1 // Copyright 2011 the V8 project authors. All rights reserved. 2 // Redistribution and use in source and binary forms, with or without 3 // modification, are permitted provided that the following conditions are 4 // met: 5 // 6 // * Redistributions of source code must retain the above copyright 7 // notice, this list of conditions and the following disclaimer. 8 // * Redistributions in binary form must reproduce the above 9 // copyright notice, this list of conditions and the following 10 // disclaimer in the documentation and/or other materials provided 11 // with the distribution. 12 // * Neither the name of Google Inc. nor the names of its 13 // contributors may be used to endorse or promote products derived 14 // from this software without specific prior written permission. 15 // 16 // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 17 // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 18 // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR 19 // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT 20 // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, 21 // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT 22 // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 23 // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 24 // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 25 // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE 26 // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 27 28 // Features shared by parsing and pre-parsing scanners. 29 30 #ifndef V8_SCANNER_H_ 31 #define V8_SCANNER_H_ 32 33 #include "allocation.h" 34 #include "char-predicates.h" 35 #include "checks.h" 36 #include "globals.h" 37 #include "token.h" 38 #include "unicode-inl.h" 39 #include "utils.h" 40 41 namespace v8 { 42 namespace internal { 43 44 45 // Returns the value (0 .. 15) of a hexadecimal character c. 46 // If c is not a legal hexadecimal character, returns a value < 0. 47 inline int HexValue(uc32 c) { 48 c -= '0'; 49 if (static_cast<unsigned>(c) <= 9) return c; 50 c = (c | 0x20) - ('a' - '0'); // detect 0x11..0x16 and 0x31..0x36. 51 if (static_cast<unsigned>(c) <= 5) return c + 10; 52 return -1; 53 } 54 55 56 // --------------------------------------------------------------------- 57 // Buffered stream of UTF-16 code units, using an internal UTF-16 buffer. 58 // A code unit is a 16 bit value representing either a 16 bit code point 59 // or one part of a surrogate pair that make a single 21 bit code point. 60 61 class Utf16CharacterStream { 62 public: 63 Utf16CharacterStream() : pos_(0) { } 64 virtual ~Utf16CharacterStream() { } 65 66 // Returns and advances past the next UTF-16 code unit in the input 67 // stream. If there are no more code units, it returns a negative 68 // value. 69 inline uc32 Advance() { 70 if (buffer_cursor_ < buffer_end_ || ReadBlock()) { 71 pos_++; 72 return static_cast<uc32>(*(buffer_cursor_++)); 73 } 74 // Note: currently the following increment is necessary to avoid a 75 // parser problem! The scanner treats the final kEndOfInput as 76 // a code unit with a position, and does math relative to that 77 // position. 78 pos_++; 79 80 return kEndOfInput; 81 } 82 83 // Return the current position in the code unit stream. 84 // Starts at zero. 85 inline unsigned pos() const { return pos_; } 86 87 // Skips forward past the next code_unit_count UTF-16 code units 88 // in the input, or until the end of input if that comes sooner. 89 // Returns the number of code units actually skipped. If less 90 // than code_unit_count, 91 inline unsigned SeekForward(unsigned code_unit_count) { 92 unsigned buffered_chars = 93 static_cast<unsigned>(buffer_end_ - buffer_cursor_); 94 if (code_unit_count <= buffered_chars) { 95 buffer_cursor_ += code_unit_count; 96 pos_ += code_unit_count; 97 return code_unit_count; 98 } 99 return SlowSeekForward(code_unit_count); 100 } 101 102 // Pushes back the most recently read UTF-16 code unit (or negative 103 // value if at end of input), i.e., the value returned by the most recent 104 // call to Advance. 105 // Must not be used right after calling SeekForward. 106 virtual void PushBack(int32_t code_unit) = 0; 107 108 protected: 109 static const uc32 kEndOfInput = -1; 110 111 // Ensures that the buffer_cursor_ points to the code_unit at 112 // position pos_ of the input, if possible. If the position 113 // is at or after the end of the input, return false. If there 114 // are more code_units available, return true. 115 virtual bool ReadBlock() = 0; 116 virtual unsigned SlowSeekForward(unsigned code_unit_count) = 0; 117 118 const uc16* buffer_cursor_; 119 const uc16* buffer_end_; 120 unsigned pos_; 121 }; 122 123 124 class UnicodeCache { 125 // --------------------------------------------------------------------- 126 // Caching predicates used by scanners. 127 public: 128 UnicodeCache() {} 129 typedef unibrow::Utf8Decoder<512> Utf8Decoder; 130 131 StaticResource<Utf8Decoder>* utf8_decoder() { 132 return &utf8_decoder_; 133 } 134 135 bool IsIdentifierStart(unibrow::uchar c) { return kIsIdentifierStart.get(c); } 136 bool IsIdentifierPart(unibrow::uchar c) { return kIsIdentifierPart.get(c); } 137 bool IsLineTerminator(unibrow::uchar c) { return kIsLineTerminator.get(c); } 138 bool IsWhiteSpace(unibrow::uchar c) { return kIsWhiteSpace.get(c); } 139 140 private: 141 unibrow::Predicate<IdentifierStart, 128> kIsIdentifierStart; 142 unibrow::Predicate<IdentifierPart, 128> kIsIdentifierPart; 143 unibrow::Predicate<unibrow::LineTerminator, 128> kIsLineTerminator; 144 unibrow::Predicate<unibrow::WhiteSpace, 128> kIsWhiteSpace; 145 StaticResource<Utf8Decoder> utf8_decoder_; 146 147 DISALLOW_COPY_AND_ASSIGN(UnicodeCache); 148 }; 149 150 151 // ---------------------------------------------------------------------------- 152 // LiteralBuffer - Collector of chars of literals. 153 154 class LiteralBuffer { 155 public: 156 LiteralBuffer() : is_ascii_(true), position_(0), backing_store_() { } 157 158 ~LiteralBuffer() { 159 if (backing_store_.length() > 0) { 160 backing_store_.Dispose(); 161 } 162 } 163 164 INLINE(void AddChar(uint32_t code_unit)) { 165 if (position_ >= backing_store_.length()) ExpandBuffer(); 166 if (is_ascii_) { 167 if (code_unit <= unibrow::Latin1::kMaxChar) { 168 backing_store_[position_] = static_cast<byte>(code_unit); 169 position_ += kOneByteSize; 170 return; 171 } 172 ConvertToUtf16(); 173 } 174 ASSERT(code_unit < 0x10000u); 175 *reinterpret_cast<uc16*>(&backing_store_[position_]) = code_unit; 176 position_ += kUC16Size; 177 } 178 179 bool is_ascii() { return is_ascii_; } 180 181 bool is_contextual_keyword(Vector<const char> keyword) { 182 return is_ascii() && keyword.length() == position_ && 183 (memcmp(keyword.start(), backing_store_.start(), position_) == 0); 184 } 185 186 Vector<const uc16> utf16_literal() { 187 ASSERT(!is_ascii_); 188 ASSERT((position_ & 0x1) == 0); 189 return Vector<const uc16>( 190 reinterpret_cast<const uc16*>(backing_store_.start()), 191 position_ >> 1); 192 } 193 194 Vector<const char> ascii_literal() { 195 ASSERT(is_ascii_); 196 return Vector<const char>( 197 reinterpret_cast<const char*>(backing_store_.start()), 198 position_); 199 } 200 201 int length() { 202 return is_ascii_ ? position_ : (position_ >> 1); 203 } 204 205 void Reset() { 206 position_ = 0; 207 is_ascii_ = true; 208 } 209 210 private: 211 static const int kInitialCapacity = 16; 212 static const int kGrowthFactory = 4; 213 static const int kMinConversionSlack = 256; 214 static const int kMaxGrowth = 1 * MB; 215 inline int NewCapacity(int min_capacity) { 216 int capacity = Max(min_capacity, backing_store_.length()); 217 int new_capacity = Min(capacity * kGrowthFactory, capacity + kMaxGrowth); 218 return new_capacity; 219 } 220 221 void ExpandBuffer() { 222 Vector<byte> new_store = Vector<byte>::New(NewCapacity(kInitialCapacity)); 223 OS::MemCopy(new_store.start(), backing_store_.start(), position_); 224 backing_store_.Dispose(); 225 backing_store_ = new_store; 226 } 227 228 void ConvertToUtf16() { 229 ASSERT(is_ascii_); 230 Vector<byte> new_store; 231 int new_content_size = position_ * kUC16Size; 232 if (new_content_size >= backing_store_.length()) { 233 // Ensure room for all currently read code units as UC16 as well 234 // as the code unit about to be stored. 235 new_store = Vector<byte>::New(NewCapacity(new_content_size)); 236 } else { 237 new_store = backing_store_; 238 } 239 uint8_t* src = backing_store_.start(); 240 uc16* dst = reinterpret_cast<uc16*>(new_store.start()); 241 for (int i = position_ - 1; i >= 0; i--) { 242 dst[i] = src[i]; 243 } 244 if (new_store.start() != backing_store_.start()) { 245 backing_store_.Dispose(); 246 backing_store_ = new_store; 247 } 248 position_ = new_content_size; 249 is_ascii_ = false; 250 } 251 252 bool is_ascii_; 253 int position_; 254 Vector<byte> backing_store_; 255 256 DISALLOW_COPY_AND_ASSIGN(LiteralBuffer); 257 }; 258 259 260 // ---------------------------------------------------------------------------- 261 // JavaScript Scanner. 262 263 class Scanner { 264 public: 265 // Scoped helper for literal recording. Automatically drops the literal 266 // if aborting the scanning before it's complete. 267 class LiteralScope { 268 public: 269 explicit LiteralScope(Scanner* self) 270 : scanner_(self), complete_(false) { 271 scanner_->StartLiteral(); 272 } 273 ~LiteralScope() { 274 if (!complete_) scanner_->DropLiteral(); 275 } 276 void Complete() { 277 scanner_->TerminateLiteral(); 278 complete_ = true; 279 } 280 281 private: 282 Scanner* scanner_; 283 bool complete_; 284 }; 285 286 // Representation of an interval of source positions. 287 struct Location { 288 Location(int b, int e) : beg_pos(b), end_pos(e) { } 289 Location() : beg_pos(0), end_pos(0) { } 290 291 bool IsValid() const { 292 return beg_pos >= 0 && end_pos >= beg_pos; 293 } 294 295 static Location invalid() { return Location(-1, -1); } 296 297 int beg_pos; 298 int end_pos; 299 }; 300 301 // -1 is outside of the range of any real source code. 302 static const int kNoOctalLocation = -1; 303 304 explicit Scanner(UnicodeCache* scanner_contants); 305 306 void Initialize(Utf16CharacterStream* source); 307 308 // Returns the next token and advances input. 309 Token::Value Next(); 310 // Returns the current token again. 311 Token::Value current_token() { return current_.token; } 312 // Returns the location information for the current token 313 // (the token last returned by Next()). 314 Location location() const { return current_.location; } 315 // Returns the literal string, if any, for the current token (the 316 // token last returned by Next()). The string is 0-terminated. 317 // Literal strings are collected for identifiers, strings, and 318 // numbers. 319 // These functions only give the correct result if the literal 320 // was scanned between calls to StartLiteral() and TerminateLiteral(). 321 Vector<const char> literal_ascii_string() { 322 ASSERT_NOT_NULL(current_.literal_chars); 323 return current_.literal_chars->ascii_literal(); 324 } 325 Vector<const uc16> literal_utf16_string() { 326 ASSERT_NOT_NULL(current_.literal_chars); 327 return current_.literal_chars->utf16_literal(); 328 } 329 bool is_literal_ascii() { 330 ASSERT_NOT_NULL(current_.literal_chars); 331 return current_.literal_chars->is_ascii(); 332 } 333 bool is_literal_contextual_keyword(Vector<const char> keyword) { 334 ASSERT_NOT_NULL(current_.literal_chars); 335 return current_.literal_chars->is_contextual_keyword(keyword); 336 } 337 int literal_length() const { 338 ASSERT_NOT_NULL(current_.literal_chars); 339 return current_.literal_chars->length(); 340 } 341 342 bool literal_contains_escapes() const { 343 Location location = current_.location; 344 int source_length = (location.end_pos - location.beg_pos); 345 if (current_.token == Token::STRING) { 346 // Subtract delimiters. 347 source_length -= 2; 348 } 349 return current_.literal_chars->length() != source_length; 350 } 351 352 // Similar functions for the upcoming token. 353 354 // One token look-ahead (past the token returned by Next()). 355 Token::Value peek() const { return next_.token; } 356 357 Location peek_location() const { return next_.location; } 358 359 // Returns the literal string for the next token (the token that 360 // would be returned if Next() were called). 361 Vector<const char> next_literal_ascii_string() { 362 ASSERT_NOT_NULL(next_.literal_chars); 363 return next_.literal_chars->ascii_literal(); 364 } 365 Vector<const uc16> next_literal_utf16_string() { 366 ASSERT_NOT_NULL(next_.literal_chars); 367 return next_.literal_chars->utf16_literal(); 368 } 369 bool is_next_literal_ascii() { 370 ASSERT_NOT_NULL(next_.literal_chars); 371 return next_.literal_chars->is_ascii(); 372 } 373 bool is_next_contextual_keyword(Vector<const char> keyword) { 374 ASSERT_NOT_NULL(next_.literal_chars); 375 return next_.literal_chars->is_contextual_keyword(keyword); 376 } 377 int next_literal_length() const { 378 ASSERT_NOT_NULL(next_.literal_chars); 379 return next_.literal_chars->length(); 380 } 381 382 UnicodeCache* unicode_cache() { return unicode_cache_; } 383 384 static const int kCharacterLookaheadBufferSize = 1; 385 386 // Scans octal escape sequence. Also accepts "\0" decimal escape sequence. 387 uc32 ScanOctalEscape(uc32 c, int length); 388 389 // Returns the location of the last seen octal literal. 390 Location octal_position() const { return octal_pos_; } 391 void clear_octal_position() { octal_pos_ = Location::invalid(); } 392 393 // Seek forward to the given position. This operation does not 394 // work in general, for instance when there are pushed back 395 // characters, but works for seeking forward until simple delimiter 396 // tokens, which is what it is used for. 397 void SeekForward(int pos); 398 399 bool HarmonyScoping() const { 400 return harmony_scoping_; 401 } 402 void SetHarmonyScoping(bool scoping) { 403 harmony_scoping_ = scoping; 404 } 405 bool HarmonyModules() const { 406 return harmony_modules_; 407 } 408 void SetHarmonyModules(bool modules) { 409 harmony_modules_ = modules; 410 } 411 bool HarmonyNumericLiterals() const { 412 return harmony_numeric_literals_; 413 } 414 void SetHarmonyNumericLiterals(bool numeric_literals) { 415 harmony_numeric_literals_ = numeric_literals; 416 } 417 418 // Returns true if there was a line terminator before the peek'ed token, 419 // possibly inside a multi-line comment. 420 bool HasAnyLineTerminatorBeforeNext() const { 421 return has_line_terminator_before_next_ || 422 has_multiline_comment_before_next_; 423 } 424 425 // Scans the input as a regular expression pattern, previous 426 // character(s) must be /(=). Returns true if a pattern is scanned. 427 bool ScanRegExpPattern(bool seen_equal); 428 // Returns true if regexp flags are scanned (always since flags can 429 // be empty). 430 bool ScanRegExpFlags(); 431 432 private: 433 // The current and look-ahead token. 434 struct TokenDesc { 435 Token::Value token; 436 Location location; 437 LiteralBuffer* literal_chars; 438 }; 439 440 // Call this after setting source_ to the input. 441 void Init() { 442 // Set c0_ (one character ahead) 443 STATIC_ASSERT(kCharacterLookaheadBufferSize == 1); 444 Advance(); 445 // Initialize current_ to not refer to a literal. 446 current_.literal_chars = NULL; 447 } 448 449 // Literal buffer support 450 inline void StartLiteral() { 451 LiteralBuffer* free_buffer = (current_.literal_chars == &literal_buffer1_) ? 452 &literal_buffer2_ : &literal_buffer1_; 453 free_buffer->Reset(); 454 next_.literal_chars = free_buffer; 455 } 456 457 INLINE(void AddLiteralChar(uc32 c)) { 458 ASSERT_NOT_NULL(next_.literal_chars); 459 next_.literal_chars->AddChar(c); 460 } 461 462 // Complete scanning of a literal. 463 inline void TerminateLiteral() { 464 // Does nothing in the current implementation. 465 } 466 467 // Stops scanning of a literal and drop the collected characters, 468 // e.g., due to an encountered error. 469 inline void DropLiteral() { 470 next_.literal_chars = NULL; 471 } 472 473 inline void AddLiteralCharAdvance() { 474 AddLiteralChar(c0_); 475 Advance(); 476 } 477 478 // Low-level scanning support. 479 void Advance() { c0_ = source_->Advance(); } 480 void PushBack(uc32 ch) { 481 source_->PushBack(c0_); 482 c0_ = ch; 483 } 484 485 inline Token::Value Select(Token::Value tok) { 486 Advance(); 487 return tok; 488 } 489 490 inline Token::Value Select(uc32 next, Token::Value then, Token::Value else_) { 491 Advance(); 492 if (c0_ == next) { 493 Advance(); 494 return then; 495 } else { 496 return else_; 497 } 498 } 499 500 uc32 ScanHexNumber(int expected_length); 501 502 // Scans a single JavaScript token. 503 void Scan(); 504 505 bool SkipWhiteSpace(); 506 Token::Value SkipSingleLineComment(); 507 Token::Value SkipMultiLineComment(); 508 // Scans a possible HTML comment -- begins with '<!'. 509 Token::Value ScanHtmlComment(); 510 511 void ScanDecimalDigits(); 512 Token::Value ScanNumber(bool seen_period); 513 Token::Value ScanIdentifierOrKeyword(); 514 Token::Value ScanIdentifierSuffix(LiteralScope* literal); 515 516 Token::Value ScanString(); 517 518 // Scans an escape-sequence which is part of a string and adds the 519 // decoded character to the current literal. Returns true if a pattern 520 // is scanned. 521 bool ScanEscape(); 522 // Decodes a Unicode escape-sequence which is part of an identifier. 523 // If the escape sequence cannot be decoded the result is kBadChar. 524 uc32 ScanIdentifierUnicodeEscape(); 525 // Scans a Unicode escape-sequence and adds its characters, 526 // uninterpreted, to the current literal. Used for parsing RegExp 527 // flags. 528 bool ScanLiteralUnicodeEscape(); 529 530 // Return the current source position. 531 int source_pos() { 532 return source_->pos() - kCharacterLookaheadBufferSize; 533 } 534 535 UnicodeCache* unicode_cache_; 536 537 // Buffers collecting literal strings, numbers, etc. 538 LiteralBuffer literal_buffer1_; 539 LiteralBuffer literal_buffer2_; 540 541 TokenDesc current_; // desc for current token (as returned by Next()) 542 TokenDesc next_; // desc for next token (one token look-ahead) 543 544 // Input stream. Must be initialized to an Utf16CharacterStream. 545 Utf16CharacterStream* source_; 546 547 548 // Start position of the octal literal last scanned. 549 Location octal_pos_; 550 551 // One Unicode character look-ahead; c0_ < 0 at the end of the input. 552 uc32 c0_; 553 554 // Whether there is a line terminator whitespace character after 555 // the current token, and before the next. Does not count newlines 556 // inside multiline comments. 557 bool has_line_terminator_before_next_; 558 // Whether there is a multi-line comment that contains a 559 // line-terminator after the current token, and before the next. 560 bool has_multiline_comment_before_next_; 561 // Whether we scan 'let' as a keyword for harmony block-scoped let bindings. 562 bool harmony_scoping_; 563 // Whether we scan 'module', 'import', 'export' as keywords. 564 bool harmony_modules_; 565 // Whether we scan 0o777 and 0b111 as numbers. 566 bool harmony_numeric_literals_; 567 }; 568 569 } } // namespace v8::internal 570 571 #endif // V8_SCANNER_H_ 572