p6naveen /
project-1
Duplicate code is one of the most pungent code smells. A rule that is often used is to re-structure code once it is duplicated in three or more places.
Common duplication problems, and corresponding solutions are:
Complex classes like Parser 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. You can also have a look at the cohesion graph to spot any un-connected, or weakly-connected components.
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 Parser, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 21 | class Parser |
||
| 22 | { |
||
| 23 | const BLOCK_SCALAR_HEADER_PATTERN = '(?P<separator>\||>)(?P<modifiers>\+|\-|\d+|\+\d+|\-\d+|\d+\+|\d+\-)?(?P<comments> +#.*)?'; |
||
| 24 | |||
| 25 | private $offset = 0; |
||
| 26 | private $lines = array(); |
||
| 27 | private $currentLineNb = -1; |
||
| 28 | private $currentLine = ''; |
||
| 29 | private $refs = array(); |
||
| 30 | |||
| 31 | /** |
||
| 32 | * Constructor. |
||
| 33 | * |
||
| 34 | * @param int $offset The offset of YAML document (used for line numbers in error messages) |
||
| 35 | */ |
||
| 36 | public function __construct($offset = 0) |
||
| 37 | { |
||
| 38 | $this->offset = $offset; |
||
| 39 | } |
||
| 40 | |||
| 41 | /** |
||
| 42 | * Parses a YAML string to a PHP value. |
||
| 43 | * |
||
| 44 | * @param string $value A YAML string |
||
| 45 | * @param bool $exceptionOnInvalidType true if an exception must be thrown on invalid types (a PHP resource or object), false otherwise |
||
| 46 | * @param bool $objectSupport true if object support is enabled, false otherwise |
||
| 47 | * @param bool $objectForMap true if maps should return a stdClass instead of array() |
||
| 48 | * |
||
| 49 | * @return mixed A PHP value |
||
| 50 | * |
||
| 51 | * @throws ParseException If the YAML is not valid |
||
| 52 | */ |
||
| 53 | public function parse($value, $exceptionOnInvalidType = false, $objectSupport = false, $objectForMap = false) |
||
| 54 | { |
||
| 55 | if (!preg_match('//u', $value)) { |
||
| 56 | throw new ParseException('The YAML value does not appear to be valid UTF-8.'); |
||
| 57 | } |
||
| 58 | $this->currentLineNb = -1; |
||
| 59 | $this->currentLine = ''; |
||
| 60 | $value = $this->cleanup($value); |
||
| 61 | $this->lines = explode("\n", $value); |
||
| 62 | |||
| 63 | if (2 /* MB_OVERLOAD_STRING */ & (int) ini_get('mbstring.func_overload')) { |
||
| 64 | $mbEncoding = mb_internal_encoding(); |
||
| 65 | mb_internal_encoding('UTF-8'); |
||
| 66 | } |
||
| 67 | |||
| 68 | $data = array(); |
||
| 69 | $context = null; |
||
| 70 | $allowOverwrite = false; |
||
| 71 | while ($this->moveToNextLine()) { |
||
| 72 | if ($this->isCurrentLineEmpty()) { |
||
| 73 | continue; |
||
| 74 | } |
||
| 75 | |||
| 76 | // tab? |
||
| 77 | if ("\t" === $this->currentLine[0]) { |
||
| 78 | throw new ParseException('A YAML file cannot contain tabs as indentation.', $this->getRealCurrentLineNb() + 1, $this->currentLine); |
||
| 79 | } |
||
| 80 | |||
| 81 | $isRef = $mergeNode = false; |
||
| 82 | if (preg_match('#^\-((?P<leadspaces>\s+)(?P<value>.+?))?\s*$#u', $this->currentLine, $values)) { |
||
| 83 | if ($context && 'mapping' == $context) { |
||
| 84 | throw new ParseException('You cannot define a sequence item when in a mapping'); |
||
| 85 | } |
||
| 86 | $context = 'sequence'; |
||
| 87 | |||
| 88 | if (isset($values['value']) && preg_match('#^&(?P<ref>[^ ]+) *(?P<value>.*)#u', $values['value'], $matches)) { |
||
| 89 | $isRef = $matches['ref']; |
||
| 90 | $values['value'] = $matches['value']; |
||
| 91 | } |
||
| 92 | |||
| 93 | // array |
||
| 94 | if (!isset($values['value']) || '' == trim($values['value'], ' ') || 0 === strpos(ltrim($values['value'], ' '), '#')) { |
||
| 95 | $c = $this->getRealCurrentLineNb() + 1; |
||
| 96 | $parser = new self($c); |
||
| 97 | $parser->refs = &$this->refs; |
||
| 98 | $data[] = $parser->parse($this->getNextEmbedBlock(null, true), $exceptionOnInvalidType, $objectSupport, $objectForMap); |
||
| 99 | } else { |
||
| 100 | if (isset($values['leadspaces']) |
||
| 101 | && preg_match('#^(?P<key>'.Inline::REGEX_QUOTED_STRING.'|[^ \'"\{\[].*?) *\:(\s+(?P<value>.+?))?\s*$#u', $values['value'], $matches) |
||
| 102 | ) { |
||
| 103 | // this is a compact notation element, add to next block and parse |
||
| 104 | $c = $this->getRealCurrentLineNb(); |
||
| 105 | $parser = new self($c); |
||
| 106 | $parser->refs = &$this->refs; |
||
| 107 | |||
| 108 | $block = $values['value']; |
||
| 109 | if ($this->isNextLineIndented()) { |
||
| 110 | $block .= "\n".$this->getNextEmbedBlock($this->getCurrentLineIndentation() + strlen($values['leadspaces']) + 1); |
||
| 111 | } |
||
| 112 | |||
| 113 | $data[] = $parser->parse($block, $exceptionOnInvalidType, $objectSupport, $objectForMap); |
||
| 114 | } else { |
||
| 115 | $data[] = $this->parseValue($values['value'], $exceptionOnInvalidType, $objectSupport, $objectForMap, $context); |
||
| 116 | } |
||
| 117 | } |
||
| 118 | if ($isRef) { |
||
| 119 | $this->refs[$isRef] = end($data); |
||
| 120 | } |
||
| 121 | } elseif (preg_match('#^(?P<key>'.Inline::REGEX_QUOTED_STRING.'|[^ \'"\[\{].*?) *\:(\s+(?P<value>.+?))?\s*$#u', $this->currentLine, $values) && (false === strpos($values['key'], ' #') || in_array($values['key'][0], array('"', "'")))) { |
||
| 122 | if ($context && 'sequence' == $context) { |
||
| 123 | throw new ParseException('You cannot define a mapping item when in a sequence'); |
||
| 124 | } |
||
| 125 | $context = 'mapping'; |
||
| 126 | |||
| 127 | // force correct settings |
||
| 128 | Inline::parse(null, $exceptionOnInvalidType, $objectSupport, $objectForMap, $this->refs); |
||
| 129 | try { |
||
| 130 | $key = Inline::parseScalar($values['key']); |
||
| 131 | } catch (ParseException $e) { |
||
| 132 | $e->setParsedLine($this->getRealCurrentLineNb() + 1); |
||
| 133 | $e->setSnippet($this->currentLine); |
||
| 134 | |||
| 135 | throw $e; |
||
| 136 | } |
||
| 137 | |||
| 138 | // Convert float keys to strings, to avoid being converted to integers by PHP |
||
| 139 | if (is_float($key)) { |
||
| 140 | $key = (string) $key; |
||
| 141 | } |
||
| 142 | |||
| 143 | if ('<<' === $key) { |
||
| 144 | $mergeNode = true; |
||
| 145 | $allowOverwrite = true; |
||
| 146 | if (isset($values['value']) && 0 === strpos($values['value'], '*')) { |
||
| 147 | $refName = substr($values['value'], 1); |
||
| 148 | if (!array_key_exists($refName, $this->refs)) { |
||
| 149 | throw new ParseException(sprintf('Reference "%s" does not exist.', $refName), $this->getRealCurrentLineNb() + 1, $this->currentLine); |
||
| 150 | } |
||
| 151 | |||
| 152 | $refValue = $this->refs[$refName]; |
||
| 153 | |||
| 154 | if (!is_array($refValue)) { |
||
| 155 | throw new ParseException('YAML merge keys used with a scalar value instead of an array.', $this->getRealCurrentLineNb() + 1, $this->currentLine); |
||
| 156 | } |
||
| 157 | |||
| 158 | foreach ($refValue as $key => $value) { |
||
| 159 | if (!isset($data[$key])) { |
||
| 160 | $data[$key] = $value; |
||
| 161 | } |
||
| 162 | } |
||
| 163 | } else { |
||
| 164 | if (isset($values['value']) && $values['value'] !== '') { |
||
| 165 | $value = $values['value']; |
||
| 166 | } else { |
||
| 167 | $value = $this->getNextEmbedBlock(); |
||
| 168 | } |
||
| 169 | $c = $this->getRealCurrentLineNb() + 1; |
||
| 170 | $parser = new self($c); |
||
| 171 | $parser->refs = &$this->refs; |
||
| 172 | $parsed = $parser->parse($value, $exceptionOnInvalidType, $objectSupport, $objectForMap); |
||
| 173 | |||
| 174 | if (!is_array($parsed)) { |
||
| 175 | throw new ParseException('YAML merge keys used with a scalar value instead of an array.', $this->getRealCurrentLineNb() + 1, $this->currentLine); |
||
| 176 | } |
||
| 177 | |||
| 178 | if (isset($parsed[0])) { |
||
| 179 | // If the value associated with the merge key is a sequence, then this sequence is expected to contain mapping nodes |
||
| 180 | // and each of these nodes is merged in turn according to its order in the sequence. Keys in mapping nodes earlier |
||
| 181 | // in the sequence override keys specified in later mapping nodes. |
||
| 182 | foreach ($parsed as $parsedItem) { |
||
| 183 | if (!is_array($parsedItem)) { |
||
| 184 | throw new ParseException('Merge items must be arrays.', $this->getRealCurrentLineNb() + 1, $parsedItem); |
||
| 185 | } |
||
| 186 | |||
| 187 | foreach ($parsedItem as $key => $value) { |
||
| 188 | if (!isset($data[$key])) { |
||
| 189 | $data[$key] = $value; |
||
| 190 | } |
||
| 191 | } |
||
| 192 | } |
||
| 193 | } else { |
||
| 194 | // If the value associated with the key is a single mapping node, each of its key/value pairs is inserted into the |
||
| 195 | // current mapping, unless the key already exists in it. |
||
| 196 | foreach ($parsed as $key => $value) { |
||
| 197 | if (!isset($data[$key])) { |
||
| 198 | $data[$key] = $value; |
||
| 199 | } |
||
| 200 | } |
||
| 201 | } |
||
| 202 | } |
||
| 203 | } elseif (isset($values['value']) && preg_match('#^&(?P<ref>[^ ]+) *(?P<value>.*)#u', $values['value'], $matches)) { |
||
| 204 | $isRef = $matches['ref']; |
||
| 205 | $values['value'] = $matches['value']; |
||
| 206 | } |
||
| 207 | |||
| 208 | if ($mergeNode) { |
||
| 209 | // Merge keys |
||
| 210 | } elseif (!isset($values['value']) || '' == trim($values['value'], ' ') || 0 === strpos(ltrim($values['value'], ' '), '#')) { |
||
| 211 | // hash |
||
| 212 | // if next line is less indented or equal, then it means that the current value is null |
||
| 213 | if (!$this->isNextLineIndented() && !$this->isNextLineUnIndentedCollection()) { |
||
| 214 | // Spec: Keys MUST be unique; first one wins. |
||
| 215 | // But overwriting is allowed when a merge node is used in current block. |
||
| 216 | if ($allowOverwrite || !isset($data[$key])) { |
||
| 217 | $data[$key] = null; |
||
| 218 | } |
||
| 219 | } else { |
||
| 220 | $c = $this->getRealCurrentLineNb() + 1; |
||
| 221 | $parser = new self($c); |
||
| 222 | $parser->refs = &$this->refs; |
||
| 223 | $value = $parser->parse($this->getNextEmbedBlock(), $exceptionOnInvalidType, $objectSupport, $objectForMap); |
||
| 224 | // Spec: Keys MUST be unique; first one wins. |
||
| 225 | // But overwriting is allowed when a merge node is used in current block. |
||
| 226 | if ($allowOverwrite || !isset($data[$key])) { |
||
| 227 | $data[$key] = $value; |
||
| 228 | } |
||
| 229 | } |
||
| 230 | } else { |
||
| 231 | $value = $this->parseValue($values['value'], $exceptionOnInvalidType, $objectSupport, $objectForMap, $context); |
||
| 232 | // Spec: Keys MUST be unique; first one wins. |
||
| 233 | // But overwriting is allowed when a merge node is used in current block. |
||
| 234 | if ($allowOverwrite || !isset($data[$key])) { |
||
| 235 | $data[$key] = $value; |
||
| 236 | } |
||
| 237 | } |
||
| 238 | if ($isRef) { |
||
| 239 | $this->refs[$isRef] = $data[$key]; |
||
| 240 | } |
||
| 241 | } else { |
||
| 242 | // multiple documents are not supported |
||
| 243 | if ('---' === $this->currentLine) { |
||
| 244 | throw new ParseException('Multiple documents are not supported.'); |
||
| 245 | } |
||
| 246 | |||
| 247 | // 1-liner optionally followed by newline(s) |
||
| 248 | if (is_string($value) && $this->lines[0] === trim($value)) { |
||
| 249 | try { |
||
| 250 | $value = Inline::parse($this->lines[0], $exceptionOnInvalidType, $objectSupport, $objectForMap, $this->refs); |
||
| 251 | } catch (ParseException $e) { |
||
| 252 | $e->setParsedLine($this->getRealCurrentLineNb() + 1); |
||
| 253 | $e->setSnippet($this->currentLine); |
||
| 254 | |||
| 255 | throw $e; |
||
| 256 | } |
||
| 257 | |||
| 258 | if (is_array($value)) { |
||
| 259 | $first = reset($value); |
||
| 260 | if (is_string($first) && 0 === strpos($first, '*')) { |
||
| 261 | $data = array(); |
||
| 262 | foreach ($value as $alias) { |
||
| 263 | $data[] = $this->refs[substr($alias, 1)]; |
||
| 264 | } |
||
| 265 | $value = $data; |
||
| 266 | } |
||
| 267 | } |
||
| 268 | |||
| 269 | if (isset($mbEncoding)) { |
||
| 270 | mb_internal_encoding($mbEncoding); |
||
| 271 | } |
||
| 272 | |||
| 273 | return $value; |
||
| 274 | } |
||
| 275 | |||
| 276 | switch (preg_last_error()) { |
||
| 277 | case PREG_INTERNAL_ERROR: |
||
| 278 | $error = 'Internal PCRE error.'; |
||
| 279 | break; |
||
| 280 | case PREG_BACKTRACK_LIMIT_ERROR: |
||
| 281 | $error = 'pcre.backtrack_limit reached.'; |
||
| 282 | break; |
||
| 283 | case PREG_RECURSION_LIMIT_ERROR: |
||
| 284 | $error = 'pcre.recursion_limit reached.'; |
||
| 285 | break; |
||
| 286 | case PREG_BAD_UTF8_ERROR: |
||
| 287 | $error = 'Malformed UTF-8 data.'; |
||
| 288 | break; |
||
| 289 | case PREG_BAD_UTF8_OFFSET_ERROR: |
||
| 290 | $error = 'Offset doesn\'t correspond to the begin of a valid UTF-8 code point.'; |
||
| 291 | break; |
||
| 292 | default: |
||
| 293 | $error = 'Unable to parse.'; |
||
| 294 | } |
||
| 295 | |||
| 296 | throw new ParseException($error, $this->getRealCurrentLineNb() + 1, $this->currentLine); |
||
| 297 | } |
||
| 298 | } |
||
| 299 | |||
| 300 | if (isset($mbEncoding)) { |
||
| 301 | mb_internal_encoding($mbEncoding); |
||
| 302 | } |
||
| 303 | |||
| 304 | if ($objectForMap && !is_object($data)) { |
||
| 305 | $data = (object) $data; |
||
| 306 | } |
||
| 307 | |||
| 308 | return empty($data) ? null : $data; |
||
| 309 | } |
||
| 310 | |||
| 311 | /** |
||
| 312 | * Returns the current line number (takes the offset into account). |
||
| 313 | * |
||
| 314 | * @return int The current line number |
||
| 315 | */ |
||
| 316 | private function getRealCurrentLineNb() |
||
| 317 | { |
||
| 318 | return $this->currentLineNb + $this->offset; |
||
| 319 | } |
||
| 320 | |||
| 321 | /** |
||
| 322 | * Returns the current line indentation. |
||
| 323 | * |
||
| 324 | * @return int The current line indentation |
||
| 325 | */ |
||
| 326 | private function getCurrentLineIndentation() |
||
| 327 | { |
||
| 328 | return strlen($this->currentLine) - strlen(ltrim($this->currentLine, ' ')); |
||
| 329 | } |
||
| 330 | |||
| 331 | /** |
||
| 332 | * Returns the next embed block of YAML. |
||
| 333 | * |
||
| 334 | * @param int $indentation The indent level at which the block is to be read, or null for default |
||
| 335 | * @param bool $inSequence True if the enclosing data structure is a sequence |
||
| 336 | * |
||
| 337 | * @return string A YAML string |
||
| 338 | * |
||
| 339 | * @throws ParseException When indentation problem are detected |
||
| 340 | */ |
||
| 341 | private function getNextEmbedBlock($indentation = null, $inSequence = false) |
||
| 342 | { |
||
| 343 | $oldLineIndentation = $this->getCurrentLineIndentation(); |
||
| 344 | $blockScalarIndentations = array(); |
||
| 345 | |||
| 346 | if ($this->isBlockScalarHeader()) { |
||
| 347 | $blockScalarIndentations[] = $this->getCurrentLineIndentation(); |
||
| 348 | } |
||
| 349 | |||
| 350 | if (!$this->moveToNextLine()) { |
||
| 351 | return; |
||
| 352 | } |
||
| 353 | |||
| 354 | if (null === $indentation) { |
||
| 355 | $newIndent = $this->getCurrentLineIndentation(); |
||
| 356 | |||
| 357 | $unindentedEmbedBlock = $this->isStringUnIndentedCollectionItem(); |
||
| 358 | |||
| 359 | if (!$this->isCurrentLineEmpty() && 0 === $newIndent && !$unindentedEmbedBlock) { |
||
| 360 | throw new ParseException('Indentation problem.', $this->getRealCurrentLineNb() + 1, $this->currentLine); |
||
| 361 | } |
||
| 362 | } else { |
||
| 363 | $newIndent = $indentation; |
||
| 364 | } |
||
| 365 | |||
| 366 | $data = array(); |
||
| 367 | if ($this->getCurrentLineIndentation() >= $newIndent) { |
||
| 368 | $data[] = substr($this->currentLine, $newIndent); |
||
| 369 | } else { |
||
| 370 | $this->moveToPreviousLine(); |
||
| 371 | |||
| 372 | return; |
||
| 373 | } |
||
| 374 | |||
| 375 | if ($inSequence && $oldLineIndentation === $newIndent && isset($data[0][0]) && '-' === $data[0][0]) { |
||
| 376 | // the previous line contained a dash but no item content, this line is a sequence item with the same indentation |
||
| 377 | // and therefore no nested list or mapping |
||
| 378 | $this->moveToPreviousLine(); |
||
| 379 | |||
| 380 | return; |
||
| 381 | } |
||
| 382 | |||
| 383 | $isItUnindentedCollection = $this->isStringUnIndentedCollectionItem(); |
||
| 384 | |||
| 385 | if (empty($blockScalarIndentations) && $this->isBlockScalarHeader()) { |
||
| 386 | $blockScalarIndentations[] = $this->getCurrentLineIndentation(); |
||
| 387 | } |
||
| 388 | |||
| 389 | $previousLineIndentation = $this->getCurrentLineIndentation(); |
||
| 390 | |||
| 391 | while ($this->moveToNextLine()) { |
||
| 392 | $indent = $this->getCurrentLineIndentation(); |
||
| 393 | |||
| 394 | // terminate all block scalars that are more indented than the current line |
||
| 395 | if (!empty($blockScalarIndentations) && $indent < $previousLineIndentation && trim($this->currentLine) !== '') { |
||
| 396 | foreach ($blockScalarIndentations as $key => $blockScalarIndentation) { |
||
| 397 | if ($blockScalarIndentation >= $this->getCurrentLineIndentation()) { |
||
| 398 | unset($blockScalarIndentations[$key]); |
||
| 399 | } |
||
| 400 | } |
||
| 401 | } |
||
| 402 | |||
| 403 | if (empty($blockScalarIndentations) && !$this->isCurrentLineComment() && $this->isBlockScalarHeader()) { |
||
| 404 | $blockScalarIndentations[] = $this->getCurrentLineIndentation(); |
||
| 405 | } |
||
| 406 | |||
| 407 | $previousLineIndentation = $indent; |
||
| 408 | |||
| 409 | if ($isItUnindentedCollection && !$this->isStringUnIndentedCollectionItem() && $newIndent === $indent) { |
||
| 410 | $this->moveToPreviousLine(); |
||
| 411 | break; |
||
| 412 | } |
||
| 413 | |||
| 414 | if ($this->isCurrentLineBlank()) { |
||
| 415 | $data[] = substr($this->currentLine, $newIndent); |
||
| 416 | continue; |
||
| 417 | } |
||
| 418 | |||
| 419 | // we ignore "comment" lines only when we are not inside a scalar block |
||
| 420 | if (empty($blockScalarIndentations) && $this->isCurrentLineComment()) { |
||
| 421 | continue; |
||
| 422 | } |
||
| 423 | |||
| 424 | if ($indent >= $newIndent) { |
||
| 425 | $data[] = substr($this->currentLine, $newIndent); |
||
| 426 | } elseif (0 == $indent) { |
||
| 427 | $this->moveToPreviousLine(); |
||
| 428 | |||
| 429 | break; |
||
| 430 | } else { |
||
| 431 | throw new ParseException('Indentation problem.', $this->getRealCurrentLineNb() + 1, $this->currentLine); |
||
| 432 | } |
||
| 433 | } |
||
| 434 | |||
| 435 | return implode("\n", $data); |
||
| 436 | } |
||
| 437 | |||
| 438 | /** |
||
| 439 | * Moves the parser to the next line. |
||
| 440 | * |
||
| 441 | * @return bool |
||
| 442 | */ |
||
| 443 | private function moveToNextLine() |
||
| 444 | { |
||
| 445 | if ($this->currentLineNb >= count($this->lines) - 1) { |
||
| 446 | return false; |
||
| 447 | } |
||
| 448 | |||
| 449 | $this->currentLine = $this->lines[++$this->currentLineNb]; |
||
| 450 | |||
| 451 | return true; |
||
| 452 | } |
||
| 453 | |||
| 454 | /** |
||
| 455 | * Moves the parser to the previous line. |
||
| 456 | */ |
||
| 457 | private function moveToPreviousLine() |
||
| 458 | { |
||
| 459 | $this->currentLine = $this->lines[--$this->currentLineNb]; |
||
| 460 | } |
||
| 461 | |||
| 462 | /** |
||
| 463 | * Parses a YAML value. |
||
| 464 | * |
||
| 465 | * @param string $value A YAML value |
||
| 466 | * @param bool $exceptionOnInvalidType True if an exception must be thrown on invalid types false otherwise |
||
| 467 | * @param bool $objectSupport True if object support is enabled, false otherwise |
||
| 468 | * @param bool $objectForMap true if maps should return a stdClass instead of array() |
||
| 469 | * @param string $context The parser context (either sequence or mapping) |
||
| 470 | * |
||
| 471 | * @return mixed A PHP value |
||
| 472 | * |
||
| 473 | * @throws ParseException When reference does not exist |
||
| 474 | */ |
||
| 475 | private function parseValue($value, $exceptionOnInvalidType, $objectSupport, $objectForMap, $context) |
||
| 476 | { |
||
| 477 | if (0 === strpos($value, '*')) { |
||
| 478 | if (false !== $pos = strpos($value, '#')) { |
||
| 479 | $value = substr($value, 1, $pos - 2); |
||
| 480 | } else { |
||
| 481 | $value = substr($value, 1); |
||
| 482 | } |
||
| 483 | |||
| 484 | if (!array_key_exists($value, $this->refs)) { |
||
| 485 | throw new ParseException(sprintf('Reference "%s" does not exist.', $value), $this->currentLine); |
||
| 486 | } |
||
| 487 | |||
| 488 | return $this->refs[$value]; |
||
| 489 | } |
||
| 490 | |||
| 491 | if (preg_match('/^'.self::BLOCK_SCALAR_HEADER_PATTERN.'$/', $value, $matches)) { |
||
| 492 | $modifiers = isset($matches['modifiers']) ? $matches['modifiers'] : ''; |
||
| 493 | |||
| 494 | return $this->parseBlockScalar($matches['separator'], preg_replace('#\d+#', '', $modifiers), (int) abs($modifiers)); |
||
| 495 | } |
||
| 496 | |||
| 497 | try { |
||
| 498 | $parsedValue = Inline::parse($value, $exceptionOnInvalidType, $objectSupport, $objectForMap, $this->refs); |
||
| 499 | |||
| 500 | if ('mapping' === $context && '"' !== $value[0] && "'" !== $value[0] && '[' !== $value[0] && '{' !== $value[0] && '!' !== $value[0] && false !== strpos($parsedValue, ': ')) { |
||
| 501 | throw new ParseException('A colon cannot be used in an unquoted mapping value.'); |
||
| 502 | } |
||
| 503 | |||
| 504 | return $parsedValue; |
||
| 505 | } catch (ParseException $e) { |
||
| 506 | $e->setParsedLine($this->getRealCurrentLineNb() + 1); |
||
| 507 | $e->setSnippet($this->currentLine); |
||
| 508 | |||
| 509 | throw $e; |
||
| 510 | } |
||
| 511 | } |
||
| 512 | |||
| 513 | /** |
||
| 514 | * Parses a block scalar. |
||
| 515 | * |
||
| 516 | * @param string $style The style indicator that was used to begin this block scalar (| or >) |
||
| 517 | * @param string $chomping The chomping indicator that was used to begin this block scalar (+ or -) |
||
| 518 | * @param int $indentation The indentation indicator that was used to begin this block scalar |
||
| 519 | * |
||
| 520 | * @return string The text value |
||
| 521 | */ |
||
| 522 | private function parseBlockScalar($style, $chomping = '', $indentation = 0) |
||
| 523 | { |
||
| 524 | $notEOF = $this->moveToNextLine(); |
||
| 525 | if (!$notEOF) { |
||
| 526 | return ''; |
||
| 527 | } |
||
| 528 | |||
| 529 | $isCurrentLineBlank = $this->isCurrentLineBlank(); |
||
| 530 | $blockLines = array(); |
||
| 531 | |||
| 532 | // leading blank lines are consumed before determining indentation |
||
| 533 | while ($notEOF && $isCurrentLineBlank) { |
||
| 534 | // newline only if not EOF |
||
| 535 | if ($notEOF = $this->moveToNextLine()) { |
||
| 536 | $blockLines[] = ''; |
||
| 537 | $isCurrentLineBlank = $this->isCurrentLineBlank(); |
||
| 538 | } |
||
| 539 | } |
||
| 540 | |||
| 541 | // determine indentation if not specified |
||
| 542 | if (0 === $indentation) { |
||
| 543 | if (preg_match('/^ +/', $this->currentLine, $matches)) { |
||
| 544 | $indentation = strlen($matches[0]); |
||
| 545 | } |
||
| 546 | } |
||
| 547 | |||
| 548 | if ($indentation > 0) { |
||
| 549 | $pattern = sprintf('/^ {%d}(.*)$/', $indentation); |
||
| 550 | |||
| 551 | while ( |
||
| 552 | $notEOF && ( |
||
| 553 | $isCurrentLineBlank || |
||
| 554 | preg_match($pattern, $this->currentLine, $matches) |
||
| 555 | ) |
||
| 556 | ) { |
||
| 557 | if ($isCurrentLineBlank && strlen($this->currentLine) > $indentation) { |
||
| 558 | $blockLines[] = substr($this->currentLine, $indentation); |
||
| 559 | } elseif ($isCurrentLineBlank) { |
||
| 560 | $blockLines[] = ''; |
||
| 561 | } else { |
||
| 562 | $blockLines[] = $matches[1]; |
||
| 563 | } |
||
| 564 | |||
| 565 | // newline only if not EOF |
||
| 566 | if ($notEOF = $this->moveToNextLine()) { |
||
| 567 | $isCurrentLineBlank = $this->isCurrentLineBlank(); |
||
| 568 | } |
||
| 569 | } |
||
| 570 | } elseif ($notEOF) { |
||
| 571 | $blockLines[] = ''; |
||
| 572 | } |
||
| 573 | |||
| 574 | if ($notEOF) { |
||
| 575 | $blockLines[] = ''; |
||
| 576 | $this->moveToPreviousLine(); |
||
| 577 | } |
||
| 578 | |||
| 579 | // folded style |
||
| 580 | if ('>' === $style) { |
||
| 581 | $text = ''; |
||
| 582 | $previousLineIndented = false; |
||
| 583 | $previousLineBlank = false; |
||
| 584 | |||
| 585 | for ($i = 0; $i < count($blockLines); ++$i) { |
||
| 586 | if ('' === $blockLines[$i]) { |
||
| 587 | $text .= "\n"; |
||
| 588 | $previousLineIndented = false; |
||
| 589 | $previousLineBlank = true; |
||
| 590 | } elseif (' ' === $blockLines[$i][0]) { |
||
| 591 | $text .= "\n".$blockLines[$i]; |
||
| 592 | $previousLineIndented = true; |
||
| 593 | $previousLineBlank = false; |
||
| 594 | } elseif ($previousLineIndented) { |
||
| 595 | $text .= "\n".$blockLines[$i]; |
||
| 596 | $previousLineIndented = false; |
||
| 597 | $previousLineBlank = false; |
||
| 598 | } elseif ($previousLineBlank || 0 === $i) { |
||
| 599 | $text .= $blockLines[$i]; |
||
| 600 | $previousLineIndented = false; |
||
| 601 | $previousLineBlank = false; |
||
| 602 | } else { |
||
| 603 | $text .= ' '.$blockLines[$i]; |
||
| 604 | $previousLineIndented = false; |
||
| 605 | $previousLineBlank = false; |
||
| 606 | } |
||
| 607 | } |
||
| 608 | } else { |
||
| 609 | $text = implode("\n", $blockLines); |
||
| 610 | } |
||
| 611 | |||
| 612 | // deal with trailing newlines |
||
| 613 | if ('' === $chomping) { |
||
| 614 | $text = preg_replace('/\n+$/', "\n", $text); |
||
| 615 | } elseif ('-' === $chomping) { |
||
| 616 | $text = preg_replace('/\n+$/', '', $text); |
||
| 617 | } |
||
| 618 | |||
| 619 | return $text; |
||
| 620 | } |
||
| 621 | |||
| 622 | /** |
||
| 623 | * Returns true if the next line is indented. |
||
| 624 | * |
||
| 625 | * @return bool Returns true if the next line is indented, false otherwise |
||
| 626 | */ |
||
| 627 | private function isNextLineIndented() |
||
| 628 | { |
||
| 629 | $currentIndentation = $this->getCurrentLineIndentation(); |
||
| 630 | $EOF = !$this->moveToNextLine(); |
||
| 631 | |||
| 632 | while (!$EOF && $this->isCurrentLineEmpty()) { |
||
| 633 | $EOF = !$this->moveToNextLine(); |
||
| 634 | } |
||
| 635 | |||
| 636 | if ($EOF) { |
||
| 637 | return false; |
||
| 638 | } |
||
| 639 | |||
| 640 | $ret = false; |
||
| 641 | if ($this->getCurrentLineIndentation() > $currentIndentation) { |
||
| 642 | $ret = true; |
||
| 643 | } |
||
| 644 | |||
| 645 | $this->moveToPreviousLine(); |
||
| 646 | |||
| 647 | return $ret; |
||
| 648 | } |
||
| 649 | |||
| 650 | /** |
||
| 651 | * Returns true if the current line is blank or if it is a comment line. |
||
| 652 | * |
||
| 653 | * @return bool Returns true if the current line is empty or if it is a comment line, false otherwise |
||
| 654 | */ |
||
| 655 | private function isCurrentLineEmpty() |
||
| 656 | { |
||
| 657 | return $this->isCurrentLineBlank() || $this->isCurrentLineComment(); |
||
| 658 | } |
||
| 659 | |||
| 660 | /** |
||
| 661 | * Returns true if the current line is blank. |
||
| 662 | * |
||
| 663 | * @return bool Returns true if the current line is blank, false otherwise |
||
| 664 | */ |
||
| 665 | private function isCurrentLineBlank() |
||
| 666 | { |
||
| 667 | return '' == trim($this->currentLine, ' '); |
||
| 668 | } |
||
| 669 | |||
| 670 | /** |
||
| 671 | * Returns true if the current line is a comment line. |
||
| 672 | * |
||
| 673 | * @return bool Returns true if the current line is a comment line, false otherwise |
||
| 674 | */ |
||
| 675 | private function isCurrentLineComment() |
||
| 676 | { |
||
| 677 | //checking explicitly the first char of the trim is faster than loops or strpos |
||
| 678 | $ltrimmedLine = ltrim($this->currentLine, ' '); |
||
| 679 | |||
| 680 | return '' !== $ltrimmedLine && $ltrimmedLine[0] === '#'; |
||
| 681 | } |
||
| 682 | |||
| 683 | /** |
||
| 684 | * Cleanups a YAML string to be parsed. |
||
| 685 | * |
||
| 686 | * @param string $value The input YAML string |
||
| 687 | * |
||
| 688 | * @return string A cleaned up YAML string |
||
| 689 | */ |
||
| 690 | private function cleanup($value) |
||
| 691 | { |
||
| 692 | $value = str_replace(array("\r\n", "\r"), "\n", $value); |
||
| 693 | |||
| 694 | // strip YAML header |
||
| 695 | $count = 0; |
||
| 696 | $value = preg_replace('#^\%YAML[: ][\d\.]+.*\n#u', '', $value, -1, $count); |
||
| 697 | $this->offset += $count; |
||
| 698 | |||
| 699 | // remove leading comments |
||
| 700 | $trimmedValue = preg_replace('#^(\#.*?\n)+#s', '', $value, -1, $count); |
||
| 701 | if ($count == 1) { |
||
| 702 | // items have been removed, update the offset |
||
| 703 | $this->offset += substr_count($value, "\n") - substr_count($trimmedValue, "\n"); |
||
| 704 | $value = $trimmedValue; |
||
| 705 | } |
||
| 706 | |||
| 707 | // remove start of the document marker (---) |
||
| 708 | $trimmedValue = preg_replace('#^\-\-\-.*?\n#s', '', $value, -1, $count); |
||
| 709 | if ($count == 1) { |
||
| 710 | // items have been removed, update the offset |
||
| 711 | $this->offset += substr_count($value, "\n") - substr_count($trimmedValue, "\n"); |
||
| 712 | $value = $trimmedValue; |
||
| 713 | |||
| 714 | // remove end of the document marker (...) |
||
| 715 | $value = preg_replace('#\.\.\.\s*$#', '', $value); |
||
| 716 | } |
||
| 717 | |||
| 718 | return $value; |
||
| 719 | } |
||
| 720 | |||
| 721 | /** |
||
| 722 | * Returns true if the next line starts unindented collection. |
||
| 723 | * |
||
| 724 | * @return bool Returns true if the next line starts unindented collection, false otherwise |
||
| 725 | */ |
||
| 726 | private function isNextLineUnIndentedCollection() |
||
| 727 | { |
||
| 728 | $currentIndentation = $this->getCurrentLineIndentation(); |
||
| 729 | $notEOF = $this->moveToNextLine(); |
||
| 730 | |||
| 731 | while ($notEOF && $this->isCurrentLineEmpty()) { |
||
| 732 | $notEOF = $this->moveToNextLine(); |
||
| 733 | } |
||
| 734 | |||
| 735 | if (false === $notEOF) { |
||
| 736 | return false; |
||
| 737 | } |
||
| 738 | |||
| 739 | $ret = false; |
||
| 740 | if ( |
||
| 741 | $this->getCurrentLineIndentation() == $currentIndentation |
||
| 742 | && |
||
| 743 | $this->isStringUnIndentedCollectionItem() |
||
| 744 | ) { |
||
| 745 | $ret = true; |
||
| 746 | } |
||
| 747 | |||
| 748 | $this->moveToPreviousLine(); |
||
| 749 | |||
| 750 | return $ret; |
||
| 751 | } |
||
| 752 | |||
| 753 | /** |
||
| 754 | * Returns true if the string is un-indented collection item. |
||
| 755 | * |
||
| 756 | * @return bool Returns true if the string is un-indented collection item, false otherwise |
||
| 757 | */ |
||
| 758 | private function isStringUnIndentedCollectionItem() |
||
| 759 | { |
||
| 760 | return 0 === strpos($this->currentLine, '- '); |
||
| 761 | } |
||
| 762 | |||
| 763 | /** |
||
| 764 | * Tests whether or not the current line is the header of a block scalar. |
||
| 765 | * |
||
| 766 | * @return bool |
||
| 767 | */ |
||
| 768 | private function isBlockScalarHeader() |
||
| 769 | { |
||
| 770 | return (bool) preg_match('~'.self::BLOCK_SCALAR_HEADER_PATTERN.'$~', $this->currentLine); |
||
| 771 | } |
||
| 772 | } |
||
| 773 |
