digiaonline /
graphql-php
| Total Complexity | 83 |
| Total Lines | 521 |
| Duplicated Lines | 0 % |
| Changes | 0 | ||
Complex classes like TokenReader often do a lot of different things. To break such a class down, we need to identify a cohesive component within that class. A common approach to find such a component is to look for fields/methods that share the same prefixes, or suffixes.
Once you have determined the fields that belong together, you can apply the Extract Class refactoring. If the component makes sense as a sub-class, Extract Subclass is also a candidate, and is often faster.
While breaking up the class, it is a good idea to analyze how other classes use TokenReader, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 7 | class TokenReader implements TokenReaderInterface |
||
| 8 | { |
||
| 9 | |||
| 10 | /** |
||
| 11 | * The lexer owning this token reader. |
||
| 12 | * |
||
| 13 | * @var LexerInterface |
||
| 14 | */ |
||
| 15 | protected $lexer; |
||
| 16 | |||
| 17 | /** |
||
| 18 | * @inheritdoc |
||
| 19 | */ |
||
| 20 | public function setLexer(LexerInterface $lexer) |
||
| 21 | { |
||
| 22 | $this->lexer = $lexer; |
||
| 23 | return $this; |
||
| 24 | } |
||
| 25 | |||
| 26 | /** |
||
| 27 | * @inheritdoc |
||
| 28 | */ |
||
| 29 | public function getLexer(): LexerInterface |
||
| 32 | } |
||
| 33 | |||
| 34 | /** |
||
| 35 | * @inheritdoc |
||
| 36 | * @throws SyntaxErrorException |
||
| 37 | */ |
||
| 38 | public function read(string $body, int $bodyLength, int $code, int $pos, int $line, int $col, Token $prev): ?Token |
||
| 39 | { |
||
| 40 | switch ($code) { |
||
| 41 | case 33: // ! |
||
| 42 | return $this->readBang($pos, $line, $col, $prev); |
||
| 43 | case 35: // # |
||
| 44 | return $this->readComment($pos, $line, $col, $prev); |
||
| 45 | case 36: // $ |
||
| 46 | return $this->readDollar($pos, $line, $col, $prev); |
||
| 47 | case 38: // & |
||
| 48 | return $this->readAmp($pos, $line, $col, $prev); |
||
| 49 | case 58: // : |
||
| 50 | return $this->readColon($pos, $line, $col, $prev); |
||
| 51 | case 61: // = |
||
| 52 | return $this->readEquals($pos, $line, $col, $prev); |
||
| 53 | case 64: // @ |
||
| 54 | return $this->readAt($pos, $line, $col, $prev); |
||
| 55 | case 124: // | |
||
| 56 | return $this->readPipe($pos, $line, $col, $prev); |
||
| 57 | case 40: |
||
| 58 | case 41: // ( or )~ |
||
| 59 | return $this->readParenthesis($code, $pos, $line, $col, $prev); |
||
| 60 | case 91: |
||
| 61 | case 93: // [ or ] |
||
| 62 | return $this->readBracket($code, $pos, $line, $col, $prev); |
||
| 63 | case 123: |
||
| 64 | case 125: // { or } |
||
| 65 | return $this->readBrace($code, $pos, $line, $col, $prev); |
||
| 66 | } |
||
| 67 | |||
| 68 | // Int: -?(0|[1-9][0-9]*) |
||
| 69 | // Float: -?(0|[1-9][0-9]*)(\.[0-9]+)?((E|e)(+|-)?[0-9]+)? |
||
| 70 | if (isNumber($code)) { |
||
| 71 | return $this->readNumber($code, $pos, $line, $col, $prev); |
||
| 72 | } |
||
| 73 | |||
| 74 | if (isLetter($code) || isUnderscore($code)) { |
||
| 75 | return $this->readName($pos, $line, $col, $prev); |
||
| 76 | } |
||
| 77 | |||
| 78 | // Spread: ... |
||
| 79 | if ($bodyLength > 3 && isSpread($body, $code, $pos)) { |
||
| 80 | return $this->readSpread($pos, $line, $col, $prev); |
||
| 81 | } |
||
| 82 | |||
| 83 | // String: "([^"\\\u000A\u000D]|(\\(u[0-9a-fA-F]{4}|["\\/bfnrt])))*" |
||
| 84 | if ($bodyLength > 2 && isString($body, $code, $pos)) { |
||
| 85 | return $this->readString($pos, $line, $col, $prev); |
||
| 86 | } |
||
| 87 | |||
| 88 | // Block String: """("?"?(\\"""|\\(?!=""")|[^"\\]))*""" |
||
| 89 | if ($bodyLength > 6 && isTripleQuote($body, $code, $pos)) { |
||
| 90 | return $this->readBlockString($pos, $line, $col, $prev); |
||
| 91 | } |
||
| 92 | |||
| 93 | return null; |
||
| 94 | } |
||
| 95 | |||
| 96 | /** |
||
| 97 | * @param int $pos |
||
| 98 | * @param int $line |
||
| 99 | * @param int $col |
||
| 100 | * @param Token $prev |
||
| 101 | * @return Token |
||
| 102 | */ |
||
| 103 | protected function readName(int $pos, int $line, int $col, Token $prev): Token |
||
| 104 | { |
||
| 105 | $body = $this->lexer->getBody(); |
||
| 106 | $bodyLength = \mb_strlen($body); |
||
| 107 | $start = $pos; |
||
| 108 | $pos = $start + 1; |
||
| 109 | |||
| 110 | while ($pos !== $bodyLength && ($code = charCodeAt($body, $pos)) !== null && isAlphaNumeric($code)) { |
||
| 111 | ++$pos; |
||
| 112 | } |
||
| 113 | |||
| 114 | return new Token(TokenKindEnum::NAME, $start, $pos, $line, $col, $prev, sliceString($body, $start, $pos)); |
||
| 115 | } |
||
| 116 | |||
| 117 | /** |
||
| 118 | * @param int $pos |
||
| 119 | * @param int $line |
||
| 120 | * @param int $col |
||
| 121 | * @param Token $prev |
||
| 122 | * @return Token |
||
| 123 | * @throws SyntaxErrorException |
||
| 124 | */ |
||
| 125 | protected function readBlockString(int $pos, int $line, int $col, Token $prev): Token |
||
| 126 | { |
||
| 127 | $body = $this->lexer->getBody(); |
||
| 128 | $bodyLength = \mb_strlen($body); |
||
| 129 | $start = $pos; |
||
| 130 | $pos = $start + 3; |
||
| 131 | $chunkStart = $pos; |
||
| 132 | $rawValue = ''; |
||
| 133 | |||
| 134 | while ($pos < $bodyLength && ($code = charCodeAt($body, $pos)) !== null) { |
||
| 135 | // Closing Triple-Quote (""") |
||
| 136 | if (isTripleQuote($body, $code, $pos)) { |
||
| 137 | $rawValue .= sliceString($body, $chunkStart, $pos); |
||
| 138 | return new Token( |
||
| 139 | TokenKindEnum::BLOCK_STRING, |
||
| 140 | $start, |
||
| 141 | $pos + 3, |
||
| 142 | $line, |
||
| 143 | $col, |
||
| 144 | $prev, |
||
| 145 | blockStringValue($rawValue) |
||
| 146 | ); |
||
| 147 | } |
||
| 148 | |||
| 149 | if (isSourceCharacter($code) && !isLineTerminator($code)) { |
||
| 150 | throw new SyntaxErrorException( |
||
| 151 | $this->lexer->getSource(), |
||
| 152 | $pos, |
||
| 153 | \sprintf('Invalid character within String: %s', printCharCode($code)) |
||
| 154 | ); |
||
| 155 | } |
||
| 156 | |||
| 157 | if (isEscapedTripleQuote($body, $code, $pos)) { |
||
| 158 | $rawValue .= sliceString($body, $chunkStart, $pos) . '"""'; |
||
| 159 | $pos += 4; |
||
| 160 | $chunkStart = $pos; |
||
| 161 | } else { |
||
| 162 | ++$pos; |
||
| 163 | } |
||
| 164 | } |
||
| 165 | |||
| 166 | throw new SyntaxErrorException($this->lexer->getSource(), $pos, 'Unterminated string.'); |
||
| 167 | } |
||
| 168 | |||
| 169 | /** |
||
| 170 | * @param int $code |
||
| 171 | * @param int $pos |
||
| 172 | * @param int $line |
||
| 173 | * @param int $col |
||
| 174 | * @param Token $prev |
||
| 175 | * @return Token |
||
| 176 | * @throws SyntaxErrorException |
||
| 177 | */ |
||
| 178 | protected function readNumber(int $code, int $pos, int $line, int $col, Token $prev): Token |
||
| 179 | { |
||
| 180 | $body = $this->lexer->getBody(); |
||
| 181 | $start = $pos; |
||
| 182 | $isFloat = false; |
||
| 183 | |||
| 184 | if ($code === 45) { |
||
| 185 | // - |
||
| 186 | $code = charCodeAt($body, ++$pos); |
||
| 187 | } |
||
| 188 | |||
| 189 | if ($code === 48) { |
||
| 190 | // 0 |
||
| 191 | $code = charCodeAt($body, ++$pos); |
||
| 192 | if (isNumber($code)) { |
||
| 193 | throw new SyntaxErrorException( |
||
| 194 | $this->lexer->getSource(), |
||
| 195 | $pos, |
||
| 196 | \sprintf('Invalid number, unexpected digit after 0: %s.', printCharCode($code)) |
||
| 197 | ); |
||
| 198 | } |
||
| 199 | } else { |
||
| 200 | $pos = $this->readDigits($code, $pos); |
||
| 201 | $code = charCodeAt($body, $pos); |
||
| 202 | } |
||
| 203 | |||
| 204 | if ($code === 46) { |
||
| 205 | // . |
||
| 206 | $isFloat = true; |
||
| 207 | $code = charCodeAt($body, ++$pos); |
||
| 208 | $pos = $this->readDigits($code, $pos); |
||
| 209 | $code = charCodeAt($body, $pos); |
||
| 210 | } |
||
| 211 | |||
| 212 | if ($code === 69 || $code === 101) { |
||
| 213 | // e or E |
||
| 214 | $isFloat = true; |
||
| 215 | $code = charCodeAt($body, ++$pos); |
||
| 216 | |||
| 217 | if ($code === 43 || $code === 45) { |
||
| 218 | // + or - |
||
| 219 | $code = charCodeAt($body, ++$pos); |
||
| 220 | } |
||
| 221 | |||
| 222 | $pos = $this->readDigits($code, $pos); |
||
| 223 | } |
||
| 224 | |||
| 225 | return new Token( |
||
| 226 | $isFloat ? TokenKindEnum::FLOAT : TokenKindEnum::INT, |
||
| 227 | $start, |
||
| 228 | $pos, |
||
| 229 | $line, |
||
| 230 | $col, |
||
| 231 | $prev, |
||
| 232 | sliceString($body, $start, $pos) |
||
| 233 | ); |
||
| 234 | } |
||
| 235 | |||
| 236 | /** |
||
| 237 | * @param int $pos |
||
| 238 | * @param int $line |
||
| 239 | * @param int $col |
||
| 240 | * @param Token $prev |
||
| 241 | * @return Token |
||
| 242 | * @throws SyntaxErrorException |
||
| 243 | */ |
||
| 244 | protected function readString(int $pos, int $line, int $col, Token $prev): Token |
||
| 245 | { |
||
| 246 | $body = $this->lexer->getBody(); |
||
| 247 | $bodyLength = \mb_strlen($body); |
||
| 248 | $start = $pos; |
||
| 249 | $pos = $start + 1; |
||
| 250 | $chunkStart = $pos; |
||
| 251 | $value = ''; |
||
| 252 | while ($pos < $bodyLength && ($code = charCodeAt($body, $pos)) !== null && !isLineTerminator($code)) { |
||
| 253 | // Closing Quote (") |
||
| 254 | if ($code === 34) { |
||
| 255 | $value .= sliceString($body, $chunkStart, $pos); |
||
| 256 | return new Token(TokenKindEnum::STRING, $start, $pos + 1, $line, $col, $prev, $value); |
||
| 257 | } |
||
| 258 | |||
| 259 | if (isSourceCharacter($code)) { |
||
| 260 | throw new SyntaxErrorException( |
||
| 261 | $this->lexer->getSource(), |
||
| 262 | $pos, |
||
| 263 | \sprintf('Invalid character within String: %s', printCharCode($code)) |
||
| 264 | ); |
||
| 265 | } |
||
| 266 | |||
| 267 | ++$pos; |
||
| 268 | |||
| 269 | if ($code === 92) { |
||
| 270 | // \ |
||
| 271 | $value .= sliceString($body, $chunkStart, $pos - 1); |
||
| 272 | $code = charCodeAt($body, $pos); |
||
| 273 | |||
| 274 | switch ($code) { |
||
| 275 | case 34: |
||
| 276 | $value .= '"'; |
||
| 277 | break; |
||
| 278 | case 47: |
||
| 279 | $value .= '/'; |
||
| 280 | break; |
||
| 281 | case 92: |
||
| 282 | $value .= '\\'; |
||
| 283 | break; |
||
| 284 | case 98: |
||
| 285 | $value .= '\b'; |
||
| 286 | break; |
||
| 287 | case 102: |
||
| 288 | $value .= '\f'; |
||
| 289 | break; |
||
| 290 | case 110: |
||
| 291 | $value .= '\n'; |
||
| 292 | break; |
||
| 293 | case 114: |
||
| 294 | $value .= '\r'; |
||
| 295 | break; |
||
| 296 | case 116: |
||
| 297 | $value .= '\t'; |
||
| 298 | break; |
||
| 299 | case 117: |
||
| 300 | // u |
||
| 301 | $unicodeString = sliceString($body, $pos - 1, $pos + 5); |
||
| 302 | $charCode = uniCharCode( |
||
| 303 | charCodeAt($body, $pos + 1), |
||
| 304 | charCodeAt($body, $pos + 2), |
||
| 305 | charCodeAt($body, $pos + 3), |
||
| 306 | charCodeAt($body, $pos + 4) |
||
| 307 | ); |
||
| 308 | if ($charCode < 0) { |
||
| 309 | throw new SyntaxErrorException( |
||
| 310 | $this->lexer->getSource(), |
||
| 311 | $pos, |
||
| 312 | \sprintf( |
||
| 313 | 'Invalid character escape sequence: %s', |
||
| 314 | $unicodeString |
||
| 315 | ) |
||
| 316 | ); |
||
| 317 | } |
||
| 318 | $value .= $unicodeString; |
||
| 319 | $pos += 4; |
||
| 320 | break; |
||
| 321 | default: |
||
| 322 | throw new SyntaxErrorException( |
||
| 323 | $this->lexer->getSource(), |
||
| 324 | $pos, |
||
| 325 | \sprintf('Invalid character escape sequence: \\%s', \chr($code)) |
||
| 326 | ); |
||
| 327 | } |
||
| 328 | |||
| 329 | ++$pos; |
||
| 330 | |||
| 331 | $chunkStart = $pos; |
||
| 332 | } |
||
| 333 | } |
||
| 334 | |||
| 335 | throw new SyntaxErrorException($this->lexer->getSource(), $pos, 'Unterminated string.'); |
||
| 336 | } |
||
| 337 | |||
| 338 | /** |
||
| 339 | * @param int $pos |
||
| 340 | * @param int $line |
||
| 341 | * @param int $col |
||
| 342 | * @param Token $prev |
||
| 343 | * @return Token |
||
| 344 | */ |
||
| 345 | protected function readSpread(int $pos, int $line, int $col, Token $prev): Token |
||
| 346 | { |
||
| 347 | return new Token(TokenKindEnum::SPREAD, $pos, $pos + 3, $line, $col, $prev); |
||
| 348 | } |
||
| 349 | |||
| 350 | /** |
||
| 351 | * @param int $pos |
||
| 352 | * @param int $line |
||
| 353 | * @param int $col |
||
| 354 | * @param Token $prev |
||
| 355 | * @return Token |
||
| 356 | */ |
||
| 357 | protected function readDollar(int $pos, int $line, int $col, Token $prev): Token |
||
| 358 | { |
||
| 359 | return new Token(TokenKindEnum::DOLLAR, $pos, $pos + 1, $line, $col, $prev); |
||
| 360 | } |
||
| 361 | |||
| 362 | /** |
||
| 363 | * @param int $pos |
||
| 364 | * @param int $line |
||
| 365 | * @param int $col |
||
| 366 | * @param Token $prev |
||
| 367 | * @return Token |
||
| 368 | */ |
||
| 369 | protected function readPipe(int $pos, int $line, int $col, Token $prev): Token |
||
| 370 | { |
||
| 371 | return new Token(TokenKindEnum::PIPE, $pos, $pos + 1, $line, $col, $prev); |
||
| 372 | } |
||
| 373 | |||
| 374 | /** |
||
| 375 | * @param int $code |
||
| 376 | * @param int $pos |
||
| 377 | * @param int $line |
||
| 378 | * @param int $col |
||
| 379 | * @param Token $prev |
||
| 380 | * @return Token |
||
| 381 | */ |
||
| 382 | protected function readParenthesis(int $code, int $pos, int $line, int $col, Token $prev): Token |
||
| 383 | { |
||
| 384 | return $code === 40 |
||
| 385 | ? new Token(TokenKindEnum::PAREN_L, $pos, $pos + 1, $line, $col, $prev) |
||
| 386 | : new Token(TokenKindEnum::PAREN_R, $pos, $pos + 1, $line, $col, $prev); |
||
| 387 | } |
||
| 388 | |||
| 389 | /** |
||
| 390 | * @param int $pos |
||
| 391 | * @param int $line |
||
| 392 | * @param int $col |
||
| 393 | * @param Token $prev |
||
| 394 | * @return Token |
||
| 395 | */ |
||
| 396 | protected function readEquals(int $pos, int $line, int $col, Token $prev): Token |
||
| 397 | { |
||
| 398 | return new Token(TokenKindEnum::EQUALS, $pos, $pos + 1, $line, $col, $prev); |
||
| 399 | } |
||
| 400 | |||
| 401 | /** |
||
| 402 | * @param int $pos |
||
| 403 | * @param int $line |
||
| 404 | * @param int $col |
||
| 405 | * @param Token $prev |
||
| 406 | * @return Token |
||
| 407 | */ |
||
| 408 | protected function readAt(int $pos, int $line, int $col, Token $prev): Token |
||
| 409 | { |
||
| 410 | return new Token(TokenKindEnum::AT, $pos, $pos + 1, $line, $col, $prev); |
||
| 411 | } |
||
| 412 | |||
| 413 | /** |
||
| 414 | * @param int $pos |
||
| 415 | * @param int $line |
||
| 416 | * @param int $col |
||
| 417 | * @param Token $prev |
||
| 418 | * @return Token |
||
| 419 | */ |
||
| 420 | protected function readComment(int $pos, int $line, int $col, Token $prev): Token |
||
| 421 | { |
||
| 422 | $body = $this->lexer->getBody(); |
||
| 423 | $start = $pos; |
||
| 424 | |||
| 425 | do { |
||
| 426 | $code = charCodeAt($body, ++$pos); |
||
| 427 | } while ($code !== null && ($code > 0x001f || $code === 0x0009)); // SourceCharacter but not LineTerminator |
||
| 428 | |||
| 429 | return new Token( |
||
| 430 | TokenKindEnum::COMMENT, |
||
| 431 | $start, |
||
| 432 | $pos, |
||
| 433 | $line, |
||
| 434 | $col, |
||
| 435 | $prev, |
||
| 436 | sliceString($body, $start + 1, $pos) |
||
| 437 | ); |
||
| 438 | } |
||
| 439 | |||
| 440 | /** |
||
| 441 | * @param int $pos |
||
| 442 | * @param int $line |
||
| 443 | * @param int $col |
||
| 444 | * @param Token $prev |
||
| 445 | * @return Token |
||
| 446 | */ |
||
| 447 | protected function readColon(int $pos, int $line, int $col, Token $prev): Token |
||
| 448 | { |
||
| 449 | return new Token(TokenKindEnum::COLON, $pos, $pos + 1, $line, $col, $prev); |
||
| 450 | } |
||
| 451 | |||
| 452 | /** |
||
| 453 | * @param int $pos |
||
| 454 | * @param int $line |
||
| 455 | * @param int $col |
||
| 456 | * @param Token $prev |
||
| 457 | * @return Token |
||
| 458 | */ |
||
| 459 | protected function readAmp(int $pos, int $line, int $col, Token $prev): Token |
||
| 460 | { |
||
| 461 | return new Token(TokenKindEnum::AMP, $pos, $pos + 1, $line, $col, $prev); |
||
| 462 | } |
||
| 463 | |||
| 464 | /** |
||
| 465 | * @param int $pos |
||
| 466 | * @param int $line |
||
| 467 | * @param int $col |
||
| 468 | * @param Token $prev |
||
| 469 | * @return Token |
||
| 470 | */ |
||
| 471 | protected function readBang(int $pos, int $line, int $col, Token $prev): Token |
||
| 472 | { |
||
| 473 | return new Token(TokenKindEnum::BANG, $pos, $pos + 1, $line, $col, $prev); |
||
| 474 | } |
||
| 475 | |||
| 476 | /** |
||
| 477 | * @param int $code |
||
| 478 | * @param int $pos |
||
| 479 | * @param int $line |
||
| 480 | * @param int $col |
||
| 481 | * @param Token $prev |
||
| 482 | * @return Token |
||
| 483 | */ |
||
| 484 | protected function readBrace(int $code, int $pos, int $line, int $col, Token $prev): Token |
||
| 485 | { |
||
| 486 | return $code === 123 |
||
| 487 | ? new Token(TokenKindEnum::BRACE_L, $pos, $pos + 1, $line, $col, $prev) |
||
| 488 | : new Token(TokenKindEnum::BRACE_R, $pos, $pos + 1, $line, $col, $prev); |
||
| 489 | } |
||
| 490 | |||
| 491 | /** |
||
| 492 | * @param int $code |
||
| 493 | * @param int $pos |
||
| 494 | * @param int $line |
||
| 495 | * @param int $col |
||
| 496 | * @param Token $prev |
||
| 497 | * @return Token |
||
| 498 | */ |
||
| 499 | protected function readBracket(int $code, int $pos, int $line, int $col, Token $prev): Token |
||
| 504 | } |
||
| 505 | |||
| 506 | /** |
||
| 507 | * @param int $code |
||
| 508 | * @param int $pos |
||
| 509 | * @return int |
||
| 510 | * @throws SyntaxErrorException |
||
| 511 | */ |
||
| 512 | protected function readDigits(int $code, int $pos): int |
||
| 513 | { |
||
| 514 | $body = $this->lexer->getBody(); |
||
| 515 | |||
| 528 | ); |
||
| 529 | } |
||
| 530 | } |
||
| 531 |
