 p++; | |
| 761 linnum = loc.linnum; | |
| 762 while (1) | |
| 763 { | |
| 764 while (1) | |
| 765 { unsigned char c = *p; | |
| 766 switch (c) | |
| 767 { | |
| 768 case '/': | |
| 769 break; | |
| 770 | |
| 771 case '\n': | |
| 772 loc.linnum++; | |
| 773 p++; | |
| 774 continue; | |
| 775 | |
| 776 case '\r': | |
| 777 p++; | |
| 778 if (*p != '\n') | |
| 779 loc.linnum++; | |
| 780 continue; | |
| 781 | |
| 782 case 0: | |
| 783 case 0x1A: | |
| 784 error("unterminated /* */ comment"); | |
| 785 p = end; | |
| 786 t->value = TOKeof; | |
| 787 return; | |
| 788 | |
| 789 default: | |
| 790 if (c & 0x80) | |
| 791 { unsigned u = decodeUTF(); | |
| 792 if (u == PS || u == LS) | |
| 793 loc.linnum++; | |
| 794 } | |
| 795 p++; | |
| 796 continue; | |
| 797 } | |
| 798 break; | |
| 799 } | |
| 800 p++; | |
| 801 if (p[-2] == '*' && p - 3 != t->ptr) | |
| 802 break; | |
| 803 } | |
| 804 if (commentToken) | |
| 805 { | |
| 806 t->value = TOKcomment; | |
| 807 return; | |
| 808 } | |
| 809 else if (doDocComment && t->ptr[2] == '*' && p - 4 != t->ptr) | |
| 810 { // if /** but not /**/ | |
| 811 getDocComment(t, lastLine == linnum); | |
| 812 } | |
| 813 continue; | |
| 814 | |
| 815 case '/': // do // style comments | |
| 816 linnum = loc.linnum; | |
| 817 while (1) | |
| 818 { unsigned char c = *++p; | |
| 819 switch (c) | |
| 820 { | |
| 821 case '\n': | |
| 822 break; | |
| 823 | |
| 824 case '\r': | |
| 825 if (p[1] == '\n') | |
| 826 p++; | |
| 827 break; | |
| 828 | |
| 829 case 0: | |
| 830 case 0x1A: | |
| 831 if (commentToken) | |
| 832 { | |
| 833 p = end; | |
| 834 t->value = TOKcomment; | |
| 835 return; | |
| 836 } | |
| 837 if (doDocComment && t->ptr[2] == '/') | |
| 838 getDocComment(t, lastLine == linnum); | |
| 839 p = end; | |
| 840 t->value = TOKeof; | |
| 841 return; | |
| 842 | |
| 843 default: | |
| 844 if (c & 0x80) | |
| 845 { unsigned u = decodeUTF(); | |
| 846 if (u == PS || u == LS) | |
| 847 break; | |
| 848 } | |
| 849 continue; | |
| 850 } | |
| 851 break; | |
| 852 } | |
| 853 | |
| 854 if (commentToken) | |
| 855 { | |
| 856 p++; | |
| 857 loc.linnum++; | |
| 858 t->value = TOKcomment; | |
| 859 return; | |
| 860 } | |
| 861 if (doDocComment && t->ptr[2] == '/') | |
| 862 getDocComment(t, lastLine == linnum); | |
| 863 | |
| 864 p++; | |
| 865 loc.linnum++; | |
| 866 continue; | |
| 867 | |
| 868 case '+': | |
| 869 { int nest; | |
| 870 | |
| 871 linnum = loc.linnum; | |
| 872 p++; | |
| 873 nest = 1; | |
| 874 while (1) | |
| 875 { unsigned char c = *p; | |
| 876 switch (c) | |
| 877 { | |
| 878 case '/': | |
| 879 p++; | |
| 880 if (*p == '+') | |
| 881 { | |
| 882 p++; | |
| 883 nest++; | |
| 884 } | |
| 885 continue; | |
| 886 | |
| 887 case '+': | |
| 888 p++; | |
| 889 if (*p == '/') | |
| 890 { | |
| 891 p++; | |
| 892 if (--nest == 0) | |
| 893 break; | |
| 894 } | |
| 895 continue; | |
| 896 | |
| 897 case '\r': | |
| 898 p++; | |
| 899 if (*p != '\n') | |
| 900 loc.linnum++; | |
| 901 continue; | |
| 902 | |
| 903 case '\n': | |
| 904 loc.linnum++; | |
| 905 p++; | |
| 906 continue; | |
| 907 | |
| 908 case 0: | |
| 909 case 0x1A: | |
| 910 error("unterminated /+ +/ comment"); | |
| 911 p = end; | |
| 912 t->value = TOKeof; | |
| 913 return; | |
| 914 | |
| 915 default: | |
| 916 if (c & 0x80) | |
| 917 { unsigned u = decodeUTF(); | |
| 918 if (u == PS || u == LS) | |
| 919 loc.linnum++; | |
| 920 } | |
| 921 p++; | |
| 922 continue; | |
| 923 } | |
| 924 break; | |
| 925 } | |
| 926 if (commentToken) | |
| 927 { | |
| 928 t->value = TOKcomment; | |
| 929 return; | |
| 930 } | |
| 931 if (doDocComment && t->ptr[2] == '+' && p - 4 != t->ptr) | |
| 932 { // if /++ but not /++/ | |
| 933 getDocComment(t, lastLine == linnum); | |
| 934 } | |
| 935 continue; | |
| 936 } | |
| 937 } | |
| 938 t->value = TOKdiv; | |
| 939 return; | |
| 940 | |
| 941 case '.': | |
| 942 p++; | |
| 943 if (isdigit(*p)) | |
| 944 { /* Note that we don't allow ._1 and ._ as being | |
| 945 * valid floating point numbers. | |
| 946 */ | |
| 947 p--; | |
| 948 t->value = inreal(t); | |
| 949 } | |
| 950 else if (p[0] == '.') | |
| 951 { | |
| 952 if (p[1] == '.') | |
| 953 { p += 2; | |
| 954 t->value = TOKdotdotdot; | |
| 955 } | |
| 956 else | |
| 957 { p++; | |
| 958 t->value = TOKslice; | |
| 959 } | |
| 960 } | |
| 961 else | |
| 962 t->value = TOKdot; | |
| 963 return; | |
| 964 | |
| 965 case '&': | |
| 966 p++; | |
| 967 if (*p == '=') | |
| 968 { p++; | |
| 969 t->value = TOKandass; | |
| 970 } | |
| 971 else if (*p == '&') | |
| 972 { p++; | |
| 973 t->value = TOKandand; | |
| 974 } | |
| 975 else | |
| 976 t->value = TOKand; | |
| 977 return; | |
| 978 | |
| 979 case '|': | |
| 980 p++; | |
| 981 if (*p == '=') | |
| 982 { p++; | |
| 983 t->value = TOKorass; | |
| 984 } | |
| 985 else if (*p == '|') | |
| 986 { p++; | |
| 987 t->value = TOKoror; | |
| 988 } | |
| 989 else | |
| 990 t->value = TOKor; | |
| 991 return; | |
| 992 | |
| 993 case '-': | |
| 994 p++; | |
| 995 if (*p == '=') | |
| 996 { p++; | |
| 997 t->value = TOKminass; | |
| 998 } | |
| 999 #if 0 | |
| 1000 else if (*p == '>') | |
| 1001 { p++; | |
| 1002 t->value = TOKarrow; | |
| 1003 } | |
| 1004 #endif | |
| 1005 else if (*p == '-') | |
| 1006 { p++; | |
| 1007 t->value = TOKminusminus; | |
| 1008 } | |
| 1009 else | |
| 1010 t->value = TOKmin; | |
| 1011 return; | |
| 1012 | |
| 1013 case '+': | |
| 1014 p++; | |
| 1015 if (*p == '=') | |
| 1016 { p++; | |
| 1017 t->value = TOKaddass; | |
| 1018 } | |
| 1019 else if (*p == '+') | |
| 1020 { p++; | |
| 1021 t->value = TOKplusplus; | |
| 1022 } | |
| 1023 else | |
| 1024 t->value = TOKadd; | |
| 1025 return; | |
| 1026 | |
| 1027 case '<': | |
| 1028 p++; | |
| 1029 if (*p == '=') | |
| 1030 { p++; | |
| 1031 t->value = TOKle; // <= | |
| 1032 } | |
| 1033 else if (*p == '<') | |
| 1034 { p++; | |
| 1035 if (*p == '=') | |
| 1036 { p++; | |
| 1037 t->value = TOKshlass; // <<= | |
| 1038 } | |
| 1039 else | |
| 1040 t->value = TOKshl; // << | |
| 1041 } | |
| 1042 else if (*p == '>') | |
| 1043 { p++; | |
| 1044 if (*p == '=') | |
| 1045 { p++; | |
| 1046 t->value = TOKleg; // <>= | |
| 1047 } | |
| 1048 else | |
| 1049 t->value = TOKlg; // <> | |
| 1050 } | |
| 1051 else | |
| 1052 t->value = TOKlt; // < | |
| 1053 return; | |
| 1054 | |
| 1055 case '>': | |
| 1056 p++; | |
| 1057 if (*p == '=') | |
| 1058 { p++; | |
| 1059 t->value = TOKge; // >= | |
| 1060 } | |
| 1061 else if (*p == '>') | |
| 1062 { p++; | |
| 1063 if (*p == '=') | |
| 1064 { p++; | |
| 1065 t->value = TOKshrass; // >>= | |
| 1066 } | |
| 1067 else if (*p == '>') | |
| 1068 { p++; | |
| 1069 if (*p == '=') | |
| 1070 { p++; | |
| 1071 t->value = TOKushrass; // >>>= | |
| 1072 } | |
| 1073 else | |
| 1074 t->value = TOKushr; // >>> | |
| 1075 } | |
| 1076 else | |
| 1077 t->value = TOKshr; // >> | |
| 1078 } | |
| 1079 else | |
| 1080 t->value = TOKgt; // > | |
| 1081 return; | |
| 1082 | |
| 1083 case '!': | |
| 1084 p++; | |
| 1085 if (*p == '=') | |
| 1086 { p++; | |
| 1087 if (*p == '=' && global.params.Dversion == 1) | |
| 1088 { p++; | |
| 1089 t->value = TOKnotidentity; // !== | |
| 1090 } | |
| 1091 else | |
| 1092 t->value = TOKnotequal; // != | |
| 1093 } | |
| 1094 else if (*p == '<') | |
| 1095 { p++; | |
| 1096 if (*p == '>') | |
| 1097 { p++; | |
| 1098 if (*p == '=') | |
| 1099 { p++; | |
| 1100 t->value = TOKunord; // !<>= | |
| 1101 } | |
| 1102 else | |
| 1103 t->value = TOKue; // !<> | |
| 1104 } | |
| 1105 else if (*p == '=') | |
| 1106 { p++; | |
| 1107 t->value = TOKug; // !<= | |
| 1108 } | |
| 1109 else | |
| 1110 t->value = TOKuge; // !< | |
| 1111 } | |
| 1112 else if (*p == '>') | |
| 1113 { p++; | |
| 1114 if (*p == '=') | |
| 1115 { p++; | |
| 1116 t->value = TOKul; // !>= | |
| 1117 } | |
| 1118 else | |
| 1119 t->value = TOKule; // !> | |
| 1120 } | |
| 1121 else | |
| 1122 t->value = TOKnot; // ! | |
| 1123 return; | |
| 1124 | |
| 1125 case '=': | |
| 1126 p++; | |
| 1127 if (*p == '=') | |
| 1128 { p++; | |
| 1129 if (*p == '=' && global.params.Dversion == 1) | |
| 1130 { p++; | |
| 1131 t->value = TOKidentity; // === | |
| 1132 } | |
| 1133 else | |
| 1134 t->value = TOKequal; // == | |
| 1135 } | |
| 1136 else | |
| 1137 t->value = TOKassign; // = | |
| 1138 return; | |
| 1139 | |
| 1140 case '~': | |
| 1141 p++; | |
| 1142 if (*p == '=') | |
| 1143 { p++; | |
| 1144 t->value = TOKcatass; // ~= | |
| 1145 } | |
| 1146 else | |
| 1147 t->value = TOKtilde; // ~ | |
| 1148 return; | |
| 1149 | |
| 1150 #define SINGLE(c,tok) case c: p++; t->value = tok; return; | |
| 1151 | |
| 1152 SINGLE('(', TOKlparen) | |
| 1153 SINGLE(')', TOKrparen) | |
| 1154 SINGLE('[', TOKlbracket) | |
| 1155 SINGLE(']', TOKrbracket) | |
| 1156 SINGLE('{', TOKlcurly) | |
| 1157 SINGLE('}', TOKrcurly) | |
| 1158 SINGLE('?', TOKquestion) | |
| 1159 SINGLE(',', TOKcomma) | |
| 1160 SINGLE(';', TOKsemicolon) | |
| 1161 SINGLE(':', TOKcolon) | |
| 1162 SINGLE('$', TOKdollar) | |
| 1163 | |
| 1164 #undef SINGLE | |
| 1165 | |
| 1166 #define DOUBLE(c1,tok1,c2,tok2) \ | |
| 1167 case c1: \ | |
| 1168 p++; \ | |
| 1169 if (*p == c2) \ | |
| 1170 { p++; \ | |
| 1171 t->value = tok2; \ | |
| 1172 } \ | |
| 1173 else \ | |
| 1174 t->value = tok1; \ | |
| 1175 return; | |
| 1176 | |
| 1177 DOUBLE('*', TOKmul, '=', TOKmulass) | |
| 1178 DOUBLE('%', TOKmod, '=', TOKmodass) | |
| 1179 DOUBLE('^', TOKxor, '=', TOKxorass) | |
| 1180 | |
| 1181 #undef DOUBLE | |
| 1182 | |
| 1183 case '#': | |
| 1184 p++; | |
| 1185 pragma(); | |
| 1186 continue; | |
| 1187 | |
| 1188 default: | |
| 1189 { unsigned char c = *p; | |
| 1190 | |
| 1191 if (c & 0x80) | |
| 1192 { unsigned u = decodeUTF(); | |
| 1193 | |
| 1194 // Check for start of unicode identifier | |
| 1195 if (isUniAlpha(u)) | |
| 1196 goto case_ident; | |
| 1197 | |
| 1198 if (u == PS || u == LS) | |
| 1199 { | |
| 1200 loc.linnum++; | |
| 1201 p++; | |
| 1202 continue; | |
| 1203 } | |
| 1204 } | |
| 1205 if (isprint(c)) | |
| 1206 error("unsupported char '%c'", c); | |
| 1207 else | |
| 1208 error("unsupported char 0x%02x", c); | |
| 1209 p++; | |
| 1210 continue; | |
| 1211 } | |
| 1212 } | |
| 1213 } | |
| 1214 } | |
| 1215 | |
| 1216 /******************************************* | |
| 1217 * Parse escape sequence. | |
| 1218 */ | |
| 1219 | |
| 1220 unsigned Lexer::escapeSequence() | |
| 1221 { unsigned c; | |
| 1222 int n; | |
| 1223 int ndigits; | |
| 1224 | |
| 1225 c = *p; | |
| 1226 switch (c) | |
| 1227 { | |
| 1228 case '\'': | |
| 1229 case '"': | |
| 1230 case '?': | |
| 1231 case '\\': | |
| 1232 Lconsume: | |
| 1233 p++; | |
| 1234 break; | |
| 1235 | |
| 1236 case 'a': c = 7; goto Lconsume; | |
| 1237 case 'b': c = 8; goto Lconsume; | |
| 1238 case 'f': c = 12; goto Lconsume; | |
| 1239 case 'n': c = 10; goto Lconsume; | |
| 1240 case 'r': c = 13; goto Lconsume; | |
| 1241 case 't': c = 9; goto Lconsume; | |
| 1242 case 'v': c = 11; goto Lconsume; | |
| 1243 | |
| 1244 case 'u': | |
| 1245 ndigits = 4; | |
| 1246 goto Lhex; | |
| 1247 case 'U': | |
| 1248 ndigits = 8; | |
| 1249 goto Lhex; | |
| 1250 case 'x': | |
| 1251 ndigits = 2; | |
| 1252 Lhex: | |
| 1253 p++; | |
| 1254 c = *p; | |
| 1255 if (ishex(c)) | |
| 1256 { unsigned v; | |
| 1257 | |
| 1258 n = 0; | |
| 1259 v = 0; | |
| 1260 while (1) | |
| 1261 { | |
| 1262 if (isdigit(c)) | |
| 1263 c -= '0'; | |
| 1264 else if (islower(c)) | |
| 1265 c -= 'a' - 10; | |
| 1266 else | |
| 1267 c -= 'A' - 10; | |
| 1268 v = v * 16 + c; | |
| 1269 c = *++p; | |
| 1270 if (++n == ndigits) | |
| 1271 break; | |
| 1272 if (!ishex(c)) | |
| 1273 { error("escape hex sequence has %d hex digits instead of %d", n, ndigits); | |
| 1274 break; | |
| 1275 } | |
| 1276 } | |
| 1277 if (ndigits != 2 && !utf_isValidDchar(v)) | |
| 1278 error("invalid UTF character \\U%08x", v); | |
| 1279 c = v; | |
| 1280 } | |
| 1281 else | |
| 1282 error("undefined escape hex sequence \\%c\n",c); | |
| 1283 break; | |
| 1284 | |
| 1285 case '&': // named character entity | |
| 1286 for (unsigned char *idstart = ++p; 1; p++) | |
| 1287 { | |
| 1288 switch (*p) | |
| 1289 { | |
| 1290 case ';': | |
| 1291 c = HtmlNamedEntity(idstart, p - idstart); | |
| 1292 if (c == ~0) | |
| 1293 { error("unnamed character entity &%.*s;", (int)(p - idstart), idstart); | |
| 1294 c = ' '; | |
| 1295 } | |
| 1296 p++; | |
| 1297 break; | |
| 1298 | |
| 1299 default: | |
| 1300 if (isalpha(*p) || | |
| 1301 (p != idstart + 1 && isdigit(*p))) | |
| 1302 continue; | |
| 1303 error("unterminated named entity"); | |
| 1304 break; | |
| 1305 } | |
| 1306 break; | |
| 1307 } | |
| 1308 break; | |
| 1309 | |
| 1310 case 0: | |
| 1311 case 0x1A: // end of file | |
| 1312 c = '\\'; | |
| 1313 break; | |
| 1314 | |
| 1315 default: | |
| 1316 if (isoctal(c)) | |
| 1317 { unsigned v; | |
| 1318 | |
| 1319 n = 0; | |
| 1320 v = 0; | |
| 1321 do | |
| 1322 { | |
| 1323 v = v * 8 + (c - '0'); | |
| 1324 c = *++p; | |
| 1325 } while (++n < 3 && isoctal(c)); | |
| 1326 c = v; | |
| 1327 if (c > 0xFF) | |
| 1328 error("0%03o is larger than a byte", c); | |
| 1329 } | |
| 1330 else | |
| 1331 error("undefined escape sequence \\%c\n",c); | |
| 1332 break; | |
| 1333 } | |
| 1334 return c; | |
| 1335 } | |
| 1336 | |
| 1337 /************************************** | |
| 1338 */ | |
| 1339 | |
| 1340 TOK Lexer::wysiwygStringConstant(Token *t, int tc) | |
| 1341 { unsigned c; | |
| 1342 Loc start = loc; | |
| 1343 | |
| 1344 p++; | |
| 1345 stringbuffer.reset(); | |
| 1346 while (1) | |
| 1347 { | |
| 1348 c = *p++; | |
| 1349 switch (c) | |
| 1350 { | |
| 1351 case '\n': | |
| 1352 loc.linnum++; | |
| 1353 break; | |
| 1354 | |
| 1355 case '\r': | |
| 1356 if (*p == '\n') | |
| 1357 continue; // ignore | |
| 1358 c = '\n'; // treat EndOfLine as \n character | |
| 1359 loc.linnum++; | |
| 1360 break; | |
| 1361 | |
| 1362 case 0: | |
| 1363 case 0x1A: | |
| 1364 error("unterminated string constant starting at %s", start.toChars()); | |
| 1365 t->ustring = (unsigned char *)""; | |
| 1366 t->len = 0; | |
| 1367 t->postfix = 0; | |
| 1368 return TOKstring; | |
| 1369 | |
| 1370 case '"': | |
| 1371 case '`': | |
| 1372 if (c == tc) | |
| 1373 { | |
| 1374 t->len = stringbuffer.offset; | |
| 1375 stringbuffer.writeByte(0); | |
| 1376 t->ustring = (unsigned char *)mem.malloc(stringbuffer.offset); | |
| 1377 memcpy(t->ustring, stringbuffer.data, stringbuffer.offset); | |
| 1378 stringPostfix(t); | |
| 1379 return TOKstring; | |
| 1380 } | |
| 1381 break; | |
| 1382 | |
| 1383 default: | |
| 1384 if (c & 0x80) | |
| 1385 { p--; | |
| 1386 unsigned u = decodeUTF(); | |
| 1387 p++; | |
| 1388 if (u == PS || u == LS) | |
| 1389 loc.linnum++; | |
| 1390 stringbuffer.writeUTF8(u); | |
| 1391 continue; | |
| 1392 } | |
| 1393 break; | |
| 1394 } | |
| 1395 stringbuffer.writeByte(c); | |
| 1396 } | |
| 1397 } | |
| 1398 | |
| 1399 /************************************** | |
| 1400 * Lex hex strings: | |
| 1401 * x"0A ae 34FE BD" | |
| 1402 */ | |
| 1403 | |
| 1404 TOK Lexer::hexStringConstant(Token *t) | |
| 1405 { unsigned c; | |
| 1406 Loc start = loc; | |
| 1407 unsigned n = 0; | |
| 1408 unsigned v; | |
| 1409 | |
| 1410 p++; | |
| 1411 stringbuffer.reset(); | |
| 1412 while (1) | |
| 1413 { | |
| 1414 c = *p++; | |
| 1415 switch (c) | |
| 1416 { | |
| 1417 case ' ': | |
| 1418 case '\t': | |
| 1419 case '\v': | |
| 1420 case '\f': | |
| 1421 continue; // skip white space | |
| 1422 | |
| 1423 case '\r': | |
| 1424 if (*p == '\n') | |
| 1425 continue; // ignore | |
| 1426 // Treat isolated '\r' as if it were a '\n' | |
| 1427 case '\n': | |
| 1428 loc.linnum++; | |
| 1429 continue; | |
| 1430 | |
| 1431 case 0: | |
| 1432 case 0x1A: | |
| 1433 error("unterminated string constant starting at %s", start.toChars()); | |
| 1434 t->ustring = (unsigned char *)""; | |
| 1435 t->len = 0; | |
| 1436 t->postfix = 0; | |
| 1437 return TOKstring; | |
| 1438 | |
| 1439 case '"': | |
| 1440 if (n & 1) | |
| 1441 { error("odd number (%d) of hex characters in hex string", n); | |
| 1442 stringbuffer.writeByte(v); | |
| 1443 } | |
| 1444 t->len = stringbuffer.offset; | |
| 1445 stringbuffer.writeByte(0); | |
| 1446 t->ustring = (unsigned char *)mem.malloc(stringbuffer.offset); | |
| 1447 memcpy(t->ustring, stringbuffer.data, stringbuffer.offset); | |
| 1448 stringPostfix(t); | |
| 1449 return TOKstring; | |
| 1450 | |
| 1451 default: | |
| 1452 if (c >= '0' && c <= '9') | |
| 1453 c -= '0'; | |
| 1454 else if (c >= 'a' && c <= 'f') | |
| 1455 c -= 'a' - 10; | |
| 1456 else if (c >= 'A' && c <= 'F') | |
| 1457 c -= 'A' - 10; | |
| 1458 else if (c & 0x80) | |
| 1459 { p--; | |
| 1460 unsigned u = decodeUTF(); | |
| 1461 p++; | |
| 1462 if (u == PS || u == LS) | |
| 1463 loc.linnum++; | |
| 1464 else | |
| 1465 error("non-hex character \\u%x", u); | |
| 1466 } | |
| 1467 else | |
| 1468 error("non-hex character '%c'", c); | |
| 1469 if (n & 1) | |
| 1470 { v = (v << 4) | c; | |
| 1471 stringbuffer.writeByte(v); | |
| 1472 } | |
| 1473 else | |
| 1474 v = c; | |
| 1475 n++; | |
| 1476 break; | |
| 1477 } | |
| 1478 } | |
| 1479 } | |
| 1480 | |
| 1481 | |
| 336 | 1482 #if DMDV2 |
| 159 | 1483 /************************************** |
| 1484 * Lex delimited strings: | |
| 1485 * q"(foo(xxx))" // "foo(xxx)" | |
| 1486 * q"[foo(]" // "foo(" | |
| 1487 * q"/foo]/" // "foo]" | |
| 1488 * q"HERE | |
| 1489 * foo | |
| 1490 * HERE" // "foo\n" | |
| 1491 * Input: | |
| 1492 * p is on the " | |
| 1493 */ | |
| 1494 | |
| 1495 TOK Lexer::delimitedStringConstant(Token *t) | |
| 1496 { unsigned c; | |
| 1497 Loc start = loc; | |
| 1498 unsigned delimleft = 0; | |
| 1499 unsigned delimright = 0; | |
| 1500 unsigned nest = 1; | |
| 1501 unsigned nestcount; | |
| 1502 Identifier *hereid = NULL; | |
| 1503 unsigned blankrol = 0; | |
| 1504 unsigned startline = 0; | |
| 1505 | |
| 1506 p++; | |
| 1507 stringbuffer.reset(); | |
| 1508 while (1) | |
| 1509 { | |
| 1510 c = *p++; | |
| 1511 //printf("c = '%c'\n", c); | |
| 1512 switch (c) | |
| 1513 { | |
| 1514 case '\n': | |
| 1515 Lnextline: | |
| 1516 loc.linnum++; | |
| 1517 startline = 1; | |
| 1518 if (blankrol) | |
| 1519 { blankrol = 0; | |
| 1520 continue; | |
| 1521 } | |
| 1522 if (hereid) | |
| 1523 { | |
| 1524 stringbuffer.writeUTF8(c); | |
| 1525 continue; | |
| 1526 } | |
| 1527 break; | |
| 1528 | |
| 1529 case '\r': | |
| 1530 if (*p == '\n') | |
| 1531 continue; // ignore | |
| 1532 c = '\n'; // treat EndOfLine as \n character | |
| 1533 goto Lnextline; | |
| 1534 | |
| 1535 case 0: | |
| 1536 case 0x1A: | |
| 1537 goto Lerror; | |
| 1538 | |
| 1539 default: | |
| 1540 if (c & 0x80) | |
| 1541 { p--; | |
| 1542 c = decodeUTF(); | |
| 1543 p++; | |
| 1544 if (c == PS || c == LS) | |
| 1545 goto Lnextline; | |
| 1546 } | |
| 1547 break; | |
| 1548 } | |
| 1549 if (delimleft == 0) | |
| 1550 { delimleft = c; | |
| 1551 nest = 1; | |
| 1552 nestcount = 1; | |
| 1553 if (c == '(') | |
| 1554 delimright = ')'; | |
| 1555 else if (c == '{') | |
| 1556 delimright = '}'; | |
| 1557 else if (c == '[') | |
| 1558 delimright = ']'; | |
| 1559 else if (c == '<') | |
| 1560 delimright = '>'; | |
| 1561 else if (isalpha(c) || c == '_' || (c >= 0x80 && isUniAlpha(c))) | |
| 1562 { // Start of identifier; must be a heredoc | |
| 1563 Token t; | |
| 1564 p--; | |
| 1565 scan(&t); // read in heredoc identifier | |
| 1566 if (t.value != TOKidentifier) | |
| 1567 { error("identifier expected for heredoc, not %s", t.toChars()); | |
| 1568 delimright = c; | |
| 1569 } | |
| 1570 else | |
| 1571 { hereid = t.ident; | |
| 1572 //printf("hereid = '%s'\n", hereid->toChars()); | |
| 1573 blankrol = 1; | |
| 1574 } | |
| 1575 nest = 0; | |
| 1576 } | |
| 1577 else | |
| 1578 { delimright = c; | |
| 1579 nest = 0; | |
| 1580 } | |
| 1581 } | |
| 1582 else | |
| 1583 { | |
| 1584 if (blankrol) | |
| 1585 { error("heredoc rest of line should be blank"); | |
| 1586 blankrol = 0; | |
| 1587 continue; | |
| 1588 } | |
| 1589 if (nest == 1) | |
| 1590 { | |
| 1591 if (c == delimleft) | |
| 1592 nestcount++; | |
| 1593 else if (c == delimright) | |
| 1594 { nestcount--; | |
| 1595 if (nestcount == 0) | |
| 1596 goto Ldone; | |
| 1597 } | |
| 1598 } | |
| 1599 else if (c == delimright) | |
| 1600 goto Ldone; | |
| 1601 if (startline && isalpha(c)) | |
| 1602 { Token t; | |
| 1603 unsigned char *psave = p; | |
| 1604 p--; | |
| 1605 scan(&t); // read in possible heredoc identifier | |
| 1606 //printf("endid = '%s'\n", t.ident->toChars()); | |
| 1607 if (t.value == TOKidentifier && t.ident->equals(hereid)) | |
| 1608 { /* should check that rest of line is blank | |
| 1609 */ | |
| 1610 goto Ldone; | |
| 1611 } | |
| 1612 p = psave; | |
| 1613 } | |
| 1614 stringbuffer.writeUTF8(c); | |
| 1615 startline = 0; | |
| 1616 } | |
| 1617 } | |
| 1618 | |
| 1619 Ldone: | |
| 1620 if (*p == '"') | |
| 1621 p++; | |
| 1622 else | |
| 1623 error("delimited string must end in %c\"", delimright); | |
| 1624 t->len = stringbuffer.offset; | |
| 1625 stringbuffer.writeByte(0); | |
| 1626 t->ustring = (unsigned char *)mem.malloc(stringbuffer.offset); | |
| 1627 memcpy(t->ustring, stringbuffer.data, stringbuffer.offset); | |
| 1628 stringPostfix(t); | |
| 1629 return TOKstring; | |
| 1630 | |
| 1631 Lerror: | |
| 1632 error("unterminated string constant starting at %s", start.toChars()); | |
| 1633 t->ustring = (unsigned char *)""; | |
| 1634 t->len = 0; | |
| 1635 t->postfix = 0; | |
| 1636 return TOKstring; | |
| 1637 } | |
| 1638 | |
| 1639 /************************************** | |
| 1640 * Lex delimited strings: | |
| 1641 * q{ foo(xxx) } // " foo(xxx) " | |
| 1642 * q{foo(} // "foo(" | |
| 1643 * q{{foo}"}"} // "{foo}"}"" | |
| 1644 * Input: | |
| 1645 * p is on the q | |
| 1646 */ | |
| 1647 | |
| 1648 TOK Lexer::tokenStringConstant(Token *t) | |
| 1649 { | |
| 1650 unsigned nest = 1; | |
| 1651 Loc start = loc; | |
| 1652 unsigned char *pstart = ++p; | |
| 1653 | |
| 1654 while (1) | |
| 1655 { Token tok; | |
| 1656 | |
| 1657 scan(&tok); | |
| 1658 switch (tok.value) | |
| 1659 { | |
| 1660 case TOKlcurly: | |
| 1661 nest++; | |
| 1662 continue; | |
| 1663 | |
| 1664 case TOKrcurly: | |
| 1665 if (--nest == 0) | |
| 1666 goto Ldone; | |
| 1667 continue; | |
| 1668 | |
| 1669 case TOKeof: | |
| 1670 goto Lerror; | |
| 1671 | |
| 1672 default: | |
| 1673 continue; | |
| 1674 } | |
| 1675 } | |
| 1676 | |
| 1677 Ldone: | |
| 1678 t->len = p - 1 - pstart; | |
| 1679 t->ustring = (unsigned char *)mem.malloc(t->len + 1); | |
| 1680 memcpy(t->ustring, pstart, t->len); | |
| 1681 t->ustring[t->len] = 0; | |
| 1682 stringPostfix(t); | |
| 1683 return TOKstring; | |
| 1684 | |
| 1685 Lerror: | |
| 1686 error("unterminated token string constant starting at %s", start.toChars()); | |
| 1687 t->ustring = (unsigned char *)""; | |
| 1688 t->len = 0; | |
| 1689 t->postfix = 0; | |
| 1690 return TOKstring; | |
| 1691 } | |
| 1692 | |
| 1693 #endif | |
| 1694 | |
| 1695 | |
| 1696 /************************************** | |
| 1697 */ | |
| 1698 | |
| 1699 TOK Lexer::escapeStringConstant(Token *t, int wide) | |
| 1700 { unsigned c; | |
| 1701 Loc start = loc; | |
| 1702 | |
| 1703 p++; | |
| 1704 stringbuffer.reset(); | |
| 1705 while (1) | |
| 1706 { | |
| 1707 c = *p++; | |
| 1708 switch (c) | |
| 1709 { | |
| 1710 case '\\': | |
| 1711 switch (*p) | |
| 1712 { | |
| 1713 case 'u': | |
| 1714 case 'U': | |
| 1715 case '&': | |
| 1716 c = escapeSequence(); | |
| 1717 stringbuffer.writeUTF8(c); | |
| 1718 continue; | |
| 1719 | |
| 1720 default: | |
| 1721 c = escapeSequence(); | |
| 1722 break; | |
| 1723 } | |
| 1724 break; | |
| 1725 | |
| 1726 case '\n': | |
| 1727 loc.linnum++; | |
| 1728 break; | |
| 1729 | |
| 1730 case '\r': | |
| 1731 if (*p == '\n') | |
| 1732 continue; // ignore | |
| 1733 c = '\n'; // treat EndOfLine as \n character | |
| 1734 loc.linnum++; | |
| 1735 break; | |
| 1736 | |
| 1737 case '"': | |
| 1738 t->len = stringbuffer.offset; | |
| 1739 stringbuffer.writeByte(0); | |
| 1740 t->ustring = (unsigned char *)mem.malloc(stringbuffer.offset); | |
| 1741 memcpy(t->ustring, stringbuffer.data, stringbuffer.offset); | |
| 1742 stringPostfix(t); | |
| 1743 return TOKstring; | |
| 1744 | |
| 1745 case 0: | |
| 1746 case 0x1A: | |
| 1747 p--; | |
| 1748 error("unterminated string constant starting at %s", start.toChars()); | |
| 1749 t->ustring = (unsigned char *)""; | |
| 1750 t->len = 0; | |
| 1751 t->postfix = 0; | |
| 1752 return TOKstring; | |
| 1753 | |
| 1754 default: | |
| 1755 if (c & 0x80) | |
| 1756 { | |
| 1757 p--; | |
| 1758 c = decodeUTF(); | |
| 1759 if (c == LS || c == PS) | |
| 1760 { c = '\n'; | |
| 1761 loc.linnum++; | |
| 1762 } | |
| 1763 p++; | |
| 1764 stringbuffer.writeUTF8(c); | |
| 1765 continue; | |
| 1766 } | |
| 1767 break; | |
| 1768 } | |
| 1769 stringbuffer.writeByte(c); | |
| 1770 } | |
| 1771 } | |
| 1772 | |
| 1773 /************************************** | |
| 1774 */ | |
| 1775 | |
| 1776 TOK Lexer::charConstant(Token *t, int wide) | |
| 1777 { | |
| 1778 unsigned c; | |
| 1779 TOK tk = TOKcharv; | |
| 1780 | |
| 1781 //printf("Lexer::charConstant\n"); | |
| 1782 p++; | |
| 1783 c = *p++; | |
| 1784 switch (c) | |
| 1785 { | |
| 1786 case '\\': | |
| 1787 switch (*p) | |
| 1788 { | |
| 1789 case 'u': | |
| 1790 t->uns64value = escapeSequence(); | |
| 1791 tk = TOKwcharv; | |
| 1792 break; | |
| 1793 | |
| 1794 case 'U': | |
| 1795 case '&': | |
| 1796 t->uns64value = escapeSequence(); | |
| 1797 tk = TOKdcharv; | |
| 1798 break; | |
| 1799 | |
| 1800 default: | |
| 1801 t->uns64value = escapeSequence(); | |
| 1802 break; | |
| 1803 } | |
| 1804 break; | |
| 1805 | |
| 1806 case '\n': | |
| 1807 L1: | |
| 1808 loc.linnum++; | |
| 1809 case '\r': | |
| 1810 case 0: | |
| 1811 case 0x1A: | |
| 1812 case '\'': | |
| 1813 error("unterminated character constant"); | |
| 1814 return tk; | |
| 1815 | |
| 1816 default: | |
| 1817 if (c & 0x80) | |
| 1818 { | |
| 1819 p--; | |
| 1820 c = decodeUTF(); | |
| 1821 p++; | |
| 1822 if (c == LS || c == PS) | |
| 1823 goto L1; | |
| 1824 if (c < 0xD800 || (c >= 0xE000 && c < 0xFFFE)) | |
| 1825 tk = TOKwcharv; | |
| 1826 else | |
| 1827 tk = TOKdcharv; | |
| 1828 } | |
| 1829 t->uns64value = c; | |
| 1830 break; | |
| 1831 } | |
| 1832 | |
| 1833 if (*p != '\'') | |
| 1834 { error("unterminated character constant"); | |
| 1835 return tk; | |
| 1836 } | |
| 1837 p++; | |
| 1838 return tk; | |
| 1839 } | |
| 1840 | |
| 1841 /*************************************** | |
| 1842 * Get postfix of string literal. | |
| 1843 */ | |
| 1844 | |
| 1845 void Lexer::stringPostfix(Token *t) | |
| 1846 { | |
| 1847 switch (*p) | |
| 1848 { | |
| 1849 case 'c': | |
| 1850 case 'w': | |
| 1851 case 'd': | |
| 1852 t->postfix = *p; | |
| 1853 p++; | |
| 1854 break; | |
| 1855 | |
| 1856 default: | |
| 1857 t->postfix = 0; | |
| 1858 break; | |
| 1859 } | |
| 1860 } | |
| 1861 | |
| 1862 /*************************************** | |
| 1863 * Read \u or \U unicode sequence | |
| 1864 * Input: | |
| 1865 * u 'u' or 'U' | |
| 1866 */ | |
| 1867 | |
| 1868 #if 0 | |
| 1869 unsigned Lexer::wchar(unsigned u) | |
| 1870 { | |
| 1871 unsigned value; | |
| 1872 unsigned n; | |
| 1873 unsigned char c; | |
| 1874 unsigned nchars; | |
| 1875 | |
| 1876 nchars = (u == 'U') ? 8 : 4; | |
| 1877 value = 0; | |
| 1878 for (n = 0; 1; n++) | |
| 1879 { | |
| 1880 ++p; | |
| 1881 if (n == nchars) | |
| 1882 break; | |
| 1883 c = *p; | |
| 1884 if (!ishex(c)) | |
| 1885 { error("\\%c sequence must be followed by %d hex characters", u, nchars); | |
| 1886 break; | |
| 1887 } | |
| 1888 if (isdigit(c)) | |
| 1889 c -= '0'; | |
| 1890 else if (islower(c)) | |
| 1891 c -= 'a' - 10; | |
| 1892 else | |
| 1893 c -= 'A' - 10; | |
| 1894 value <<= 4; | |
| 1895 value |= c; | |
| 1896 } | |
| 1897 return value; | |
| 1898 } | |
| 1899 #endif | |
| 1900 | |
| 1901 /************************************** | |
| 1902 * Read in a number. | |
| 1903 * If it's an integer, store it in tok.TKutok.Vlong. | |
| 1904 * integers can be decimal, octal or hex | |
| 1905 * Handle the suffixes U, UL, LU, L, etc. | |
| 1906 * If it's double, store it in tok.TKutok.Vdouble. | |
| 1907 * Returns: | |
| 1908 * TKnum | |
| 1909 * TKdouble,... | |
| 1910 */ | |
| 1911 | |
| 1912 TOK Lexer::number(Token *t) | |
| 1913 { | |
| 1914 // We use a state machine to collect numbers | |
| 1915 enum STATE { STATE_initial, STATE_0, STATE_decimal, STATE_octal, STATE_octale, | |
| 1916 STATE_hex, STATE_binary, STATE_hex0, STATE_binary0, | |
| 1917 STATE_hexh, STATE_error }; | |
| 1918 enum STATE state; | |
| 1919 | |
| 1920 enum FLAGS | |
| 1921 { FLAGS_decimal = 1, // decimal | |
| 1922 FLAGS_unsigned = 2, // u or U suffix | |
| 1923 FLAGS_long = 4, // l or L suffix | |
| 1924 }; | |
| 1925 enum FLAGS flags = FLAGS_decimal; | |
| 1926 | |
| 1927 int i; | |
| 1928 int base; | |
| 1929 unsigned c; | |
| 1930 unsigned char *start; | |
| 1931 TOK result; | |
| 1932 | |
| 1933 //printf("Lexer::number()\n"); | |
| 1934 state = STATE_initial; | |
| 1935 base = 0; | |
| 1936 stringbuffer.reset(); | |
| 1937 start = p; | |
| 1938 while (1) | |
| 1939 { | |
| 1940 c = *p; | |
| 1941 switch (state) | |
| 1942 { | |
| 1943 case STATE_initial: // opening state | |
| 1944 if (c == '0') | |
| 1945 state = STATE_0; | |
| 1946 else | |
| 1947 state = STATE_decimal; | |
| 1948 break; | |
| 1949 | |
| 1950 case STATE_0: | |
| 1951 flags = (FLAGS) (flags & ~FLAGS_decimal); | |
| 1952 switch (c) | |
| 1953 { | |
| 1954 #if ZEROH | |
| 1955 case 'H': // 0h | |
| 1956 case 'h': | |
| 1957 goto hexh; | |
| 1958 #endif | |
| 1959 case 'X': | |
| 1960 case 'x': | |
| 1961 state = STATE_hex0; | |
| 1962 break; | |
| 1963 | |
| 1964 case '.': | |
| 1965 if (p[1] == '.') // .. is a separate token | |
| 1966 goto done; | |
| 1967 case 'i': | |
| 1968 case 'f': | |
| 1969 case 'F': | |
| 1970 goto real; | |
| 1971 #if ZEROH | |
| 1972 case 'E': | |
| 1973 case 'e': | |
| 1974 goto case_hex; | |
| 1975 #endif | |
| 1976 case 'B': | |
| 1977 case 'b': | |
| 1978 state = STATE_binary0; | |
| 1979 break; | |
| 1980 | |
| 1981 case '0': case '1': case '2': case '3': | |
| 1982 case '4': case '5': case '6': case '7': | |
| 1983 state = STATE_octal; | |
| 1984 break; | |
| 1985 | |
| 1986 #if ZEROH | |
| 1987 case '8': case '9': case 'A': | |
| 1988 case 'C': case 'D': case 'F': | |
| 1989 case 'a': case 'c': case 'd': case 'f': | |
| 1990 case_hex: | |
| 1991 state = STATE_hexh; | |
| 1992 break; | |
| 1993 #endif | |
| 1994 case '_': | |
| 1995 state = STATE_octal; | |
| 1996 p++; | |
| 1997 continue; | |
| 1998 | |
| 1999 case 'L': | |
| 2000 if (p[1] == 'i') | |
| 2001 goto real; | |
| 2002 goto done; | |
| 2003 | |
| 2004 default: | |
| 2005 goto done; | |
| 2006 } | |
| 2007 break; | |
| 2008 | |
| 2009 case STATE_decimal: // reading decimal number | |
| 2010 if (!isdigit(c)) | |
| 2011 { | |
| 2012 #if ZEROH | |
| 2013 if (ishex(c) | |
| 2014 || c == 'H' || c == 'h' | |
| 2015 ) | |
| 2016 goto hexh; | |
| 2017 #endif | |
| 2018 if (c == '_') // ignore embedded _ | |
| 2019 { p++; | |
| 2020 continue; | |
| 2021 } | |
| 2022 if (c == '.' && p[1] != '.') | |
| 2023 goto real; | |
| 2024 else if (c == 'i' || c == 'f' || c == 'F' || | |
| 2025 c == 'e' || c == 'E') | |
| 2026 { | |
| 2027 real: // It's a real number. Back up and rescan as a real | |
| 2028 p = start; | |
| 2029 return inreal(t); | |
| 2030 } | |
| 2031 else if (c == 'L' && p[1] == 'i') | |
| 2032 goto real; | |
| 2033 goto done; | |
| 2034 } | |
| 2035 break; | |
| 2036 | |
| 2037 case STATE_hex0: // reading hex number | |
| 2038 case STATE_hex: | |
| 2039 if (!ishex(c)) | |
| 2040 { | |
| 2041 if (c == '_') // ignore embedded _ | |
| 2042 { p++; | |
| 2043 continue; | |
| 2044 } | |
| 2045 if (c == '.' && p[1] != '.') | |
| 2046 goto real; | |
| 2047 if (c == 'P' || c == 'p' || c == 'i') | |
| 2048 goto real; | |
| 2049 if (state == STATE_hex0) | |
| 2050 error("Hex digit expected, not '%c'", c); | |
| 2051 goto done; | |
| 2052 } | |
| 2053 state = STATE_hex; | |
| 2054 break; | |
| 2055 | |
| 2056 #if ZEROH | |
| 2057 hexh: | |
| 2058 state = STATE_hexh; | |
| 2059 case STATE_hexh: // parse numbers like 0FFh | |
| 2060 if (!ishex(c)) | |
| 2061 { | |
| 2062 if (c == 'H' || c == 'h') | |
| 2063 { | |
| 2064 p++; | |
| 2065 base = 16; | |
| 2066 goto done; | |
| 2067 } | |
| 2068 else | |
| 2069 { | |
| 2070 // Check for something like 1E3 or 0E24 | |
| 2071 if (memchr((char *)stringbuffer.data, 'E', stringbuffer.offset) || | |
| 2072 memchr((char *)stringbuffer.data, 'e', stringbuffer.offset)) | |
| 2073 goto real; | |
| 2074 error("Hex digit expected, not '%c'", c); | |
| 2075 goto done; | |
| 2076 } | |
| 2077 } | |
| 2078 break; | |
| 2079 #endif | |
| 2080 | |
| 2081 case STATE_octal: // reading octal number | |
| 2082 case STATE_octale: // reading octal number with non-octal digits | |
| 2083 if (!isoctal(c)) | |
| 2084 { | |
| 2085 #if ZEROH | |
| 2086 if (ishex(c) | |
| 2087 || c == 'H' || c == 'h' | |
| 2088 ) | |
| 2089 goto hexh; | |
| 2090 #endif | |
| 2091 if (c == '_') // ignore embedded _ | |
| 2092 { p++; | |
| 2093 continue; | |
| 2094 } | |
| 2095 if (c == '.' && p[1] != '.') | |
| 2096 goto real; | |
| 2097 if (c == 'i') | |
| 2098 goto real; | |
| 2099 if (isdigit(c)) | |
| 2100 { | |
| 2101 state = STATE_octale; | |
| 2102 } | |
| 2103 else | |
| 2104 goto done; | |
| 2105 } | |
| 2106 break; | |
| 2107 | |
| 2108 case STATE_binary0: // starting binary number | |
| 2109 case STATE_binary: // reading binary number | |
| 2110 if (c != '0' && c != '1') | |
| 2111 { | |
| 2112 #if ZEROH | |
| 2113 if (ishex(c) | |
| 2114 || c == 'H' || c == 'h' | |
| 2115 ) | |
| 2116 goto hexh; | |
| 2117 #endif | |
| 2118 if (c == '_') // ignore embedded _ | |
| 2119 { p++; | |
| 2120 continue; | |
| 2121 } | |
| 2122 if (state == STATE_binary0) | |
| 2123 { error("binary digit expected"); | |
| 2124 state = STATE_error; | |
| 2125 break; | |
| 2126 } | |
| 2127 else | |
| 2128 goto done; | |
| 2129 } | |
| 2130 state = STATE_binary; | |
| 2131 break; | |
| 2132 | |
| 2133 case STATE_error: // for error recovery | |
| 2134 if (!isdigit(c)) // scan until non-digit | |
| 2135 goto done; | |
| 2136 break; | |
| 2137 | |
| 2138 default: | |
| 2139 assert(0); | |
| 2140 } | |
| 2141 stringbuffer.writeByte(c); | |
| 2142 p++; | |
| 2143 } | |
| 2144 done: | |
| 2145 stringbuffer.writeByte(0); // terminate string | |
| 2146 if (state == STATE_octale) | |
| 2147 error("Octal digit expected"); | |
| 2148 | |
| 2149 uinteger_t n; // unsigned >=64 bit integer type | |
| 2150 | |
| 2151 if (stringbuffer.offset == 2 && (state == STATE_decimal || state == STATE_0)) | |
| 2152 n = stringbuffer.data[0] - '0'; | |
| 2153 else | |
| 2154 { | |
| 2155 // Convert string to integer | |
| 2156 #if __DMC__ | |
| 2157 errno = 0; | |
| 2158 n = strtoull((char *)stringbuffer.data,NULL,base); | |
| 2159 if (errno == ERANGE) | |
| 2160 error("integer overflow"); | |
| 2161 #else | |
| 2162 // Not everybody implements strtoull() | |
| 2163 char *p = (char *)stringbuffer.data; | |
| 2164 int r = 10, d; | |
| 2165 | |
| 2166 if (*p == '0') | |
| 2167 { | |
| 2168 if (p[1] == 'x' || p[1] == 'X') | |
| 2169 p += 2, r = 16; | |
| 2170 else if (p[1] == 'b' || p[1] == 'B') | |
| 2171 p += 2, r = 2; | |
| 2172 else if (isdigit(p[1])) | |
| 2173 p += 1, r = 8; | |
| 2174 } | |
| 2175 | |
| 2176 n = 0; | |
| 2177 while (1) | |
| 2178 { | |
| 2179 if (*p >= '0' && *p <= '9') | |
| 2180 d = *p - '0'; | |
| 2181 else if (*p >= 'a' && *p <= 'z') | |
| 2182 d = *p - 'a' + 10; | |
| 2183 else if (*p >= 'A' && *p <= 'Z') | |
| 2184 d = *p - 'A' + 10; | |
| 2185 else | |
| 2186 break; | |
| 2187 if (d >= r) | |
| 2188 break; | |
| 2189 if (n && n * r + d <= n) | |
| 2190 { | |
| 2191 error ("integer overflow"); | |
| 2192 break; | |
| 2193 } | |
| 2194 | |
| 2195 n = n * r + d; | |
| 2196 p++; | |
| 2197 } | |
| 2198 #endif | |
| 2199 if (sizeof(n) > 8 && | |
| 2200 n > 0xFFFFFFFFFFFFFFFFULL) // if n needs more than 64 bits | |
| 2201 error("integer overflow"); | |
| 2202 } | |
| 2203 | |
| 2204 // Parse trailing 'u', 'U', 'l' or 'L' in any combination | |
| 2205 while (1) | |
| 2206 { unsigned char f; | |
| 2207 | |
| 2208 switch (*p) | |
| 2209 { case 'U': | |
| 2210 case 'u': | |
| 2211 f = FLAGS_unsigned; | |
| 2212 goto L1; | |
| 2213 | |
| 2214 case 'l': | |
| 2215 if (1 || !global.params.useDeprecated) | |
| 2216 error("'l' suffix is deprecated, use 'L' instead"); | |
| 2217 case 'L': | |
| 2218 f = FLAGS_long; | |
| 2219 L1: | |
| 2220 p++; | |
| 2221 if (flags & f) | |
| 2222 error("unrecognized token"); | |
| 2223 flags = (FLAGS) (flags | f); | |
| 2224 continue; | |
| 2225 default: | |
| 2226 break; | |
| 2227 } | |
| 2228 break; | |
| 2229 } | |
| 2230 | |
| 2231 switch (flags) | |
| 2232 { | |
| 2233 case 0: | |
| 2234 /* Octal or Hexadecimal constant. | |
| 2235 * First that fits: int, uint, long, ulong | |
| 2236 */ | |
| 2237 if (n & 0x8000000000000000LL) | |
| 2238 result = TOKuns64v; | |
| 2239 else if (n & 0xFFFFFFFF00000000LL) | |
| 2240 result = TOKint64v; | |
| 2241 else if (n & 0x80000000) | |
| 2242 result = TOKuns32v; | |
| 2243 else | |
| 2244 result = TOKint32v; | |
| 2245 break; | |
| 2246 | |
| 2247 case FLAGS_decimal: | |
| 2248 /* First that fits: int, long, long long | |
| 2249 */ | |
| 2250 if (n & 0x8000000000000000LL) | |
| 2251 { error("signed integer overflow"); | |
| 2252 result = TOKuns64v; | |
| 2253 } | |
| 2254 else if (n & 0xFFFFFFFF80000000LL) | |
| 2255 result = TOKint64v; | |
| 2256 else | |
| 2257 result = TOKint32v; | |
| 2258 break; | |
| 2259 | |
| 2260 case FLAGS_unsigned: | |
| 2261 case FLAGS_decimal | FLAGS_unsigned: | |
| 2262 /* First that fits: uint, ulong | |
| 2263 */ | |
| 2264 if (n & 0xFFFFFFFF00000000LL) | |
| 2265 result = TOKuns64v; | |
| 2266 else | |
| 2267 result = TOKuns32v; | |
| 2268 break; | |
| 2269 | |
| 2270 case FLAGS_decimal | FLAGS_long: | |
| 2271 if (n & 0x8000000000000000LL) | |
| 2272 { error("signed integer overflow"); | |
| 2273 result = TOKuns64v; | |
| 2274 } | |
| 2275 else | |
| 2276 result = TOKint64v; | |
| 2277 break; | |
| 2278 | |
| 2279 case FLAGS_long: | |
| 2280 if (n & 0x8000000000000000LL) | |
| 2281 result = TOKuns64v; | |
| 2282 else | |
| 2283 result = TOKint64v; | |
| 2284 break; | |
| 2285 | |
| 2286 case FLAGS_unsigned | FLAGS_long: | |
| 2287 case FLAGS_decimal | FLAGS_unsigned | FLAGS_long: | |
| 2288 result = TOKuns64v; | |
| 2289 break; | |
| 2290 | |
| 2291 default: | |
| 2292 #ifdef DEBUG | |
| 2293 printf("%x\n",flags); | |
| 2294 #endif | |
| 2295 assert(0); | |
| 2296 } | |
| 2297 t->uns64value = n; | |
| 2298 return result; | |
| 2299 } | |
| 2300 | |
| 2301 /************************************** | |
| 2302 * Read in characters, converting them to real. | |
| 2303 * Bugs: | |
| 2304 * Exponent overflow not detected. | |
| 2305 * Too much requested precision is not detected. | |
| 2306 */ | |
| 2307 | |
| 2308 TOK Lexer::inreal(Token *t) | |
| 2309 #ifdef __DMC__ | |
| 2310 __in | |
| 2311 { | |
| 2312 assert(*p == '.' || isdigit(*p)); | |
| 2313 } | |
| 2314 __out (result) | |
| 2315 { | |
| 2316 switch (result) | |
| 2317 { | |
| 2318 case TOKfloat32v: | |
| 2319 case TOKfloat64v: | |
| 2320 case TOKfloat80v: | |
| 2321 case TOKimaginary32v: | |
| 2322 case TOKimaginary64v: | |
| 2323 case TOKimaginary80v: | |
| 2324 break; | |
| 2325 | |
| 2326 default: | |
| 2327 assert(0); | |
| 2328 } | |
| 2329 } | |
| 2330 __body | |
| 2331 #endif /* __DMC__ */ | |
| 2332 { int dblstate; | |
| 2333 unsigned c; | |
| 2334 char hex; // is this a hexadecimal-floating-constant? | |
| 2335 TOK result; | |
| 2336 | |
| 2337 //printf("Lexer::inreal()\n"); | |
| 2338 stringbuffer.reset(); | |
| 2339 dblstate = 0; | |
| 2340 hex = 0; | |
| 2341 Lnext: | |
| 2342 while (1) | |
| 2343 { | |
| 2344 // Get next char from input | |
| 2345 c = *p++; | |
| 2346 //printf("dblstate = %d, c = '%c'\n", dblstate, c); | |
| 2347 while (1) | |
| 2348 { | |
| 2349 switch (dblstate) | |
| 2350 { | |
| 2351 case 0: // opening state | |
| 2352 if (c == '0') | |
| 2353 dblstate = 9; | |
| 2354 else if (c == '.') | |
| 2355 dblstate = 3; | |
| 2356 else | |
| 2357 dblstate = 1; | |
| 2358 break; | |
| 2359 | |
| 2360 case 9: | |
| 2361 dblstate = 1; | |
| 2362 if (c == 'X' || c == 'x') | |
| 2363 { hex++; | |
| 2364 break; | |
| 2365 } | |
| 2366 case 1: // digits to left of . | |
| 2367 case 3: // digits to right of . | |
| 2368 case 7: // continuing exponent digits | |
| 2369 if (!isdigit(c) && !(hex && isxdigit(c))) | |
| 2370 { | |
| 2371 if (c == '_') | |
| 2372 goto Lnext; // ignore embedded '_' | |
| 2373 dblstate++; | |
| 2374 continue; | |
| 2375 } | |
| 2376 break; | |
| 2377 | |
| 2378 case 2: // no more digits to left of . | |
| 2379 if (c == '.') | |
| 2380 { dblstate++; | |
| 2381 break; | |
| 2382 } | |
| 2383 case 4: // no more digits to right of . | |
| 2384 if ((c == 'E' || c == 'e') || | |
| 2385 hex && (c == 'P' || c == 'p')) | |
| 2386 { dblstate = 5; | |
| 2387 hex = 0; // exponent is always decimal | |
| 2388 break; | |
| 2389 } | |
| 2390 if (hex) | |
| 2391 error("binary-exponent-part required"); | |
| 2392 goto done; | |
| 2393 | |
| 2394 case 5: // looking immediately to right of E | |
| 2395 dblstate++; | |
| 2396 if (c == '-' || c == '+') | |
| 2397 break; | |
| 2398 case 6: // 1st exponent digit expected | |
| 2399 if (!isdigit(c)) | |
| 2400 error("exponent expected"); | |
| 2401 dblstate++; | |
| 2402 break; | |
| 2403 | |
| 2404 case 8: // past end of exponent digits | |
| 2405 goto done; | |
| 2406 } | |
| 2407 break; | |
| 2408 } | |
| 2409 stringbuffer.writeByte(c); | |
| 2410 } | |
| 2411 done: | |
| 2412 p--; | |
| 2413 | |
| 2414 stringbuffer.writeByte(0); | |
| 2415 | |
| 2416 #if _WIN32 && __DMC__ | |
| 2417 char *save = __locale_decpoint; | |
| 2418 __locale_decpoint = "."; | |
| 2419 #endif | |
| 2420 #ifdef IN_GCC | |
| 2421 t->float80value = real_t::parse((char *)stringbuffer.data, real_t::LongDouble); | |
| 2422 #else | |
| 2423 t->float80value = strtold((char *)stringbuffer.data, NULL); | |
| 2424 #endif | |
| 2425 errno = 0; | |
| 2426 switch (*p) | |
| 2427 { | |
| 2428 case 'F': | |
| 2429 case 'f': | |
| 2430 #ifdef IN_GCC | |
| 2431 real_t::parse((char *)stringbuffer.data, real_t::Float); | |
| 2432 #else | |
| 2433 strtof((char *)stringbuffer.data, NULL); | |
| 2434 #endif | |
| 2435 result = TOKfloat32v; | |
| 2436 p++; | |
| 2437 break; | |
| 2438 | |
| 2439 default: | |
| 2440 #ifdef IN_GCC | |
| 2441 real_t::parse((char *)stringbuffer.data, real_t::Double); | |
| 2442 #else | |
| 2443 strtod((char *)stringbuffer.data, NULL); | |
| 2444 #endif | |
| 2445 result = TOKfloat64v; | |
| 2446 break; | |
| 2447 | |
| 2448 case 'l': | |
| 2449 if (!global.params.useDeprecated) | |
| 2450 error("'l' suffix is deprecated, use 'L' instead"); | |
| 2451 case 'L': | |
| 2452 result = TOKfloat80v; | |
| 2453 p++; | |
| 2454 break; | |
| 2455 } | |
| 2456 if (*p == 'i' || *p == 'I') | |
| 2457 { | |
| 2458 if (!global.params.useDeprecated && *p == 'I') | |
| 2459 error("'I' suffix is deprecated, use 'i' instead"); | |
| 2460 p++; | |
| 2461 switch (result) | |
| 2462 { | |
| 2463 case TOKfloat32v: | |
| 2464 result = TOKimaginary32v; | |
| 2465 break; | |
| 2466 case TOKfloat64v: | |
| 2467 result = TOKimaginary64v; | |
| 2468 break; | |
| 2469 case TOKfloat80v: | |
| 2470 result = TOKimaginary80v; | |
| 2471 break; | |
| 2472 } | |
| 2473 } | |
| 2474 #if _WIN32 && __DMC__ | |
| 2475 __locale_decpoint = save; | |
| 2476 #endif | |
| 2477 if (errno == ERANGE) | |
| 2478 error("number is not representable"); | |
| 2479 return result; | |
| 2480 } | |
| 2481 | |
| 2482 /********************************************* | |
| 2483 * Do pragma. | |
| 2484 * Currently, the only pragma supported is: | |
| 2485 * #line linnum [filespec] | |
| 2486 */ | |
| 2487 | |
| 2488 void Lexer::pragma() | |
| 2489 { | |
| 2490 Token tok; | |
| 2491 int linnum; | |
| 2492 char *filespec = NULL; | |
| 2493 Loc loc = this->loc; | |
| 2494 | |
| 2495 scan(&tok); | |
| 2496 if (tok.value != TOKidentifier || tok.ident != Id::line) | |
| 2497 goto Lerr; | |
| 2498 | |
| 2499 scan(&tok); | |
| 2500 if (tok.value == TOKint32v || tok.value == TOKint64v) | |
| 2501 linnum = tok.uns64value - 1; | |
| 2502 else | |
| 2503 goto Lerr; | |
| 2504 | |
| 2505 while (1) | |
| 2506 { | |
| 2507 switch (*p) | |
| 2508 { | |
| 2509 case 0: | |
| 2510 case 0x1A: | |
| 2511 case '\n': | |
| 2512 Lnewline: | |
| 2513 this->loc.linnum = linnum; | |
| 2514 if (filespec) | |
| 2515 this->loc.filename = filespec; | |
| 2516 return; | |
| 2517 | |
| 2518 case '\r': | |
| 2519 p++; | |
| 2520 if (*p != '\n') | |
| 2521 { p--; | |
| 2522 goto Lnewline; | |
| 2523 } | |
| 2524 continue; | |
| 2525 | |
| 2526 case ' ': | |
| 2527 case '\t': | |
| 2528 case '\v': | |
| 2529 case '\f': | |
| 2530 p++; | |
| 2531 continue; // skip white space | |
| 2532 | |
| 2533 case '_': | |
| 2534 if (mod && memcmp(p, "__FILE__", 8) == 0) | |
| 2535 { | |
| 2536 p += 8; | |
| 2537 filespec = mem.strdup(loc.filename ? loc.filename : mod->ident->toChars()); | |
| 2538 } | |
| 2539 continue; | |
| 2540 | |
| 2541 case '"': | |
| 2542 if (filespec) | |
| 2543 goto Lerr; | |
| 2544 stringbuffer.reset(); | |
| 2545 p++; | |
| 2546 while (1) | |
| 2547 { unsigned c; | |
| 2548 | |
| 2549 c = *p; | |
| 2550 switch (c) | |
| 2551 { | |
| 2552 case '\n': | |
| 2553 case '\r': | |
| 2554 case 0: | |
| 2555 case 0x1A: | |
| 2556 goto Lerr; | |
| 2557 | |
| 2558 case '"': | |
| 2559 stringbuffer.writeByte(0); | |
| 2560 filespec = mem.strdup((char *)stringbuffer.data); | |
| 2561 p++; | |
| 2562 break; | |
| 2563 | |
| 2564 default: | |
| 2565 if (c & 0x80) | |
| 2566 { unsigned u = decodeUTF(); | |
| 2567 if (u == PS || u == LS) | |
| 2568 goto Lerr; | |
| 2569 } | |
| 2570 stringbuffer.writeByte(c); | |
| 2571 p++; | |
| 2572 continue; | |
| 2573 } | |
| 2574 break; | |
| 2575 } | |
| 2576 continue; | |
| 2577 | |
| 2578 default: | |
| 2579 if (*p & 0x80) | |
| 2580 { unsigned u = decodeUTF(); | |
| 2581 if (u == PS || u == LS) | |
| 2582 goto Lnewline; | |
| 2583 } | |
| 2584 goto Lerr; | |
| 2585 } | |
| 2586 } | |
| 2587 | |
| 2588 Lerr: | |
| 2589 error(loc, "#line integer [\"filespec\"]\\n expected"); | |
| 2590 } | |
| 2591 | |
| 2592 | |
| 2593 /******************************************** | |
| 2594 * Decode UTF character. | |
| 2595 * Issue error messages for invalid sequences. | |
| 2596 * Return decoded character, advance p to last character in UTF sequence. | |
| 2597 */ | |
| 2598 | |
| 2599 unsigned Lexer::decodeUTF() | |
| 2600 { | |
| 2601 dchar_t u; | |
| 2602 unsigned char c; | |
| 2603 unsigned char *s = p; | |
| 2604 size_t len; | |
| 2605 size_t idx; | |
| 2606 char *msg; | |
| 2607 | |
| 2608 c = *s; | |
| 2609 assert(c & 0x80); | |
| 2610 | |
| 2611 // Check length of remaining string up to 6 UTF-8 characters | |
| 2612 for (len = 1; len < 6 && s[len]; len++) | |
| 2613 ; | |
| 2614 | |
| 2615 idx = 0; | |
| 2616 msg = utf_decodeChar(s, len, &idx, &u); | |
| 2617 p += idx - 1; | |
| 2618 if (msg) | |
| 2619 { | |
| 2620 error("%s", msg); | |
| 2621 } | |
| 2622 return u; | |
| 2623 } | |
| 2624 | |
| 2625 | |
| 2626 /*************************************************** | |
| 2627 * Parse doc comment embedded between t->ptr and p. | |
| 2628 * Remove trailing blanks and tabs from lines. | |
| 2629 * Replace all newlines with \n. | |
| 2630 * Remove leading comment character from each line. | |
| 2631 * Decide if it's a lineComment or a blockComment. | |
| 2632 * Append to previous one for this token. | |
| 2633 */ | |
| 2634 | |
| 2635 void Lexer::getDocComment(Token *t, unsigned lineComment) | |
| 2636 { | |
| 2637 OutBuffer buf; | |
| 2638 unsigned char ct = t->ptr[2]; | |
| 2639 unsigned char *q = t->ptr + 3; // start of comment text | |
| 2640 int linestart = 0; | |
| 2641 | |
| 2642 unsigned char *qend = p; | |
| 2643 if (ct == '*' || ct == '+') | |
| 2644 qend -= 2; | |
| 2645 | |
| 2646 /* Scan over initial row of ****'s or ++++'s or ////'s | |
| 2647 */ | |
| 2648 for (; q < qend; q++) | |
| 2649 { | |
| 2650 if (*q != ct) | |
| 2651 break; | |
| 2652 } | |
| 2653 | |
| 2654 /* Remove trailing row of ****'s or ++++'s | |
| 2655 */ | |
| 2656 if (ct != '/') | |
| 2657 { | |
| 2658 for (; q < qend; qend--) | |
| 2659 { | |
| 2660 if (qend[-1] != ct) | |
| 2661 break; | |
| 2662 } | |
| 2663 } | |
| 2664 | |
| 2665 for (; q < qend; q++) | |
| 2666 { | |
| 2667 unsigned char c = *q; | |
| 2668 | |
| 2669 switch (c) | |
| 2670 { | |
| 2671 case '*': | |
| 2672 case '+': | |
| 2673 if (linestart && c == ct) | |
| 2674 { linestart = 0; | |
| 2675 /* Trim preceding whitespace up to preceding \n | |
| 2676 */ | |
| 2677 while (buf.offset && (buf.data[buf.offset - 1] == ' ' || buf.data[buf.offset - 1] == '\t')) | |
| 2678 buf.offset--; | |
| 2679 continue; | |
| 2680 } | |
| 2681 break; | |
| 2682 | |
| 2683 case ' ': | |
| 2684 case '\t': | |
| 2685 break; | |
| 2686 | |
| 2687 case '\r': | |
| 2688 if (q[1] == '\n') | |
| 2689 continue; // skip the \r | |
| 2690 goto Lnewline; | |
| 2691 | |
| 2692 default: | |
| 2693 if (c == 226) | |
| 2694 { | |
| 2695 // If LS or PS | |
| 2696 if (q[1] == 128 && | |
| 2697 (q[2] == 168 || q[2] == 169)) | |
| 2698 { | |
| 2699 q += 2; | |
| 2700 goto Lnewline; | |
| 2701 } | |
| 2702 } | |
| 2703 linestart = 0; | |
| 2704 break; | |
| 2705 | |
| 2706 Lnewline: | |
| 2707 c = '\n'; // replace all newlines with \n | |
| 2708 case '\n': | |
| 2709 linestart = 1; | |
| 2710 | |
| 2711 /* Trim trailing whitespace | |
| 2712 */ | |
| 2713 while (buf.offset && (buf.data[buf.offset - 1] == ' ' || buf.data[buf.offset - 1] == '\t')) | |
| 2714 buf.offset--; | |
| 2715 | |
| 2716 break; | |
| 2717 } | |
| 2718 buf.writeByte(c); | |
| 2719 } | |
| 2720 | |
| 2721 // Always end with a newline | |
| 2722 if (!buf.offset || buf.data[buf.offset - 1] != '\n') | |
| 2723 buf.writeByte('\n'); | |
| 2724 | |
| 2725 buf.writeByte(0); | |
| 2726 | |
| 2727 // It's a line comment if the start of the doc comment comes | |
| 2728 // after other non-whitespace on the same line. | |
| 2729 unsigned char** dc = (lineComment && anyToken) | |
| 2730 ? &t->lineComment | |
| 2731 : &t->blockComment; | |
| 2732 | |
| 2733 // Combine with previous doc comment, if any | |
| 2734 if (*dc) | |
| 2735 *dc = combineComments(*dc, (unsigned char *)buf.data); | |
| 2736 else | |
| 2737 *dc = (unsigned char *)buf.extractData(); | |
| 2738 } | |
| 2739 | |
| 2740 /******************************************** | |
| 2741 * Combine two document comments into one. | |
| 2742 */ | |
| 2743 | |
| 2744 unsigned char *Lexer::combineComments(unsigned char *c1, unsigned char *c2) | |
| 2745 { | |
| 2746 unsigned char *c = c2; | |
| 2747 | |
| 2748 if (c1) | |
| 2749 { c = c1; | |
| 2750 if (c2) | |
| 2751 { size_t len1 = strlen((char *)c1); | |
| 2752 size_t len2 = strlen((char *)c2); | |
| 2753 | |
| 2754 c = (unsigned char *)mem.malloc(len1 + 1 + len2 + 1); | |
| 2755 memcpy(c, c1, len1); | |
| 2756 c[len1] = '\n'; | |
| 2757 memcpy(c + len1 + 1, c2, len2); | |
| 2758 c[len1 + 1 + len2] = 0; | |
| 2759 } | |
| 2760 } | |
| 2761 return c; | |
| 2762 } | |
| 2763 | |
| 2764 /******************************************** | |
| 2765 * Create an identifier in the string table. | |
| 2766 */ | |
| 2767 | |
| 2768 Identifier *Lexer::idPool(const char *s) | |
| 2769 { | |
| 2770 size_t len = strlen(s); | |
| 2771 StringValue *sv = stringtable.update(s, len); | |
| 2772 Identifier *id = (Identifier *) sv->ptrvalue; | |
| 2773 if (!id) | |
| 2774 { | |
| 2775 id = new Identifier(sv->lstring.string, TOKidentifier); | |
| 2776 sv->ptrvalue = id; | |
| 2777 } | |
| 2778 return id; | |
| 2779 } | |
| 2780 | |
| 2781 /********************************************* | |
| 2782 * Create a unique identifier using the prefix s. | |
| 2783 */ | |
| 2784 | |
| 2785 Identifier *Lexer::uniqueId(const char *s, int num) | |
| 2786 { char buffer[32]; | |
| 2787 size_t slen = strlen(s); | |
| 2788 | |
| 2789 assert(slen + sizeof(num) * 3 + 1 <= sizeof(buffer)); | |
| 2790 sprintf(buffer, "%s%d", s, num); | |
| 2791 return idPool(buffer); | |
| 2792 } | |
| 2793 | |
| 2794 Identifier *Lexer::uniqueId(const char *s) | |
| 2795 { | |
| 2796 static int num; | |
| 2797 return uniqueId(s, ++num); | |
| 2798 } | |
| 2799 | |
| 2800 /**************************************** | |
| 2801 */ | |
| 2802 | |
| 2803 struct Keyword | |
| 2804 { char *name; | |
| 2805 enum TOK value; | |
| 2806 }; | |
| 2807 | |
| 2808 static Keyword keywords[] = | |
| 2809 { | |
| 2810 // { "", TOK }, | |
| 2811 | |
| 2812 { "this", TOKthis }, | |
| 2813 { "super", TOKsuper }, | |
| 2814 { "assert", TOKassert }, | |
| 2815 { "null", TOKnull }, | |
| 2816 { "true", TOKtrue }, | |
| 2817 { "false", TOKfalse }, | |
| 2818 { "cast", TOKcast }, | |
| 2819 { "new", TOKnew }, | |
| 2820 { "delete", TOKdelete }, | |
| 2821 { "throw", TOKthrow }, | |
| 2822 { "module", TOKmodule }, | |
| 2823 { "pragma", TOKpragma }, | |
| 2824 { "typeof", TOKtypeof }, | |
| 2825 { "typeid", TOKtypeid }, | |
| 2826 | |
| 2827 { "template", TOKtemplate }, | |
| 2828 | |
| 2829 { "void", TOKvoid }, | |
| 2830 { "byte", TOKint8 }, | |
| 2831 { "ubyte", TOKuns8 }, | |
| 2832 { "short", TOKint16 }, | |
| 2833 { "ushort", TOKuns16 }, | |
| 2834 { "int", TOKint32 }, | |
| 2835 { "uint", TOKuns32 }, | |
| 2836 { "long", TOKint64 }, | |
| 2837 { "ulong", TOKuns64 }, | |
| 2838 { "cent", TOKcent, }, | |
| 2839 { "ucent", TOKucent, }, | |
| 2840 { "float", TOKfloat32 }, | |
| 2841 { "double", TOKfloat64 }, | |
| 2842 { "real", TOKfloat80 }, | |
| 2843 | |
| 2844 { "bool", TOKbool }, | |
| 2845 { "char", TOKchar }, | |
| 2846 { "wchar", TOKwchar }, | |
| 2847 { "dchar", TOKdchar }, | |
| 2848 | |
| 2849 { "ifloat", TOKimaginary32 }, | |
| 2850 { "idouble", TOKimaginary64 }, | |
| 2851 { "ireal", TOKimaginary80 }, | |
| 2852 | |
| 2853 { "cfloat", TOKcomplex32 }, | |
| 2854 { "cdouble", TOKcomplex64 }, | |
| 2855 { "creal", TOKcomplex80 }, | |
| 2856 | |
| 2857 { "delegate", TOKdelegate }, | |
| 2858 { "function", TOKfunction }, | |
| 2859 | |
| 2860 { "is", TOKis }, | |
| 2861 { "if", TOKif }, | |
| 2862 { "else", TOKelse }, | |
| 2863 { "while", TOKwhile }, | |
| 2864 { "for", TOKfor }, | |
| 2865 { "do", TOKdo }, | |
| 2866 { "switch", TOKswitch }, | |
| 2867 { "case", TOKcase }, | |
| 2868 { "default", TOKdefault }, | |
| 2869 { "break", TOKbreak }, | |
| 2870 { "continue", TOKcontinue }, | |
| 2871 { "synchronized", TOKsynchronized }, | |
| 2872 { "return", TOKreturn }, | |
| 2873 { "goto", TOKgoto }, | |
| 2874 { "try", TOKtry }, | |
| 2875 { "catch", TOKcatch }, | |
| 2876 { "finally", TOKfinally }, | |
| 2877 { "with", TOKwith }, | |
| 2878 { "asm", TOKasm }, | |
| 2879 { "foreach", TOKforeach }, | |
| 2880 { "foreach_reverse", TOKforeach_reverse }, | |
| 2881 { "scope", TOKscope }, | |
| 2882 | |
| 2883 { "struct", TOKstruct }, | |
| 2884 { "class", TOKclass }, | |
| 2885 { "interface", TOKinterface }, | |
| 2886 { "union", TOKunion }, | |
| 2887 { "enum", TOKenum }, | |
| 2888 { "import", TOKimport }, | |
| 2889 { "mixin", TOKmixin }, | |
| 2890 { "static", TOKstatic }, | |
| 2891 { "final", TOKfinal }, | |
| 2892 { "const", TOKconst }, | |
| 2893 { "typedef", TOKtypedef }, | |
| 2894 { "alias", TOKalias }, | |
| 2895 { "override", TOKoverride }, | |
| 2896 { "abstract", TOKabstract }, | |
| 2897 { "volatile", TOKvolatile }, | |
| 2898 { "debug", TOKdebug }, | |
| 2899 { "deprecated", TOKdeprecated }, | |
| 2900 { "in", TOKin }, | |
| 2901 { "out", TOKout }, | |
| 2902 { "inout", TOKinout }, | |
| 2903 { "lazy", TOKlazy }, | |
| 2904 { "auto", TOKauto }, | |
| 2905 | |
| 2906 { "align", TOKalign }, | |
| 2907 { "extern", TOKextern }, | |
| 2908 { "private", TOKprivate }, | |
| 2909 { "package", TOKpackage }, | |
| 2910 { "protected", TOKprotected }, | |
| 2911 { "public", TOKpublic }, | |
| 2912 { "export", TOKexport }, | |
| 2913 | |
| 2914 { "body", TOKbody }, | |
| 2915 { "invariant", TOKinvariant }, | |
| 2916 { "unittest", TOKunittest }, | |
| 2917 { "version", TOKversion }, | |
| 2918 //{ "manifest", TOKmanifest }, | |
| 2919 | |
| 2920 // Added after 1.0 | |
| 2921 { "ref", TOKref }, | |
| 2922 { "macro", TOKmacro }, | |
| 336 | 2923 #if DMDV2 |
| 159 | 2924 { "pure", TOKpure }, |
| 2925 { "nothrow", TOKnothrow }, | |
| 336 | 2926 { "__thread", TOKtls }, |
| 159 | 2927 { "__traits", TOKtraits }, |
| 2928 { "__overloadset", TOKoverloadset }, | |
| 336 | 2929 { "__FILE__", TOKfile }, |
| 2930 { "__LINE__", TOKline }, | |
| 159 | 2931 #endif |
| 2932 }; | |
| 2933 | |
| 2934 int Token::isKeyword() | |
| 2935 { | |
| 2936 for (unsigned u = 0; u < sizeof(keywords) / sizeof(keywords[0]); u++) | |
| 2937 { | |
| 2938 if (keywords[u].value == value) | |
| 2939 return 1; | |
| 2940 } | |
| 2941 return 0; | |
| 2942 } | |
| 2943 | |
| 2944 void Lexer::initKeywords() | |
| 2945 { StringValue *sv; | |
| 2946 unsigned u; | |
| 2947 enum TOK v; | |
| 2948 unsigned nkeywords = sizeof(keywords) / sizeof(keywords[0]); | |
| 2949 | |
| 2950 if (global.params.Dversion == 1) | |
| 2951 nkeywords -= 2; | |
| 2952 | |
| 2953 cmtable_init(); | |
| 2954 | |
| 2955 for (u = 0; u < nkeywords; u++) | |
| 2956 { char *s; | |
| 2957 | |
| 2958 //printf("keyword[%d] = '%s'\n",u, keywords[u].name); | |
| 2959 s = keywords[u].name; | |
| 2960 v = keywords[u].value; | |
| 2961 sv = stringtable.insert(s, strlen(s)); | |
| 2962 sv->ptrvalue = (void *) new Identifier(sv->lstring.string,v); | |
| 2963 | |
| 2964 //printf("tochars[%d] = '%s'\n",v, s); | |
| 2965 Token::tochars[v] = s; | |
| 2966 } | |
| 2967 | |
| 2968 Token::tochars[TOKeof] = "EOF"; | |
| 2969 Token::tochars[TOKlcurly] = "{"; | |
| 2970 Token::tochars[TOKrcurly] = "}"; | |
| 2971 Token::tochars[TOKlparen] = "("; | |
| 2972 Token::tochars[TOKrparen] = ")"; | |
| 2973 Token::tochars[TOKlbracket] = "["; | |
| 2974 Token::tochars[TOKrbracket] = "]"; | |
| 2975 Token::tochars[TOKsemicolon] = ";"; | |
| 2976 Token::tochars[TOKcolon] = ":"; | |
| 2977 Token::tochars[TOKcomma] = ","; | |
| 2978 Token::tochars[TOKdot] = "."; | |
| 2979 Token::tochars[TOKxor] = "^"; | |
| 2980 Token::tochars[TOKxorass] = "^="; | |
| 2981 Token::tochars[TOKassign] = "="; | |
| 2982 Token::tochars[TOKconstruct] = "="; | |
| 336 | 2983 #if DMDV2 |
| 159 | 2984 Token::tochars[TOKblit] = "="; |
| 2985 #endif | |
| 2986 Token::tochars[TOKlt] = "<"; | |
| 2987 Token::tochars[TOKgt] = ">"; | |
| 2988 Token::tochars[TOKle] = "<="; | |
| 2989 Token::tochars[TOKge] = ">="; | |
| 2990 Token::tochars[TOKequal] = "=="; | |
| 2991 Token::tochars[TOKnotequal] = "!="; | |
| 2992 Token::tochars[TOKnotidentity] = "!is"; | |
| 2993 Token::tochars[TOKtobool] = "!!"; | |
| 2994 | |
| 2995 Token::tochars[TOKunord] = "!<>="; | |
| 2996 Token::tochars[TOKue] = "!<>"; | |
| 2997 Token::tochars[TOKlg] = "<>"; | |
| 2998 Token::tochars[TOKleg] = "<>="; | |
| 2999 Token::tochars[TOKule] = "!>"; | |
| 3000 Token::tochars[TOKul] = "!>="; | |
| 3001 Token::tochars[TOKuge] = "!<"; | |
| 3002 Token::tochars[TOKug] = "!<="; | |
| 3003 | |
| 3004 Token::tochars[TOKnot] = "!"; | |
| 3005 Token::tochars[TOKtobool] = "!!"; | |
| 3006 Token::tochars[TOKshl] = "<<"; | |
| 3007 Token::tochars[TOKshr] = ">>"; | |
| 3008 Token::tochars[TOKushr] = ">>>"; | |
| 3009 Token::tochars[TOKadd] = "+"; | |
| 3010 Token::tochars[TOKmin] = "-"; | |
| 3011 Token::tochars[TOKmul] = "*"; | |
| 3012 Token::tochars[TOKdiv] = "/"; | |
| 3013 Token::tochars[TOKmod] = "%"; | |
| 3014 Token::tochars[TOKslice] = ".."; | |
| 3015 Token::tochars[TOKdotdotdot] = "..."; | |
| 3016 Token::tochars[TOKand] = "&"; | |
| 3017 Token::tochars[TOKandand] = "&&"; | |
| 3018 Token::tochars[TOKor] = "|"; | |
| 3019 Token::tochars[TOKoror] = "||"; | |
| 3020 Token::tochars[TOKarray] = "[]"; | |
| 3021 Token::tochars[TOKindex] = "[i]"; | |
| 3022 Token::tochars[TOKaddress] = "&"; | |
| 3023 Token::tochars[TOKstar] = "*"; | |
| 3024 Token::tochars[TOKtilde] = "~"; | |
| 3025 Token::tochars[TOKdollar] = "$"; | |
| 3026 Token::tochars[TOKcast] = "cast"; | |
| 3027 Token::tochars[TOKplusplus] = "++"; | |
| 3028 Token::tochars[TOKminusminus] = "--"; | |
| 3029 Token::tochars[TOKtype] = "type"; | |
| 3030 Token::tochars[TOKquestion] = "?"; | |
| 3031 Token::tochars[TOKneg] = "-"; | |
| 3032 Token::tochars[TOKuadd] = "+"; | |
| 3033 Token::tochars[TOKvar] = "var"; | |
| 3034 Token::tochars[TOKaddass] = "+="; | |
| 3035 Token::tochars[TOKminass] = "-="; | |
| 3036 Token::tochars[TOKmulass] = "*="; | |
| 3037 Token::tochars[TOKdivass] = "/="; | |
| 3038 Token::tochars[TOKmodass] = "%="; | |
| 3039 Token::tochars[TOKshlass] = "<<="; | |
| 3040 Token::tochars[TOKshrass] = ">>="; | |
| 3041 Token::tochars[TOKushrass] = ">>>="; | |
| 3042 Token::tochars[TOKandass] = "&="; | |
| 3043 Token::tochars[TOKorass] = "|="; | |
| 3044 Token::tochars[TOKcatass] = "~="; | |
| 3045 Token::tochars[TOKcat] = "~"; | |
| 3046 Token::tochars[TOKcall] = "call"; | |
| 3047 Token::tochars[TOKidentity] = "is"; | |
| 3048 Token::tochars[TOKnotidentity] = "!is"; | |
| 3049 | |
| 3050 Token::tochars[TOKorass] = "|="; | |
| 3051 Token::tochars[TOKidentifier] = "identifier"; | |
| 3052 | |
| 3053 // For debugging | |
| 3054 Token::tochars[TOKdotexp] = "dotexp"; | |
| 3055 Token::tochars[TOKdotti] = "dotti"; | |
| 3056 Token::tochars[TOKdotvar] = "dotvar"; | |
| 3057 Token::tochars[TOKdottype] = "dottype"; | |
| 3058 Token::tochars[TOKsymoff] = "symoff"; | |
| 3059 Token::tochars[TOKtypedot] = "typedot"; | |
| 3060 Token::tochars[TOKarraylength] = "arraylength"; | |
| 3061 Token::tochars[TOKarrayliteral] = "arrayliteral"; | |
| 3062 Token::tochars[TOKassocarrayliteral] = "assocarrayliteral"; | |
| 3063 Token::tochars[TOKstructliteral] = "structliteral"; | |
| 3064 Token::tochars[TOKstring] = "string"; | |
| 3065 Token::tochars[TOKdsymbol] = "symbol"; | |
| 3066 Token::tochars[TOKtuple] = "tuple"; | |
| 3067 Token::tochars[TOKdeclaration] = "declaration"; | |
| 3068 Token::tochars[TOKdottd] = "dottd"; | |
| 3069 Token::tochars[TOKon_scope_exit] = "scope(exit)"; | |
| 3070 } |