1 /* 2 * Copyright (c) 2010, 2013, Oracle and/or its affiliates. All rights reserved. 3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. 4 * 5 * This code is free software; you can redistribute it and/or modify it 6 * under the terms of the GNU General Public License version 2 only, as 7 * published by the Free Software Foundation. Oracle designates this 8 * particular file as subject to the "Classpath" exception as provided 9 * by Oracle in the LICENSE file that accompanied this code. 10 * 11 * This code is distributed in the hope that it will be useful, but WITHOUT 12 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or 13 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License 14 * version 2 for more details (a copy is included in the LICENSE file that 15 * accompanied this code). 16 * 17 * You should have received a copy of the GNU General Public License version 18 * 2 along with this work; if not, write to the Free Software Foundation, 19 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. 20 * 21 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA 22 * or visit www.oracle.com if you need additional information or have any 23 * questions. 24 */ 25 26 package jdk.nashorn.internal.parser; 27 28 import static jdk.nashorn.internal.parser.TokenType.ADD; 29 import static jdk.nashorn.internal.parser.TokenType.DECIMAL; 30 import static jdk.nashorn.internal.parser.TokenType.EOF; 31 import static jdk.nashorn.internal.parser.TokenType.EOL; 32 import static jdk.nashorn.internal.parser.TokenType.ERROR; 33 import static jdk.nashorn.internal.parser.TokenType.ESCSTRING; 34 import static jdk.nashorn.internal.parser.TokenType.EXECSTRING; 35 import static jdk.nashorn.internal.parser.TokenType.FLOATING; 36 import static jdk.nashorn.internal.parser.TokenType.HEXADECIMAL; 37 import static jdk.nashorn.internal.parser.TokenType.LBRACE; 38 import static jdk.nashorn.internal.parser.TokenType.LPAREN; 39 import static jdk.nashorn.internal.parser.TokenType.OCTAL; 40 import static jdk.nashorn.internal.parser.TokenType.RBRACE; 41 import static jdk.nashorn.internal.parser.TokenType.REGEX; 42 import static jdk.nashorn.internal.parser.TokenType.RPAREN; 43 import static jdk.nashorn.internal.parser.TokenType.STRING; 44 import static jdk.nashorn.internal.parser.TokenType.XML; 45 46 import jdk.nashorn.internal.runtime.ECMAErrors; 47 import jdk.nashorn.internal.runtime.ErrorManager; 48 import jdk.nashorn.internal.runtime.JSErrorType; 49 import jdk.nashorn.internal.runtime.JSType; 50 import jdk.nashorn.internal.runtime.ParserException; 51 import jdk.nashorn.internal.runtime.Source; 52 import jdk.nashorn.internal.runtime.options.Options; 53 54 /** 55 * Responsible for converting source content into a stream of tokens. 56 * 57 */ 58 @SuppressWarnings("fallthrough") 59 public class Lexer extends Scanner { 60 private static final long MIN_INT_L = Integer.MIN_VALUE; 61 private static final long MAX_INT_L = Integer.MAX_VALUE; 62 63 private static final boolean XML_LITERALS = Options.getBooleanProperty("nashorn.lexer.xmlliterals"); 64 65 /** Content source. */ 66 private final Source source; 67 68 /** Buffered stream for tokens. */ 69 private final TokenStream stream; 70 71 /** True if here and edit strings are supported. */ 72 private final boolean scripting; 73 74 /** True if a nested scan. (scan to completion, no EOF.) */ 75 private final boolean nested; 76 77 /** Pending new line number and position. */ 78 private int pendingLine; 79 80 /** Position of last EOL + 1. */ 81 private int linePosition; 82 83 /** Type of last token added. */ 84 private TokenType last; 85 86 private static final String JAVASCRIPT_WHITESPACE; 87 private static final String JAVASCRIPT_WHITESPACE_EOL; 88 private static final String JAVASCRIPT_WHITESPACE_IN_REGEXP; 89 90 private static final String JSON_WHITESPACE; 91 private static final String JSON_WHITESPACE_EOL; 92 93 static String unicodeEscape(final char ch) { 94 final StringBuilder sb = new StringBuilder(); 95 96 sb.append("\\u"); 97 98 final String hex = Integer.toHexString(ch); 99 for (int i = hex.length(); i < 4; i++) { 100 sb.append('0'); 101 } 102 sb.append(hex); 103 104 return sb.toString(); 105 } 106 107 static { 108 final StringBuilder ws = new StringBuilder(); 109 final StringBuilder wsEOL = new StringBuilder(); 110 final StringBuilder wsRegExp = new StringBuilder(); 111 final StringBuilder jsonWs = new StringBuilder(); 112 113 jsonWs.append((char)0x000a); 114 jsonWs.append((char)0x000d); 115 JSON_WHITESPACE_EOL = jsonWs.toString(); 116 117 jsonWs.append((char)0x0009); 118 jsonWs.append((char)0x0020); 119 JSON_WHITESPACE = jsonWs.toString(); 120 121 for (int i = 0; i <= 0xffff; i++) { 122 switch (i) { 123 case 0x000a: // line feed 124 case 0x000d: // carriage return (ctrl-m) 125 case 0x2028: // line separator 126 case 0x2029: // paragraph separator 127 wsEOL.append((char)i); 128 case 0x0009: // tab 129 case 0x0020: // ASCII space 130 case 0x000b: // tabulation line 131 case 0x000c: // ff (ctrl-l) 132 case 0x00a0: // Latin-1 space 133 case 0x1680: // Ogham space mark 134 case 0x180e: // separator, Mongolian vowel 135 case 0x2000: // en quad 136 case 0x2001: // em quad 137 case 0x2002: // en space 138 case 0x2003: // em space 139 case 0x2004: // three-per-em space 140 case 0x2005: // four-per-em space 141 case 0x2006: // six-per-em space 142 case 0x2007: // figure space 143 case 0x2008: // punctuation space 144 case 0x2009: // thin space 145 case 0x200a: // hair space 146 case 0x202f: // narrow no-break space 147 case 0x205f: // medium mathematical space 148 case 0x3000: // ideographic space 149 case 0xfeff: // byte order mark 150 ws.append((char)i); 151 152 wsRegExp.append(Lexer.unicodeEscape((char)i)); 153 break; 154 155 default: 156 break; 157 } 158 } 159 160 JAVASCRIPT_WHITESPACE = ws.toString(); 161 JAVASCRIPT_WHITESPACE_EOL = wsEOL.toString(); 162 JAVASCRIPT_WHITESPACE_IN_REGEXP = wsRegExp.toString(); 163 164 } 165 166 /** 167 * Constructor 168 * 169 * @param source the source 170 * @param stream the token stream to lex 171 */ 172 public Lexer(final Source source, final TokenStream stream) { 173 this(source, stream, false); 174 } 175 176 /** 177 * Constructor 178 * 179 * @param source the source 180 * @param stream the token stream to lex 181 * @param scripting are we in scripting mode 182 */ 183 public Lexer(final Source source, final TokenStream stream, final boolean scripting) { 184 super(source.getContent(), 1, 0, source.getLength()); 185 186 this.source = source; 187 this.stream = stream; 188 this.scripting = scripting; 189 this.nested = false; 190 this.pendingLine = 1; 191 this.last = EOL; 192 } 193 194 private Lexer(final Lexer lexer, final State state) { 195 super(lexer, state); 196 197 source = lexer.source; 198 stream = lexer.stream; 199 scripting = lexer.scripting; 200 nested = true; 201 202 pendingLine = state.pendingLine; 203 linePosition = state.linePosition; 204 last = EOL; 205 } 206 207 static class State extends Scanner.State { 208 /** Pending new line number and position. */ 209 public final int pendingLine; 210 211 /** Position of last EOL + 1. */ 212 public final int linePosition; 213 214 /** Type of last token added. */ 215 public final TokenType last; 216 217 /* 218 * Constructor. 219 */ 220 221 State(final int position, final int limit, final int line, final int pendingLine, final int linePosition, final TokenType last) { 222 super(position, limit, line); 223 224 this.pendingLine = pendingLine; 225 this.linePosition = linePosition; 226 this.last = last; 227 } 228 } 229 230 /** 231 * Save the state of the scan. 232 * 233 * @return Captured state. 234 */ 235 @Override 236 State saveState() { 237 return new State(position, limit, line, pendingLine, linePosition, last); 238 } 239 240 /** 241 * Restore the state of the scan. 242 * 243 * @param state 244 * Captured state. 245 */ 246 void restoreState(final State state) { 247 super.restoreState(state); 248 249 pendingLine = state.pendingLine; 250 linePosition = state.linePosition; 251 last = state.last; 252 } 253 254 /** 255 * Add a new token to the stream. 256 * 257 * @param type 258 * Token type. 259 * @param start 260 * Start position. 261 * @param end 262 * End position. 263 */ 264 protected void add(final TokenType type, final int start, final int end) { 265 // Record last token. 266 last = type; 267 268 // Only emit the last EOL in a cluster. 269 if (type == EOL) { 270 pendingLine = end; 271 linePosition = start; 272 } else { 273 // Write any pending EOL to stream. 274 if (pendingLine != -1) { 275 stream.put(Token.toDesc(EOL, linePosition, pendingLine)); 276 pendingLine = -1; 277 } 278 279 // Write token to stream. 280 stream.put(Token.toDesc(type, start, end - start)); 281 } 282 } 283 284 /** 285 * Add a new token to the stream. 286 * 287 * @param type 288 * Token type. 289 * @param start 290 * Start position. 291 */ 292 protected void add(final TokenType type, final int start) { 293 add(type, start, position); 294 } 295 296 /** 297 * Return the String of valid whitespace characters for regular 298 * expressions in JavaScript 299 * @return regexp whitespace string 300 */ 301 public static String getWhitespaceRegExp() { 302 return JAVASCRIPT_WHITESPACE_IN_REGEXP; 303 } 304 305 /** 306 * Skip end of line. 307 * 308 * @param addEOL true if EOL token should be recorded. 309 */ 310 private void skipEOL(final boolean addEOL) { 311 312 if (ch0 == '\r') { // detect \r\n pattern 313 skip(1); 314 if (ch0 == '\n') { 315 skip(1); 316 } 317 } else { // all other space, ch0 is guaranteed to be EOL or \0 318 skip(1); 319 } 320 321 // bump up line count 322 line++; 323 324 if (addEOL) { 325 // Add an EOL token. 326 add(EOL, position, line); 327 } 328 } 329 330 /** 331 * Skip over rest of line including end of line. 332 * 333 * @param addEOL true if EOL token should be recorded. 334 */ 335 private void skipLine(final boolean addEOL) { 336 // Ignore characters. 337 while (!isEOL(ch0) && !atEOF()) { 338 skip(1); 339 } 340 // Skip over end of line. 341 skipEOL(addEOL); 342 } 343 344 /** 345 * Test whether a char is valid JavaScript whitespace 346 * @param ch a char 347 * @return true if valid JavaScript whitespace 348 */ 349 public static boolean isJSWhitespace(final char ch) { 350 return JAVASCRIPT_WHITESPACE.indexOf(ch) != -1; 351 } 352 353 /** 354 * Test whether a char is valid JavaScript end of line 355 * @param ch a char 356 * @return true if valid JavaScript end of line 357 */ 358 public static boolean isJSEOL(final char ch) { 359 return JAVASCRIPT_WHITESPACE_EOL.indexOf(ch) != -1; 360 } 361 362 /** 363 * Test whether a char is valid JSON whitespace 364 * @param ch a char 365 * @return true if valid JSON whitespace 366 */ 367 public static boolean isJsonWhitespace(final char ch) { 368 return JSON_WHITESPACE.indexOf(ch) != -1; 369 } 370 371 /** 372 * Test whether a char is valid JSON end of line 373 * @param ch a char 374 * @return true if valid JSON end of line 375 */ 376 public static boolean isJsonEOL(final char ch) { 377 return JSON_WHITESPACE_EOL.indexOf(ch) != -1; 378 } 379 380 /** 381 * Test if char is a string delimiter, e.g. '\' or '"'. Also scans exec 382 * strings ('`') in scripting mode. 383 * @param ch a char 384 * @return true if string delimiter 385 */ 386 protected boolean isStringDelimiter(final char ch) { 387 return ch == '\'' || ch == '"' || (scripting && ch == '`'); 388 } 389 390 /** 391 * Test whether a char is valid JavaScript whitespace 392 * @param ch a char 393 * @return true if valid JavaScript whitespace 394 */ 395 protected boolean isWhitespace(final char ch) { 396 return Lexer.isJSWhitespace(ch); 397 } 398 399 /** 400 * Test whether a char is valid JavaScript end of line 401 * @param ch a char 402 * @return true if valid JavaScript end of line 403 */ 404 protected boolean isEOL(final char ch) { 405 return Lexer.isJSEOL(ch); 406 } 407 408 /** 409 * Skip over whitespace and detect end of line, adding EOL tokens if 410 * encountered. 411 * 412 * @param addEOL true if EOL tokens should be recorded. 413 */ 414 private void skipWhitespace(final boolean addEOL) { 415 while (isWhitespace(ch0)) { 416 if (isEOL(ch0)) { 417 skipEOL(addEOL); 418 } else { 419 skip(1); 420 } 421 } 422 } 423 424 /** 425 * Skip over comments. 426 * 427 * @return True if a comment. 428 */ 429 protected boolean skipComments() { 430 if (ch0 == '/') { 431 // Is it a // comment. 432 if (ch1 == '/') { 433 // Skip over //. 434 skip(2); 435 // Scan for EOL. 436 while (!atEOF() && !isEOL(ch0)) { 437 skip(1); 438 } 439 // Did detect a comment. 440 return true; 441 } else if (ch1 == '*') { 442 // Record beginning of comment. 443 final int start = position; 444 // Skip over /*. 445 skip(2); 446 // Scan for */. 447 while (!atEOF() && !(ch0 == '*' && ch1 == '/')) { 448 // If end of line handle else skip character. 449 if (isEOL(ch0)) { 450 skipEOL(true); 451 } else { 452 skip(1); 453 } 454 } 455 456 if (atEOF()) { 457 // TODO - Report closing */ missing in parser. 458 add(ERROR, start); 459 } else { 460 // Skip */. 461 skip(2); 462 } 463 464 // Did detect a comment. 465 return true; 466 } 467 } 468 469 if (scripting && ch0 == '#') { 470 // shell style comment 471 // Skip over #. 472 skip(1); 473 // Scan for EOL. 474 while (!atEOF() && !isEOL(ch0)) { 475 skip(1); 476 } 477 // Did detect a comment. 478 return true; 479 } 480 481 // Not a comment. 482 return false; 483 } 484 485 /** 486 * Convert a regex token to a token object. 487 * 488 * @param start Position in source content. 489 * @param length Length of regex token. 490 * @return Regex token object. 491 */ 492 public RegexToken valueOfPattern(final int start, final int length) { 493 // Save the current position. 494 final int savePosition = position; 495 // Reset to beginning of content. 496 reset(start); 497 // Buffer for recording characters. 498 final StringBuilder sb = new StringBuilder(length); 499 500 // Skip /. 501 skip(1); 502 boolean inBrackets = false; 503 // Scan for closing /, stopping at end of line. 504 while (!atEOF() && ch0 != '/' && !isEOL(ch0) || inBrackets) { 505 // Skip over escaped character. 506 if (ch0 == '\\') { 507 sb.append(ch0); 508 sb.append(ch1); 509 skip(2); 510 } else { 511 if (ch0 == '[') { 512 inBrackets = true; 513 } else if (ch0 == ']') { 514 inBrackets = false; 515 } 516 517 // Skip literal character. 518 sb.append(ch0); 519 skip(1); 520 } 521 } 522 523 // Get pattern as string. 524 final String regex = sb.toString(); 525 526 // Skip /. 527 skip(1); 528 529 // Options as string. 530 final String options = source.getString(position, scanIdentifier()); 531 532 reset(savePosition); 533 534 // Compile the pattern. 535 return new RegexToken(regex, options); 536 } 537 538 /** 539 * Return true if the given token can be the beginning of a literal. 540 * 541 * @param token a token 542 * @return true if token can start a literal. 543 */ 544 public boolean canStartLiteral(final TokenType token) { 545 return token.startsWith('/') || ((scripting || XML_LITERALS) && token.startsWith('<')); 546 } 547 548 /** 549 * Check whether the given token represents the beginning of a literal. If so scan 550 * the literal and return <tt>true</tt>, otherwise return false. 551 * 552 * @param token the token. 553 * @param startTokenType the token type. 554 * @return True if a literal beginning with startToken was found and scanned. 555 */ 556 protected boolean scanLiteral(final long token, final TokenType startTokenType) { 557 // Check if it can be a literal. 558 if (!canStartLiteral(startTokenType)) { 559 return false; 560 } 561 // We break on ambiguous tokens so if we already moved on it can't be a literal. 562 if (stream.get(stream.last()) != token) { 563 return false; 564 } 565 // Rewind to token start position 566 reset(Token.descPosition(token)); 567 568 if (ch0 == '/') { 569 return scanRegEx(); 570 } else if (ch0 == '<') { 571 if (ch1 == '<') { 572 return scanHereString(); 573 } else if (Character.isJavaIdentifierStart(ch1)) { 574 return scanXMLLiteral(); 575 } 576 } 577 578 return false; 579 } 580 581 /** 582 * Scan over regex literal. 583 * 584 * @return True if a regex literal. 585 */ 586 private boolean scanRegEx() { 587 assert ch0 == '/'; 588 // Make sure it's not a comment. 589 if (ch1 != '/' && ch1 != '*') { 590 // Record beginning of literal. 591 final int start = position; 592 // Skip /. 593 skip(1); 594 boolean inBrackets = false; 595 596 // Scan for closing /, stopping at end of line. 597 while (!atEOF() && (ch0 != '/' || inBrackets) && !isEOL(ch0)) { 598 // Skip over escaped character. 599 if (ch0 == '\\') { 600 skip(1); 601 if (isEOL(ch0)) { 602 reset(start); 603 return false; 604 } 605 skip(1); 606 } else { 607 if (ch0 == '[') { 608 inBrackets = true; 609 } else if (ch0 == ']') { 610 inBrackets = false; 611 } 612 613 // Skip literal character. 614 skip(1); 615 } 616 } 617 618 // If regex literal. 619 if (ch0 == '/') { 620 // Skip /. 621 skip(1); 622 623 // Skip over options. 624 while (!atEOF() && Character.isJavaIdentifierPart(ch0) || ch0 == '\\' && ch1 == 'u') { 625 skip(1); 626 } 627 628 // Add regex token. 629 add(REGEX, start); 630 // Regex literal detected. 631 return true; 632 } 633 634 // False start try again. 635 reset(start); 636 } 637 638 // Regex literal not detected. 639 return false; 640 } 641 642 /** 643 * Convert a digit to a integer. Can't use Character.digit since we are 644 * restricted to ASCII by the spec. 645 * 646 * @param ch Character to convert. 647 * @param base Numeric base. 648 * 649 * @return The converted digit or -1 if invalid. 650 */ 651 private static int convertDigit(final char ch, final int base) { 652 int digit; 653 654 if ('0' <= ch && ch <= '9') { 655 digit = ch - '0'; 656 } else if ('A' <= ch && ch <= 'Z') { 657 digit = ch - 'A' + 10; 658 } else if ('a' <= ch && ch <= 'z') { 659 digit = ch - 'a' + 10; 660 } else { 661 return -1; 662 } 663 664 return digit < base ? digit : -1; 665 } 666 667 668 /** 669 * Get the value of a numeric sequence. 670 * 671 * @param base Numeric base. 672 * @param max Maximum number of digits. 673 * @param skip Skip over escape first. 674 * @param check Tells whether to throw error if a digit is invalid for the given base. 675 * @param type Type of token to report against. 676 * 677 * @return Value of sequence or < 0 if no digits. 678 */ 679 private int valueOfSequence(final int base, final int max, final boolean skip, final boolean check, final TokenType type) { 680 assert base == 16 || base == 8 : "base other than 16 or 8"; 681 final boolean isHex = base == 16; 682 final int shift = isHex ? 4 : 3; 683 int value = 0; 684 685 if (skip) { 686 skip(2); 687 } 688 689 for (int i = 0; i < max; i++) { 690 final int digit = convertDigit(ch0, base); 691 692 if (digit == -1) { 693 if (check) { 694 error(Lexer.message("invalid." + (isHex ? "hex" : "octal")), type, position, limit); 695 } 696 return i == 0 ? -1 : value; 697 } 698 699 value = value << shift | digit; 700 skip(1); 701 } 702 703 return value; 704 } 705 706 /** 707 * Convert a string to a JavaScript identifier. 708 * 709 * @param start Position in source content. 710 * @param length Length of token. 711 * @return Ident string or null if an error. 712 */ 713 private String valueOfIdent(final int start, final int length) throws RuntimeException { 714 // Save the current position. 715 final int savePosition = position; 716 // End of scan. 717 final int end = start + length; 718 // Reset to beginning of content. 719 reset(start); 720 // Buffer for recording characters. 721 final StringBuilder sb = new StringBuilder(length); 722 723 // Scan until end of line or end of file. 724 while (!atEOF() && position < end && !isEOL(ch0)) { 725 // If escape character. 726 if (ch0 == '\\' && ch1 == 'u') { 727 final int ch = valueOfSequence(16, 4, true, true, TokenType.IDENT); 728 if (isWhitespace((char)ch)) { 729 return null; 730 } 731 if (ch < 0) { 732 sb.append('\\'); 733 sb.append('u'); 734 } else { 735 sb.append((char)ch); 736 } 737 } else { 738 // Add regular character. 739 sb.append(ch0); 740 skip(1); 741 } 742 } 743 744 // Restore position. 745 reset(savePosition); 746 747 return sb.toString(); 748 } 749 750 /** 751 * Scan over and identifier or keyword. Handles identifiers containing 752 * encoded Unicode chars. 753 * 754 * Example: 755 * 756 * var \u0042 = 44; 757 */ 758 private void scanIdentifierOrKeyword() { 759 // Record beginning of identifier. 760 final int start = position; 761 // Scan identifier. 762 final int length = scanIdentifier(); 763 // Check to see if it is a keyword. 764 final TokenType type = TokenLookup.lookupKeyword(content, start, length); 765 // Add keyword or identifier token. 766 add(type, start); 767 } 768 769 /** 770 * Convert a string to a JavaScript string object. 771 * 772 * @param start Position in source content. 773 * @param length Length of token. 774 * @return JavaScript string object. 775 */ 776 private String valueOfString(final int start, final int length, final boolean strict) throws RuntimeException { 777 // Save the current position. 778 final int savePosition = position; 779 // Calculate the end position. 780 final int end = start + length; 781 // Reset to beginning of string. 782 reset(start); 783 784 // Buffer for recording characters. 785 final StringBuilder sb = new StringBuilder(length); 786 787 // Scan until end of string. 788 while (position < end) { 789 // If escape character. 790 if (ch0 == '\\') { 791 skip(1); 792 793 final char next = ch0; 794 final int afterSlash = position; 795 796 skip(1); 797 798 // Special characters. 799 switch (next) { 800 case '0': 801 case '1': 802 case '2': 803 case '3': 804 case '4': 805 case '5': 806 case '6': 807 case '7': { 808 if (strict) { 809 // "\0" itself is allowed in strict mode. Only other 'real' 810 // octal escape sequences are not allowed (eg. "\02", "\31"). 811 // See section 7.8.4 String literals production EscapeSequence 812 if (next != '0' || (ch0 >= '0' && ch0 <= '9')) { 813 error(Lexer.message("strict.no.octal"), STRING, position, limit); 814 } 815 } 816 reset(afterSlash); 817 // Octal sequence. 818 final int ch = valueOfSequence(8, 3, false, false, STRING); 819 820 if (ch < 0) { 821 sb.append('\\'); 822 sb.append('x'); 823 } else { 824 sb.append((char)ch); 825 } 826 break; 827 } 828 case 'n': 829 sb.append('\n'); 830 break; 831 case 't': 832 sb.append('\t'); 833 break; 834 case 'b': 835 sb.append('\b'); 836 break; 837 case 'f': 838 sb.append('\f'); 839 break; 840 case 'r': 841 sb.append('\r'); 842 break; 843 case '\'': 844 sb.append('\''); 845 break; 846 case '\"': 847 sb.append('\"'); 848 break; 849 case '\\': 850 sb.append('\\'); 851 break; 852 case '\r': // CR | CRLF 853 if (ch0 == '\n') { 854 skip(1); 855 } 856 // fall through 857 case '\n': // LF 858 case '\u2028': // LS 859 case '\u2029': // PS 860 // continue on the next line, slash-return continues string 861 // literal 862 break; 863 case 'x': { 864 // Hex sequence. 865 final int ch = valueOfSequence(16, 2, false, true, STRING); 866 867 if (ch < 0) { 868 sb.append('\\'); 869 sb.append('x'); 870 } else { 871 sb.append((char)ch); 872 } 873 } 874 break; 875 case 'u': { 876 // Unicode sequence. 877 final int ch = valueOfSequence(16, 4, false, true, STRING); 878 879 if (ch < 0) { 880 sb.append('\\'); 881 sb.append('u'); 882 } else { 883 sb.append((char)ch); 884 } 885 } 886 break; 887 case 'v': 888 sb.append('\u000B'); 889 break; 890 // All other characters. 891 default: 892 sb.append(next); 893 break; 894 } 895 } else { 896 // Add regular character. 897 sb.append(ch0); 898 skip(1); 899 } 900 } 901 902 // Restore position. 903 reset(savePosition); 904 905 return sb.toString(); 906 } 907 908 /** 909 * Scan over a string literal. 910 */ 911 private void scanString(final boolean add) { 912 // Type of string. 913 TokenType type = STRING; 914 // Record starting quote. 915 final char quote = ch0; 916 // Skip over quote. 917 skip(1); 918 919 // Record beginning of string content. 920 final State stringState = saveState(); 921 922 // Scan until close quote or end of line. 923 while (!atEOF() && ch0 != quote && !isEOL(ch0)) { 924 // Skip over escaped character. 925 if (ch0 == '\\') { 926 type = ESCSTRING; 927 skip(1); 928 if (isEOL(ch0)) { 929 // Multiline string literal 930 skipEOL(false); 931 continue; 932 } 933 } 934 // Skip literal character. 935 skip(1); 936 } 937 938 // If close quote. 939 if (ch0 == quote) { 940 // Skip close quote. 941 skip(1); 942 } else { 943 error(Lexer.message("missing.close.quote"), STRING, position, limit); 944 } 945 946 // If not just scanning. 947 if (add) { 948 // Record end of string. 949 stringState.setLimit(position - 1); 950 951 if (scripting && !stringState.isEmpty()) { 952 switch (quote) { 953 case '`': 954 // Mark the beginning of an exec string. 955 add(EXECSTRING, stringState.position, stringState.limit); 956 // Frame edit string with left brace. 957 add(LBRACE, stringState.position, stringState.position); 958 // Process edit string. 959 editString(type, stringState); 960 // Frame edit string with right brace. 961 add(RBRACE, stringState.limit, stringState.limit); 962 break; 963 case '"': 964 // Only edit double quoted strings. 965 editString(type, stringState); 966 break; 967 case '\'': 968 // Add string token without editing. 969 add(type, stringState.position, stringState.limit); 970 break; 971 default: 972 break; 973 } 974 } else { 975 /// Add string token without editing. 976 add(type, stringState.position, stringState.limit); 977 } 978 } 979 } 980 981 /** 982 * Convert string to number. 983 * 984 * @param valueString String to convert. 985 * @param radix Numeric base. 986 * @return Converted number. 987 */ 988 private static Number valueOf(final String valueString, final int radix) throws NumberFormatException { 989 try { 990 final long value = Long.parseLong(valueString, radix); 991 if(value >= MIN_INT_L && value <= MAX_INT_L) { 992 return Integer.valueOf((int)value); 993 } 994 return Long.valueOf(value); 995 } catch (final NumberFormatException e) { 996 if (radix == 10) { 997 return Double.valueOf(valueString); 998 } 999 1000 double value = 0.0; 1001 1002 for (int i = 0; i < valueString.length(); i++) { 1003 final char ch = valueString.charAt(i); 1004 // Preverified, should always be a valid digit. 1005 final int digit = convertDigit(ch, radix); 1006 value *= radix; 1007 value += digit; 1008 } 1009 1010 return value; 1011 } 1012 } 1013 1014 /** 1015 * Convert string to number. 1016 * 1017 * @param valueString String to convert. 1018 * @return Converted number. 1019 */ 1020 private static Number valueOf(final String valueString) throws NumberFormatException { 1021 return JSType.narrowestIntegerRepresentation(Double.valueOf(valueString)); 1022 } 1023 1024 /** 1025 * Scan a number. 1026 */ 1027 private void scanNumber() { 1028 // Record beginning of number. 1029 final int start = position; 1030 // Assume value is a decimal. 1031 TokenType type = DECIMAL; 1032 1033 // First digit of number. 1034 int digit = convertDigit(ch0, 10); 1035 1036 // If number begins with 0x. 1037 if (digit == 0 && (ch1 == 'x' || ch1 == 'X') && convertDigit(ch2, 16) != -1) { 1038 // Skip over 0xN. 1039 skip(3); 1040 // Skip over remaining digits. 1041 while (convertDigit(ch0, 16) != -1) { 1042 skip(1); 1043 } 1044 1045 type = HEXADECIMAL; 1046 } else { 1047 // Check for possible octal constant. 1048 boolean octal = digit == 0; 1049 // Skip first digit if not leading '.'. 1050 if (digit != -1) { 1051 skip(1); 1052 } 1053 1054 // Skip remaining digits. 1055 while ((digit = convertDigit(ch0, 10)) != -1) { 1056 // Check octal only digits. 1057 octal = octal && digit < 8; 1058 // Skip digit. 1059 skip(1); 1060 } 1061 1062 if (octal && position - start > 1) { 1063 type = OCTAL; 1064 } else if (ch0 == '.' || ch0 == 'E' || ch0 == 'e') { 1065 // Must be a double. 1066 if (ch0 == '.') { 1067 // Skip period. 1068 skip(1); 1069 // Skip mantissa. 1070 while (convertDigit(ch0, 10) != -1) { 1071 skip(1); 1072 } 1073 } 1074 1075 // Detect exponent. 1076 if (ch0 == 'E' || ch0 == 'e') { 1077 // Skip E. 1078 skip(1); 1079 // Detect and skip exponent sign. 1080 if (ch0 == '+' || ch0 == '-') { 1081 skip(1); 1082 } 1083 // Skip exponent. 1084 while (convertDigit(ch0, 10) != -1) { 1085 skip(1); 1086 } 1087 } 1088 1089 type = FLOATING; 1090 } 1091 } 1092 1093 // Add number token. 1094 add(type, start); 1095 } 1096 1097 /** 1098 * Convert a regex token to a token object. 1099 * 1100 * @param start Position in source content. 1101 * @param length Length of regex token. 1102 * @return Regex token object. 1103 */ 1104 XMLToken valueOfXML(final int start, final int length) { 1105 return new XMLToken(source.getString(start, length)); 1106 } 1107 1108 /** 1109 * Scan over a XML token. 1110 * 1111 * @return TRUE if is an XML literal. 1112 */ 1113 private boolean scanXMLLiteral() { 1114 assert ch0 == '<' && Character.isJavaIdentifierStart(ch1); 1115 if (XML_LITERALS) { 1116 // Record beginning of xml expression. 1117 final int start = position; 1118 1119 int openCount = 0; 1120 1121 do { 1122 if (ch0 == '<') { 1123 if (ch1 == '/' && Character.isJavaIdentifierStart(ch2)) { 1124 skip(3); 1125 openCount--; 1126 } else if (Character.isJavaIdentifierStart(ch1)) { 1127 skip(2); 1128 openCount++; 1129 } else if (ch1 == '?') { 1130 skip(2); 1131 } else if (ch1 == '!' && ch2 == '-' && ch3 == '-') { 1132 skip(4); 1133 } else { 1134 reset(start); 1135 return false; 1136 } 1137 1138 while (!atEOF() && ch0 != '>') { 1139 if (ch0 == '/' && ch1 == '>') { 1140 openCount--; 1141 skip(1); 1142 break; 1143 } else if (ch0 == '\"' || ch0 == '\'') { 1144 scanString(false); 1145 } else { 1146 skip(1); 1147 } 1148 } 1149 1150 if (ch0 != '>') { 1151 reset(start); 1152 return false; 1153 } 1154 1155 skip(1); 1156 } else if (atEOF()) { 1157 reset(start); 1158 return false; 1159 } else { 1160 skip(1); 1161 } 1162 } while (openCount > 0); 1163 1164 add(XML, start); 1165 return true; 1166 } 1167 1168 return false; 1169 } 1170 1171 /** 1172 * Scan over identifier characters. 1173 * 1174 * @return Length of identifier or zero if none found. 1175 */ 1176 private int scanIdentifier() { 1177 final int start = position; 1178 1179 // Make sure first character is valid start character. 1180 if (ch0 == '\\' && ch1 == 'u') { 1181 final int ch = valueOfSequence(16, 4, true, true, TokenType.IDENT); 1182 1183 if (!Character.isJavaIdentifierStart(ch)) { 1184 error(Lexer.message("illegal.identifier.character"), TokenType.IDENT, start, position); 1185 } 1186 } else if (!Character.isJavaIdentifierStart(ch0)) { 1187 // Not an identifier. 1188 return 0; 1189 } 1190 1191 // Make sure remaining characters are valid part characters. 1192 while (!atEOF()) { 1193 if (ch0 == '\\' && ch1 == 'u') { 1194 final int ch = valueOfSequence(16, 4, true, true, TokenType.IDENT); 1195 1196 if (!Character.isJavaIdentifierPart(ch)) { 1197 error(Lexer.message("illegal.identifier.character"), TokenType.IDENT, start, position); 1198 } 1199 } else if (Character.isJavaIdentifierPart(ch0)) { 1200 skip(1); 1201 } else { 1202 break; 1203 } 1204 } 1205 1206 // Length of identifier sequence. 1207 return position - start; 1208 } 1209 1210 /** 1211 * Compare two identifiers (in content) for equality. 1212 * 1213 * @param aStart Start of first identifier. 1214 * @param aLength Length of first identifier. 1215 * @param bStart Start of second identifier. 1216 * @param bLength Length of second identifier. 1217 * @return True if equal. 1218 */ 1219 private boolean identifierEqual(final int aStart, final int aLength, final int bStart, final int bLength) { 1220 if (aLength == bLength) { 1221 for (int i = 0; i < aLength; i++) { 1222 if (content[aStart + i] != content[bStart + i]) { 1223 return false; 1224 } 1225 } 1226 1227 return true; 1228 } 1229 1230 return false; 1231 } 1232 1233 /** 1234 * Detect if a line starts with a marker identifier. 1235 * 1236 * @param identStart Start of identifier. 1237 * @param identLength Length of identifier. 1238 * @return True if detected. 1239 */ 1240 private boolean hasHereMarker(final int identStart, final int identLength) { 1241 // Skip any whitespace. 1242 skipWhitespace(false); 1243 1244 return identifierEqual(identStart, identLength, position, scanIdentifier()); 1245 } 1246 1247 /** 1248 * Lexer to service edit strings. 1249 */ 1250 private static class EditStringLexer extends Lexer { 1251 /** Type of string literals to emit. */ 1252 final TokenType stringType; 1253 1254 /* 1255 * Constructor. 1256 */ 1257 1258 EditStringLexer(final Lexer lexer, final TokenType stringType, final State stringState) { 1259 super(lexer, stringState); 1260 1261 this.stringType = stringType; 1262 } 1263 1264 /** 1265 * Lexify the contents of the string. 1266 */ 1267 @Override 1268 public void lexify() { 1269 // Record start of string position. 1270 int stringStart = position; 1271 // Indicate that the priming first string has not been emitted. 1272 boolean primed = false; 1273 1274 while (true) { 1275 // Detect end of content. 1276 if (atEOF()) { 1277 break; 1278 } 1279 1280 // Honour escapes (should be well formed.) 1281 if (ch0 == '\\' && stringType == ESCSTRING) { 1282 skip(2); 1283 1284 continue; 1285 } 1286 1287 // If start of expression. 1288 if (ch0 == '$' && ch1 == '{') { 1289 if (!primed || stringStart != position) { 1290 if (primed) { 1291 add(ADD, stringStart, stringStart + 1); 1292 } 1293 1294 add(stringType, stringStart, position); 1295 primed = true; 1296 } 1297 1298 // Skip ${ 1299 skip(2); 1300 1301 // Save expression state. 1302 final State expressionState = saveState(); 1303 1304 // Start with one open brace. 1305 int braceCount = 1; 1306 1307 // Scan for the rest of the string. 1308 while (!atEOF()) { 1309 // If closing brace. 1310 if (ch0 == '}') { 1311 // Break only only if matching brace. 1312 if (--braceCount == 0) { 1313 break; 1314 } 1315 } else if (ch0 == '{') { 1316 // Bump up the brace count. 1317 braceCount++; 1318 } 1319 1320 // Skip to next character. 1321 skip(1); 1322 } 1323 1324 // If braces don't match then report an error. 1325 if (braceCount != 0) { 1326 error(Lexer.message("edit.string.missing.brace"), LBRACE, expressionState.position - 1, 1); 1327 } 1328 1329 // Mark end of expression. 1330 expressionState.setLimit(position); 1331 // Skip closing brace. 1332 skip(1); 1333 1334 // Start next string. 1335 stringStart = position; 1336 1337 // Concatenate expression. 1338 add(ADD, expressionState.position, expressionState.position + 1); 1339 add(LPAREN, expressionState.position, expressionState.position + 1); 1340 1341 // Scan expression. 1342 final Lexer lexer = new Lexer(this, expressionState); 1343 lexer.lexify(); 1344 1345 // Close out expression parenthesis. 1346 add(RPAREN, position - 1, position); 1347 1348 continue; 1349 } 1350 1351 // Next character in string. 1352 skip(1); 1353 } 1354 1355 // If there is any unemitted string portion. 1356 if (stringStart != limit) { 1357 // Concatenate remaining string. 1358 if (primed) { 1359 add(ADD, stringStart, 1); 1360 } 1361 1362 add(stringType, stringStart, limit); 1363 } 1364 } 1365 1366 } 1367 1368 /** 1369 * Edit string for nested expressions. 1370 * 1371 * @param stringType Type of string literals to emit. 1372 * @param stringState State of lexer at start of string. 1373 */ 1374 private void editString(final TokenType stringType, final State stringState) { 1375 // Use special lexer to scan string. 1376 final EditStringLexer lexer = new EditStringLexer(this, stringType, stringState); 1377 lexer.lexify(); 1378 1379 // Need to keep lexer informed. 1380 last = stringType; 1381 } 1382 1383 /** 1384 * Scan over a here string. 1385 * 1386 * @return TRUE if is a here string. 1387 */ 1388 private boolean scanHereString() { 1389 assert ch0 == '<' && ch1 == '<'; 1390 if (scripting) { 1391 // Record beginning of here string. 1392 final State saved = saveState(); 1393 1394 // << or <<< 1395 final boolean excludeLastEOL = ch2 != '<'; 1396 1397 if (excludeLastEOL) { 1398 skip(2); 1399 } else { 1400 skip(3); 1401 } 1402 1403 // Scan identifier. 1404 final int identStart = position; 1405 final int identLength = scanIdentifier(); 1406 1407 // Check for identifier. 1408 if (identLength == 0) { 1409 // Treat as shift. 1410 restoreState(saved); 1411 1412 return false; 1413 } 1414 1415 // Record rest of line. 1416 final State restState = saveState(); 1417 skipLine(false); 1418 restState.setLimit(position); 1419 1420 // Record beginning of string. 1421 final State stringState = saveState(); 1422 int stringEnd = position; 1423 1424 // Hunt down marker. 1425 while (!atEOF()) { 1426 // Skip any whitespace. 1427 skipWhitespace(false); 1428 1429 if (hasHereMarker(identStart, identLength)) { 1430 break; 1431 } 1432 1433 skipLine(false); 1434 stringEnd = position; 1435 } 1436 1437 // Record end of string. 1438 stringState.setLimit(stringEnd); 1439 1440 // If marker is missing. 1441 if (stringState.isEmpty() || atEOF()) { 1442 error(Lexer.message("here.missing.end.marker", source.getString(identStart, identLength)), last, position, position); 1443 restoreState(saved); 1444 1445 return false; 1446 } 1447 1448 // Remove last end of line if specified. 1449 if (excludeLastEOL) { 1450 // Handles \n. 1451 if (content[stringEnd - 1] == '\n') { 1452 stringEnd--; 1453 } 1454 1455 // Handles \r and \r\n. 1456 if (content[stringEnd - 1] == '\r') { 1457 stringEnd--; 1458 } 1459 1460 // Update end of string. 1461 stringState.setLimit(stringEnd); 1462 } 1463 1464 // Edit string if appropriate. 1465 if (scripting && !stringState.isEmpty()) { 1466 editString(STRING, stringState); 1467 } else { 1468 // Add here string. 1469 add(STRING, stringState.position, stringState.limit); 1470 } 1471 1472 // Scan rest of original line. 1473 final Lexer restLexer = new Lexer(this, restState); 1474 1475 restLexer.lexify(); 1476 1477 return true; 1478 } 1479 1480 return false; 1481 } 1482 1483 /** 1484 * Breaks source content down into lex units, adding tokens to the token 1485 * stream. The routine scans until the stream buffer is full. Can be called 1486 * repeatedly until EOF is detected. 1487 */ 1488 public void lexify() { 1489 while (!stream.isFull() || nested) { 1490 // Skip over whitespace. 1491 skipWhitespace(true); 1492 1493 // Detect end of file. 1494 if (atEOF()) { 1495 if (!nested) { 1496 // Add an EOF token at the end. 1497 add(EOF, position); 1498 } 1499 1500 break; 1501 } 1502 1503 // Check for comments. Note that we don't scan for regexp and other literals here as 1504 // we may not have enough context to distinguish them from similar looking operators. 1505 // Instead we break on ambiguous operators below and let the parser decide. 1506 if (ch0 == '/' && skipComments()) { 1507 continue; 1508 } 1509 1510 if (scripting && ch0 == '#' && skipComments()) { 1511 continue; 1512 } 1513 1514 // TokenType for lookup of delimiter or operator. 1515 TokenType type; 1516 1517 if (ch0 == '.' && convertDigit(ch1, 10) != -1) { 1518 // '.' followed by digit. 1519 // Scan and add a number. 1520 scanNumber(); 1521 } else if ((type = TokenLookup.lookupOperator(ch0, ch1, ch2, ch3)) != null) { 1522 // Get the number of characters in the token. 1523 final int typeLength = type.getLength(); 1524 // Skip that many characters. 1525 skip(typeLength); 1526 // Add operator token. 1527 add(type, position - typeLength); 1528 // Some operator tokens also mark the beginning of regexp, XML, or here string literals. 1529 // We break to let the parser decide what it is. 1530 if (canStartLiteral(type)) { 1531 break; 1532 } 1533 } else if (Character.isJavaIdentifierStart(ch0) || ch0 == '\\' && ch1 == 'u') { 1534 // Scan and add identifier or keyword. 1535 scanIdentifierOrKeyword(); 1536 } else if (isStringDelimiter(ch0)) { 1537 // Scan and add a string. 1538 scanString(true); 1539 } else if (Character.isDigit(ch0)) { 1540 // Scan and add a number. 1541 scanNumber(); 1542 } else { 1543 // Don't recognize this character. 1544 skip(1); 1545 add(ERROR, position - 1); 1546 } 1547 } 1548 } 1549 1550 /** 1551 * Return value of token given its token descriptor. 1552 * 1553 * @param token Token descriptor. 1554 * @return JavaScript value. 1555 */ 1556 Object getValueOf(final long token, final boolean strict) { 1557 final int start = Token.descPosition(token); 1558 final int len = Token.descLength(token); 1559 1560 switch (Token.descType(token)) { 1561 case DECIMAL: 1562 return Lexer.valueOf(source.getString(start, len), 10); // number 1563 case OCTAL: 1564 return Lexer.valueOf(source.getString(start, len), 8); // number 1565 case HEXADECIMAL: 1566 return Lexer.valueOf(source.getString(start + 2, len - 2), 16); // number 1567 case FLOATING: 1568 return Lexer.valueOf(source.getString(start, len)); // number 1569 case STRING: 1570 return source.getString(start, len); // String 1571 case ESCSTRING: 1572 return valueOfString(start, len, strict); // String 1573 case IDENT: 1574 return valueOfIdent(start, len); // String 1575 case REGEX: 1576 return valueOfPattern(start, len); // RegexToken::LexerToken 1577 case XML: 1578 return valueOfXML(start, len); // XMLToken::LexerToken 1579 default: 1580 break; 1581 } 1582 1583 return null; 1584 } 1585 1586 private static String message(final String msgId, final String... args) { 1587 return ECMAErrors.getMessage("lexer.error." + msgId, args); 1588 } 1589 1590 /** 1591 * Generate a runtime exception 1592 * 1593 * @param message error message 1594 * @param type token type 1595 * @param start start position of lexed error 1596 * @param length length of lexed error 1597 * @throws ParserException unconditionally 1598 */ 1599 protected void error(final String message, final TokenType type, final int start, final int length) throws ParserException { 1600 final long token = Token.toDesc(type, start, length); 1601 final int pos = Token.descPosition(token); 1602 final int lineNum = source.getLine(pos); 1603 final int columnNum = source.getColumn(pos); 1604 final String formatted = ErrorManager.format(message, source, lineNum, columnNum, token); 1605 throw new ParserException(JSErrorType.SYNTAX_ERROR, formatted, source, lineNum, columnNum, token); 1606 } 1607 1608 /** 1609 * Helper class for Lexer tokens, e.g XML or RegExp tokens. 1610 * This is the abstract superclass 1611 */ 1612 public static abstract class LexerToken { 1613 private final String expression; 1614 1615 /** 1616 * Constructor 1617 * @param expression token expression 1618 */ 1619 protected LexerToken(final String expression) { 1620 this.expression = expression; 1621 } 1622 1623 /** 1624 * Get the expression 1625 * @return expression 1626 */ 1627 public String getExpression() { 1628 return expression; 1629 } 1630 } 1631 1632 /** 1633 * Temporary container for regular expressions. 1634 */ 1635 public static class RegexToken extends LexerToken { 1636 /** Options. */ 1637 private final String options; 1638 1639 /** 1640 * Constructor. 1641 * 1642 * @param expression regexp expression 1643 * @param options regexp options 1644 */ 1645 public RegexToken(final String expression, final String options) { 1646 super(expression); 1647 this.options = options; 1648 } 1649 1650 /** 1651 * Get regexp options 1652 * @return options 1653 */ 1654 public String getOptions() { 1655 return options; 1656 } 1657 1658 @Override 1659 public String toString() { 1660 return '/' + getExpression() + '/' + options; 1661 } 1662 } 1663 1664 /** 1665 * Temporary container for XML expression. 1666 */ 1667 public static class XMLToken extends LexerToken { 1668 1669 /** 1670 * Constructor. 1671 * 1672 * @param expression XML expression 1673 */ 1674 public XMLToken(final String expression) { 1675 super(expression); 1676 } 1677 } 1678 }