preparser.cc revision b8e0da25ee8efac3bb05cd6b2730aafbd96119f4
1 2// Copyright 2010 the V8 project authors. All rights reserved. 3// Redistribution and use in source and binary forms, with or without 4// modification, are permitted provided that the following conditions are 5// met: 6// 7// * Redistributions of source code must retain the above copyright 8// notice, this list of conditions and the following disclaimer. 9// * Redistributions in binary form must reproduce the above 10// copyright notice, this list of conditions and the following 11// disclaimer in the documentation and/or other materials provided 12// with the distribution. 13// * Neither the name of Google Inc. nor the names of its 14// contributors may be used to endorse or promote products derived 15// from this software without specific prior written permission. 16// 17// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 18// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 19// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR 20// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT 21// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, 22// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT 23// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 24// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 25// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 26// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE 27// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 28 29#include "../include/v8stdint.h" 30#include "unicode.h" 31#include "globals.h" 32#include "checks.h" 33#include "allocation.h" 34#include "utils.h" 35#include "list.h" 36#include "scanner-base.h" 37#include "preparse-data.h" 38#include "preparser.h" 39 40namespace v8 { 41namespace preparser { 42 43// Preparsing checks a JavaScript program and emits preparse-data that helps 44// a later parsing to be faster. 45// See preparser-data.h for the data. 46 47// The PreParser checks that the syntax follows the grammar for JavaScript, 48// and collects some information about the program along the way. 49// The grammar check is only performed in order to understand the program 50// sufficiently to deduce some information about it, that can be used 51// to speed up later parsing. Finding errors is not the goal of pre-parsing, 52// rather it is to speed up properly written and correct programs. 53// That means that contextual checks (like a label being declared where 54// it is used) are generally omitted. 55 56namespace i = ::v8::internal; 57 58#define CHECK_OK ok); \ 59 if (!*ok) return -1; \ 60 ((void)0 61#define DUMMY ) // to make indentation work 62#undef DUMMY 63 64 65void PreParser::ReportUnexpectedToken(i::Token::Value token) { 66 // We don't report stack overflows here, to avoid increasing the 67 // stack depth even further. Instead we report it after parsing is 68 // over, in ParseProgram. 69 if (token == i::Token::ILLEGAL && stack_overflow_) { 70 return; 71 } 72 i::JavaScriptScanner::Location source_location = scanner_->location(); 73 74 // Four of the tokens are treated specially 75 switch (token) { 76 case i::Token::EOS: 77 return ReportMessageAt(source_location.beg_pos, source_location.end_pos, 78 "unexpected_eos", NULL); 79 case i::Token::NUMBER: 80 return ReportMessageAt(source_location.beg_pos, source_location.end_pos, 81 "unexpected_token_number", NULL); 82 case i::Token::STRING: 83 return ReportMessageAt(source_location.beg_pos, source_location.end_pos, 84 "unexpected_token_string", NULL); 85 case i::Token::IDENTIFIER: 86 return ReportMessageAt(source_location.beg_pos, source_location.end_pos, 87 "unexpected_token_identifier", NULL); 88 default: 89 const char* name = i::Token::String(token); 90 ReportMessageAt(source_location.beg_pos, source_location.end_pos, 91 "unexpected_token", name); 92 } 93} 94 95 96PreParser::SourceElements PreParser::ParseSourceElements(int end_token, 97 bool* ok) { 98 // SourceElements :: 99 // (Statement)* <end_token> 100 101 while (peek() != end_token) { 102 ParseStatement(CHECK_OK); 103 } 104 return kUnknownSourceElements; 105} 106 107 108PreParser::Statement PreParser::ParseStatement(bool* ok) { 109 // Statement :: 110 // Block 111 // VariableStatement 112 // EmptyStatement 113 // ExpressionStatement 114 // IfStatement 115 // IterationStatement 116 // ContinueStatement 117 // BreakStatement 118 // ReturnStatement 119 // WithStatement 120 // LabelledStatement 121 // SwitchStatement 122 // ThrowStatement 123 // TryStatement 124 // DebuggerStatement 125 126 // Note: Since labels can only be used by 'break' and 'continue' 127 // statements, which themselves are only valid within blocks, 128 // iterations or 'switch' statements (i.e., BreakableStatements), 129 // labels can be simply ignored in all other cases; except for 130 // trivial labeled break statements 'label: break label' which is 131 // parsed into an empty statement. 132 133 // Keep the source position of the statement 134 switch (peek()) { 135 case i::Token::LBRACE: 136 return ParseBlock(ok); 137 138 case i::Token::CONST: 139 case i::Token::VAR: 140 return ParseVariableStatement(ok); 141 142 case i::Token::SEMICOLON: 143 Next(); 144 return kUnknownStatement; 145 146 case i::Token::IF: 147 return ParseIfStatement(ok); 148 149 case i::Token::DO: 150 return ParseDoWhileStatement(ok); 151 152 case i::Token::WHILE: 153 return ParseWhileStatement(ok); 154 155 case i::Token::FOR: 156 return ParseForStatement(ok); 157 158 case i::Token::CONTINUE: 159 return ParseContinueStatement(ok); 160 161 case i::Token::BREAK: 162 return ParseBreakStatement(ok); 163 164 case i::Token::RETURN: 165 return ParseReturnStatement(ok); 166 167 case i::Token::WITH: 168 return ParseWithStatement(ok); 169 170 case i::Token::SWITCH: 171 return ParseSwitchStatement(ok); 172 173 case i::Token::THROW: 174 return ParseThrowStatement(ok); 175 176 case i::Token::TRY: 177 return ParseTryStatement(ok); 178 179 case i::Token::FUNCTION: 180 return ParseFunctionDeclaration(ok); 181 182 case i::Token::NATIVE: 183 return ParseNativeDeclaration(ok); 184 185 case i::Token::DEBUGGER: 186 return ParseDebuggerStatement(ok); 187 188 default: 189 return ParseExpressionOrLabelledStatement(ok); 190 } 191} 192 193 194PreParser::Statement PreParser::ParseFunctionDeclaration(bool* ok) { 195 // FunctionDeclaration :: 196 // 'function' Identifier '(' FormalParameterListopt ')' '{' FunctionBody '}' 197 Expect(i::Token::FUNCTION, CHECK_OK); 198 ParseIdentifier(CHECK_OK); 199 ParseFunctionLiteral(CHECK_OK); 200 return kUnknownStatement; 201} 202 203 204// Language extension which is only enabled for source files loaded 205// through the API's extension mechanism. A native function 206// declaration is resolved by looking up the function through a 207// callback provided by the extension. 208PreParser::Statement PreParser::ParseNativeDeclaration(bool* ok) { 209 Expect(i::Token::NATIVE, CHECK_OK); 210 Expect(i::Token::FUNCTION, CHECK_OK); 211 ParseIdentifier(CHECK_OK); 212 Expect(i::Token::LPAREN, CHECK_OK); 213 bool done = (peek() == i::Token::RPAREN); 214 while (!done) { 215 ParseIdentifier(CHECK_OK); 216 done = (peek() == i::Token::RPAREN); 217 if (!done) { 218 Expect(i::Token::COMMA, CHECK_OK); 219 } 220 } 221 Expect(i::Token::RPAREN, CHECK_OK); 222 Expect(i::Token::SEMICOLON, CHECK_OK); 223 return kUnknownStatement; 224} 225 226 227PreParser::Statement PreParser::ParseBlock(bool* ok) { 228 // Block :: 229 // '{' Statement* '}' 230 231 // Note that a Block does not introduce a new execution scope! 232 // (ECMA-262, 3rd, 12.2) 233 // 234 Expect(i::Token::LBRACE, CHECK_OK); 235 while (peek() != i::Token::RBRACE) { 236 ParseStatement(CHECK_OK); 237 } 238 Expect(i::Token::RBRACE, CHECK_OK); 239 return kUnknownStatement; 240} 241 242 243PreParser::Statement PreParser::ParseVariableStatement(bool* ok) { 244 // VariableStatement :: 245 // VariableDeclarations ';' 246 247 Statement result = ParseVariableDeclarations(true, NULL, CHECK_OK); 248 ExpectSemicolon(CHECK_OK); 249 return result; 250} 251 252 253// If the variable declaration declares exactly one non-const 254// variable, then *var is set to that variable. In all other cases, 255// *var is untouched; in particular, it is the caller's responsibility 256// to initialize it properly. This mechanism is also used for the parsing 257// of 'for-in' loops. 258PreParser::Statement PreParser::ParseVariableDeclarations(bool accept_IN, 259 int* num_decl, 260 bool* ok) { 261 // VariableDeclarations :: 262 // ('var' | 'const') (Identifier ('=' AssignmentExpression)?)+[','] 263 264 if (peek() == i::Token::VAR) { 265 Consume(i::Token::VAR); 266 } else if (peek() == i::Token::CONST) { 267 Consume(i::Token::CONST); 268 } else { 269 *ok = false; 270 return 0; 271 } 272 273 // The scope of a variable/const declared anywhere inside a function 274 // is the entire function (ECMA-262, 3rd, 10.1.3, and 12.2). . 275 int nvars = 0; // the number of variables declared 276 do { 277 // Parse variable name. 278 if (nvars > 0) Consume(i::Token::COMMA); 279 ParseIdentifier(CHECK_OK); 280 nvars++; 281 if (peek() == i::Token::ASSIGN) { 282 Expect(i::Token::ASSIGN, CHECK_OK); 283 ParseAssignmentExpression(accept_IN, CHECK_OK); 284 } 285 } while (peek() == i::Token::COMMA); 286 287 if (num_decl != NULL) *num_decl = nvars; 288 return kUnknownStatement; 289} 290 291 292PreParser::Statement PreParser::ParseExpressionOrLabelledStatement( 293 bool* ok) { 294 // ExpressionStatement | LabelledStatement :: 295 // Expression ';' 296 // Identifier ':' Statement 297 298 Expression expr = ParseExpression(true, CHECK_OK); 299 if (peek() == i::Token::COLON && expr == kIdentifierExpression) { 300 Consume(i::Token::COLON); 301 return ParseStatement(ok); 302 } 303 // Parsed expression statement. 304 ExpectSemicolon(CHECK_OK); 305 return kUnknownStatement; 306} 307 308 309PreParser::Statement PreParser::ParseIfStatement(bool* ok) { 310 // IfStatement :: 311 // 'if' '(' Expression ')' Statement ('else' Statement)? 312 313 Expect(i::Token::IF, CHECK_OK); 314 Expect(i::Token::LPAREN, CHECK_OK); 315 ParseExpression(true, CHECK_OK); 316 Expect(i::Token::RPAREN, CHECK_OK); 317 ParseStatement(CHECK_OK); 318 if (peek() == i::Token::ELSE) { 319 Next(); 320 ParseStatement(CHECK_OK); 321 } 322 return kUnknownStatement; 323} 324 325 326PreParser::Statement PreParser::ParseContinueStatement(bool* ok) { 327 // ContinueStatement :: 328 // 'continue' [no line terminator] Identifier? ';' 329 330 Expect(i::Token::CONTINUE, CHECK_OK); 331 i::Token::Value tok = peek(); 332 if (!scanner_->has_line_terminator_before_next() && 333 tok != i::Token::SEMICOLON && 334 tok != i::Token::RBRACE && 335 tok != i::Token::EOS) { 336 ParseIdentifier(CHECK_OK); 337 } 338 ExpectSemicolon(CHECK_OK); 339 return kUnknownStatement; 340} 341 342 343PreParser::Statement PreParser::ParseBreakStatement(bool* ok) { 344 // BreakStatement :: 345 // 'break' [no line terminator] Identifier? ';' 346 347 Expect(i::Token::BREAK, CHECK_OK); 348 i::Token::Value tok = peek(); 349 if (!scanner_->has_line_terminator_before_next() && 350 tok != i::Token::SEMICOLON && 351 tok != i::Token::RBRACE && 352 tok != i::Token::EOS) { 353 ParseIdentifier(CHECK_OK); 354 } 355 ExpectSemicolon(CHECK_OK); 356 return kUnknownStatement; 357} 358 359 360PreParser::Statement PreParser::ParseReturnStatement(bool* ok) { 361 // ReturnStatement :: 362 // 'return' [no line terminator] Expression? ';' 363 364 // Consume the return token. It is necessary to do the before 365 // reporting any errors on it, because of the way errors are 366 // reported (underlining). 367 Expect(i::Token::RETURN, CHECK_OK); 368 369 // An ECMAScript program is considered syntactically incorrect if it 370 // contains a return statement that is not within the body of a 371 // function. See ECMA-262, section 12.9, page 67. 372 // This is not handled during preparsing. 373 374 i::Token::Value tok = peek(); 375 if (!scanner_->has_line_terminator_before_next() && 376 tok != i::Token::SEMICOLON && 377 tok != i::Token::RBRACE && 378 tok != i::Token::EOS) { 379 ParseExpression(true, CHECK_OK); 380 } 381 ExpectSemicolon(CHECK_OK); 382 return kUnknownStatement; 383} 384 385 386PreParser::Statement PreParser::ParseWithStatement(bool* ok) { 387 // WithStatement :: 388 // 'with' '(' Expression ')' Statement 389 Expect(i::Token::WITH, CHECK_OK); 390 Expect(i::Token::LPAREN, CHECK_OK); 391 ParseExpression(true, CHECK_OK); 392 Expect(i::Token::RPAREN, CHECK_OK); 393 394 scope_->EnterWith(); 395 ParseStatement(CHECK_OK); 396 scope_->LeaveWith(); 397 return kUnknownStatement; 398} 399 400 401PreParser::Statement PreParser::ParseSwitchStatement(bool* ok) { 402 // SwitchStatement :: 403 // 'switch' '(' Expression ')' '{' CaseClause* '}' 404 405 Expect(i::Token::SWITCH, CHECK_OK); 406 Expect(i::Token::LPAREN, CHECK_OK); 407 ParseExpression(true, CHECK_OK); 408 Expect(i::Token::RPAREN, CHECK_OK); 409 410 Expect(i::Token::LBRACE, CHECK_OK); 411 i::Token::Value token = peek(); 412 while (token != i::Token::RBRACE) { 413 if (token == i::Token::CASE) { 414 Expect(i::Token::CASE, CHECK_OK); 415 ParseExpression(true, CHECK_OK); 416 Expect(i::Token::COLON, CHECK_OK); 417 } else if (token == i::Token::DEFAULT) { 418 Expect(i::Token::DEFAULT, CHECK_OK); 419 Expect(i::Token::COLON, CHECK_OK); 420 } else { 421 ParseStatement(CHECK_OK); 422 } 423 token = peek(); 424 } 425 Expect(i::Token::RBRACE, CHECK_OK); 426 427 return kUnknownStatement; 428} 429 430 431PreParser::Statement PreParser::ParseDoWhileStatement(bool* ok) { 432 // DoStatement :: 433 // 'do' Statement 'while' '(' Expression ')' ';' 434 435 Expect(i::Token::DO, CHECK_OK); 436 ParseStatement(CHECK_OK); 437 Expect(i::Token::WHILE, CHECK_OK); 438 Expect(i::Token::LPAREN, CHECK_OK); 439 ParseExpression(true, CHECK_OK); 440 Expect(i::Token::RPAREN, CHECK_OK); 441 return kUnknownStatement; 442} 443 444 445PreParser::Statement PreParser::ParseWhileStatement(bool* ok) { 446 // WhileStatement :: 447 // 'while' '(' Expression ')' Statement 448 449 Expect(i::Token::WHILE, CHECK_OK); 450 Expect(i::Token::LPAREN, CHECK_OK); 451 ParseExpression(true, CHECK_OK); 452 Expect(i::Token::RPAREN, CHECK_OK); 453 ParseStatement(CHECK_OK); 454 return kUnknownStatement; 455} 456 457 458PreParser::Statement PreParser::ParseForStatement(bool* ok) { 459 // ForStatement :: 460 // 'for' '(' Expression? ';' Expression? ';' Expression? ')' Statement 461 462 Expect(i::Token::FOR, CHECK_OK); 463 Expect(i::Token::LPAREN, CHECK_OK); 464 if (peek() != i::Token::SEMICOLON) { 465 if (peek() == i::Token::VAR || peek() == i::Token::CONST) { 466 int decl_count; 467 ParseVariableDeclarations(false, &decl_count, CHECK_OK); 468 if (peek() == i::Token::IN && decl_count == 1) { 469 Expect(i::Token::IN, CHECK_OK); 470 ParseExpression(true, CHECK_OK); 471 Expect(i::Token::RPAREN, CHECK_OK); 472 473 ParseStatement(CHECK_OK); 474 return kUnknownStatement; 475 } 476 } else { 477 ParseExpression(false, CHECK_OK); 478 if (peek() == i::Token::IN) { 479 Expect(i::Token::IN, CHECK_OK); 480 ParseExpression(true, CHECK_OK); 481 Expect(i::Token::RPAREN, CHECK_OK); 482 483 ParseStatement(CHECK_OK); 484 return kUnknownStatement; 485 } 486 } 487 } 488 489 // Parsed initializer at this point. 490 Expect(i::Token::SEMICOLON, CHECK_OK); 491 492 if (peek() != i::Token::SEMICOLON) { 493 ParseExpression(true, CHECK_OK); 494 } 495 Expect(i::Token::SEMICOLON, CHECK_OK); 496 497 if (peek() != i::Token::RPAREN) { 498 ParseExpression(true, CHECK_OK); 499 } 500 Expect(i::Token::RPAREN, CHECK_OK); 501 502 ParseStatement(CHECK_OK); 503 return kUnknownStatement; 504} 505 506 507PreParser::Statement PreParser::ParseThrowStatement(bool* ok) { 508 // ThrowStatement :: 509 // 'throw' [no line terminator] Expression ';' 510 511 Expect(i::Token::THROW, CHECK_OK); 512 if (scanner_->has_line_terminator_before_next()) { 513 i::JavaScriptScanner::Location pos = scanner_->location(); 514 ReportMessageAt(pos.beg_pos, pos.end_pos, 515 "newline_after_throw", NULL); 516 *ok = false; 517 return kUnknownStatement; 518 } 519 ParseExpression(true, CHECK_OK); 520 ExpectSemicolon(CHECK_OK); 521 522 return kUnknownStatement; 523} 524 525 526PreParser::Statement PreParser::ParseTryStatement(bool* ok) { 527 // TryStatement :: 528 // 'try' Block Catch 529 // 'try' Block Finally 530 // 'try' Block Catch Finally 531 // 532 // Catch :: 533 // 'catch' '(' Identifier ')' Block 534 // 535 // Finally :: 536 // 'finally' Block 537 538 // In preparsing, allow any number of catch/finally blocks, including zero 539 // of both. 540 541 Expect(i::Token::TRY, CHECK_OK); 542 543 ParseBlock(CHECK_OK); 544 545 bool catch_or_finally_seen = false; 546 if (peek() == i::Token::CATCH) { 547 Consume(i::Token::CATCH); 548 Expect(i::Token::LPAREN, CHECK_OK); 549 ParseIdentifier(CHECK_OK); 550 Expect(i::Token::RPAREN, CHECK_OK); 551 scope_->EnterWith(); 552 ParseBlock(ok); 553 scope_->LeaveWith(); 554 if (!*ok) return kUnknownStatement; 555 catch_or_finally_seen = true; 556 } 557 if (peek() == i::Token::FINALLY) { 558 Consume(i::Token::FINALLY); 559 ParseBlock(CHECK_OK); 560 catch_or_finally_seen = true; 561 } 562 if (!catch_or_finally_seen) { 563 *ok = false; 564 } 565 return kUnknownStatement; 566} 567 568 569PreParser::Statement PreParser::ParseDebuggerStatement(bool* ok) { 570 // In ECMA-262 'debugger' is defined as a reserved keyword. In some browser 571 // contexts this is used as a statement which invokes the debugger as if a 572 // break point is present. 573 // DebuggerStatement :: 574 // 'debugger' ';' 575 576 Expect(i::Token::DEBUGGER, CHECK_OK); 577 ExpectSemicolon(CHECK_OK); 578 return kUnknownStatement; 579} 580 581 582// Precedence = 1 583PreParser::Expression PreParser::ParseExpression(bool accept_IN, bool* ok) { 584 // Expression :: 585 // AssignmentExpression 586 // Expression ',' AssignmentExpression 587 588 Expression result = ParseAssignmentExpression(accept_IN, CHECK_OK); 589 while (peek() == i::Token::COMMA) { 590 Expect(i::Token::COMMA, CHECK_OK); 591 ParseAssignmentExpression(accept_IN, CHECK_OK); 592 result = kUnknownExpression; 593 } 594 return result; 595} 596 597 598// Precedence = 2 599PreParser::Expression PreParser::ParseAssignmentExpression(bool accept_IN, 600 bool* ok) { 601 // AssignmentExpression :: 602 // ConditionalExpression 603 // LeftHandSideExpression AssignmentOperator AssignmentExpression 604 605 Expression expression = ParseConditionalExpression(accept_IN, CHECK_OK); 606 607 if (!i::Token::IsAssignmentOp(peek())) { 608 // Parsed conditional expression only (no assignment). 609 return expression; 610 } 611 612 i::Token::Value op = Next(); // Get assignment operator. 613 ParseAssignmentExpression(accept_IN, CHECK_OK); 614 615 if ((op == i::Token::ASSIGN) && (expression == kThisPropertyExpression)) { 616 scope_->AddProperty(); 617 } 618 619 return kUnknownExpression; 620} 621 622 623// Precedence = 3 624PreParser::Expression PreParser::ParseConditionalExpression(bool accept_IN, 625 bool* ok) { 626 // ConditionalExpression :: 627 // LogicalOrExpression 628 // LogicalOrExpression '?' AssignmentExpression ':' AssignmentExpression 629 630 // We start using the binary expression parser for prec >= 4 only! 631 Expression expression = ParseBinaryExpression(4, accept_IN, CHECK_OK); 632 if (peek() != i::Token::CONDITIONAL) return expression; 633 Consume(i::Token::CONDITIONAL); 634 // In parsing the first assignment expression in conditional 635 // expressions we always accept the 'in' keyword; see ECMA-262, 636 // section 11.12, page 58. 637 ParseAssignmentExpression(true, CHECK_OK); 638 Expect(i::Token::COLON, CHECK_OK); 639 ParseAssignmentExpression(accept_IN, CHECK_OK); 640 return kUnknownExpression; 641} 642 643 644int PreParser::Precedence(i::Token::Value tok, bool accept_IN) { 645 if (tok == i::Token::IN && !accept_IN) 646 return 0; // 0 precedence will terminate binary expression parsing 647 648 return i::Token::Precedence(tok); 649} 650 651 652// Precedence >= 4 653PreParser::Expression PreParser::ParseBinaryExpression(int prec, 654 bool accept_IN, 655 bool* ok) { 656 Expression result = ParseUnaryExpression(CHECK_OK); 657 for (int prec1 = Precedence(peek(), accept_IN); prec1 >= prec; prec1--) { 658 // prec1 >= 4 659 while (Precedence(peek(), accept_IN) == prec1) { 660 Next(); 661 ParseBinaryExpression(prec1 + 1, accept_IN, CHECK_OK); 662 result = kUnknownExpression; 663 } 664 } 665 return result; 666} 667 668 669PreParser::Expression PreParser::ParseUnaryExpression(bool* ok) { 670 // UnaryExpression :: 671 // PostfixExpression 672 // 'delete' UnaryExpression 673 // 'void' UnaryExpression 674 // 'typeof' UnaryExpression 675 // '++' UnaryExpression 676 // '--' UnaryExpression 677 // '+' UnaryExpression 678 // '-' UnaryExpression 679 // '~' UnaryExpression 680 // '!' UnaryExpression 681 682 i::Token::Value op = peek(); 683 if (i::Token::IsUnaryOp(op) || i::Token::IsCountOp(op)) { 684 op = Next(); 685 ParseUnaryExpression(ok); 686 return kUnknownExpression; 687 } else { 688 return ParsePostfixExpression(ok); 689 } 690} 691 692 693PreParser::Expression PreParser::ParsePostfixExpression(bool* ok) { 694 // PostfixExpression :: 695 // LeftHandSideExpression ('++' | '--')? 696 697 Expression expression = ParseLeftHandSideExpression(CHECK_OK); 698 if (!scanner_->has_line_terminator_before_next() && 699 i::Token::IsCountOp(peek())) { 700 Next(); 701 return kUnknownExpression; 702 } 703 return expression; 704} 705 706 707PreParser::Expression PreParser::ParseLeftHandSideExpression(bool* ok) { 708 // LeftHandSideExpression :: 709 // (NewExpression | MemberExpression) ... 710 711 Expression result; 712 if (peek() == i::Token::NEW) { 713 result = ParseNewExpression(CHECK_OK); 714 } else { 715 result = ParseMemberExpression(CHECK_OK); 716 } 717 718 while (true) { 719 switch (peek()) { 720 case i::Token::LBRACK: { 721 Consume(i::Token::LBRACK); 722 ParseExpression(true, CHECK_OK); 723 Expect(i::Token::RBRACK, CHECK_OK); 724 if (result == kThisExpression) { 725 result = kThisPropertyExpression; 726 } else { 727 result = kUnknownExpression; 728 } 729 break; 730 } 731 732 case i::Token::LPAREN: { 733 ParseArguments(CHECK_OK); 734 result = kUnknownExpression; 735 break; 736 } 737 738 case i::Token::PERIOD: { 739 Consume(i::Token::PERIOD); 740 ParseIdentifierName(CHECK_OK); 741 if (result == kThisExpression) { 742 result = kThisPropertyExpression; 743 } else { 744 result = kUnknownExpression; 745 } 746 break; 747 } 748 749 default: 750 return result; 751 } 752 } 753} 754 755 756PreParser::Expression PreParser::ParseNewExpression(bool* ok) { 757 // NewExpression :: 758 // ('new')+ MemberExpression 759 760 // The grammar for new expressions is pretty warped. The keyword 761 // 'new' can either be a part of the new expression (where it isn't 762 // followed by an argument list) or a part of the member expression, 763 // where it must be followed by an argument list. To accommodate 764 // this, we parse the 'new' keywords greedily and keep track of how 765 // many we have parsed. This information is then passed on to the 766 // member expression parser, which is only allowed to match argument 767 // lists as long as it has 'new' prefixes left 768 unsigned new_count = 0; 769 do { 770 Consume(i::Token::NEW); 771 new_count++; 772 } while (peek() == i::Token::NEW); 773 774 return ParseMemberWithNewPrefixesExpression(new_count, ok); 775} 776 777 778PreParser::Expression PreParser::ParseMemberExpression(bool* ok) { 779 return ParseMemberWithNewPrefixesExpression(0, ok); 780} 781 782 783PreParser::Expression PreParser::ParseMemberWithNewPrefixesExpression( 784 unsigned new_count, bool* ok) { 785 // MemberExpression :: 786 // (PrimaryExpression | FunctionLiteral) 787 // ('[' Expression ']' | '.' Identifier | Arguments)* 788 789 // Parse the initial primary or function expression. 790 Expression result = kUnknownExpression; 791 if (peek() == i::Token::FUNCTION) { 792 Consume(i::Token::FUNCTION); 793 if (peek() == i::Token::IDENTIFIER) { 794 ParseIdentifier(CHECK_OK); 795 } 796 result = ParseFunctionLiteral(CHECK_OK); 797 } else { 798 result = ParsePrimaryExpression(CHECK_OK); 799 } 800 801 while (true) { 802 switch (peek()) { 803 case i::Token::LBRACK: { 804 Consume(i::Token::LBRACK); 805 ParseExpression(true, CHECK_OK); 806 Expect(i::Token::RBRACK, CHECK_OK); 807 if (result == kThisExpression) { 808 result = kThisPropertyExpression; 809 } else { 810 result = kUnknownExpression; 811 } 812 break; 813 } 814 case i::Token::PERIOD: { 815 Consume(i::Token::PERIOD); 816 ParseIdentifierName(CHECK_OK); 817 if (result == kThisExpression) { 818 result = kThisPropertyExpression; 819 } else { 820 result = kUnknownExpression; 821 } 822 break; 823 } 824 case i::Token::LPAREN: { 825 if (new_count == 0) return result; 826 // Consume one of the new prefixes (already parsed). 827 ParseArguments(CHECK_OK); 828 new_count--; 829 result = kUnknownExpression; 830 break; 831 } 832 default: 833 return result; 834 } 835 } 836} 837 838 839PreParser::Expression PreParser::ParsePrimaryExpression(bool* ok) { 840 // PrimaryExpression :: 841 // 'this' 842 // 'null' 843 // 'true' 844 // 'false' 845 // Identifier 846 // Number 847 // String 848 // ArrayLiteral 849 // ObjectLiteral 850 // RegExpLiteral 851 // '(' Expression ')' 852 853 Expression result = kUnknownExpression; 854 switch (peek()) { 855 case i::Token::THIS: { 856 Next(); 857 result = kThisExpression; 858 break; 859 } 860 861 case i::Token::IDENTIFIER: { 862 ParseIdentifier(CHECK_OK); 863 result = kIdentifierExpression; 864 break; 865 } 866 867 case i::Token::NULL_LITERAL: 868 case i::Token::TRUE_LITERAL: 869 case i::Token::FALSE_LITERAL: 870 case i::Token::NUMBER: { 871 Next(); 872 break; 873 } 874 case i::Token::STRING: { 875 Next(); 876 result = GetStringSymbol(); 877 break; 878 } 879 880 case i::Token::ASSIGN_DIV: 881 result = ParseRegExpLiteral(true, CHECK_OK); 882 break; 883 884 case i::Token::DIV: 885 result = ParseRegExpLiteral(false, CHECK_OK); 886 break; 887 888 case i::Token::LBRACK: 889 result = ParseArrayLiteral(CHECK_OK); 890 break; 891 892 case i::Token::LBRACE: 893 result = ParseObjectLiteral(CHECK_OK); 894 break; 895 896 case i::Token::LPAREN: 897 Consume(i::Token::LPAREN); 898 parenthesized_function_ = (peek() == i::Token::FUNCTION); 899 result = ParseExpression(true, CHECK_OK); 900 Expect(i::Token::RPAREN, CHECK_OK); 901 if (result == kIdentifierExpression) result = kUnknownExpression; 902 break; 903 904 case i::Token::MOD: 905 result = ParseV8Intrinsic(CHECK_OK); 906 break; 907 908 default: { 909 Next(); 910 *ok = false; 911 return kUnknownExpression; 912 } 913 } 914 915 return result; 916} 917 918 919PreParser::Expression PreParser::ParseArrayLiteral(bool* ok) { 920 // ArrayLiteral :: 921 // '[' Expression? (',' Expression?)* ']' 922 Expect(i::Token::LBRACK, CHECK_OK); 923 while (peek() != i::Token::RBRACK) { 924 if (peek() != i::Token::COMMA) { 925 ParseAssignmentExpression(true, CHECK_OK); 926 } 927 if (peek() != i::Token::RBRACK) { 928 Expect(i::Token::COMMA, CHECK_OK); 929 } 930 } 931 Expect(i::Token::RBRACK, CHECK_OK); 932 933 scope_->NextMaterializedLiteralIndex(); 934 return kUnknownExpression; 935} 936 937 938PreParser::Expression PreParser::ParseObjectLiteral(bool* ok) { 939 // ObjectLiteral :: 940 // '{' ( 941 // ((IdentifierName | String | Number) ':' AssignmentExpression) 942 // | (('get' | 'set') (IdentifierName | String | Number) FunctionLiteral) 943 // )*[','] '}' 944 945 Expect(i::Token::LBRACE, CHECK_OK); 946 while (peek() != i::Token::RBRACE) { 947 i::Token::Value next = peek(); 948 switch (next) { 949 case i::Token::IDENTIFIER: { 950 bool is_getter = false; 951 bool is_setter = false; 952 ParseIdentifierOrGetOrSet(&is_getter, &is_setter, CHECK_OK); 953 if ((is_getter || is_setter) && peek() != i::Token::COLON) { 954 i::Token::Value name = Next(); 955 bool is_keyword = i::Token::IsKeyword(name); 956 if (name != i::Token::IDENTIFIER && 957 name != i::Token::NUMBER && 958 name != i::Token::STRING && 959 !is_keyword) { 960 *ok = false; 961 return kUnknownExpression; 962 } 963 if (!is_keyword) { 964 LogSymbol(); 965 } 966 ParseFunctionLiteral(CHECK_OK); 967 if (peek() != i::Token::RBRACE) { 968 Expect(i::Token::COMMA, CHECK_OK); 969 } 970 continue; // restart the while 971 } 972 break; 973 } 974 case i::Token::STRING: 975 Consume(next); 976 GetStringSymbol(); 977 break; 978 case i::Token::NUMBER: 979 Consume(next); 980 break; 981 default: 982 if (i::Token::IsKeyword(next)) { 983 Consume(next); 984 } else { 985 // Unexpected token. 986 *ok = false; 987 return kUnknownExpression; 988 } 989 } 990 991 Expect(i::Token::COLON, CHECK_OK); 992 ParseAssignmentExpression(true, CHECK_OK); 993 994 // TODO(1240767): Consider allowing trailing comma. 995 if (peek() != i::Token::RBRACE) Expect(i::Token::COMMA, CHECK_OK); 996 } 997 Expect(i::Token::RBRACE, CHECK_OK); 998 999 scope_->NextMaterializedLiteralIndex(); 1000 return kUnknownExpression; 1001} 1002 1003 1004PreParser::Expression PreParser::ParseRegExpLiteral(bool seen_equal, 1005 bool* ok) { 1006 if (!scanner_->ScanRegExpPattern(seen_equal)) { 1007 Next(); 1008 i::JavaScriptScanner::Location location = scanner_->location(); 1009 ReportMessageAt(location.beg_pos, location.end_pos, 1010 "unterminated_regexp", NULL); 1011 *ok = false; 1012 return kUnknownExpression; 1013 } 1014 1015 scope_->NextMaterializedLiteralIndex(); 1016 1017 if (!scanner_->ScanRegExpFlags()) { 1018 Next(); 1019 i::JavaScriptScanner::Location location = scanner_->location(); 1020 ReportMessageAt(location.beg_pos, location.end_pos, 1021 "invalid_regexp_flags", NULL); 1022 *ok = false; 1023 return kUnknownExpression; 1024 } 1025 Next(); 1026 return kUnknownExpression; 1027} 1028 1029 1030PreParser::Arguments PreParser::ParseArguments(bool* ok) { 1031 // Arguments :: 1032 // '(' (AssignmentExpression)*[','] ')' 1033 1034 Expect(i::Token::LPAREN, CHECK_OK); 1035 bool done = (peek() == i::Token::RPAREN); 1036 int argc = 0; 1037 while (!done) { 1038 ParseAssignmentExpression(true, CHECK_OK); 1039 argc++; 1040 done = (peek() == i::Token::RPAREN); 1041 if (!done) Expect(i::Token::COMMA, CHECK_OK); 1042 } 1043 Expect(i::Token::RPAREN, CHECK_OK); 1044 return argc; 1045} 1046 1047 1048PreParser::Expression PreParser::ParseFunctionLiteral(bool* ok) { 1049 // Function :: 1050 // '(' FormalParameterList? ')' '{' FunctionBody '}' 1051 1052 // Parse function body. 1053 ScopeType outer_scope_type = scope_->type(); 1054 bool inside_with = scope_->IsInsideWith(); 1055 Scope function_scope(&scope_, kFunctionScope); 1056 1057 // FormalParameterList :: 1058 // '(' (Identifier)*[','] ')' 1059 Expect(i::Token::LPAREN, CHECK_OK); 1060 bool done = (peek() == i::Token::RPAREN); 1061 while (!done) { 1062 ParseIdentifier(CHECK_OK); 1063 done = (peek() == i::Token::RPAREN); 1064 if (!done) { 1065 Expect(i::Token::COMMA, CHECK_OK); 1066 } 1067 } 1068 Expect(i::Token::RPAREN, CHECK_OK); 1069 1070 Expect(i::Token::LBRACE, CHECK_OK); 1071 int function_block_pos = scanner_->location().beg_pos; 1072 1073 // Determine if the function will be lazily compiled. 1074 // Currently only happens to top-level functions. 1075 // Optimistically assume that all top-level functions are lazily compiled. 1076 bool is_lazily_compiled = (outer_scope_type == kTopLevelScope && 1077 !inside_with && allow_lazy_ && 1078 !parenthesized_function_); 1079 parenthesized_function_ = false; 1080 1081 if (is_lazily_compiled) { 1082 log_->PauseRecording(); 1083 ParseSourceElements(i::Token::RBRACE, ok); 1084 log_->ResumeRecording(); 1085 if (!*ok) return kUnknownExpression; 1086 1087 Expect(i::Token::RBRACE, CHECK_OK); 1088 1089 // Position right after terminal '}'. 1090 int end_pos = scanner_->location().end_pos; 1091 log_->LogFunction(function_block_pos, end_pos, 1092 function_scope.materialized_literal_count(), 1093 function_scope.expected_properties()); 1094 } else { 1095 ParseSourceElements(i::Token::RBRACE, CHECK_OK); 1096 Expect(i::Token::RBRACE, CHECK_OK); 1097 } 1098 return kUnknownExpression; 1099} 1100 1101 1102PreParser::Expression PreParser::ParseV8Intrinsic(bool* ok) { 1103 // CallRuntime :: 1104 // '%' Identifier Arguments 1105 1106 Expect(i::Token::MOD, CHECK_OK); 1107 ParseIdentifier(CHECK_OK); 1108 ParseArguments(CHECK_OK); 1109 1110 return kUnknownExpression; 1111} 1112 1113 1114void PreParser::ExpectSemicolon(bool* ok) { 1115 // Check for automatic semicolon insertion according to 1116 // the rules given in ECMA-262, section 7.9, page 21. 1117 i::Token::Value tok = peek(); 1118 if (tok == i::Token::SEMICOLON) { 1119 Next(); 1120 return; 1121 } 1122 if (scanner_->has_line_terminator_before_next() || 1123 tok == i::Token::RBRACE || 1124 tok == i::Token::EOS) { 1125 return; 1126 } 1127 Expect(i::Token::SEMICOLON, ok); 1128} 1129 1130 1131void PreParser::LogSymbol() { 1132 int identifier_pos = scanner_->location().beg_pos; 1133 if (scanner_->is_literal_ascii()) { 1134 log_->LogAsciiSymbol(identifier_pos, scanner_->literal_ascii_string()); 1135 } else { 1136 log_->LogUC16Symbol(identifier_pos, scanner_->literal_uc16_string()); 1137 } 1138} 1139 1140 1141PreParser::Identifier PreParser::GetIdentifierSymbol() { 1142 LogSymbol(); 1143 return kUnknownIdentifier; 1144} 1145 1146 1147PreParser::Expression PreParser::GetStringSymbol() { 1148 LogSymbol(); 1149 return kUnknownExpression; 1150} 1151 1152 1153PreParser::Identifier PreParser::ParseIdentifier(bool* ok) { 1154 Expect(i::Token::IDENTIFIER, ok); 1155 if (!*ok) return kUnknownIdentifier; 1156 return GetIdentifierSymbol(); 1157} 1158 1159 1160PreParser::Identifier PreParser::ParseIdentifierName(bool* ok) { 1161 i::Token::Value next = Next(); 1162 if (i::Token::IsKeyword(next)) { 1163 int pos = scanner_->location().beg_pos; 1164 const char* keyword = i::Token::String(next); 1165 log_->LogAsciiSymbol(pos, i::Vector<const char>(keyword, 1166 i::StrLength(keyword))); 1167 return kUnknownExpression; 1168 } 1169 if (next == i::Token::IDENTIFIER) { 1170 return GetIdentifierSymbol(); 1171 } 1172 *ok = false; 1173 return kUnknownIdentifier; 1174} 1175 1176 1177// This function reads an identifier and determines whether or not it 1178// is 'get' or 'set'. The reason for not using ParseIdentifier and 1179// checking on the output is that this involves heap allocation which 1180// we can't do during preparsing. 1181PreParser::Identifier PreParser::ParseIdentifierOrGetOrSet(bool* is_get, 1182 bool* is_set, 1183 bool* ok) { 1184 Expect(i::Token::IDENTIFIER, CHECK_OK); 1185 if (scanner_->is_literal_ascii() && scanner_->literal_length() == 3) { 1186 const char* token = scanner_->literal_ascii_string().start(); 1187 *is_get = strncmp(token, "get", 3) == 0; 1188 *is_set = !*is_get && strncmp(token, "set", 3) == 0; 1189 } 1190 return GetIdentifierSymbol(); 1191} 1192 1193#undef CHECK_OK 1194} } // v8::preparser 1195