dlls/jscript/lex.c

   1 /*
   2  * Copyright 2008 Jacek Caban for CodeWeavers
   3  *
   4  * This library is free software; you can redistribute it and/or
   5  * modify it under the terms of the GNU Lesser General Public
   6  * License as published by the Free Software Foundation; either
   7  * version 2.1 of the License, or (at your option) any later version.
   8  *
   9  * This library is distributed in the hope that it will be useful,
  10  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  11  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  12  * Lesser General Public License for more details.
  13  *
  14  * You should have received a copy of the GNU Lesser General Public
  15  * License along with this library; if not, write to the Free Software
  16  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
  17  */
  18
  19 #include "config.h"
  20 #include "wine/port.h"
  21
  22 #include <limits.h>
  23
  24 #include "jscript.h"
  25 #include "activscp.h"
  26 #include "objsafe.h"
  27 #include "engine.h"
  28
  29 #include "parser.tab.h"
  30
  31 #include "wine/debug.h"
  32 #include "wine/unicode.h"
  33
  34 WINE_DEFAULT_DEBUG_CHANNEL(jscript);
  35
  36 static const WCHAR breakW[] = {'b','r','e','a','k',0};
  37 static const WCHAR caseW[] = {'c','a','s','e',0};
  38 static const WCHAR catchW[] = {'c','a','t','c','h',0};
  39 static const WCHAR continueW[] = {'c','o','n','t','i','n','u','e',0};
  40 static const WCHAR defaultW[] = {'d','e','f','a','u','l','t',0};
  41 static const WCHAR deleteW[] = {'d','e','l','e','t','e',0};
  42 static const WCHAR doW[] = {'d','o',0};
  43 static const WCHAR elseW[] = {'e','l','s','e',0};
  44 static const WCHAR falseW[] = {'f','a','l','s','e',0};
  45 static const WCHAR finallyW[] = {'f','i','n','a','l','l','y',0};
  46 static const WCHAR forW[] = {'f','o','r',0};
  47 static const WCHAR functionW[] = {'f','u','n','c','t','i','o','n',0};
  48 static const WCHAR ifW[] = {'i','f',0};
  49 static const WCHAR inW[] = {'i','n',0};
  50 static const WCHAR instanceofW[] = {'i','n','s','t','a','n','c','e','o','f',0};
  51 static const WCHAR newW[] = {'n','e','w',0};
  52 static const WCHAR nullW[] = {'n','u','l','l',0};
  53 static const WCHAR returnW[] = {'r','e','t','u','r','n',0};
  54 static const WCHAR switchW[] = {'s','w','i','t','c','h',0};
  55 static const WCHAR thisW[] = {'t','h','i','s',0};
  56 static const WCHAR throwW[] = {'t','h','r','o','w',0};
  57 static const WCHAR trueW[] = {'t','r','u','e',0};
  58 static const WCHAR tryW[] = {'t','r','y',0};
  59 static const WCHAR typeofW[] = {'t','y','p','e','o','f',0};
  60 static const WCHAR varW[] = {'v','a','r',0};
  61 static const WCHAR voidW[] = {'v','o','i','d',0};
  62 static const WCHAR whileW[] = {'w','h','i','l','e',0};
  63 static const WCHAR withW[] = {'w','i','t','h',0};
  64
  65 static const struct {
  66     const WCHAR *word;
  67     int token;
  68     BOOL no_nl;
  69 } keywords[] = {
  70     {breakW,       kBREAK, TRUE},
  71     {caseW,        kCASE},
  72     {catchW,       kCATCH},
  73     {continueW,    kCONTINUE, TRUE},
  74     {defaultW,     kDEFAULT},
  75     {deleteW,      kDELETE},
  76     {doW,          kDO},
  77     {elseW,        kELSE},
  78     {falseW,       kFALSE},
  79     {finallyW,     kFINALLY},
  80     {forW,         kFOR},
  81     {functionW,    kFUNCTION},
  82     {ifW,          kIF},
  83     {inW,          kIN},
  84     {instanceofW,  kINSTANCEOF},
  85     {newW,         kNEW},
  86     {nullW,        kNULL},
  87     {returnW,      kRETURN, TRUE},
  88     {switchW,      kSWITCH},
  89     {thisW,        kTHIS},
  90     {throwW,       kTHROW},
  91     {trueW,        kTRUE},
  92     {tryW,         kTRY},
  93     {typeofW,      kTYPEOF},
  94     {varW,         kVAR},
  95     {voidW,        kVOID},
  96     {whileW,       kWHILE},
  97     {withW,        kWITH}
  98 };
  99
 100 static int lex_error(parser_ctx_t *ctx, HRESULT hres)
 101 {
 102     ctx->hres = hres;
 103     ctx->lexer_error = TRUE;
 104     return -1;
 105 }
 106
 107 /* ECMA-262 3rd Edition    7.6 */
 108 static BOOL is_identifier_char(WCHAR c)
 109 {
 110     return isalnumW(c) || c == '$' || c == '_' || c == '\\';
 111 }
 112
 113 static int check_keyword(parser_ctx_t *ctx, const WCHAR *word, const WCHAR **lval)
 114 {
 115     const WCHAR *p1 = ctx->ptr;
 116     const WCHAR *p2 = word;
 117
 118     while(p1 < ctx->end && *p2) {
 119         if(*p1 != *p2)
 120             return *p1 - *p2;
 121         p1++;
 122         p2++;
 123     }
 124
 125     if(*p2 || (p1 < ctx->end && is_identifier_char(*p1)))
 126         return 1;
 127
 128     if(lval)
 129         *lval = ctx->ptr;
 130     ctx->ptr = p1;
 131     return 0;
 132 }
 133
 134 /* ECMA-262 3rd Edition    7.3 */
 135 static BOOL is_endline(WCHAR c)
 136 {
 137     return c == '\n' || c == '\r' || c == 0x2028 || c == 0x2029;
 138 }
 139
 140 static int hex_to_int(WCHAR c)
 141 {
 142     if('0' <= c && c <= '9')
 143         return c-'0';
 144
 145     if('a' <= c && c <= 'f')
 146         return c-'a'+10;
 147
 148     if('A' <= c && c <= 'F')
 149         return c-'A'+10;
 150
 151     return -1;
 152 }
 153
 154 static int check_keywords(parser_ctx_t *ctx, const WCHAR **lval)
 155 {
 156     int min = 0, max = sizeof(keywords)/sizeof(keywords[0])-1, r, i;
 157
 158     while(min <= max) {
 159         i = (min+max)/2;
 160
 161         r = check_keyword(ctx, keywords[i].word, lval);
 162         if(!r) {
 163             ctx->implicit_nl_semicolon = keywords[i].no_nl;
 164             return keywords[i].token;
 165         }
 166
 167         if(r > 0)
 168             min = i+1;
 169         else
 170             max = i-1;
 171     }
 172
 173     return 0;
 174 }
 175
 176 static BOOL skip_html_comment(parser_ctx_t *ctx)
 177 {
 178     const WCHAR html_commentW[] = {'<','!','-','-',0};
 179
 180     if(!ctx->is_html || ctx->ptr+3 >= ctx->end ||
 181         memcmp(ctx->ptr, html_commentW, sizeof(WCHAR)*4))
 182         return FALSE;
 183
 184     ctx->nl = TRUE;
 185     while(ctx->ptr < ctx->end && !is_endline(*ctx->ptr++));
 186
 187     return TRUE;
 188 }
 189
 190 static BOOL skip_comment(parser_ctx_t *ctx)
 191 {
 192     if(ctx->ptr+1 >= ctx->end)
 193         return FALSE;
 194
 195     if(*ctx->ptr != '/') {
 196         if(*ctx->ptr == '@' && ctx->ptr+2 < ctx->end && ctx->ptr[1] == '*' && ctx->ptr[2] == '/') {
 197             ctx->ptr += 3;
 198             return TRUE;
 199         }
 200
 201         return FALSE;
 202     }
 203
 204     switch(ctx->ptr[1]) {
 205     case '*':
 206         ctx->ptr += 2;
 207         if(ctx->ptr+2 < ctx->end && *ctx->ptr == '@' && is_identifier_char(ctx->ptr[1]))
 208             return FALSE;
 209         while(ctx->ptr+1 < ctx->end && (ctx->ptr[0] != '*' || ctx->ptr[1] != '/'))
 210             ctx->ptr++;
 211
 212         if(ctx->ptr[0] == '*' && ctx->ptr[1] == '/') {
 213             ctx->ptr += 2;
 214         }else {
 215             WARN("unexpected end of file (missing end of comment)\n");
 216             ctx->ptr = ctx->end;
 217         }
 218         break;
 219     case '/':
 220         ctx->ptr += 2;
 221         if(ctx->ptr+2 < ctx->end && *ctx->ptr == '@' && is_identifier_char(ctx->ptr[1]))
 222             return FALSE;
 223         while(ctx->ptr < ctx->end && !is_endline(*ctx->ptr))
 224             ctx->ptr++;
 225         break;
 226     default:
 227         return FALSE;
 228     }
 229
 230     return TRUE;
 231 }
 232
 233 static BOOL unescape(WCHAR *str)
 234 {
 235     WCHAR *pd, *p, c;
 236     int i;
 237
 238     pd = p = str;
 239     while(*p) {
 240         if(*p != '\\') {
 241             *pd++ = *p++;
 242             continue;
 243         }
 244
 245         p++;
 246
 247         switch(*p) {
 248         case '\'':
 249         case '\"':
 250         case '\\':
 251             c = *p;
 252             break;
 253         case 'b':
 254             c = '\b';
 255             break;
 256         case 't':
 257             c = '\t';
 258             break;
 259         case 'n':
 260             c = '\n';
 261             break;
 262         case 'f':
 263             c = '\f';
 264             break;
 265         case 'r':
 266             c = '\r';
 267             break;
 268         case 'x':
 269             i = hex_to_int(*++p);
 270             if(i == -1)
 271                 return FALSE;
 272             c = i << 4;
 273
 274             i = hex_to_int(*++p);
 275             if(i == -1)
 276                 return FALSE;
 277             c += i;
 278             break;
 279         case 'u':
 280             i = hex_to_int(*++p);
 281             if(i == -1)
 282                 return FALSE;
 283             c = i << 12;
 284
 285             i = hex_to_int(*++p);
 286             if(i == -1)
 287                 return FALSE;
 288             c += i << 8;
 289
 290             i = hex_to_int(*++p);
 291             if(i == -1)
 292                 return FALSE;
 293             c += i << 4;
 294
 295             i = hex_to_int(*++p);
 296             if(i == -1)
 297                 return FALSE;
 298             c += i;
 299             break;
 300         default:
 301             if(isdigitW(*p)) {
 302                 c = *p++ - '0';
 303                 if(isdigitW(*p)) {
 304                     c = c*8 + (*p++ - '0');
 305                     if(isdigitW(*p))
 306                         c = c*8 + (*p++ - '0');
 307                 }
 308                 p--;
 309             }
 310             else
 311                 c = *p;
 312         }
 313
 314         *pd++ = c;
 315         p++;
 316     }
 317
 318     *pd = 0;
 319     return TRUE;
 320 }
 321
 322 static int parse_identifier(parser_ctx_t *ctx, const WCHAR **ret)
 323 {
 324     const WCHAR *ptr = ctx->ptr++;
 325     WCHAR *wstr;
 326     int len;
 327
 328     while(ctx->ptr < ctx->end && is_identifier_char(*ctx->ptr))
 329         ctx->ptr++;
 330
 331     len = ctx->ptr-ptr;
 332
 333     *ret = wstr = parser_alloc(ctx, (len+1)*sizeof(WCHAR));
 334     memcpy(wstr, ptr, len*sizeof(WCHAR));
 335     wstr[len] = 0;
 336
 337     /* FIXME: unescape */
 338     return tIdentifier;
 339 }
 340
 341 static int parse_string_literal(parser_ctx_t *ctx, const WCHAR **ret, WCHAR endch)
 342 {
 343     const WCHAR *ptr = ++ctx->ptr;
 344     WCHAR *wstr;
 345     int len;
 346
 347     while(ctx->ptr < ctx->end && *ctx->ptr != endch) {
 348         if(*ctx->ptr++ == '\\')
 349             ctx->ptr++;
 350     }
 351
 352     if(ctx->ptr == ctx->end)
 353         return lex_error(ctx, JS_E_UNTERMINATED_STRING);
 354
 355     len = ctx->ptr-ptr;
 356
 357     *ret = wstr = parser_alloc(ctx, (len+1)*sizeof(WCHAR));
 358     memcpy(wstr, ptr, len*sizeof(WCHAR));
 359     wstr[len] = 0;
 360
 361     ctx->ptr++;
 362
 363     if(!unescape(wstr)) {
 364         WARN("unescape failed\n");
 365         return lex_error(ctx, E_FAIL);
 366     }
 367
 368     return tStringLiteral;
 369 }
 370
 371 static literal_t *new_double_literal(parser_ctx_t *ctx, DOUBLE d)
 372 {
 373     literal_t *ret = parser_alloc(ctx, sizeof(literal_t));
 374
 375     ret->type = LT_DOUBLE;
 376     ret->u.dval = d;
 377     return ret;
 378 }
 379
 380 literal_t *new_boolean_literal(parser_ctx_t *ctx, BOOL bval)
 381 {
 382     literal_t *ret = parser_alloc(ctx, sizeof(literal_t));
 383
 384     ret->type = LT_BOOL;
 385     ret->u.bval = bval;
 386
 387     return ret;
 388 }
 389
 390 static int parse_double_literal(parser_ctx_t *ctx, LONG int_part, literal_t **literal)
 391 {
 392     LONGLONG d, hlp;
 393     int exp = 0;
 394
 395     d = int_part;
 396     while(ctx->ptr < ctx->end && isdigitW(*ctx->ptr)) {
 397         hlp = d*10 + *(ctx->ptr++) - '0';
 398         if(d>MAXLONGLONG/10 || hlp<0) {
 399             exp++;
 400             break;
 401         }
 402         else
 403             d = hlp;
 404     }
 405     while(ctx->ptr < ctx->end && isdigitW(*ctx->ptr)) {
 406         exp++;
 407         ctx->ptr++;
 408     }
 409
 410     if(*ctx->ptr == '.') {
 411         ctx->ptr++;
 412
 413         while(ctx->ptr < ctx->end && isdigitW(*ctx->ptr)) {
 414             hlp = d*10 + *(ctx->ptr++) - '0';
 415             if(d>MAXLONGLONG/10 || hlp<0)
 416                 break;
 417
 418             d = hlp;
 419             exp--;
 420         }
 421         while(ctx->ptr < ctx->end && isdigitW(*ctx->ptr))
 422             ctx->ptr++;
 423     }
 424
 425     if(ctx->ptr < ctx->end && (*ctx->ptr == 'e' || *ctx->ptr == 'E')) {
 426         int sign = 1, e = 0;
 427
 428         ctx->ptr++;
 429         if(ctx->ptr < ctx->end) {
 430             if(*ctx->ptr == '+') {
 431                 ctx->ptr++;
 432             }else if(*ctx->ptr == '-') {
 433                 sign = -1;
 434                 ctx->ptr++;
 435             }else if(!isdigitW(*ctx->ptr)) {
 436                 WARN("Expected exponent part\n");
 437                 return lex_error(ctx, E_FAIL);
 438             }
 439         }
 440
 441         if(ctx->ptr == ctx->end) {
 442             WARN("unexpected end of file\n");
 443             return lex_error(ctx, E_FAIL);
 444         }
 445
 446         while(ctx->ptr < ctx->end && isdigitW(*ctx->ptr)) {
 447             if(e > INT_MAX/10 || (e = e*10 + *ctx->ptr++ - '0')<0)
 448                 e = INT_MAX;
 449         }
 450         e *= sign;
 451
 452         if(exp<0 && e<0 && e+exp>0) exp = INT_MIN;
 453         else if(exp>0 && e>0 && e+exp<0) exp = INT_MAX;
 454         else exp += e;
 455     }
 456
 457     *literal = new_double_literal(ctx, exp>=0 ? d*pow(10, exp) : d/pow(10, -exp));
 458     return tNumericLiteral;
 459 }
 460
 461 static int parse_numeric_literal(parser_ctx_t *ctx, literal_t **literal)
 462 {
 463     LONG l, d;
 464
 465     l = *ctx->ptr++ - '0';
 466     if(!l) {
 467         if(*ctx->ptr == 'x' || *ctx->ptr == 'X') {
 468             if(++ctx->ptr == ctx->end) {
 469                 ERR("unexpected end of file\n");
 470                 return 0;
 471             }
 472
 473             while(ctx->ptr < ctx->end && (d = hex_to_int(*ctx->ptr)) != -1) {
 474                 l = l*16 + d;
 475                 ctx->ptr++;
 476             }
 477
 478             if(ctx->ptr < ctx->end && is_identifier_char(*ctx->ptr)) {
 479                 WARN("unexpected identifier char\n");
 480                 return lex_error(ctx, E_FAIL);
 481             }
 482
 483             *literal = new_double_literal(ctx, l);
 484             return tNumericLiteral;
 485         }
 486
 487         if(isdigitW(*ctx->ptr)) {
 488             unsigned base = 8;
 489             const WCHAR *ptr;
 490             double val = 0;
 491
 492             for(ptr = ctx->ptr; ptr < ctx->end && isdigitW(*ptr); ptr++) {
 493                 if(*ptr > '7') {
 494                     base = 10;
 495                     break;
 496                 }
 497             }
 498
 499             do {
 500                 val = val*base + *ctx->ptr-'0';
 501             }while(++ctx->ptr < ctx->end && isdigitW(*ctx->ptr));
 502
 503             /* FIXME: Do we need it here? */
 504             if(ctx->ptr < ctx->end && (is_identifier_char(*ctx->ptr) || *ctx->ptr == '.')) {
 505                 WARN("wrong char after octal literal: '%c'\n", *ctx->ptr);
 506                 return lex_error(ctx, JS_E_MISSING_SEMICOLON);
 507             }
 508
 509             *literal = new_double_literal(ctx, val);
 510             return tNumericLiteral;
 511         }
 512
 513         if(is_identifier_char(*ctx->ptr)) {
 514             WARN("wrong char after zero\n");
 515             return lex_error(ctx, E_FAIL);
 516         }
 517
 518     }
 519
 520     return parse_double_literal(ctx, l, literal);
 521 }
 522
 523 static int next_token(parser_ctx_t *ctx, void *lval)
 524 {
 525     do {
 526         while(ctx->ptr < ctx->end && isspaceW(*ctx->ptr)) {
 527             if(is_endline(*ctx->ptr++))
 528                 ctx->nl = TRUE;
 529         }
 530         if(ctx->ptr == ctx->end)
 531             return tEOF;
 532     }while(skip_comment(ctx) || skip_html_comment(ctx));
 533
 534     if(ctx->implicit_nl_semicolon) {
 535         if(ctx->nl)
 536             return ';';
 537         ctx->implicit_nl_semicolon = FALSE;
 538     }
 539
 540     if(isalphaW(*ctx->ptr)) {
 541         int ret = check_keywords(ctx, lval);
 542         if(ret)
 543             return ret;
 544
 545         return parse_identifier(ctx, lval);
 546     }
 547
 548     if(isdigitW(*ctx->ptr))
 549         return parse_numeric_literal(ctx, lval);
 550
 551     switch(*ctx->ptr) {
 552     case '{':
 553     case '(':
 554     case ')':
 555     case '[':
 556     case ']':
 557     case ';':
 558     case ',':
 559     case '~':
 560     case '?':
 561     case ':':
 562         return *ctx->ptr++;
 563
 564     case '}':
 565         *(const WCHAR**)lval = ctx->ptr++;
 566         return '}';
 567
 568     case '.':
 569         if(++ctx->ptr < ctx->end && isdigitW(*ctx->ptr))
 570             return parse_double_literal(ctx, 0, lval);
 571         return '.';
 572
 573     case '<':
 574         if(++ctx->ptr == ctx->end) {
 575             *(int*)lval = EXPR_LESS;
 576             return tRelOper;
 577         }
 578
 579         switch(*ctx->ptr) {
 580         case '=':  /* <= */
 581             ctx->ptr++;
 582             *(int*)lval = EXPR_LESSEQ;
 583             return tRelOper;
 584         case '<':  /* << */
 585             if(++ctx->ptr < ctx->end && *ctx->ptr == '=') { /* <<= */
 586                 ctx->ptr++;
 587                 *(int*)lval = EXPR_ASSIGNLSHIFT;
 588                 return tAssignOper;
 589             }
 590             *(int*)lval = EXPR_LSHIFT;
 591             return tShiftOper;
 592         default: /* < */
 593             *(int*)lval = EXPR_LESS;
 594             return tRelOper;
 595         }
 596
 597     case '>':
 598         if(++ctx->ptr == ctx->end) { /* > */
 599             *(int*)lval = EXPR_GREATER;
 600             return tRelOper;
 601         }
 602
 603         switch(*ctx->ptr) {
 604         case '=':  /* >= */
 605             ctx->ptr++;
 606             *(int*)lval = EXPR_GREATEREQ;
 607             return tRelOper;
 608         case '>':  /* >> */
 609             if(++ctx->ptr < ctx->end) {
 610                 if(*ctx->ptr == '=') {  /* >>= */
 611                     ctx->ptr++;
 612                     *(int*)lval = EXPR_ASSIGNRSHIFT;
 613                     return tAssignOper;
 614                 }
 615                 if(*ctx->ptr == '>') {  /* >>> */
 616                     if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* >>>= */
 617                         ctx->ptr++;
 618                         *(int*)lval = EXPR_ASSIGNRRSHIFT;
 619                         return tAssignOper;
 620                     }
 621                     *(int*)lval = EXPR_RRSHIFT;
 622                     return tRelOper;
 623                 }
 624             }
 625             *(int*)lval = EXPR_RSHIFT;
 626             return tShiftOper;
 627         default:
 628             *(int*)lval = EXPR_GREATER;
 629             return tRelOper;
 630         }
 631
 632     case '+':
 633         ctx->ptr++;
 634         if(ctx->ptr < ctx->end) {
 635             switch(*ctx->ptr) {
 636             case '+':  /* ++ */
 637                 ctx->ptr++;
 638                 return tINC;
 639             case '=':  /* += */
 640                 ctx->ptr++;
 641                 *(int*)lval = EXPR_ASSIGNADD;
 642                 return tAssignOper;
 643             }
 644         }
 645         return '+';
 646
 647     case '-':
 648         ctx->ptr++;
 649         if(ctx->ptr < ctx->end) {
 650             switch(*ctx->ptr) {
 651             case '-':  /* -- or --> */
 652                 ctx->ptr++;
 653                 if(ctx->is_html && ctx->nl && ctx->ptr < ctx->end && *ctx->ptr == '>') {
 654                     ctx->ptr++;
 655                     return tHTMLCOMMENT;
 656                 }
 657                 return tDEC;
 658             case '=':  /* -= */
 659                 ctx->ptr++;
 660                 *(int*)lval = EXPR_ASSIGNSUB;
 661                 return tAssignOper;
 662             }
 663         }
 664         return '-';
 665
 666     case '*':
 667         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') { /* *= */
 668             ctx->ptr++;
 669             *(int*)lval = EXPR_ASSIGNMUL;
 670             return tAssignOper;
 671         }
 672         return '*';
 673
 674     case '%':
 675         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') { /* %= */
 676             ctx->ptr++;
 677             *(int*)lval = EXPR_ASSIGNMOD;
 678             return tAssignOper;
 679         }
 680         return '%';
 681
 682     case '&':
 683         if(++ctx->ptr < ctx->end) {
 684             switch(*ctx->ptr) {
 685             case '=':  /* &= */
 686                 ctx->ptr++;
 687                 *(int*)lval = EXPR_ASSIGNAND;
 688                 return tAssignOper;
 689             case '&':  /* && */
 690                 ctx->ptr++;
 691                 return tANDAND;
 692             }
 693         }
 694         return '&';
 695
 696     case '|':
 697         if(++ctx->ptr < ctx->end) {
 698             switch(*ctx->ptr) {
 699             case '=':  /* |= */
 700                 ctx->ptr++;
 701                 *(int*)lval = EXPR_ASSIGNOR;
 702                 return tAssignOper;
 703             case '|':  /* || */
 704                 ctx->ptr++;
 705                 return tOROR;
 706             }
 707         }
 708         return '|';
 709
 710     case '^':
 711         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* ^= */
 712             ctx->ptr++;
 713             *(int*)lval = EXPR_ASSIGNXOR;
 714             return tAssignOper;
 715         }
 716         return '^';
 717
 718     case '!':
 719         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* != */
 720             if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* !== */
 721                 ctx->ptr++;
 722                 *(int*)lval = EXPR_NOTEQEQ;
 723                 return tEqOper;
 724             }
 725             *(int*)lval = EXPR_NOTEQ;
 726             return tEqOper;
 727         }
 728         return '!';
 729
 730     case '=':
 731         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* == */
 732             if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* === */
 733                 ctx->ptr++;
 734                 *(int*)lval = EXPR_EQEQ;
 735                 return tEqOper;
 736             }
 737             *(int*)lval = EXPR_EQ;
 738             return tEqOper;
 739         }
 740         return '=';
 741
 742     case '/':
 743         if(++ctx->ptr < ctx->end) {
 744             if(*ctx->ptr == '=') {  /* /= */
 745                 ctx->ptr++;
 746                 *(int*)lval = EXPR_ASSIGNDIV;
 747                 return kDIVEQ;
 748             }
 749         }
 750         return '/';
 751
 752     case '\"':
 753     case '\'':
 754         return parse_string_literal(ctx, lval, *ctx->ptr);
 755
 756     case '_':
 757     case '$':
 758         return parse_identifier(ctx, lval);
 759
 760     case '@':
 761         return '@';
 762     }
 763
 764     WARN("unexpected char '%c' %d\n", *ctx->ptr, *ctx->ptr);
 765     return 0;
 766 }
 767
 768 struct _cc_var_t {
 769     BOOL is_num;
 770     union {
 771         BOOL b;
 772         DOUBLE n;
 773     } u;
 774     struct _cc_var_t *next;
 775     unsigned name_len;
 776     WCHAR name[0];
 777 };
 778
 779 void release_cc(cc_ctx_t *cc)
 780 {
 781     cc_var_t *iter, *next;
 782
 783     for(iter = cc->vars; iter; iter = next) {
 784         next = iter->next;
 785         heap_free(iter);
 786     }
 787
 788     heap_free(cc);
 789 }
 790
 791 static BOOL add_cc_var(cc_ctx_t *cc, const WCHAR *name, cc_var_t *v)
 792 {
 793     cc_var_t *new_v;
 794     unsigned len;
 795
 796     len = strlenW(name);
 797
 798     new_v = heap_alloc(sizeof(cc_var_t) + (len+1)*sizeof(WCHAR));
 799     if(!new_v)
 800         return FALSE;
 801
 802     memcpy(new_v, v, sizeof(*v));
 803     memcpy(new_v->name, name, (len+1)*sizeof(WCHAR));
 804     new_v->name_len = len;
 805     new_v->next = cc->vars;
 806     cc->vars = new_v;
 807     return TRUE;
 808 }
 809
 810 static cc_var_t *find_cc_var(cc_ctx_t *cc, const WCHAR *name, unsigned name_len)
 811 {
 812     cc_var_t *iter;
 813
 814     for(iter = cc->vars; iter; iter = iter->next) {
 815         if(iter->name_len == name_len && !memcmp(iter->name, name, name_len*sizeof(WCHAR)))
 816             return iter;
 817     }
 818
 819     return NULL;
 820 }
 821
 822 static int init_cc(parser_ctx_t *ctx)
 823 {
 824     cc_ctx_t *cc;
 825     cc_var_t v;
 826
 827     static const WCHAR _win32W[] = {'_','w','i','n','3','2',0};
 828     static const WCHAR _win64W[] = {'_','w','i','n','6','4',0};
 829     static const WCHAR _x86W[] = {'_','x','8','6',0};
 830     static const WCHAR _amd64W[] = {'_','a','m','d','6','4',0};
 831     static const WCHAR _jscriptW[] = {'_','j','s','c','r','i','p','t',0};
 832     static const WCHAR _jscript_buildW[] = {'_','j','s','c','r','i','p','t','_','b','u','i','l','d',0};
 833     static const WCHAR _jscript_versionW[] = {'_','j','s','c','r','i','p','t','_','v','e','r','s','i','o','n',0};
 834
 835     if(ctx->script->cc)
 836         return 0;
 837
 838     cc = heap_alloc(sizeof(cc_ctx_t));
 839     if(!cc)
 840         return lex_error(ctx, E_OUTOFMEMORY);
 841
 842     cc->vars = NULL;
 843     v.is_num = FALSE;
 844     v.u.b = TRUE;
 845     if(!add_cc_var(cc, _jscriptW, &v)
 846        || !add_cc_var(cc, sizeof(void*) == 8 ? _win64W : _win32W, &v)
 847        || !add_cc_var(cc, sizeof(void*) == 8 ? _amd64W : _x86W, &v)) {
 848         release_cc(cc);
 849         return lex_error(ctx, E_OUTOFMEMORY);
 850     }
 851
 852     v.is_num = TRUE;
 853     v.u.n = JSCRIPT_BUILD_VERSION;
 854     if(!add_cc_var(cc, _jscript_buildW, &v)) {
 855         release_cc(cc);
 856         return lex_error(ctx, E_OUTOFMEMORY);
 857     }
 858
 859     v.u.n = JSCRIPT_MAJOR_VERSION + (DOUBLE)JSCRIPT_MINOR_VERSION/10.0;
 860     if(!add_cc_var(cc, _jscript_versionW, &v)) {
 861         release_cc(cc);
 862         return lex_error(ctx, E_OUTOFMEMORY);
 863     }
 864
 865     ctx->script->cc = cc;
 866     return 0;
 867 }
 868
 869 static int cc_token(parser_ctx_t *ctx, void *lval)
 870 {
 871     unsigned id_len = 0;
 872     cc_var_t *var;
 873
 874     static const WCHAR cc_onW[] = {'c','c','_','o','n',0};
 875     static const WCHAR setW[] = {'s','e','t',0};
 876     static const WCHAR elifW[] = {'e','l','i','f',0};
 877     static const WCHAR endW[] = {'e','n','d',0};
 878
 879     ctx->ptr++;
 880
 881     if(!check_keyword(ctx, cc_onW, NULL))
 882         return init_cc(ctx);
 883
 884     if(!check_keyword(ctx, setW, NULL)) {
 885         FIXME("@set not implemented\n");
 886         return lex_error(ctx, E_NOTIMPL);
 887     }
 888
 889     if(!check_keyword(ctx, ifW, NULL)) {
 890         FIXME("@if not implemented\n");
 891         return lex_error(ctx, E_NOTIMPL);
 892     }
 893
 894     if(!check_keyword(ctx, elifW, NULL)) {
 895         FIXME("@elif not implemented\n");
 896         return lex_error(ctx, E_NOTIMPL);
 897     }
 898
 899     if(!check_keyword(ctx, elseW, NULL)) {
 900         FIXME("@else not implemented\n");
 901         return lex_error(ctx, E_NOTIMPL);
 902     }
 903
 904     if(!check_keyword(ctx, endW, NULL)) {
 905         FIXME("@end not implemented\n");
 906         return lex_error(ctx, E_NOTIMPL);
 907     }
 908
 909     if(!ctx->script->cc)
 910         return lex_error(ctx, JS_E_DISABLED_CC);
 911
 912     while(ctx->ptr+id_len < ctx->end && is_identifier_char(ctx->ptr[id_len]))
 913         id_len++;
 914     if(!id_len)
 915         return '@';
 916
 917     TRACE("var %s\n", debugstr_wn(ctx->ptr, id_len));
 918
 919     var = find_cc_var(ctx->script->cc, ctx->ptr, id_len);
 920     ctx->ptr += id_len;
 921     if(!var || var->is_num) {
 922         *(literal_t**)lval = new_double_literal(ctx, var ? var->u.n : NAN);
 923         return tNumericLiteral;
 924     }
 925
 926     *(literal_t**)lval = new_boolean_literal(ctx, var->u.b);
 927     return tBooleanLiteral;
 928 }
 929
 930 int parser_lex(void *lval, parser_ctx_t *ctx)
 931 {
 932     int ret;
 933
 934     ctx->nl = ctx->ptr == ctx->begin;
 935
 936     do {
 937         ret = next_token(ctx, lval);
 938     } while(ret == '@' && !(ret = cc_token(ctx, lval)));
 939
 940     return ret;
 941 }
 942
 943 literal_t *parse_regexp(parser_ctx_t *ctx)
 944 {
 945     const WCHAR *re, *flags_ptr;
 946     BOOL in_class = FALSE;
 947     DWORD re_len, flags;
 948     literal_t *ret;
 949     HRESULT hres;
 950
 951     TRACE("\n");
 952
 953     while(*--ctx->ptr != '/');
 954
 955     /* Simple regexp pre-parser; '/' if used in char class does not terminate regexp literal */
 956     re = ++ctx->ptr;
 957     while(ctx->ptr < ctx->end) {
 958         if(*ctx->ptr == '\\') {
 959             if(++ctx->ptr == ctx->end)
 960                 break;
 961         }else if(in_class) {
 962             if(*ctx->ptr == '\n')
 963                 break;
 964             if(*ctx->ptr == ']')
 965                 in_class = FALSE;
 966         }else {
 967             if(*ctx->ptr == '/')
 968                 break;
 969
 970             if(*ctx->ptr == '[')
 971                 in_class = TRUE;
 972         }
 973         ctx->ptr++;
 974     }
 975
 976     if(ctx->ptr == ctx->end || *ctx->ptr != '/') {
 977         WARN("pre-parsing failed\n");
 978         return NULL;
 979     }
 980
 981     re_len = ctx->ptr-re;
 982
 983     flags_ptr = ++ctx->ptr;
 984     while(ctx->ptr < ctx->end && isalnumW(*ctx->ptr))
 985         ctx->ptr++;
 986
 987     hres = parse_regexp_flags(flags_ptr, ctx->ptr-flags_ptr, &flags);
 988     if(FAILED(hres))
 989         return NULL;
 990
 991     ret = parser_alloc(ctx, sizeof(literal_t));
 992     ret->type = LT_REGEXP;
 993     ret->u.regexp.str = re;
 994     ret->u.regexp.str_len = re_len;
 995     ret->u.regexp.flags = flags;
 996     return ret;
 997 }