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 static 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 static 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 static BOOL parse_double_literal(parser_ctx_t *ctx, LONG int_part, double *ret)
 410 {
 411     LONGLONG d, hlp;
 412     int exp = 0;
 413
 414     d = int_part;
 415     while(ctx->ptr < ctx->end && isdigitW(*ctx->ptr)) {
 416         hlp = d*10 + *(ctx->ptr++) - '0';
 417         if(d>MAXLONGLONG/10 || hlp<0) {
 418             exp++;
 419             break;
 420         }
 421         else
 422             d = hlp;
 423     }
 424     while(ctx->ptr < ctx->end && isdigitW(*ctx->ptr)) {
 425         exp++;
 426         ctx->ptr++;
 427     }
 428
 429     if(*ctx->ptr == '.') {
 430         ctx->ptr++;
 431
 432         while(ctx->ptr < ctx->end && isdigitW(*ctx->ptr)) {
 433             hlp = d*10 + *(ctx->ptr++) - '0';
 434             if(d>MAXLONGLONG/10 || hlp<0)
 435                 break;
 436
 437             d = hlp;
 438             exp--;
 439         }
 440         while(ctx->ptr < ctx->end && isdigitW(*ctx->ptr))
 441             ctx->ptr++;
 442     }
 443
 444     if(ctx->ptr < ctx->end && (*ctx->ptr == 'e' || *ctx->ptr == 'E')) {
 445         int sign = 1, e = 0;
 446
 447         ctx->ptr++;
 448         if(ctx->ptr < ctx->end) {
 449             if(*ctx->ptr == '+') {
 450                 ctx->ptr++;
 451             }else if(*ctx->ptr == '-') {
 452                 sign = -1;
 453                 ctx->ptr++;
 454             }else if(!isdigitW(*ctx->ptr)) {
 455                 WARN("Expected exponent part\n");
 456                 lex_error(ctx, E_FAIL);
 457                 return FALSE;
 458             }
 459         }
 460
 461         if(ctx->ptr == ctx->end) {
 462             WARN("unexpected end of file\n");
 463             lex_error(ctx, E_FAIL);
 464             return FALSE;
 465         }
 466
 467         while(ctx->ptr < ctx->end && isdigitW(*ctx->ptr)) {
 468             if(e > INT_MAX/10 || (e = e*10 + *ctx->ptr++ - '0')<0)
 469                 e = INT_MAX;
 470         }
 471         e *= sign;
 472
 473         if(exp<0 && e<0 && e+exp>0) exp = INT_MIN;
 474         else if(exp>0 && e>0 && e+exp<0) exp = INT_MAX;
 475         else exp += e;
 476     }
 477
 478     if(is_identifier_char(*ctx->ptr)) {
 479         WARN("wrong char after zero\n");
 480         lex_error(ctx, JS_E_MISSING_SEMICOLON);
 481         return FALSE;
 482     }
 483
 484     *ret = exp>=0 ? d*pow(10, exp) : d/pow(10, -exp);
 485     return TRUE;
 486 }
 487
 488 static BOOL parse_numeric_literal(parser_ctx_t *ctx, double *ret)
 489 {
 490     LONG l, d;
 491
 492     l = *ctx->ptr++ - '0';
 493     if(!l) {
 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     return parse_double_literal(ctx, l, ret);
 550 }
 551
 552 static int next_token(parser_ctx_t *ctx, void *lval)
 553 {
 554     do {
 555         if(!skip_spaces(ctx))
 556             return tEOF;
 557     }while(skip_comment(ctx) || skip_html_comment(ctx));
 558
 559     if(ctx->implicit_nl_semicolon) {
 560         if(ctx->nl)
 561             return ';';
 562         ctx->implicit_nl_semicolon = FALSE;
 563     }
 564
 565     if(isalphaW(*ctx->ptr)) {
 566         int ret = check_keywords(ctx, lval);
 567         if(ret)
 568             return ret;
 569
 570         return parse_identifier(ctx, lval);
 571     }
 572
 573     if(isdigitW(*ctx->ptr)) {
 574         double n;
 575
 576         if(!parse_numeric_literal(ctx, &n))
 577             return -1;
 578
 579         *(literal_t**)lval = new_double_literal(ctx, n);
 580         return tNumericLiteral;
 581     }
 582
 583     switch(*ctx->ptr) {
 584     case '{':
 585     case '(':
 586     case ')':
 587     case '[':
 588     case ']':
 589     case ';':
 590     case ',':
 591     case '~':
 592     case '?':
 593     case ':':
 594         return *ctx->ptr++;
 595
 596     case '}':
 597         *(const WCHAR**)lval = ctx->ptr++;
 598         return '}';
 599
 600     case '.':
 601         if(++ctx->ptr < ctx->end && isdigitW(*ctx->ptr)) {
 602             double n;
 603             if(!parse_double_literal(ctx, 0, &n))
 604                 return -1;
 605             *(literal_t**)lval = new_double_literal(ctx, n);
 606             return tNumericLiteral;
 607         }
 608         return '.';
 609
 610     case '<':
 611         if(++ctx->ptr == ctx->end) {
 612             *(int*)lval = EXPR_LESS;
 613             return tRelOper;
 614         }
 615
 616         switch(*ctx->ptr) {
 617         case '=':  /* <= */
 618             ctx->ptr++;
 619             *(int*)lval = EXPR_LESSEQ;
 620             return tRelOper;
 621         case '<':  /* << */
 622             if(++ctx->ptr < ctx->end && *ctx->ptr == '=') { /* <<= */
 623                 ctx->ptr++;
 624                 *(int*)lval = EXPR_ASSIGNLSHIFT;
 625                 return tAssignOper;
 626             }
 627             *(int*)lval = EXPR_LSHIFT;
 628             return tShiftOper;
 629         default: /* < */
 630             *(int*)lval = EXPR_LESS;
 631             return tRelOper;
 632         }
 633
 634     case '>':
 635         if(++ctx->ptr == ctx->end) { /* > */
 636             *(int*)lval = EXPR_GREATER;
 637             return tRelOper;
 638         }
 639
 640         switch(*ctx->ptr) {
 641         case '=':  /* >= */
 642             ctx->ptr++;
 643             *(int*)lval = EXPR_GREATEREQ;
 644             return tRelOper;
 645         case '>':  /* >> */
 646             if(++ctx->ptr < ctx->end) {
 647                 if(*ctx->ptr == '=') {  /* >>= */
 648                     ctx->ptr++;
 649                     *(int*)lval = EXPR_ASSIGNRSHIFT;
 650                     return tAssignOper;
 651                 }
 652                 if(*ctx->ptr == '>') {  /* >>> */
 653                     if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* >>>= */
 654                         ctx->ptr++;
 655                         *(int*)lval = EXPR_ASSIGNRRSHIFT;
 656                         return tAssignOper;
 657                     }
 658                     *(int*)lval = EXPR_RRSHIFT;
 659                     return tRelOper;
 660                 }
 661             }
 662             *(int*)lval = EXPR_RSHIFT;
 663             return tShiftOper;
 664         default:
 665             *(int*)lval = EXPR_GREATER;
 666             return tRelOper;
 667         }
 668
 669     case '+':
 670         ctx->ptr++;
 671         if(ctx->ptr < ctx->end) {
 672             switch(*ctx->ptr) {
 673             case '+':  /* ++ */
 674                 ctx->ptr++;
 675                 return tINC;
 676             case '=':  /* += */
 677                 ctx->ptr++;
 678                 *(int*)lval = EXPR_ASSIGNADD;
 679                 return tAssignOper;
 680             }
 681         }
 682         return '+';
 683
 684     case '-':
 685         ctx->ptr++;
 686         if(ctx->ptr < ctx->end) {
 687             switch(*ctx->ptr) {
 688             case '-':  /* -- or --> */
 689                 ctx->ptr++;
 690                 if(ctx->is_html && ctx->nl && ctx->ptr < ctx->end && *ctx->ptr == '>') {
 691                     ctx->ptr++;
 692                     return tHTMLCOMMENT;
 693                 }
 694                 return tDEC;
 695             case '=':  /* -= */
 696                 ctx->ptr++;
 697                 *(int*)lval = EXPR_ASSIGNSUB;
 698                 return tAssignOper;
 699             }
 700         }
 701         return '-';
 702
 703     case '*':
 704         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') { /* *= */
 705             ctx->ptr++;
 706             *(int*)lval = EXPR_ASSIGNMUL;
 707             return tAssignOper;
 708         }
 709         return '*';
 710
 711     case '%':
 712         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') { /* %= */
 713             ctx->ptr++;
 714             *(int*)lval = EXPR_ASSIGNMOD;
 715             return tAssignOper;
 716         }
 717         return '%';
 718
 719     case '&':
 720         if(++ctx->ptr < ctx->end) {
 721             switch(*ctx->ptr) {
 722             case '=':  /* &= */
 723                 ctx->ptr++;
 724                 *(int*)lval = EXPR_ASSIGNAND;
 725                 return tAssignOper;
 726             case '&':  /* && */
 727                 ctx->ptr++;
 728                 return tANDAND;
 729             }
 730         }
 731         return '&';
 732
 733     case '|':
 734         if(++ctx->ptr < ctx->end) {
 735             switch(*ctx->ptr) {
 736             case '=':  /* |= */
 737                 ctx->ptr++;
 738                 *(int*)lval = EXPR_ASSIGNOR;
 739                 return tAssignOper;
 740             case '|':  /* || */
 741                 ctx->ptr++;
 742                 return tOROR;
 743             }
 744         }
 745         return '|';
 746
 747     case '^':
 748         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* ^= */
 749             ctx->ptr++;
 750             *(int*)lval = EXPR_ASSIGNXOR;
 751             return tAssignOper;
 752         }
 753         return '^';
 754
 755     case '!':
 756         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* != */
 757             if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* !== */
 758                 ctx->ptr++;
 759                 *(int*)lval = EXPR_NOTEQEQ;
 760                 return tEqOper;
 761             }
 762             *(int*)lval = EXPR_NOTEQ;
 763             return tEqOper;
 764         }
 765         return '!';
 766
 767     case '=':
 768         if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* == */
 769             if(++ctx->ptr < ctx->end && *ctx->ptr == '=') {  /* === */
 770                 ctx->ptr++;
 771                 *(int*)lval = EXPR_EQEQ;
 772                 return tEqOper;
 773             }
 774             *(int*)lval = EXPR_EQ;
 775             return tEqOper;
 776         }
 777         return '=';
 778
 779     case '/':
 780         if(++ctx->ptr < ctx->end) {
 781             if(*ctx->ptr == '=') {  /* /= */
 782                 ctx->ptr++;
 783                 *(int*)lval = EXPR_ASSIGNDIV;
 784                 return kDIVEQ;
 785             }
 786         }
 787         return '/';
 788
 789     case '\"':
 790     case '\'':
 791         return parse_string_literal(ctx, lval, *ctx->ptr);
 792
 793     case '_':
 794     case '$':
 795         return parse_identifier(ctx, lval);
 796
 797     case '@':
 798         return '@';
 799     }
 800
 801     WARN("unexpected char '%c' %d\n", *ctx->ptr, *ctx->ptr);
 802     return 0;
 803 }
 804
 805 struct _cc_var_t {
 806     ccval_t val;
 807     struct _cc_var_t *next;
 808     unsigned name_len;
 809     WCHAR name[0];
 810 };
 811
 812 void release_cc(cc_ctx_t *cc)
 813 {
 814     cc_var_t *iter, *next;
 815
 816     for(iter = cc->vars; iter; iter = next) {
 817         next = iter->next;
 818         heap_free(iter);
 819     }
 820
 821     heap_free(cc);
 822 }
 823
 824 static BOOL new_cc_var(cc_ctx_t *cc, const WCHAR *name, int len, ccval_t v)
 825 {
 826     cc_var_t *new_v;
 827
 828     if(len == -1)
 829         len = strlenW(name);
 830
 831     new_v = heap_alloc(sizeof(cc_var_t) + (len+1)*sizeof(WCHAR));
 832     if(!new_v)
 833         return FALSE;
 834
 835     new_v->val = v;
 836     memcpy(new_v->name, name, (len+1)*sizeof(WCHAR));
 837     new_v->name_len = len;
 838     new_v->next = cc->vars;
 839     cc->vars = new_v;
 840     return TRUE;
 841 }
 842
 843 static cc_var_t *find_cc_var(cc_ctx_t *cc, const WCHAR *name, unsigned name_len)
 844 {
 845     cc_var_t *iter;
 846
 847     for(iter = cc->vars; iter; iter = iter->next) {
 848         if(iter->name_len == name_len && !memcmp(iter->name, name, name_len*sizeof(WCHAR)))
 849             return iter;
 850     }
 851
 852     return NULL;
 853 }
 854
 855 static BOOL init_cc(parser_ctx_t *ctx)
 856 {
 857     cc_ctx_t *cc;
 858
 859     static const WCHAR _win32W[] = {'_','w','i','n','3','2',0};
 860     static const WCHAR _win64W[] = {'_','w','i','n','6','4',0};
 861     static const WCHAR _x86W[] = {'_','x','8','6',0};
 862     static const WCHAR _amd64W[] = {'_','a','m','d','6','4',0};
 863     static const WCHAR _jscriptW[] = {'_','j','s','c','r','i','p','t',0};
 864     static const WCHAR _jscript_buildW[] = {'_','j','s','c','r','i','p','t','_','b','u','i','l','d',0};
 865     static const WCHAR _jscript_versionW[] = {'_','j','s','c','r','i','p','t','_','v','e','r','s','i','o','n',0};
 866
 867     if(ctx->script->cc)
 868         return TRUE;
 869
 870     cc = heap_alloc(sizeof(cc_ctx_t));
 871     if(!cc) {
 872         lex_error(ctx, E_OUTOFMEMORY);
 873         return FALSE;
 874     }
 875
 876     cc->vars = NULL;
 877
 878     if(!new_cc_var(cc, _jscriptW, -1, ccval_bool(TRUE))
 879        || !new_cc_var(cc, sizeof(void*) == 8 ? _win64W : _win32W, -1, ccval_bool(TRUE))
 880        || !new_cc_var(cc, sizeof(void*) == 8 ? _amd64W : _x86W, -1, ccval_bool(TRUE))
 881        || !new_cc_var(cc, _jscript_versionW, -1, ccval_num(JSCRIPT_MAJOR_VERSION + (DOUBLE)JSCRIPT_MINOR_VERSION/10.0))
 882        || !new_cc_var(cc, _jscript_buildW, -1, ccval_num(JSCRIPT_BUILD_VERSION))) {
 883         release_cc(cc);
 884         lex_error(ctx, E_OUTOFMEMORY);
 885         return FALSE;
 886     }
 887
 888     ctx->script->cc = cc;
 889     return TRUE;
 890 }
 891
 892 static BOOL parse_cc_identifier(parser_ctx_t *ctx, const WCHAR **ret, unsigned *ret_len)
 893 {
 894     if(*ctx->ptr != '@') {
 895         lex_error(ctx, JS_E_EXPECTED_AT);
 896         return FALSE;
 897     }
 898
 899     if(!is_identifier_first_char(*++ctx->ptr)) {
 900         lex_error(ctx, JS_E_EXPECTED_IDENTIFIER);
 901         return FALSE;
 902     }
 903
 904     *ret = ctx->ptr;
 905     while(++ctx->ptr < ctx->end && is_identifier_char(*ctx->ptr));
 906     *ret_len = ctx->ptr - *ret;
 907     return TRUE;
 908 }
 909
 910 int try_parse_ccval(parser_ctx_t *ctx, ccval_t *r)
 911 {
 912     if(!skip_spaces(ctx))
 913         return -1;
 914
 915     if(isdigitW(*ctx->ptr)) {
 916         double n;
 917
 918         if(!parse_numeric_literal(ctx, &n))
 919             return -1;
 920
 921         *r = ccval_num(n);
 922         return 1;
 923     }
 924
 925     if(*ctx->ptr == '@') {
 926         const WCHAR *ident;
 927         unsigned ident_len;
 928         cc_var_t *cc_var;
 929
 930         if(!parse_cc_identifier(ctx, &ident, &ident_len))
 931             return -1;
 932
 933         cc_var = find_cc_var(ctx->script->cc, ident, ident_len);
 934         *r = cc_var ? cc_var->val : ccval_num(NAN);
 935         return 1;
 936     }
 937
 938     if(!check_keyword(ctx, trueW, NULL)) {
 939         *r = ccval_bool(TRUE);
 940         return 1;
 941     }
 942
 943     if(!check_keyword(ctx, falseW, NULL)) {
 944         *r = ccval_bool(FALSE);
 945         return 1;
 946     }
 947
 948     return 0;
 949 }
 950
 951 static int skip_code(parser_ctx_t *ctx, BOOL exec_else)
 952 {
 953     int if_depth = 1;
 954     const WCHAR *ptr;
 955
 956     while(1) {
 957         ptr = strchrW(ctx->ptr, '@');
 958         if(!ptr) {
 959             WARN("No @end\n");
 960             return lex_error(ctx, JS_E_EXPECTED_CCEND);
 961         }
 962         ctx->ptr = ptr+1;
 963
 964         if(!check_keyword(ctx, endW, NULL)) {
 965             if(--if_depth)
 966                 continue;
 967             return 0;
 968         }
 969
 970         if(exec_else && !check_keyword(ctx, elifW, NULL)) {
 971             if(if_depth > 1)
 972                 continue;
 973
 974             if(!skip_spaces(ctx) || *ctx->ptr != '(')
 975                 return lex_error(ctx, JS_E_MISSING_LBRACKET);
 976
 977             if(!parse_cc_expr(ctx))
 978                 return -1;
 979
 980             if(!get_ccbool(ctx->ccval))
 981                 continue; /* skip block of code */
 982
 983             /* continue parsing */
 984             ctx->cc_if_depth++;
 985             return 0;
 986         }
 987
 988         if(exec_else && !check_keyword(ctx, elseW, NULL)) {
 989             if(if_depth > 1)
 990                 continue;
 991
 992             /* parse else block */
 993             ctx->cc_if_depth++;
 994             return 0;
 995         }
 996
 997         if(!check_keyword(ctx, ifW, NULL)) {
 998             if_depth++;
 999             continue;
1000         }
1001
1002         ctx->ptr++;
1003     }
1004 }
1005
1006 static int cc_token(parser_ctx_t *ctx, void *lval)
1007 {
1008     unsigned id_len = 0;
1009     cc_var_t *var;
1010
1011     static const WCHAR cc_onW[] = {'c','c','_','o','n',0};
1012     static const WCHAR setW[] = {'s','e','t',0};
1013
1014     ctx->ptr++;
1015
1016     if(!check_keyword(ctx, cc_onW, NULL))
1017         return init_cc(ctx) ? 0 : -1;
1018
1019     if(!check_keyword(ctx, setW, NULL)) {
1020         const WCHAR *ident;
1021         unsigned ident_len;
1022         cc_var_t *var;
1023
1024         if(!init_cc(ctx))
1025             return -1;
1026
1027         if(!skip_spaces(ctx))
1028             return lex_error(ctx, JS_E_EXPECTED_AT);
1029
1030         if(!parse_cc_identifier(ctx, &ident, &ident_len))
1031             return -1;
1032
1033         if(!skip_spaces(ctx) || *ctx->ptr != '=')
1034             return lex_error(ctx, JS_E_EXPECTED_ASSIGN);
1035         ctx->ptr++;
1036
1037         if(!parse_cc_expr(ctx)) {
1038             WARN("parsing CC expression failed\n");
1039             return -1;
1040         }
1041
1042         var = find_cc_var(ctx->script->cc, ident, ident_len);
1043         if(var) {
1044             var->val = ctx->ccval;
1045         }else {
1046             if(!new_cc_var(ctx->script->cc, ident, ident_len, ctx->ccval))
1047                 return lex_error(ctx, E_OUTOFMEMORY);
1048         }
1049
1050         return 0;
1051     }
1052
1053     if(!check_keyword(ctx, ifW, NULL)) {
1054         if(!init_cc(ctx))
1055             return -1;
1056
1057         if(!skip_spaces(ctx) || *ctx->ptr != '(')
1058             return lex_error(ctx, JS_E_MISSING_LBRACKET);
1059
1060         if(!parse_cc_expr(ctx))
1061             return -1;
1062
1063         if(get_ccbool(ctx->ccval)) {
1064             /* continue parsing block inside if */
1065             ctx->cc_if_depth++;
1066             return 0;
1067         }
1068
1069         return skip_code(ctx, TRUE);
1070     }
1071
1072     if(!check_keyword(ctx, elifW, NULL) || !check_keyword(ctx, elseW, NULL)) {
1073         if(!ctx->cc_if_depth)
1074             return lex_error(ctx, JS_E_SYNTAX);
1075
1076         return skip_code(ctx, FALSE);
1077     }
1078
1079     if(!check_keyword(ctx, endW, NULL)) {
1080         if(!ctx->cc_if_depth)
1081             return lex_error(ctx, JS_E_SYNTAX);
1082
1083         ctx->cc_if_depth--;
1084         return 0;
1085     }
1086
1087     if(!ctx->script->cc)
1088         return lex_error(ctx, JS_E_DISABLED_CC);
1089
1090     while(ctx->ptr+id_len < ctx->end && is_identifier_char(ctx->ptr[id_len]))
1091         id_len++;
1092     if(!id_len)
1093         return '@';
1094
1095     TRACE("var %s\n", debugstr_wn(ctx->ptr, id_len));
1096
1097     var = find_cc_var(ctx->script->cc, ctx->ptr, id_len);
1098     ctx->ptr += id_len;
1099     if(!var || var->val.is_num) {
1100         *(literal_t**)lval = new_double_literal(ctx, var ? var->val.u.n : NAN);
1101         return tNumericLiteral;
1102     }
1103
1104     *(literal_t**)lval = new_boolean_literal(ctx, var->val.u.b);
1105     return tBooleanLiteral;
1106 }
1107
1108 int parser_lex(void *lval, parser_ctx_t *ctx)
1109 {
1110     int ret;
1111
1112     ctx->nl = ctx->ptr == ctx->begin;
1113
1114     do {
1115         ret = next_token(ctx, lval);
1116     } while(ret == '@' && !(ret = cc_token(ctx, lval)));
1117
1118     return ret;
1119 }
1120
1121 literal_t *parse_regexp(parser_ctx_t *ctx)
1122 {
1123     const WCHAR *re, *flags_ptr;
1124     BOOL in_class = FALSE;
1125     DWORD re_len, flags;
1126     literal_t *ret;
1127     HRESULT hres;
1128
1129     TRACE("\n");
1130
1131     while(*--ctx->ptr != '/');
1132
1133     /* Simple regexp pre-parser; '/' if used in char class does not terminate regexp literal */
1134     re = ++ctx->ptr;
1135     while(ctx->ptr < ctx->end) {
1136         if(*ctx->ptr == '\\') {
1137             if(++ctx->ptr == ctx->end)
1138                 break;
1139         }else if(in_class) {
1140             if(*ctx->ptr == '\n')
1141                 break;
1142             if(*ctx->ptr == ']')
1143                 in_class = FALSE;
1144         }else {
1145             if(*ctx->ptr == '/')
1146                 break;
1147
1148             if(*ctx->ptr == '[')
1149                 in_class = TRUE;
1150         }
1151         ctx->ptr++;
1152     }
1153
1154     if(ctx->ptr == ctx->end || *ctx->ptr != '/') {
1155         WARN("pre-parsing failed\n");
1156         return NULL;
1157     }
1158
1159     re_len = ctx->ptr-re;
1160
1161     flags_ptr = ++ctx->ptr;
1162     while(ctx->ptr < ctx->end && isalnumW(*ctx->ptr))
1163         ctx->ptr++;
1164
1165     hres = parse_regexp_flags(flags_ptr, ctx->ptr-flags_ptr, &flags);
1166     if(FAILED(hres))
1167         return NULL;
1168
1169     ret = parser_alloc(ctx, sizeof(literal_t));
1170     ret->type = LT_REGEXP;
1171     ret->u.regexp.str = re;
1172     ret->u.regexp.str_len = re_len;
1173     ret->u.regexp.flags = flags;
1174     return ret;
1175 }