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 #include "parser.h"
  29
  30 #include "parser.tab.h"
  31
  32 #include "wine/debug.h"
  33 #include "wine/unicode.h"
  34
  35 WINE_DEFAULT_DEBUG_CHANNEL(jscript);
  36
  37 static const WCHAR breakW[] = {'b','r','e','a','k',0};
  38 static const WCHAR caseW[] = {'c','a','s','e',0};
  39 static const WCHAR catchW[] = {'c','a','t','c','h',0};
  40 static const WCHAR continueW[] = {'c','o','n','t','i','n','u','e',0};
  41 static const WCHAR defaultW[] = {'d','e','f','a','u','l','t',0};
  42 static const WCHAR deleteW[] = {'d','e','l','e','t','e',0};
  43 static const WCHAR doW[] = {'d','o',0};
  44 static const WCHAR elseW[] = {'e','l','s','e',0};
  45 static const WCHAR falseW[] = {'f','a','l','s','e',0};
  46 static const WCHAR finallyW[] = {'f','i','n','a','l','l','y',0};
  47 static const WCHAR forW[] = {'f','o','r',0};
  48 static const WCHAR functionW[] = {'f','u','n','c','t','i','o','n',0};
  49 static const WCHAR ifW[] = {'i','f',0};
  50 static const WCHAR inW[] = {'i','n',0};
  51 static const WCHAR instanceofW[] = {'i','n','s','t','a','n','c','e','o','f',0};
  52 static const WCHAR newW[] = {'n','e','w',0};
  53 static const WCHAR nullW[] = {'n','u','l','l',0};
  54 static const WCHAR returnW[] = {'r','e','t','u','r','n',0};
  55 static const WCHAR switchW[] = {'s','w','i','t','c','h',0};
  56 static const WCHAR thisW[] = {'t','h','i','s',0};
  57 static const WCHAR throwW[] = {'t','h','r','o','w',0};
  58 static const WCHAR trueW[] = {'t','r','u','e',0};
  59 static const WCHAR tryW[] = {'t','r','y',0};
  60 static const WCHAR typeofW[] = {'t','y','p','e','o','f',0};
  61 static const WCHAR varW[] = {'v','a','r',0};
  62 static const WCHAR voidW[] = {'v','o','i','d',0};
  63 static const WCHAR whileW[] = {'w','h','i','l','e',0};
  64 static const WCHAR withW[] = {'w','i','t','h',0};
  65
  66 static const WCHAR elifW[] = {'e','l','i','f',0};
  67 static const WCHAR endW[] = {'e','n','d',0};
  68
  69 static const struct {
  70     const WCHAR *word;
  71     int token;
  72     BOOL no_nl;
  73 } keywords[] = {
  74     {breakW,       kBREAK, TRUE},
  75     {caseW,        kCASE},
  76     {catchW,       kCATCH},
  77     {continueW,    kCONTINUE, TRUE},
  78     {defaultW,     kDEFAULT},
  79     {deleteW,      kDELETE},
  80     {doW,          kDO},
  81     {elseW,        kELSE},
  82     {falseW,       kFALSE},
  83     {finallyW,     kFINALLY},
  84     {forW,         kFOR},
  85     {functionW,    kFUNCTION},
  86     {ifW,          kIF},
  87     {inW,          kIN},
  88     {instanceofW,  kINSTANCEOF},
  89     {newW,         kNEW},
  90     {nullW,        kNULL},
  91     {returnW,      kRETURN, TRUE},
  92     {switchW,      kSWITCH},
  93     {thisW,        kTHIS},
  94     {throwW,       kTHROW},
  95     {trueW,        kTRUE},
  96     {tryW,         kTRY},
  97     {typeofW,      kTYPEOF},
  98     {varW,         kVAR},
  99     {voidW,        kVOID},
 100     {whileW,       kWHILE},
 101     {withW,        kWITH}
 102 };
 103
 104 static int lex_error(parser_ctx_t *ctx, HRESULT hres)
 105 {
 106     ctx->hres = hres;
 107     ctx->lexer_error = TRUE;
 108     return -1;
 109 }
 110
 111 /* ECMA-262 3rd Edition    7.6 */
 112 BOOL is_identifier_char(WCHAR c)
 113 {
 114     return isalnumW(c) || c == '$' || c == '_' || c == '\\';
 115 }
 116
 117 static BOOL is_identifier_first_char(WCHAR c)
 118 {
 119     return isalphaW(c) || c == '$' || c == '_' || c == '\\';
 120 }
 121
 122 static int check_keyword(parser_ctx_t *ctx, const WCHAR *word, const WCHAR **lval)
 123 {
 124     const WCHAR *p1 = ctx->ptr;
 125     const WCHAR *p2 = word;
 126
 127     while(p1 < ctx->end && *p2) {
 128         if(*p1 != *p2)
 129             return *p1 - *p2;
 130         p1++;
 131         p2++;
 132     }
 133
 134     if(*p2 || (p1 < ctx->end && is_identifier_char(*p1)))
 135         return 1;
 136
 137     if(lval)
 138         *lval = ctx->ptr;
 139     ctx->ptr = p1;
 140     return 0;
 141 }
 142
 143 /* ECMA-262 3rd Edition    7.3 */
 144 static BOOL is_endline(WCHAR c)
 145 {
 146     return c == '\n' || c == '\r' || c == 0x2028 || c == 0x2029;
 147 }
 148
 149 static int hex_to_int(WCHAR c)
 150 {
 151     if('0' <= c && c <= '9')
 152         return c-'0';
 153
 154     if('a' <= c && c <= 'f')
 155         return c-'a'+10;
 156
 157     if('A' <= c && c <= 'F')
 158         return c-'A'+10;
 159
 160     return -1;
 161 }
 162
 163 static int check_keywords(parser_ctx_t *ctx, const WCHAR **lval)
 164 {
 165     int min = 0, max = sizeof(keywords)/sizeof(keywords[0])-1, r, i;
 166
 167     while(min <= max) {
 168         i = (min+max)/2;
 169
 170         r = check_keyword(ctx, keywords[i].word, lval);
 171         if(!r) {
 172             ctx->implicit_nl_semicolon = keywords[i].no_nl;
 173             return keywords[i].token;
 174         }
 175
 176         if(r > 0)
 177             min = i+1;
 178         else
 179             max = i-1;
 180     }
 181
 182     return 0;
 183 }
 184
 185 static BOOL skip_html_comment(parser_ctx_t *ctx)
 186 {
 187     const WCHAR html_commentW[] = {'<','!','-','-',0};
 188
 189     if(!ctx->is_html || ctx->ptr+3 >= ctx->end ||
 190         memcmp(ctx->ptr, html_commentW, sizeof(WCHAR)*4))
 191         return FALSE;
 192
 193     ctx->nl = TRUE;
 194     while(ctx->ptr < ctx->end && !is_endline(*ctx->ptr++));
 195
 196     return TRUE;
 197 }
 198
 199 static BOOL skip_comment(parser_ctx_t *ctx)
 200 {
 201     if(ctx->ptr+1 >= ctx->end)
 202         return FALSE;
 203
 204     if(*ctx->ptr != '/') {
 205         if(*ctx->ptr == '@' && ctx->ptr+2 < ctx->end && ctx->ptr[1] == '*' && ctx->ptr[2] == '/') {
 206             ctx->ptr += 3;
 207             return TRUE;
 208         }
 209
 210         return FALSE;
 211     }
 212
 213     switch(ctx->ptr[1]) {
 214     case '*':
 215         ctx->ptr += 2;
 216         if(ctx->ptr+2 < ctx->end && *ctx->ptr == '@' && is_identifier_char(ctx->ptr[1]))
 217             return FALSE;
 218         while(ctx->ptr+1 < ctx->end && (ctx->ptr[0] != '*' || ctx->ptr[1] != '/'))
 219             ctx->ptr++;
 220
 221         if(ctx->ptr[0] == '*' && ctx->ptr[1] == '/') {
 222             ctx->ptr += 2;
 223         }else {
 224             WARN("unexpected end of file (missing end of comment)\n");
 225             ctx->ptr = ctx->end;
 226         }
 227         break;
 228     case '/':
 229         ctx->ptr += 2;
 230         if(ctx->ptr+2 < ctx->end && *ctx->ptr == '@' && is_identifier_char(ctx->ptr[1]))
 231             return FALSE;
 232         while(ctx->ptr < ctx->end && !is_endline(*ctx->ptr))
 233             ctx->ptr++;
 234         break;
 235     default:
 236         return FALSE;
 237     }
 238
 239     return TRUE;
 240 }
 241
 242 static BOOL skip_spaces(parser_ctx_t *ctx)
 243 {
 244     while(ctx->ptr < ctx->end && (isspaceW(*ctx->ptr) || *ctx->ptr == 0xFEFF /* UTF16 BOM */)) {
 245         if(is_endline(*ctx->ptr++))
 246             ctx->nl = TRUE;
 247     }
 248
 249     return ctx->ptr != ctx->end;
 250 }
 251
 252 BOOL unescape(WCHAR *str)
 253 {
 254     WCHAR *pd, *p, c;
 255     int i;
 256
 257     pd = p = str;
 258     while(*p) {
 259         if(*p != '\\') {
 260             *pd++ = *p++;
 261             continue;
 262         }
 263
 264         p++;
 265
 266         switch(*p) {
 267         case '\'':
 268         case '\"':
 269         case '\\':
 270             c = *p;
 271             break;
 272         case 'b':
 273             c = '\b';
 274             break;
 275         case 't':
 276             c = '\t';
 277             break;
 278         case 'n':
 279             c = '\n';
 280             break;
 281         case 'f':
 282             c = '\f';
 283             break;
 284         case 'r':
 285             c = '\r';
 286             break;
 287         case 'x':
 288             i = hex_to_int(*++p);
 289             if(i == -1)
 290                 return FALSE;
 291             c = i << 4;
 292
 293             i = hex_to_int(*++p);
 294             if(i == -1)
 295                 return FALSE;
 296             c += i;
 297             break;
 298         case 'u':
 299             i = hex_to_int(*++p);
 300             if(i == -1)
 301                 return FALSE;
 302             c = i << 12;
 303
 304             i = hex_to_int(*++p);
 305             if(i == -1)
 306                 return FALSE;
 307             c += i << 8;
 308
 309             i = hex_to_int(*++p);
 310             if(i == -1)
 311                 return FALSE;
 312             c += i << 4;
 313
 314             i = hex_to_int(*++p);
 315             if(i == -1)
 316                 return FALSE;
 317             c += i;
 318             break;
 319         default:
 320             if(isdigitW(*p)) {
 321                 c = *p++ - '0';
 322                 if(isdigitW(*p)) {
 323                     c = c*8 + (*p++ - '0');
 324                     if(isdigitW(*p))
 325                         c = c*8 + (*p++ - '0');
 326                 }
 327                 p--;
 328             }
 329             else
 330                 c = *p;
 331         }
 332
 333         *pd++ = c;
 334         p++;
 335     }
 336
 337     *pd = 0;
 338     return TRUE;
 339 }
 340
 341 static int parse_identifier(parser_ctx_t *ctx, const WCHAR **ret)
 342 {
 343     const WCHAR *ptr = ctx->ptr++;
 344     WCHAR *wstr;
 345     int len;
 346
 347     while(ctx->ptr < ctx->end && is_identifier_char(*ctx->ptr))
 348         ctx->ptr++;
 349
 350     len = ctx->ptr-ptr;
 351
 352     *ret = wstr = parser_alloc(ctx, (len+1)*sizeof(WCHAR));
 353     memcpy(wstr, ptr, len*sizeof(WCHAR));
 354     wstr[len] = 0;
 355
 356     /* FIXME: unescape */
 357     return tIdentifier;
 358 }
 359
 360 static int parse_string_literal(parser_ctx_t *ctx, const WCHAR **ret, WCHAR endch)
 361 {
 362     const WCHAR *ptr = ++ctx->ptr;
 363     WCHAR *wstr;
 364     int len;
 365
 366     while(ctx->ptr < ctx->end && *ctx->ptr != endch) {
 367         if(*ctx->ptr++ == '\\')
 368             ctx->ptr++;
 369     }
 370
 371     if(ctx->ptr == ctx->end)
 372         return lex_error(ctx, JS_E_UNTERMINATED_STRING);
 373
 374     len = ctx->ptr-ptr;
 375
 376     *ret = wstr = parser_alloc(ctx, (len+1)*sizeof(WCHAR));
 377     memcpy(wstr, ptr, len*sizeof(WCHAR));
 378     wstr[len] = 0;
 379
 380     ctx->ptr++;
 381
 382     if(!unescape(wstr)) {
 383         WARN("unescape failed\n");
 384         return lex_error(ctx, E_FAIL);
 385     }
 386
 387     return tStringLiteral;
 388 }
 389
 390 static literal_t *new_double_literal(parser_ctx_t *ctx, DOUBLE d)
 391 {
 392     literal_t *ret = parser_alloc(ctx, sizeof(literal_t));
 393
 394     ret->type = LT_DOUBLE;
 395     ret->u.dval = d;
 396     return ret;
 397 }
 398
 399 literal_t *new_boolean_literal(parser_ctx_t *ctx, BOOL bval)
 400 {
 401     literal_t *ret = parser_alloc(ctx, sizeof(literal_t));
 402
 403     ret->type = LT_BOOL;
 404     ret->u.bval = bval;
 405
 406     return ret;
 407 }
 408
 409 HRESULT parse_decimal(const WCHAR **iter, const WCHAR *end, double *ret)
 410 {
 411     const WCHAR *ptr = *iter;
 412     LONGLONG d = 0, hlp;
 413     int exp = 0;
 414
 415     while(ptr < end && isdigitW(*ptr)) {
 416         hlp = d*10 + *(ptr++) - '0';
 417         if(d>MAXLONGLONG/10 || hlp<0) {
 418             exp++;
 419             break;
 420         }
 421         else
 422             d = hlp;
 423     }
 424     while(ptr < end && isdigitW(*ptr)) {
 425         exp++;
 426         ptr++;
 427     }
 428
 429     if(*ptr == '.') {
 430         ptr++;
 431
 432         while(ptr < end && isdigitW(*ptr)) {
 433             hlp = d*10 + *(ptr++) - '0';
 434             if(d>MAXLONGLONG/10 || hlp<0)
 435                 break;
 436
 437             d = hlp;
 438             exp--;
 439         }
 440         while(ptr < end && isdigitW(*ptr))
 441             ptr++;
 442     }
 443
 444     if(ptr < end && (*ptr == 'e' || *ptr == 'E')) {
 445         int sign = 1, e = 0;
 446
 447         if(++ptr < end) {
 448             if(*ptr == '+') {
 449                 ptr++;
 450             }else if(*ptr == '-') {
 451                 sign = -1;
 452                 ptr++;
 453             }else if(!isdigitW(*ptr)) {
 454                 WARN("Expected exponent part\n");
 455                 return E_FAIL;
 456             }
 457         }
 458
 459         if(ptr == end) {
 460             WARN("unexpected end of file\n");
 461             return E_FAIL;
 462         }
 463
 464         while(ptr < end && isdigitW(*ptr)) {
 465             if(e > INT_MAX/10 || (e = e*10 + *ptr++ - '0')<0)
 466                 e = INT_MAX;
 467         }
 468         e *= sign;
 469
 470         if(exp<0 && e<0 && e+exp>0) exp = INT_MIN;
 471         else if(exp>0 && e>0 && e+exp<0) exp = INT_MAX;
 472         else exp += e;
 473     }
 474
 475     if(is_identifier_char(*ptr)) {
 476         WARN("wrong char after zero\n");
 477         return JS_E_MISSING_SEMICOLON;
 478     }
 479
 480     *ret = exp>=0 ? d*pow(10, exp) : d/pow(10, -exp);
 481     *iter = ptr;
 482     return S_OK;
 483 }
 484
 485 static BOOL parse_numeric_literal(parser_ctx_t *ctx, double *ret)
 486 {
 487     HRESULT hres;
 488
 489     if(*ctx->ptr == '0') {
 490         LONG d, l = 0;
 491
 492         ctx->ptr++;
 493
 494         if(*ctx->ptr == 'x' || *ctx->ptr == 'X') {
 495             if(++ctx->ptr == ctx->end) {
 496                 ERR("unexpected end of file\n");
 497                 return FALSE;
 498             }
 499
 500             while(ctx->ptr < ctx->end && (d = hex_to_int(*ctx->ptr)) != -1) {
 501                 l = l*16 + d;
 502                 ctx->ptr++;
 503             }
 504
 505             if(ctx->ptr < ctx->end && is_identifier_char(*ctx->ptr)) {
 506                 WARN("unexpected identifier char\n");
 507                 lex_error(ctx, JS_E_MISSING_SEMICOLON);
 508                 return FALSE;
 509             }
 510
 511             *ret = l;
 512             return TRUE;
 513         }
 514
 515         if(isdigitW(*ctx->ptr)) {
 516             unsigned base = 8;
 517             const WCHAR *ptr;
 518             double val = 0;
 519
 520             for(ptr = ctx->ptr; ptr < ctx->end && isdigitW(*ptr); ptr++) {
 521                 if(*ptr > '7') {
 522                     base = 10;
 523                     break;
 524                 }
 525             }
 526
 527             do {
 528                 val = val*base + *ctx->ptr-'0';
 529             }while(++ctx->ptr < ctx->end && isdigitW(*ctx->ptr));
 530
 531             /* FIXME: Do we need it here? */
 532             if(ctx->ptr < ctx->end && (is_identifier_char(*ctx->ptr) || *ctx->ptr == '.')) {
 533                 WARN("wrong char after octal literal: '%c'\n", *ctx->ptr);
 534                 lex_error(ctx, JS_E_MISSING_SEMICOLON);
 535                 return FALSE;
 536             }
 537
 538             *ret = val;
 539             return TRUE;
 540         }
 541
 542         if(is_identifier_char(*ctx->ptr)) {
 543             WARN("wrong char after zero\n");
 544             lex_error(ctx, JS_E_MISSING_SEMICOLON);
 545             return FALSE;
 546         }
 547     }
 548
 549     hres = parse_decimal(&ctx->ptr, ctx->end, ret);
 550     if(FAILED(hres)) {
 551         lex_error(ctx, hres);
 552         return FALSE;
 553     }
 554
 555     return TRUE;
 556 }
 557
 558 static int next_token(parser_ctx_t *ctx, void *lval)
 559 {
 560     do {
 561         if(!skip_spaces(ctx))
 562             return tEOF;
 563     }while(skip_comment(ctx) || skip_html_comment(ctx));
 564
 565     if(ctx->implicit_nl_semicolon) {
 566         if(ctx->nl)
 567             return ';';
 568         ctx->implicit_nl_semicolon = FALSE;
 569     }
 570
 571     if(isalphaW(*ctx->ptr)) {
 572         int ret = check_keywords(ctx, lval);
 573         if(ret)
 574             return ret;
 575
 576         return parse_identifier(ctx, lval);
 577     }
 578
 579     if(isdigitW(*ctx->ptr)) {
 580         double n;
 581
 582         if(!parse_numeric_literal(ctx, &n))
 583             return -1;
 584
 585         *(literal_t**)lval = new_double_literal(ctx, n);
 586         return tNumericLiteral;
 587     }
 588
 589     switch(*ctx->ptr) {
 590     case '{':
 591     case '(':
 592     case ')':
 593     case '[':
 594     case ']':
 595     case ';':
 596     case ',':
 597     case '~':
 598     case '?':
 599         return *ctx->ptr++;
 600
 601     case '}':
 602         *(const WCHAR**)lval = ctx->ptr++;
 603         return '}';
 604
 605     case '.':
 606         if(ctx->ptr+1 < ctx->end && isdigitW(ctx->ptr[1])) {
 607             double n;
 608             HRESULT hres;
 609             hres = parse_decimal(&ctx->ptr, ctx->end, &n);
 610             if(FAILED(hres)) {
 611                 lex_error(ctx, hres);
 612                 return -1;
 613             }
 614             *(literal_t**)lval = new_double_literal(ctx, n);
 615             return tNumericLiteral;
 616         }
 617         ctx->ptr++;
 618         return '.';
 619
 620     case '<':
 621         if(++ctx->ptr == ctx->end) {
 622             *(int*)lval = EXPR_LESS;
 623             return tRelOper;
 624         }
 625
 626         switch(*ctx->ptr) {
 627         case '=':  /* <= */
 628             ctx->ptr++;
 629             *(int*)lval = EXPR_LESSEQ;
 630             return tRelOper;
 631         case '<':  /* << */
 632             if(++ctx->ptr < ctx->end && *ctx->ptr == '=') { /* <<= */
 633                 ctx->ptr++;
 634                 *(int*)lval = EXPR_ASSIGNLSHIFT;
 635                 return tAssignOper;
 636             }
 637             *(int*)lval = EXPR_LSHIFT;
 638             return tShiftOper;
 639         default: /* < */
 640             *(int*)lval = EXPR_LESS;
 641             return tRelOper;
 642         }
 643
 644     case '>':
 645         if(++ctx->ptr == ctx->end) { /* > */
 646             *(int*)lval = EXPR_GREATER;
 647             return tRelOper;
 648         }
 649
 650         switch(*ctx->ptr) {
 651         case '=':  /* >= */
 652             ctx->ptr++;
 653             *(int*)lval = EXPR_GREATEREQ;
 654             return tRelOper;
 655         case '>':  /* >> */
 656             if(++ctx->ptr < ctx->end) {
 657                 if(*ctx->ptr == '=') {  /* >>= */
 658                     ctx->ptr++;
 659                     *(int*)lval = EXPR_ASSIGNRSHIFT;
 660                     return tAssignOper;
 661                 }
 662                 if(*ctx->ptr == '>') {  /* >>> */
 663                     if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* >>>= */
 664                         ctx->ptr++;
 665                         *(int*)lval = EXPR_ASSIGNRRSHIFT;
 666                         return tAssignOper;
 667                     }
 668                     *(int*)lval = EXPR_RRSHIFT;
 669                     return tRelOper;
 670                 }
 671             }
 672             *(int*)lval = EXPR_RSHIFT;
 673             return tShiftOper;
 674         default:
 675             *(int*)lval = EXPR_GREATER;
 676             return tRelOper;
 677         }
 678
 679     case '+':
 680         ctx->ptr++;
 681         if(ctx->ptr < ctx->end) {
 682             switch(*ctx->ptr) {
 683             case '+':  /* ++ */
 684                 ctx->ptr++;
 685                 return tINC;
 686             case '=':  /* += */
 687                 ctx->ptr++;
 688                 *(int*)lval = EXPR_ASSIGNADD;
 689                 return tAssignOper;
 690             }
 691         }
 692         return '+';
 693
 694     case '-':
 695         ctx->ptr++;
 696         if(ctx->ptr < ctx->end) {
 697             switch(*ctx->ptr) {
 698             case '-':  /* -- or --> */
 699                 ctx->ptr++;
 700                 if(ctx->is_html && ctx->nl && ctx->ptr < ctx->end && *ctx->ptr == '>') {
 701                     ctx->ptr++;
 702                     return tHTMLCOMMENT;
 703                 }
 704                 return tDEC;
 705             case '=':  /* -= */
 706                 ctx->ptr++;
 707                 *(int*)lval = EXPR_ASSIGNSUB;
 708                 return tAssignOper;
 709             }
 710         }
 711         return '-';
 712
 713     case '*':
 714         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') { /* *= */
 715             ctx->ptr++;
 716             *(int*)lval = EXPR_ASSIGNMUL;
 717             return tAssignOper;
 718         }
 719         return '*';
 720
 721     case '%':
 722         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') { /* %= */
 723             ctx->ptr++;
 724             *(int*)lval = EXPR_ASSIGNMOD;
 725             return tAssignOper;
 726         }
 727         return '%';
 728
 729     case '&':
 730         if(++ctx->ptr < ctx->end) {
 731             switch(*ctx->ptr) {
 732             case '=':  /* &= */
 733                 ctx->ptr++;
 734                 *(int*)lval = EXPR_ASSIGNAND;
 735                 return tAssignOper;
 736             case '&':  /* && */
 737                 ctx->ptr++;
 738                 return tANDAND;
 739             }
 740         }
 741         return '&';
 742
 743     case '|':
 744         if(++ctx->ptr < ctx->end) {
 745             switch(*ctx->ptr) {
 746             case '=':  /* |= */
 747                 ctx->ptr++;
 748                 *(int*)lval = EXPR_ASSIGNOR;
 749                 return tAssignOper;
 750             case '|':  /* || */
 751                 ctx->ptr++;
 752                 return tOROR;
 753             }
 754         }
 755         return '|';
 756
 757     case '^':
 758         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* ^= */
 759             ctx->ptr++;
 760             *(int*)lval = EXPR_ASSIGNXOR;
 761             return tAssignOper;
 762         }
 763         return '^';
 764
 765     case '!':
 766         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* != */
 767             if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* !== */
 768                 ctx->ptr++;
 769                 *(int*)lval = EXPR_NOTEQEQ;
 770                 return tEqOper;
 771             }
 772             *(int*)lval = EXPR_NOTEQ;
 773             return tEqOper;
 774         }
 775         return '!';
 776
 777     case '=':
 778         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* == */
 779             if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* === */
 780                 ctx->ptr++;
 781                 *(int*)lval = EXPR_EQEQ;
 782                 return tEqOper;
 783             }
 784             *(int*)lval = EXPR_EQ;
 785             return tEqOper;
 786         }
 787         return '=';
 788
 789     case '/':
 790         if(++ctx->ptr < ctx->end) {
 791             if(*ctx->ptr == '=') {  /* /= */
 792                 ctx->ptr++;
 793                 *(int*)lval = EXPR_ASSIGNDIV;
 794                 return kDIVEQ;
 795             }
 796         }
 797         return '/';
 798
 799     case ':':
 800         if(++ctx->ptr < ctx->end && *ctx->ptr == ':') {
 801             ctx->ptr++;
 802             return kDCOL;
 803         }
 804         return ':';
 805
 806     case '\"':
 807     case '\'':
 808         return parse_string_literal(ctx, lval, *ctx->ptr);
 809
 810     case '_':
 811     case '$':
 812         return parse_identifier(ctx, lval);
 813
 814     case '@':
 815         return '@';
 816     }
 817
 818     WARN("unexpected char '%c' %d\n", *ctx->ptr, *ctx->ptr);
 819     return 0;
 820 }
 821
 822 struct _cc_var_t {
 823     ccval_t val;
 824     struct _cc_var_t *next;
 825     unsigned name_len;
 826     WCHAR name[0];
 827 };
 828
 829 void release_cc(cc_ctx_t *cc)
 830 {
 831     cc_var_t *iter, *next;
 832
 833     for(iter = cc->vars; iter; iter = next) {
 834         next = iter->next;
 835         heap_free(iter);
 836     }
 837
 838     heap_free(cc);
 839 }
 840
 841 static BOOL new_cc_var(cc_ctx_t *cc, const WCHAR *name, int len, ccval_t v)
 842 {
 843     cc_var_t *new_v;
 844
 845     if(len == -1)
 846         len = strlenW(name);
 847
 848     new_v = heap_alloc(sizeof(cc_var_t) + (len+1)*sizeof(WCHAR));
 849     if(!new_v)
 850         return FALSE;
 851
 852     new_v->val = v;
 853     memcpy(new_v->name, name, (len+1)*sizeof(WCHAR));
 854     new_v->name_len = len;
 855     new_v->next = cc->vars;
 856     cc->vars = new_v;
 857     return TRUE;
 858 }
 859
 860 static cc_var_t *find_cc_var(cc_ctx_t *cc, const WCHAR *name, unsigned name_len)
 861 {
 862     cc_var_t *iter;
 863
 864     for(iter = cc->vars; iter; iter = iter->next) {
 865         if(iter->name_len == name_len && !memcmp(iter->name, name, name_len*sizeof(WCHAR)))
 866             return iter;
 867     }
 868
 869     return NULL;
 870 }
 871
 872 static BOOL init_cc(parser_ctx_t *ctx)
 873 {
 874     cc_ctx_t *cc;
 875
 876     static const WCHAR _win32W[] = {'_','w','i','n','3','2',0};
 877     static const WCHAR _win64W[] = {'_','w','i','n','6','4',0};
 878     static const WCHAR _x86W[] = {'_','x','8','6',0};
 879     static const WCHAR _amd64W[] = {'_','a','m','d','6','4',0};
 880     static const WCHAR _jscriptW[] = {'_','j','s','c','r','i','p','t',0};
 881     static const WCHAR _jscript_buildW[] = {'_','j','s','c','r','i','p','t','_','b','u','i','l','d',0};
 882     static const WCHAR _jscript_versionW[] = {'_','j','s','c','r','i','p','t','_','v','e','r','s','i','o','n',0};
 883
 884     if(ctx->script->cc)
 885         return TRUE;
 886
 887     cc = heap_alloc(sizeof(cc_ctx_t));
 888     if(!cc) {
 889         lex_error(ctx, E_OUTOFMEMORY);
 890         return FALSE;
 891     }
 892
 893     cc->vars = NULL;
 894
 895     if(!new_cc_var(cc, _jscriptW, -1, ccval_bool(TRUE))
 896        || !new_cc_var(cc, sizeof(void*) == 8 ? _win64W : _win32W, -1, ccval_bool(TRUE))
 897        || !new_cc_var(cc, sizeof(void*) == 8 ? _amd64W : _x86W, -1, ccval_bool(TRUE))
 898        || !new_cc_var(cc, _jscript_versionW, -1, ccval_num(JSCRIPT_MAJOR_VERSION + (DOUBLE)JSCRIPT_MINOR_VERSION/10.0))
 899        || !new_cc_var(cc, _jscript_buildW, -1, ccval_num(JSCRIPT_BUILD_VERSION))) {
 900         release_cc(cc);
 901         lex_error(ctx, E_OUTOFMEMORY);
 902         return FALSE;
 903     }
 904
 905     ctx->script->cc = cc;
 906     return TRUE;
 907 }
 908
 909 static BOOL parse_cc_identifier(parser_ctx_t *ctx, const WCHAR **ret, unsigned *ret_len)
 910 {
 911     if(*ctx->ptr != '@') {
 912         lex_error(ctx, JS_E_EXPECTED_AT);
 913         return FALSE;
 914     }
 915
 916     if(!is_identifier_first_char(*++ctx->ptr)) {
 917         lex_error(ctx, JS_E_EXPECTED_IDENTIFIER);
 918         return FALSE;
 919     }
 920
 921     *ret = ctx->ptr;
 922     while(++ctx->ptr < ctx->end && is_identifier_char(*ctx->ptr));
 923     *ret_len = ctx->ptr - *ret;
 924     return TRUE;
 925 }
 926
 927 int try_parse_ccval(parser_ctx_t *ctx, ccval_t *r)
 928 {
 929     if(!skip_spaces(ctx))
 930         return -1;
 931
 932     if(isdigitW(*ctx->ptr)) {
 933         double n;
 934
 935         if(!parse_numeric_literal(ctx, &n))
 936             return -1;
 937
 938         *r = ccval_num(n);
 939         return 1;
 940     }
 941
 942     if(*ctx->ptr == '@') {
 943         const WCHAR *ident;
 944         unsigned ident_len;
 945         cc_var_t *cc_var;
 946
 947         if(!parse_cc_identifier(ctx, &ident, &ident_len))
 948             return -1;
 949
 950         cc_var = find_cc_var(ctx->script->cc, ident, ident_len);
 951         *r = cc_var ? cc_var->val : ccval_num(NAN);
 952         return 1;
 953     }
 954
 955     if(!check_keyword(ctx, trueW, NULL)) {
 956         *r = ccval_bool(TRUE);
 957         return 1;
 958     }
 959
 960     if(!check_keyword(ctx, falseW, NULL)) {
 961         *r = ccval_bool(FALSE);
 962         return 1;
 963     }
 964
 965     return 0;
 966 }
 967
 968 static int skip_code(parser_ctx_t *ctx, BOOL exec_else)
 969 {
 970     int if_depth = 1;
 971     const WCHAR *ptr;
 972
 973     while(1) {
 974         ptr = strchrW(ctx->ptr, '@');
 975         if(!ptr) {
 976             WARN("No @end\n");
 977             return lex_error(ctx, JS_E_EXPECTED_CCEND);
 978         }
 979         ctx->ptr = ptr+1;
 980
 981         if(!check_keyword(ctx, endW, NULL)) {
 982             if(--if_depth)
 983                 continue;
 984             return 0;
 985         }
 986
 987         if(exec_else && !check_keyword(ctx, elifW, NULL)) {
 988             if(if_depth > 1)
 989                 continue;
 990
 991             if(!skip_spaces(ctx) || *ctx->ptr != '(')
 992                 return lex_error(ctx, JS_E_MISSING_LBRACKET);
 993
 994             if(!parse_cc_expr(ctx))
 995                 return -1;
 996
 997             if(!get_ccbool(ctx->ccval))
 998                 continue; /* skip block of code */
 999
1000             /* continue parsing */
1001             ctx->cc_if_depth++;
1002             return 0;
1003         }
1004
1005         if(exec_else && !check_keyword(ctx, elseW, NULL)) {
1006             if(if_depth > 1)
1007                 continue;
1008
1009             /* parse else block */
1010             ctx->cc_if_depth++;
1011             return 0;
1012         }
1013
1014         if(!check_keyword(ctx, ifW, NULL)) {
1015             if_depth++;
1016             continue;
1017         }
1018
1019         ctx->ptr++;
1020     }
1021 }
1022
1023 static int cc_token(parser_ctx_t *ctx, void *lval)
1024 {
1025     unsigned id_len = 0;
1026     cc_var_t *var;
1027
1028     static const WCHAR cc_onW[] = {'c','c','_','o','n',0};
1029     static const WCHAR setW[] = {'s','e','t',0};
1030
1031     ctx->ptr++;
1032
1033     if(!check_keyword(ctx, cc_onW, NULL))
1034         return init_cc(ctx) ? 0 : -1;
1035
1036     if(!check_keyword(ctx, setW, NULL)) {
1037         const WCHAR *ident;
1038         unsigned ident_len;
1039         cc_var_t *var;
1040
1041         if(!init_cc(ctx))
1042             return -1;
1043
1044         if(!skip_spaces(ctx))
1045             return lex_error(ctx, JS_E_EXPECTED_AT);
1046
1047         if(!parse_cc_identifier(ctx, &ident, &ident_len))
1048             return -1;
1049
1050         if(!skip_spaces(ctx) || *ctx->ptr != '=')
1051             return lex_error(ctx, JS_E_EXPECTED_ASSIGN);
1052         ctx->ptr++;
1053
1054         if(!parse_cc_expr(ctx)) {
1055             WARN("parsing CC expression failed\n");
1056             return -1;
1057         }
1058
1059         var = find_cc_var(ctx->script->cc, ident, ident_len);
1060         if(var) {
1061             var->val = ctx->ccval;
1062         }else {
1063             if(!new_cc_var(ctx->script->cc, ident, ident_len, ctx->ccval))
1064                 return lex_error(ctx, E_OUTOFMEMORY);
1065         }
1066
1067         return 0;
1068     }
1069
1070     if(!check_keyword(ctx, ifW, NULL)) {
1071         if(!init_cc(ctx))
1072             return -1;
1073
1074         if(!skip_spaces(ctx) || *ctx->ptr != '(')
1075             return lex_error(ctx, JS_E_MISSING_LBRACKET);
1076
1077         if(!parse_cc_expr(ctx))
1078             return -1;
1079
1080         if(get_ccbool(ctx->ccval)) {
1081             /* continue parsing block inside if */
1082             ctx->cc_if_depth++;
1083             return 0;
1084         }
1085
1086         return skip_code(ctx, TRUE);
1087     }
1088
1089     if(!check_keyword(ctx, elifW, NULL) || !check_keyword(ctx, elseW, NULL)) {
1090         if(!ctx->cc_if_depth)
1091             return lex_error(ctx, JS_E_SYNTAX);
1092
1093         return skip_code(ctx, FALSE);
1094     }
1095
1096     if(!check_keyword(ctx, endW, NULL)) {
1097         if(!ctx->cc_if_depth)
1098             return lex_error(ctx, JS_E_SYNTAX);
1099
1100         ctx->cc_if_depth--;
1101         return 0;
1102     }
1103
1104     if(!ctx->script->cc)
1105         return lex_error(ctx, JS_E_DISABLED_CC);
1106
1107     while(ctx->ptr+id_len < ctx->end && is_identifier_char(ctx->ptr[id_len]))
1108         id_len++;
1109     if(!id_len)
1110         return '@';
1111
1112     TRACE("var %s\n", debugstr_wn(ctx->ptr, id_len));
1113
1114     var = find_cc_var(ctx->script->cc, ctx->ptr, id_len);
1115     ctx->ptr += id_len;
1116     if(!var || var->val.is_num) {
1117         *(literal_t**)lval = new_double_literal(ctx, var ? var->val.u.n : NAN);
1118         return tNumericLiteral;
1119     }
1120
1121     *(literal_t**)lval = new_boolean_literal(ctx, var->val.u.b);
1122     return tBooleanLiteral;
1123 }
1124
1125 int parser_lex(void *lval, parser_ctx_t *ctx)
1126 {
1127     int ret;
1128
1129     ctx->nl = ctx->ptr == ctx->begin;
1130
1131     do {
1132         ret = next_token(ctx, lval);
1133     } while(ret == '@' && !(ret = cc_token(ctx, lval)));
1134
1135     return ret;
1136 }
1137
1138 literal_t *parse_regexp(parser_ctx_t *ctx)
1139 {
1140     const WCHAR *re, *flags_ptr;
1141     BOOL in_class = FALSE;
1142     DWORD re_len, flags;
1143     literal_t *ret;
1144     HRESULT hres;
1145
1146     TRACE("\n");
1147
1148     while(*--ctx->ptr != '/');
1149
1150     /* Simple regexp pre-parser; '/' if used in char class does not terminate regexp literal */
1151     re = ++ctx->ptr;
1152     while(ctx->ptr < ctx->end) {
1153         if(*ctx->ptr == '\\') {
1154             if(++ctx->ptr == ctx->end)
1155                 break;
1156         }else if(in_class) {
1157             if(*ctx->ptr == '\n')
1158                 break;
1159             if(*ctx->ptr == ']')
1160                 in_class = FALSE;
1161         }else {
1162             if(*ctx->ptr == '/')
1163                 break;
1164
1165             if(*ctx->ptr == '[')
1166                 in_class = TRUE;
1167         }
1168         ctx->ptr++;
1169     }
1170
1171     if(ctx->ptr == ctx->end || *ctx->ptr != '/') {
1172         WARN("pre-parsing failed\n");
1173         return NULL;
1174     }
1175
1176     re_len = ctx->ptr-re;
1177
1178     flags_ptr = ++ctx->ptr;
1179     while(ctx->ptr < ctx->end && isalnumW(*ctx->ptr))
1180         ctx->ptr++;
1181
1182     hres = parse_regexp_flags(flags_ptr, ctx->ptr-flags_ptr, &flags);
1183     if(FAILED(hres))
1184         return NULL;
1185
1186     ret = parser_alloc(ctx, sizeof(literal_t));
1187     ret->type = LT_REGEXP;
1188     ret->u.regexp.str = re;
1189     ret->u.regexp.str_len = re_len;
1190     ret->u.regexp.flags = flags;
1191     return ret;
1192 }