ager@chromium.org | 9258b6b | 2008-09-11 09:11:10 +0000 | [diff] [blame] | 1 | // Copyright 2006-2008 the V8 project authors. All rights reserved. |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 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 | #include "v8.h" |
| 29 | |
| 30 | #include "ast.h" |
| 31 | #include "scanner.h" |
| 32 | |
kasperl@chromium.org | 71affb5 | 2009-05-26 05:44:31 +0000 | [diff] [blame] | 33 | namespace v8 { |
| 34 | namespace internal { |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 35 | |
| 36 | // ---------------------------------------------------------------------------- |
| 37 | // Character predicates |
| 38 | |
| 39 | |
| 40 | unibrow::Predicate<IdentifierStart, 128> Scanner::kIsIdentifierStart; |
| 41 | unibrow::Predicate<IdentifierPart, 128> Scanner::kIsIdentifierPart; |
| 42 | unibrow::Predicate<unibrow::LineTerminator, 128> Scanner::kIsLineTerminator; |
| 43 | unibrow::Predicate<unibrow::WhiteSpace, 128> Scanner::kIsWhiteSpace; |
| 44 | |
| 45 | |
| 46 | StaticResource<Scanner::Utf8Decoder> Scanner::utf8_decoder_; |
| 47 | |
| 48 | |
| 49 | // ---------------------------------------------------------------------------- |
| 50 | // UTF8Buffer |
| 51 | |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 52 | UTF8Buffer::UTF8Buffer() { |
| 53 | static const int kInitialCapacity = 1 * KB; |
| 54 | data_ = NewArray<char>(kInitialCapacity); |
| 55 | limit_ = ComputeLimit(data_, kInitialCapacity); |
| 56 | Reset(); |
| 57 | ASSERT(Capacity() == kInitialCapacity && pos() == 0); |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 58 | } |
| 59 | |
| 60 | |
| 61 | UTF8Buffer::~UTF8Buffer() { |
| 62 | DeleteArray(data_); |
| 63 | } |
| 64 | |
| 65 | |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 66 | void UTF8Buffer::AddCharSlow(uc32 c) { |
| 67 | static const int kCapacityGrowthLimit = 1 * MB; |
| 68 | if (cursor_ > limit_) { |
| 69 | int old_capacity = Capacity(); |
| 70 | int old_position = pos(); |
| 71 | int new_capacity = |
| 72 | Min(old_capacity * 2, old_capacity + kCapacityGrowthLimit); |
| 73 | char* new_data = NewArray<char>(new_capacity); |
| 74 | memcpy(new_data, data_, old_position); |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 75 | DeleteArray(data_); |
| 76 | data_ = new_data; |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 77 | cursor_ = new_data + old_position; |
| 78 | limit_ = ComputeLimit(new_data, new_capacity); |
| 79 | ASSERT(Capacity() == new_capacity && pos() == old_position); |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 80 | } |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 81 | if (static_cast<unsigned>(c) <= unibrow::Utf8::kMaxOneByteChar) { |
| 82 | *cursor_++ = c; // Common case: 7-bit ASCII. |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 83 | } else { |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 84 | cursor_ += unibrow::Utf8::Encode(cursor_, c); |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 85 | } |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 86 | ASSERT(pos() <= Capacity()); |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 87 | } |
| 88 | |
| 89 | |
| 90 | // ---------------------------------------------------------------------------- |
| 91 | // UTF16Buffer |
| 92 | |
| 93 | |
| 94 | UTF16Buffer::UTF16Buffer() |
| 95 | : pos_(0), |
| 96 | pushback_buffer_(0), |
| 97 | last_(0), |
| 98 | stream_(NULL) { } |
| 99 | |
| 100 | |
| 101 | void UTF16Buffer::Initialize(Handle<String> data, |
| 102 | unibrow::CharacterStream* input) { |
| 103 | data_ = data; |
| 104 | pos_ = 0; |
| 105 | stream_ = input; |
| 106 | } |
| 107 | |
| 108 | |
| 109 | Handle<String> UTF16Buffer::SubString(int start, int end) { |
| 110 | return internal::SubString(data_, start, end); |
| 111 | } |
| 112 | |
| 113 | |
| 114 | void UTF16Buffer::PushBack(uc32 ch) { |
| 115 | pushback_buffer()->Add(last_); |
| 116 | last_ = ch; |
| 117 | pos_--; |
| 118 | } |
| 119 | |
| 120 | |
| 121 | uc32 UTF16Buffer::Advance() { |
| 122 | // NOTE: It is of importance to Persian / Farsi resources that we do |
| 123 | // *not* strip format control characters in the scanner; see |
| 124 | // |
| 125 | // https://bugzilla.mozilla.org/show_bug.cgi?id=274152 |
| 126 | // |
| 127 | // So, even though ECMA-262, section 7.1, page 11, dictates that we |
| 128 | // must remove Unicode format-control characters, we do not. This is |
| 129 | // in line with how IE and SpiderMonkey handles it. |
| 130 | if (!pushback_buffer()->is_empty()) { |
| 131 | pos_++; |
| 132 | return last_ = pushback_buffer()->RemoveLast(); |
| 133 | } else if (stream_->has_more()) { |
| 134 | pos_++; |
| 135 | uc32 next = stream_->GetNext(); |
| 136 | return last_ = next; |
| 137 | } else { |
| 138 | // note: currently the following increment is necessary to avoid a |
| 139 | // test-parser problem! |
| 140 | pos_++; |
| 141 | return last_ = static_cast<uc32>(-1); |
| 142 | } |
| 143 | } |
| 144 | |
| 145 | |
| 146 | void UTF16Buffer::SeekForward(int pos) { |
| 147 | pos_ = pos; |
| 148 | ASSERT(pushback_buffer()->is_empty()); |
| 149 | stream_->Seek(pos); |
| 150 | } |
| 151 | |
| 152 | |
| 153 | // ---------------------------------------------------------------------------- |
| 154 | // Scanner |
| 155 | |
| 156 | Scanner::Scanner(bool pre) : stack_overflow_(false), is_pre_parsing_(pre) { |
| 157 | Token::Initialize(); |
| 158 | } |
| 159 | |
| 160 | |
| 161 | void Scanner::Init(Handle<String> source, unibrow::CharacterStream* stream, |
| 162 | int position) { |
| 163 | // Initialize the source buffer. |
| 164 | source_.Initialize(source, stream); |
| 165 | position_ = position; |
| 166 | |
| 167 | // Reset literals buffer |
| 168 | literals_.Reset(); |
| 169 | |
| 170 | // Set c0_ (one character ahead) |
| 171 | ASSERT(kCharacterLookaheadBufferSize == 1); |
| 172 | Advance(); |
| 173 | |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 174 | // Skip initial whitespace allowing HTML comment ends just like |
| 175 | // after a newline and scan first token. |
| 176 | has_line_terminator_before_next_ = true; |
| 177 | SkipWhiteSpace(); |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 178 | Scan(); |
| 179 | } |
| 180 | |
| 181 | |
| 182 | Handle<String> Scanner::SubString(int start, int end) { |
| 183 | return source_.SubString(start - position_, end - position_); |
| 184 | } |
| 185 | |
| 186 | |
| 187 | Token::Value Scanner::Next() { |
| 188 | // BUG 1215673: Find a thread safe way to set a stack limit in |
| 189 | // pre-parse mode. Otherwise, we cannot safely pre-parse from other |
| 190 | // threads. |
| 191 | current_ = next_; |
| 192 | // Check for stack-overflow before returning any tokens. |
| 193 | StackLimitCheck check; |
| 194 | if (check.HasOverflowed()) { |
| 195 | stack_overflow_ = true; |
| 196 | next_.token = Token::ILLEGAL; |
| 197 | } else { |
| 198 | Scan(); |
| 199 | } |
| 200 | return current_.token; |
| 201 | } |
| 202 | |
| 203 | |
| 204 | void Scanner::StartLiteral() { |
| 205 | next_.literal_pos = literals_.pos(); |
| 206 | } |
| 207 | |
| 208 | |
| 209 | void Scanner::AddChar(uc32 c) { |
| 210 | literals_.AddChar(c); |
| 211 | } |
| 212 | |
| 213 | |
| 214 | void Scanner::TerminateLiteral() { |
| 215 | next_.literal_end = literals_.pos(); |
| 216 | AddChar(0); |
| 217 | } |
| 218 | |
| 219 | |
| 220 | void Scanner::AddCharAdvance() { |
| 221 | AddChar(c0_); |
| 222 | Advance(); |
| 223 | } |
| 224 | |
| 225 | |
| 226 | void Scanner::Advance() { |
| 227 | c0_ = source_.Advance(); |
| 228 | } |
| 229 | |
| 230 | |
| 231 | void Scanner::PushBack(uc32 ch) { |
| 232 | source_.PushBack(ch); |
| 233 | c0_ = ch; |
| 234 | } |
| 235 | |
| 236 | |
ager@chromium.org | 3bf7b91 | 2008-11-17 09:09:45 +0000 | [diff] [blame] | 237 | static inline bool IsByteOrderMark(uc32 c) { |
| 238 | // The Unicode value U+FFFE is guaranteed never to be assigned as a |
| 239 | // Unicode character; this implies that in a Unicode context the |
| 240 | // 0xFF, 0xFE byte pattern can only be interpreted as the U+FEFF |
| 241 | // character expressed in little-endian byte order (since it could |
| 242 | // not be a U+FFFE character expressed in big-endian byte |
| 243 | // order). Nevertheless, we check for it to be compatible with |
| 244 | // Spidermonkey. |
| 245 | return c == 0xFEFF || c == 0xFFFE; |
| 246 | } |
| 247 | |
| 248 | |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 249 | bool Scanner::SkipWhiteSpace() { |
| 250 | int start_position = source_pos(); |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 251 | |
| 252 | while (true) { |
ager@chromium.org | 3bf7b91 | 2008-11-17 09:09:45 +0000 | [diff] [blame] | 253 | // We treat byte-order marks (BOMs) as whitespace for better |
| 254 | // compatibility with Spidermonkey and other JavaScript engines. |
| 255 | while (kIsWhiteSpace.get(c0_) || IsByteOrderMark(c0_)) { |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 256 | // IsWhiteSpace() includes line terminators! |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 257 | if (kIsLineTerminator.get(c0_)) { |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 258 | // Ignore line terminators, but remember them. This is necessary |
| 259 | // for automatic semicolon insertion. |
| 260 | has_line_terminator_before_next_ = true; |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 261 | } |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 262 | Advance(); |
| 263 | } |
| 264 | |
| 265 | // If there is an HTML comment end '-->' at the beginning of a |
| 266 | // line (with only whitespace in front of it), we treat the rest |
| 267 | // of the line as a comment. This is in line with the way |
| 268 | // SpiderMonkey handles it. |
| 269 | if (c0_ == '-' && has_line_terminator_before_next_) { |
| 270 | Advance(); |
| 271 | if (c0_ == '-') { |
| 272 | Advance(); |
| 273 | if (c0_ == '>') { |
| 274 | // Treat the rest of the line as a comment. |
| 275 | SkipSingleLineComment(); |
| 276 | // Continue skipping white space after the comment. |
| 277 | continue; |
| 278 | } |
| 279 | PushBack('-'); // undo Advance() |
| 280 | } |
| 281 | PushBack('-'); // undo Advance() |
| 282 | } |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 283 | // Return whether or not we skipped any characters. |
| 284 | return source_pos() != start_position; |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 285 | } |
| 286 | } |
| 287 | |
| 288 | |
| 289 | Token::Value Scanner::SkipSingleLineComment() { |
| 290 | Advance(); |
| 291 | |
| 292 | // The line terminator at the end of the line is not considered |
| 293 | // to be part of the single-line comment; it is recognized |
| 294 | // separately by the lexical grammar and becomes part of the |
| 295 | // stream of input elements for the syntactic grammar (see |
| 296 | // ECMA-262, section 7.4, page 12). |
| 297 | while (c0_ >= 0 && !kIsLineTerminator.get(c0_)) { |
| 298 | Advance(); |
| 299 | } |
| 300 | |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 301 | return Token::WHITESPACE; |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 302 | } |
| 303 | |
| 304 | |
| 305 | Token::Value Scanner::SkipMultiLineComment() { |
| 306 | ASSERT(c0_ == '*'); |
| 307 | Advance(); |
| 308 | |
| 309 | while (c0_ >= 0) { |
| 310 | char ch = c0_; |
| 311 | Advance(); |
| 312 | // If we have reached the end of the multi-line comment, we |
| 313 | // consume the '/' and insert a whitespace. This way all |
| 314 | // multi-line comments are treated as whitespace - even the ones |
| 315 | // containing line terminators. This contradicts ECMA-262, section |
| 316 | // 7.4, page 12, that says that multi-line comments containing |
| 317 | // line terminators should be treated as a line terminator, but it |
| 318 | // matches the behaviour of SpiderMonkey and KJS. |
| 319 | if (ch == '*' && c0_ == '/') { |
| 320 | c0_ = ' '; |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 321 | return Token::WHITESPACE; |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 322 | } |
| 323 | } |
| 324 | |
| 325 | // Unterminated multi-line comment. |
| 326 | return Token::ILLEGAL; |
| 327 | } |
| 328 | |
| 329 | |
| 330 | Token::Value Scanner::ScanHtmlComment() { |
| 331 | // Check for <!-- comments. |
| 332 | ASSERT(c0_ == '!'); |
| 333 | Advance(); |
| 334 | if (c0_ == '-') { |
| 335 | Advance(); |
| 336 | if (c0_ == '-') return SkipSingleLineComment(); |
| 337 | PushBack('-'); // undo Advance() |
| 338 | } |
| 339 | PushBack('!'); // undo Advance() |
| 340 | ASSERT(c0_ == '!'); |
| 341 | return Token::LT; |
| 342 | } |
| 343 | |
| 344 | |
| 345 | void Scanner::Scan() { |
| 346 | Token::Value token; |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 347 | has_line_terminator_before_next_ = false; |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 348 | do { |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 349 | // Remember the position of the next token |
| 350 | next_.location.beg_pos = source_pos(); |
| 351 | |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 352 | switch (c0_) { |
| 353 | case ' ': |
| 354 | case '\t': |
| 355 | Advance(); |
| 356 | token = Token::WHITESPACE; |
| 357 | break; |
| 358 | |
| 359 | case '\n': |
| 360 | Advance(); |
| 361 | has_line_terminator_before_next_ = true; |
| 362 | token = Token::WHITESPACE; |
| 363 | break; |
| 364 | |
| 365 | case '"': case '\'': |
| 366 | token = ScanString(); |
| 367 | break; |
| 368 | |
| 369 | case '<': |
| 370 | // < <= << <<= <!-- |
| 371 | Advance(); |
| 372 | if (c0_ == '=') { |
| 373 | token = Select(Token::LTE); |
| 374 | } else if (c0_ == '<') { |
| 375 | token = Select('=', Token::ASSIGN_SHL, Token::SHL); |
| 376 | } else if (c0_ == '!') { |
| 377 | token = ScanHtmlComment(); |
| 378 | } else { |
| 379 | token = Token::LT; |
| 380 | } |
| 381 | break; |
| 382 | |
| 383 | case '>': |
| 384 | // > >= >> >>= >>> >>>= |
| 385 | Advance(); |
| 386 | if (c0_ == '=') { |
| 387 | token = Select(Token::GTE); |
| 388 | } else if (c0_ == '>') { |
| 389 | // >> >>= >>> >>>= |
| 390 | Advance(); |
| 391 | if (c0_ == '=') { |
| 392 | token = Select(Token::ASSIGN_SAR); |
| 393 | } else if (c0_ == '>') { |
| 394 | token = Select('=', Token::ASSIGN_SHR, Token::SHR); |
| 395 | } else { |
| 396 | token = Token::SAR; |
| 397 | } |
| 398 | } else { |
| 399 | token = Token::GT; |
| 400 | } |
| 401 | break; |
| 402 | |
| 403 | case '=': |
| 404 | // = == === |
| 405 | Advance(); |
| 406 | if (c0_ == '=') { |
| 407 | token = Select('=', Token::EQ_STRICT, Token::EQ); |
| 408 | } else { |
| 409 | token = Token::ASSIGN; |
| 410 | } |
| 411 | break; |
| 412 | |
| 413 | case '!': |
| 414 | // ! != !== |
| 415 | Advance(); |
| 416 | if (c0_ == '=') { |
| 417 | token = Select('=', Token::NE_STRICT, Token::NE); |
| 418 | } else { |
| 419 | token = Token::NOT; |
| 420 | } |
| 421 | break; |
| 422 | |
| 423 | case '+': |
| 424 | // + ++ += |
| 425 | Advance(); |
| 426 | if (c0_ == '+') { |
| 427 | token = Select(Token::INC); |
| 428 | } else if (c0_ == '=') { |
| 429 | token = Select(Token::ASSIGN_ADD); |
| 430 | } else { |
| 431 | token = Token::ADD; |
| 432 | } |
| 433 | break; |
| 434 | |
| 435 | case '-': |
| 436 | // - -- --> -= |
| 437 | Advance(); |
| 438 | if (c0_ == '-') { |
| 439 | Advance(); |
| 440 | if (c0_ == '>' && has_line_terminator_before_next_) { |
| 441 | // For compatibility with SpiderMonkey, we skip lines that |
| 442 | // start with an HTML comment end '-->'. |
| 443 | token = SkipSingleLineComment(); |
| 444 | } else { |
| 445 | token = Token::DEC; |
| 446 | } |
| 447 | } else if (c0_ == '=') { |
| 448 | token = Select(Token::ASSIGN_SUB); |
| 449 | } else { |
| 450 | token = Token::SUB; |
| 451 | } |
| 452 | break; |
| 453 | |
| 454 | case '*': |
| 455 | // * *= |
| 456 | token = Select('=', Token::ASSIGN_MUL, Token::MUL); |
| 457 | break; |
| 458 | |
| 459 | case '%': |
| 460 | // % %= |
| 461 | token = Select('=', Token::ASSIGN_MOD, Token::MOD); |
| 462 | break; |
| 463 | |
| 464 | case '/': |
| 465 | // / // /* /= |
| 466 | Advance(); |
| 467 | if (c0_ == '/') { |
| 468 | token = SkipSingleLineComment(); |
| 469 | } else if (c0_ == '*') { |
| 470 | token = SkipMultiLineComment(); |
| 471 | } else if (c0_ == '=') { |
| 472 | token = Select(Token::ASSIGN_DIV); |
| 473 | } else { |
| 474 | token = Token::DIV; |
| 475 | } |
| 476 | break; |
| 477 | |
| 478 | case '&': |
| 479 | // & && &= |
| 480 | Advance(); |
| 481 | if (c0_ == '&') { |
| 482 | token = Select(Token::AND); |
| 483 | } else if (c0_ == '=') { |
| 484 | token = Select(Token::ASSIGN_BIT_AND); |
| 485 | } else { |
| 486 | token = Token::BIT_AND; |
| 487 | } |
| 488 | break; |
| 489 | |
| 490 | case '|': |
| 491 | // | || |= |
| 492 | Advance(); |
| 493 | if (c0_ == '|') { |
| 494 | token = Select(Token::OR); |
| 495 | } else if (c0_ == '=') { |
| 496 | token = Select(Token::ASSIGN_BIT_OR); |
| 497 | } else { |
| 498 | token = Token::BIT_OR; |
| 499 | } |
| 500 | break; |
| 501 | |
| 502 | case '^': |
| 503 | // ^ ^= |
| 504 | token = Select('=', Token::ASSIGN_BIT_XOR, Token::BIT_XOR); |
| 505 | break; |
| 506 | |
| 507 | case '.': |
| 508 | // . Number |
| 509 | Advance(); |
| 510 | if (IsDecimalDigit(c0_)) { |
| 511 | token = ScanNumber(true); |
| 512 | } else { |
| 513 | token = Token::PERIOD; |
| 514 | } |
| 515 | break; |
| 516 | |
| 517 | case ':': |
| 518 | token = Select(Token::COLON); |
| 519 | break; |
| 520 | |
| 521 | case ';': |
| 522 | token = Select(Token::SEMICOLON); |
| 523 | break; |
| 524 | |
| 525 | case ',': |
| 526 | token = Select(Token::COMMA); |
| 527 | break; |
| 528 | |
| 529 | case '(': |
| 530 | token = Select(Token::LPAREN); |
| 531 | break; |
| 532 | |
| 533 | case ')': |
| 534 | token = Select(Token::RPAREN); |
| 535 | break; |
| 536 | |
| 537 | case '[': |
| 538 | token = Select(Token::LBRACK); |
| 539 | break; |
| 540 | |
| 541 | case ']': |
| 542 | token = Select(Token::RBRACK); |
| 543 | break; |
| 544 | |
| 545 | case '{': |
| 546 | token = Select(Token::LBRACE); |
| 547 | break; |
| 548 | |
| 549 | case '}': |
| 550 | token = Select(Token::RBRACE); |
| 551 | break; |
| 552 | |
| 553 | case '?': |
| 554 | token = Select(Token::CONDITIONAL); |
| 555 | break; |
| 556 | |
| 557 | case '~': |
| 558 | token = Select(Token::BIT_NOT); |
| 559 | break; |
| 560 | |
| 561 | default: |
| 562 | if (kIsIdentifierStart.get(c0_)) { |
| 563 | token = ScanIdentifier(); |
| 564 | } else if (IsDecimalDigit(c0_)) { |
| 565 | token = ScanNumber(false); |
| 566 | } else if (SkipWhiteSpace()) { |
| 567 | token = Token::WHITESPACE; |
| 568 | } else if (c0_ < 0) { |
| 569 | token = Token::EOS; |
| 570 | } else { |
| 571 | token = Select(Token::ILLEGAL); |
| 572 | } |
| 573 | break; |
| 574 | } |
| 575 | |
| 576 | // Continue scanning for tokens as long as we're just skipping |
| 577 | // whitespace. |
| 578 | } while (token == Token::WHITESPACE); |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 579 | |
| 580 | next_.location.end_pos = source_pos(); |
| 581 | next_.token = token; |
| 582 | } |
| 583 | |
| 584 | |
| 585 | void Scanner::SeekForward(int pos) { |
| 586 | source_.SeekForward(pos - 1); |
| 587 | Advance(); |
| 588 | Scan(); |
| 589 | } |
| 590 | |
| 591 | |
| 592 | uc32 Scanner::ScanHexEscape(uc32 c, int length) { |
| 593 | ASSERT(length <= 4); // prevent overflow |
| 594 | |
| 595 | uc32 digits[4]; |
| 596 | uc32 x = 0; |
| 597 | for (int i = 0; i < length; i++) { |
| 598 | digits[i] = c0_; |
| 599 | int d = HexValue(c0_); |
| 600 | if (d < 0) { |
| 601 | // According to ECMA-262, 3rd, 7.8.4, page 18, these hex escapes |
| 602 | // should be illegal, but other JS VMs just return the |
| 603 | // non-escaped version of the original character. |
| 604 | |
| 605 | // Push back digits read, except the last one (in c0_). |
| 606 | for (int j = i-1; j >= 0; j--) { |
| 607 | PushBack(digits[j]); |
| 608 | } |
ager@chromium.org | 6f10e41 | 2009-02-13 10:11:16 +0000 | [diff] [blame] | 609 | // Notice: No handling of error - treat it as "\u"->"u". |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 610 | return c; |
| 611 | } |
| 612 | x = x * 16 + d; |
| 613 | Advance(); |
| 614 | } |
| 615 | |
| 616 | return x; |
| 617 | } |
| 618 | |
| 619 | |
| 620 | // Octal escapes of the forms '\0xx' and '\xxx' are not a part of |
| 621 | // ECMA-262. Other JS VMs support them. |
| 622 | uc32 Scanner::ScanOctalEscape(uc32 c, int length) { |
| 623 | uc32 x = c - '0'; |
| 624 | for (int i = 0; i < length; i++) { |
| 625 | int d = c0_ - '0'; |
| 626 | if (d < 0 || d > 7) break; |
| 627 | int nx = x * 8 + d; |
| 628 | if (nx >= 256) break; |
| 629 | x = nx; |
| 630 | Advance(); |
| 631 | } |
| 632 | return x; |
| 633 | } |
| 634 | |
| 635 | |
| 636 | void Scanner::ScanEscape() { |
| 637 | uc32 c = c0_; |
| 638 | Advance(); |
| 639 | |
| 640 | // Skip escaped newlines. |
| 641 | if (kIsLineTerminator.get(c)) { |
| 642 | // Allow CR+LF newlines in multiline string literals. |
| 643 | if (IsCarriageReturn(c) && IsLineFeed(c0_)) Advance(); |
| 644 | // Allow LF+CR newlines in multiline string literals. |
| 645 | if (IsLineFeed(c) && IsCarriageReturn(c0_)) Advance(); |
| 646 | return; |
| 647 | } |
| 648 | |
| 649 | switch (c) { |
| 650 | case '\'': // fall through |
| 651 | case '"' : // fall through |
| 652 | case '\\': break; |
| 653 | case 'b' : c = '\b'; break; |
| 654 | case 'f' : c = '\f'; break; |
| 655 | case 'n' : c = '\n'; break; |
| 656 | case 'r' : c = '\r'; break; |
| 657 | case 't' : c = '\t'; break; |
| 658 | case 'u' : c = ScanHexEscape(c, 4); break; |
| 659 | case 'v' : c = '\v'; break; |
| 660 | case 'x' : c = ScanHexEscape(c, 2); break; |
| 661 | case '0' : // fall through |
| 662 | case '1' : // fall through |
| 663 | case '2' : // fall through |
| 664 | case '3' : // fall through |
| 665 | case '4' : // fall through |
| 666 | case '5' : // fall through |
| 667 | case '6' : // fall through |
| 668 | case '7' : c = ScanOctalEscape(c, 2); break; |
| 669 | } |
| 670 | |
| 671 | // According to ECMA-262, 3rd, 7.8.4 (p 18ff) these |
| 672 | // should be illegal, but they are commonly handled |
| 673 | // as non-escaped characters by JS VMs. |
| 674 | AddChar(c); |
| 675 | } |
| 676 | |
| 677 | |
| 678 | Token::Value Scanner::ScanString() { |
| 679 | uc32 quote = c0_; |
| 680 | Advance(); // consume quote |
| 681 | |
| 682 | StartLiteral(); |
| 683 | while (c0_ != quote && c0_ >= 0 && !kIsLineTerminator.get(c0_)) { |
| 684 | uc32 c = c0_; |
| 685 | Advance(); |
| 686 | if (c == '\\') { |
| 687 | if (c0_ < 0) return Token::ILLEGAL; |
| 688 | ScanEscape(); |
| 689 | } else { |
| 690 | AddChar(c); |
| 691 | } |
| 692 | } |
| 693 | if (c0_ != quote) { |
| 694 | return Token::ILLEGAL; |
| 695 | } |
| 696 | TerminateLiteral(); |
| 697 | |
| 698 | Advance(); // consume quote |
| 699 | return Token::STRING; |
| 700 | } |
| 701 | |
| 702 | |
| 703 | Token::Value Scanner::Select(Token::Value tok) { |
| 704 | Advance(); |
| 705 | return tok; |
| 706 | } |
| 707 | |
| 708 | |
| 709 | Token::Value Scanner::Select(uc32 next, Token::Value then, Token::Value else_) { |
| 710 | Advance(); |
| 711 | if (c0_ == next) { |
| 712 | Advance(); |
| 713 | return then; |
| 714 | } else { |
| 715 | return else_; |
| 716 | } |
| 717 | } |
| 718 | |
| 719 | |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 720 | // Returns true if any decimal digits were scanned, returns false otherwise. |
| 721 | void Scanner::ScanDecimalDigits() { |
| 722 | while (IsDecimalDigit(c0_)) |
| 723 | AddCharAdvance(); |
| 724 | } |
| 725 | |
| 726 | |
| 727 | Token::Value Scanner::ScanNumber(bool seen_period) { |
| 728 | ASSERT(IsDecimalDigit(c0_)); // the first digit of the number or the fraction |
| 729 | |
| 730 | enum { DECIMAL, HEX, OCTAL } kind = DECIMAL; |
| 731 | |
| 732 | StartLiteral(); |
| 733 | if (seen_period) { |
| 734 | // we have already seen a decimal point of the float |
| 735 | AddChar('.'); |
| 736 | ScanDecimalDigits(); // we know we have at least one digit |
| 737 | |
| 738 | } else { |
| 739 | // if the first character is '0' we must check for octals and hex |
| 740 | if (c0_ == '0') { |
| 741 | AddCharAdvance(); |
| 742 | |
| 743 | // either 0, 0exxx, 0Exxx, 0.xxx, an octal number, or a hex number |
| 744 | if (c0_ == 'x' || c0_ == 'X') { |
| 745 | // hex number |
| 746 | kind = HEX; |
| 747 | AddCharAdvance(); |
| 748 | if (!IsHexDigit(c0_)) |
| 749 | // we must have at least one hex digit after 'x'/'X' |
| 750 | return Token::ILLEGAL; |
| 751 | while (IsHexDigit(c0_)) |
| 752 | AddCharAdvance(); |
| 753 | |
| 754 | } else if ('0' <= c0_ && c0_ <= '7') { |
| 755 | // (possible) octal number |
| 756 | kind = OCTAL; |
| 757 | while (true) { |
| 758 | if (c0_ == '8' || c0_ == '9') { |
| 759 | kind = DECIMAL; |
| 760 | break; |
| 761 | } |
| 762 | if (c0_ < '0' || '7' < c0_) break; |
| 763 | AddCharAdvance(); |
| 764 | } |
| 765 | } |
| 766 | } |
| 767 | |
| 768 | // Parse decimal digits and allow trailing fractional part. |
| 769 | if (kind == DECIMAL) { |
| 770 | ScanDecimalDigits(); // optional |
| 771 | if (c0_ == '.') { |
| 772 | AddCharAdvance(); |
| 773 | ScanDecimalDigits(); // optional |
| 774 | } |
| 775 | } |
| 776 | } |
| 777 | |
| 778 | // scan exponent, if any |
| 779 | if (c0_ == 'e' || c0_ == 'E') { |
| 780 | ASSERT(kind != HEX); // 'e'/'E' must be scanned as part of the hex number |
| 781 | if (kind == OCTAL) return Token::ILLEGAL; // no exponent for octals allowed |
| 782 | // scan exponent |
| 783 | AddCharAdvance(); |
| 784 | if (c0_ == '+' || c0_ == '-') |
| 785 | AddCharAdvance(); |
| 786 | if (!IsDecimalDigit(c0_)) |
| 787 | // we must have at least one decimal digit after 'e'/'E' |
| 788 | return Token::ILLEGAL; |
| 789 | ScanDecimalDigits(); |
| 790 | } |
| 791 | TerminateLiteral(); |
| 792 | |
| 793 | // The source character immediately following a numeric literal must |
| 794 | // not be an identifier start or a decimal digit; see ECMA-262 |
| 795 | // section 7.8.3, page 17 (note that we read only one decimal digit |
| 796 | // if the value is 0). |
| 797 | if (IsDecimalDigit(c0_) || kIsIdentifierStart.get(c0_)) |
| 798 | return Token::ILLEGAL; |
| 799 | |
| 800 | return Token::NUMBER; |
| 801 | } |
| 802 | |
| 803 | |
| 804 | uc32 Scanner::ScanIdentifierUnicodeEscape() { |
| 805 | Advance(); |
| 806 | if (c0_ != 'u') return unibrow::Utf8::kBadChar; |
| 807 | Advance(); |
| 808 | uc32 c = ScanHexEscape('u', 4); |
| 809 | // We do not allow a unicode escape sequence to start another |
| 810 | // unicode escape sequence. |
| 811 | if (c == '\\') return unibrow::Utf8::kBadChar; |
| 812 | return c; |
| 813 | } |
| 814 | |
| 815 | |
| 816 | Token::Value Scanner::ScanIdentifier() { |
| 817 | ASSERT(kIsIdentifierStart.get(c0_)); |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 818 | bool has_escapes = false; |
| 819 | |
| 820 | StartLiteral(); |
| 821 | // Scan identifier start character. |
| 822 | if (c0_ == '\\') { |
| 823 | has_escapes = true; |
| 824 | uc32 c = ScanIdentifierUnicodeEscape(); |
| 825 | // Only allow legal identifier start characters. |
| 826 | if (!kIsIdentifierStart.get(c)) return Token::ILLEGAL; |
| 827 | AddChar(c); |
| 828 | } else { |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 829 | AddChar(c0_); |
| 830 | Advance(); |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 831 | } |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 832 | |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 833 | // Scan the rest of the identifier characters. |
| 834 | while (kIsIdentifierPart.get(c0_)) { |
| 835 | if (c0_ == '\\') { |
| 836 | has_escapes = true; |
| 837 | uc32 c = ScanIdentifierUnicodeEscape(); |
| 838 | // Only allow legal identifier part characters. |
| 839 | if (!kIsIdentifierPart.get(c)) return Token::ILLEGAL; |
| 840 | AddChar(c); |
| 841 | } else { |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 842 | AddChar(c0_); |
| 843 | Advance(); |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 844 | } |
| 845 | } |
| 846 | TerminateLiteral(); |
| 847 | |
| 848 | // We don't have any 1-letter keywords (this is probably a common case). |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 849 | if ((next_.literal_end - next_.literal_pos) == 1) { |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 850 | return Token::IDENTIFIER; |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 851 | } |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 852 | |
| 853 | // If the identifier contains unicode escapes, it must not be |
| 854 | // resolved to a keyword. |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 855 | if (has_escapes) { |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 856 | return Token::IDENTIFIER; |
kasperl@chromium.org | b3284ad | 2009-05-18 06:12:45 +0000 | [diff] [blame] | 857 | } |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 858 | |
| 859 | return Token::Lookup(&literals_.data()[next_.literal_pos]); |
| 860 | } |
| 861 | |
| 862 | |
| 863 | |
| 864 | bool Scanner::IsIdentifier(unibrow::CharacterStream* buffer) { |
ager@chromium.org | 3291210 | 2009-01-16 10:38:43 +0000 | [diff] [blame] | 865 | // Checks whether the buffer contains an identifier (no escape). |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 866 | if (!buffer->has_more()) return false; |
| 867 | if (!kIsIdentifierStart.get(buffer->GetNext())) return false; |
| 868 | while (buffer->has_more()) { |
| 869 | if (!kIsIdentifierPart.get(buffer->GetNext())) return false; |
| 870 | } |
| 871 | return true; |
| 872 | } |
| 873 | |
| 874 | |
| 875 | bool Scanner::ScanRegExpPattern(bool seen_equal) { |
| 876 | // Scan: ('/' | '/=') RegularExpressionBody '/' RegularExpressionFlags |
| 877 | bool in_character_class = false; |
| 878 | |
| 879 | // Previous token is either '/' or '/=', in the second case, the |
| 880 | // pattern starts at =. |
| 881 | next_.location.beg_pos = source_pos() - (seen_equal ? 2 : 1); |
| 882 | next_.location.end_pos = source_pos() - (seen_equal ? 1 : 0); |
| 883 | |
| 884 | // Scan regular expression body: According to ECMA-262, 3rd, 7.8.5, |
| 885 | // the scanner should pass uninterpreted bodies to the RegExp |
| 886 | // constructor. |
| 887 | StartLiteral(); |
| 888 | if (seen_equal) |
| 889 | AddChar('='); |
| 890 | |
| 891 | while (c0_ != '/' || in_character_class) { |
| 892 | if (kIsLineTerminator.get(c0_) || c0_ < 0) |
| 893 | return false; |
| 894 | if (c0_ == '\\') { // escaped character |
| 895 | AddCharAdvance(); |
| 896 | if (kIsLineTerminator.get(c0_) || c0_ < 0) |
| 897 | return false; |
| 898 | AddCharAdvance(); |
| 899 | } else { // unescaped character |
| 900 | if (c0_ == '[') |
| 901 | in_character_class = true; |
| 902 | if (c0_ == ']') |
| 903 | in_character_class = false; |
| 904 | AddCharAdvance(); |
| 905 | } |
| 906 | } |
| 907 | Advance(); // consume '/' |
| 908 | |
| 909 | TerminateLiteral(); |
| 910 | |
| 911 | return true; |
| 912 | } |
| 913 | |
| 914 | bool Scanner::ScanRegExpFlags() { |
| 915 | // Scan regular expression flags. |
| 916 | StartLiteral(); |
ager@chromium.org | 6f10e41 | 2009-02-13 10:11:16 +0000 | [diff] [blame] | 917 | while (kIsIdentifierPart.get(c0_)) { |
| 918 | if (c0_ == '\\') { |
| 919 | uc32 c = ScanIdentifierUnicodeEscape(); |
| 920 | if (c != static_cast<uc32>(unibrow::Utf8::kBadChar)) { |
| 921 | // We allow any escaped character, unlike the restriction on |
| 922 | // IdentifierPart when it is used to build an IdentifierName. |
| 923 | AddChar(c); |
| 924 | continue; |
| 925 | } |
| 926 | } |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 927 | AddCharAdvance(); |
ager@chromium.org | 6f10e41 | 2009-02-13 10:11:16 +0000 | [diff] [blame] | 928 | } |
christian.plesner.hansen | 43d26ec | 2008-07-03 15:10:15 +0000 | [diff] [blame] | 929 | TerminateLiteral(); |
| 930 | |
| 931 | next_.location.end_pos = source_pos() - 1; |
| 932 | return true; |
| 933 | } |
| 934 | |
| 935 | } } // namespace v8::internal |