TYPO3 /
TYPO3.CMS
| Total Complexity | 66 |
| Total Lines | 450 |
| Duplicated Lines | 0 % |
| Changes | 0 | ||
Complex classes like LinkAnalyzer 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 LinkAnalyzer, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 35 | class LinkAnalyzer |
||
| 36 | { |
||
| 37 | |||
| 38 | /** |
||
| 39 | * Array of tables and fields to search for broken links |
||
| 40 | * |
||
| 41 | * @var array<string, array<int, string>> |
||
| 42 | */ |
||
| 43 | protected $searchFields = []; |
||
| 44 | |||
| 45 | /** |
||
| 46 | * List of page uids (rootline downwards) |
||
| 47 | * |
||
| 48 | * @var int[] |
||
| 49 | */ |
||
| 50 | protected $pids = []; |
||
| 51 | |||
| 52 | /** |
||
| 53 | * Array of tables and the number of external links they contain |
||
| 54 | * |
||
| 55 | * @var array |
||
| 56 | */ |
||
| 57 | protected $linkCounts = []; |
||
| 58 | |||
| 59 | /** |
||
| 60 | * Array of tables and the number of broken external links they contain |
||
| 61 | * |
||
| 62 | * @var array |
||
| 63 | */ |
||
| 64 | protected $brokenLinkCounts = []; |
||
| 65 | |||
| 66 | /** |
||
| 67 | * Array for hooks for own checks |
||
| 68 | * |
||
| 69 | * @var LinktypeInterface[] |
||
| 70 | */ |
||
| 71 | protected $hookObjectsArr = []; |
||
| 72 | |||
| 73 | /** |
||
| 74 | * The currently active TSconfig. Will be passed to the init function. |
||
| 75 | * |
||
| 76 | * @var array |
||
| 77 | */ |
||
| 78 | protected $tsConfig = []; |
||
| 79 | |||
| 80 | /** |
||
| 81 | * @var EventDispatcherInterface |
||
| 82 | */ |
||
| 83 | protected $eventDispatcher; |
||
| 84 | |||
| 85 | /** |
||
| 86 | * @var BrokenLinkRepository |
||
| 87 | */ |
||
| 88 | protected $brokenLinkRepository; |
||
| 89 | |||
| 90 | public function __construct(EventDispatcherInterface $eventDispatcher, BrokenLinkRepository $brokenLinkRepository) |
||
| 95 | } |
||
| 96 | |||
| 97 | /** |
||
| 98 | * Store all the needed configuration values in class variables |
||
| 99 | * |
||
| 100 | * @param array $searchFields List of fields in which to search for links |
||
| 101 | * @param int[] $pidList List of page uids in which to search for links |
||
| 102 | * @param array $tsConfig The currently active TSconfig. |
||
| 103 | */ |
||
| 104 | public function init(array $searchFields, array $pidList, $tsConfig) |
||
| 120 | } |
||
| 121 | } |
||
| 122 | |||
| 123 | /** |
||
| 124 | * Find all supported broken links and store them in tx_linkvalidator_link |
||
| 125 | * |
||
| 126 | * @param array<int,string> $linkTypes List of hook object to activate |
||
| 127 | * @param bool $considerHidden Defines whether to look into hidden fields |
||
| 128 | */ |
||
| 129 | public function getLinkStatistics(array $linkTypes = [], $considerHidden = false) |
||
| 184 | } |
||
| 185 | |||
| 186 | /** |
||
| 187 | * @param array $links |
||
| 188 | * @param array<int,string> $linkTypes |
||
| 189 | */ |
||
| 190 | protected function checkLinks(array $links, array $linkTypes) |
||
| 191 | { |
||
| 192 | foreach ($this->hookObjectsArr as $key => $hookObj) { |
||
| 193 | if (!is_array($links[$key] ?? false) || (!in_array($key, $linkTypes, true))) { |
||
| 194 | continue; |
||
| 195 | } |
||
| 196 | |||
| 197 | // Check them |
||
| 198 | foreach ($links[$key] as $entryKey => $entryValue) { |
||
| 199 | $table = $entryValue['table']; |
||
| 200 | $record = []; |
||
| 201 | $record['headline'] = BackendUtility::getRecordTitle($table, $entryValue['row']); |
||
| 202 | $record['record_pid'] = $entryValue['row']['pid']; |
||
| 203 | $record['record_uid'] = $entryValue['uid']; |
||
| 204 | $record['table_name'] = $table; |
||
| 205 | $record['link_type'] = $key; |
||
| 206 | $record['link_title'] = $entryValue['link_title'] ?? ''; |
||
| 207 | $record['field'] = $entryValue['field']; |
||
| 208 | $record['last_check'] = time(); |
||
| 209 | $typeField = $GLOBALS['TCA'][$table]['ctrl']['type'] ?? false; |
||
| 210 | if ($entryValue['row'][$typeField] ?? false) { |
||
| 211 | $record['element_type'] = $entryValue['row'][$typeField]; |
||
| 212 | } |
||
| 213 | $languageField = $GLOBALS['TCA'][$table]['ctrl']['languageField'] ?? false; |
||
| 214 | if ($languageField && isset($entryValue['row'][$languageField])) { |
||
| 215 | $record['language'] = $entryValue['row'][$languageField]; |
||
| 216 | } else { |
||
| 217 | $record['language'] = -1; |
||
| 218 | } |
||
| 219 | if (!empty($entryValue['pageAndAnchor'] ?? '')) { |
||
| 220 | // Page with anchor, e.g. 18#1580 |
||
| 221 | $url = $entryValue['pageAndAnchor']; |
||
| 222 | } else { |
||
| 223 | $url = $entryValue['substr']['tokenValue']; |
||
| 224 | } |
||
| 225 | $record['url'] = $url; |
||
| 226 | |||
| 227 | if (!($this->linkCounts[$table] ?? false)) { |
||
| 228 | $this->linkCounts[$table] = 0; |
||
| 229 | } |
||
| 230 | |||
| 231 | if (!($this->brokenLinkCounts[$table] ?? false)) { |
||
| 232 | $this->brokenLinkCounts[$table] = 0; |
||
| 233 | } |
||
| 234 | |||
| 235 | $this->linkCounts[$table]++; |
||
| 236 | $checkUrl = $hookObj->checkLink($url, $entryValue, $this); |
||
| 237 | |||
| 238 | // Broken link found |
||
| 239 | if (!$checkUrl) { |
||
| 240 | $this->brokenLinkRepository->addBrokenLink($record, false, $hookObj->getErrorParams()); |
||
| 241 | $this->brokenLinkCounts[$table]++; |
||
| 242 | } elseif (GeneralUtility::_GP('showalllinks')) { |
||
| 243 | $this->brokenLinkRepository->addBrokenLink($record, true); |
||
| 244 | $this->brokenLinkCounts[$table]++; |
||
| 245 | } |
||
| 246 | } |
||
| 247 | } |
||
| 248 | } |
||
| 249 | |||
| 250 | /** |
||
| 251 | * Recheck for broken links for one field in table for record. |
||
| 252 | * |
||
| 253 | * @param array $checkOptions |
||
| 254 | * @param string $recordUid uid of record to check |
||
| 255 | * @param string $table |
||
| 256 | * @param string $field |
||
| 257 | * @param int $timestamp - only recheck if timestamp changed |
||
| 258 | * @param bool $considerHidden |
||
| 259 | */ |
||
| 260 | public function recheckLinks( |
||
| 261 | array $checkOptions, |
||
| 262 | string $recordUid, |
||
| 263 | string $table, |
||
| 264 | string $field, |
||
| 265 | int $timestamp, |
||
| 266 | bool $considerHidden = true |
||
| 267 | ): void { |
||
| 268 | // If table is not configured, assume the extension is not installed |
||
| 269 | // and therefore no need to check it |
||
| 270 | if (!is_array($GLOBALS['TCA'][$table])) { |
||
| 271 | return; |
||
| 272 | } |
||
| 273 | |||
| 274 | // get all links for $record / $table / $field combination |
||
| 275 | $queryBuilder = GeneralUtility::makeInstance(ConnectionPool::class) |
||
| 276 | ->getQueryBuilderForTable($table); |
||
| 277 | if ($considerHidden) { |
||
| 278 | $queryBuilder->getRestrictions()->removeByType(HiddenRestriction::class); |
||
| 279 | } |
||
| 280 | |||
| 281 | $row = $queryBuilder->select('uid', 'pid', $GLOBALS['TCA'][$table]['ctrl']['label'], $field, 'tstamp') |
||
| 282 | ->from($table) |
||
| 283 | ->where( |
||
| 284 | $queryBuilder->expr()->eq( |
||
| 285 | 'uid', |
||
| 286 | $queryBuilder->createNamedParameter($recordUid, Connection::PARAM_INT) |
||
| 287 | ) |
||
| 288 | ) |
||
| 289 | ->execute() |
||
| 290 | ->fetch(); |
||
| 291 | |||
| 292 | if (!$row) { |
||
| 293 | // missing record: remove existing links |
||
| 294 | $this->brokenLinkRepository->removeBrokenLinksForRecord($table, $recordUid); |
||
|
Bug
introduced
by
|
|||
| 295 | return; |
||
| 296 | } |
||
| 297 | if (($row['tstamp'] ?? 0) && $timestamp && ((int)($row['tstamp']) < $timestamp)) { |
||
| 298 | // timestamp has not changed: no need to recheck |
||
| 299 | return; |
||
| 300 | } |
||
| 301 | $resultsLinks = []; |
||
| 302 | $this->analyzeRecord($resultsLinks, $table, [$field], $row); |
||
| 303 | if ($resultsLinks) { |
||
| 304 | // remove existing broken links from table |
||
| 305 | $this->brokenLinkRepository->removeBrokenLinksForRecord($table, $recordUid); |
||
| 306 | // find all broken links for list of links |
||
| 307 | $this->checkLinks($resultsLinks, $checkOptions); |
||
| 308 | } |
||
| 309 | } |
||
| 310 | |||
| 311 | /** |
||
| 312 | * Find all supported broken links for a specific record |
||
| 313 | * |
||
| 314 | * @param array $results Array of broken links |
||
| 315 | * @param string $table Table name of the record |
||
| 316 | * @param array $fields Array of fields to analyze |
||
| 317 | * @param array $record Record to analyze |
||
| 318 | */ |
||
| 319 | public function analyzeRecord(array &$results, $table, array $fields, array $record) |
||
| 320 | { |
||
| 321 | $event = new BeforeRecordIsAnalyzedEvent($table, $record, $fields, $this, $results); |
||
| 322 | $this->eventDispatcher->dispatch($event); |
||
| 323 | $results = $event->getResults(); |
||
| 324 | $record = $event->getRecord(); |
||
| 325 | |||
| 326 | // Put together content of all relevant fields |
||
| 327 | $htmlParser = GeneralUtility::makeInstance(HtmlParser::class); |
||
| 328 | $idRecord = $record['uid']; |
||
| 329 | // Get all references |
||
| 330 | foreach ($fields as $field) { |
||
| 331 | $conf = $GLOBALS['TCA'][$table]['columns'][$field]['config']; |
||
| 332 | $valueField = $record[$field]; |
||
| 333 | |||
| 334 | // Check if a TCA configured field has soft references defined (see TYPO3 Core API document) |
||
| 335 | if (!($conf['softref'] ?? false) || (string)$valueField === '') { |
||
| 336 | continue; |
||
| 337 | } |
||
| 338 | |||
| 339 | // Explode the list of soft references/parameters |
||
| 340 | $softRefs = BackendUtility::explodeSoftRefParserList($conf['softref']); |
||
| 341 | if ($softRefs === false) { |
||
| 342 | continue; |
||
| 343 | } |
||
| 344 | |||
| 345 | // Traverse soft references |
||
| 346 | foreach ($softRefs as $spKey => $spParams) { |
||
| 347 | /** @var \TYPO3\CMS\Core\Database\SoftReferenceIndex $softRefObj */ |
||
| 348 | $softRefObj = BackendUtility::softRefParserObj($spKey); |
||
| 349 | |||
| 350 | // If there is an object returned... |
||
| 351 | if (!is_object($softRefObj)) { |
||
| 352 | continue; |
||
| 353 | } |
||
| 354 | $softRefParams = $spParams; |
||
| 355 | if (!is_array($softRefParams)) { |
||
| 356 | // set subst such that findRef will return substitutes for urls, emails etc |
||
| 357 | $softRefParams = ['subst' => true]; |
||
| 358 | } |
||
| 359 | |||
| 360 | // Do processing |
||
| 361 | $resultArray = $softRefObj->findRef($table, $field, $idRecord, $valueField, $spKey, $softRefParams); |
||
| 362 | if (!is_array($resultArray) || !isset($resultArray['elements']) || !is_array($resultArray['elements'])) { |
||
| 363 | continue; |
||
| 364 | } |
||
| 365 | |||
| 366 | if ($spKey === 'typolink_tag') { |
||
| 367 | $this->analyzeTypoLinks($resultArray, $results, $htmlParser, $record, $field, $table); |
||
| 368 | } else { |
||
| 369 | $this->analyzeLinks($resultArray, $results, $record, $field, $table); |
||
| 370 | } |
||
| 371 | } |
||
| 372 | } |
||
| 373 | } |
||
| 374 | |||
| 375 | /** |
||
| 376 | * Find all supported broken links for a specific link list |
||
| 377 | * |
||
| 378 | * @param array $resultArray findRef parsed records |
||
| 379 | * @param array $results Array of broken links |
||
| 380 | * @param array $record UID of the current record |
||
| 381 | * @param string $field The current field |
||
| 382 | * @param string $table The current table |
||
| 383 | */ |
||
| 384 | protected function analyzeLinks(array $resultArray, array &$results, array $record, $field, $table) |
||
| 385 | { |
||
| 386 | foreach ($resultArray['elements'] as $element) { |
||
| 387 | $r = $element['subst']; |
||
| 388 | $type = ''; |
||
| 389 | $idRecord = $record['uid']; |
||
| 390 | if (empty($r)) { |
||
| 391 | continue; |
||
| 392 | } |
||
| 393 | |||
| 394 | foreach ($this->hookObjectsArr as $keyArr => $hookObj) { |
||
| 395 | $type = $hookObj->fetchType($r, $type, $keyArr); |
||
| 396 | // Store the type that was found |
||
| 397 | // This prevents overriding by internal validator |
||
| 398 | if (!empty($type)) { |
||
| 399 | $r['type'] = $type; |
||
| 400 | } |
||
| 401 | } |
||
| 402 | $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r['tokenID']]['substr'] = $r; |
||
| 403 | $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r['tokenID']]['row'] = $record; |
||
| 404 | $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r['tokenID']]['table'] = $table; |
||
| 405 | $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r['tokenID']]['field'] = $field; |
||
| 406 | $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r['tokenID']]['uid'] = $idRecord; |
||
| 407 | } |
||
| 408 | } |
||
| 409 | |||
| 410 | /** |
||
| 411 | * Find all supported broken links for a specific typoLink |
||
| 412 | * |
||
| 413 | * @param array $resultArray findRef parsed records |
||
| 414 | * @param array $results Array of broken links |
||
| 415 | * @param HtmlParser $htmlParser Instance of html parser |
||
| 416 | * @param array $record The current record |
||
| 417 | * @param string $field The current field |
||
| 418 | * @param string $table The current table |
||
| 419 | */ |
||
| 420 | protected function analyzeTypoLinks(array $resultArray, array &$results, $htmlParser, array $record, $field, $table) |
||
| 421 | { |
||
| 422 | $currentR = []; |
||
| 423 | $linkTags = $htmlParser->splitIntoBlock('a,link', $resultArray['content']); |
||
| 424 | $idRecord = $record['uid']; |
||
| 425 | $type = ''; |
||
| 426 | $title = ''; |
||
| 427 | $countLinkTags = count($linkTags); |
||
| 428 | for ($i = 1; $i < $countLinkTags; $i += 2) { |
||
| 429 | $referencedRecordType = ''; |
||
| 430 | foreach ($resultArray['elements'] as $element) { |
||
| 431 | $type = ''; |
||
| 432 | $r = $element['subst']; |
||
| 433 | if (empty($r['tokenID']) || substr_count($linkTags[$i], $r['tokenID']) === 0) { |
||
| 434 | continue; |
||
| 435 | } |
||
| 436 | |||
| 437 | // Type of referenced record |
||
| 438 | if (strpos($r['recordRef'] ?? '', 'pages') !== false) { |
||
| 439 | $currentR = $r; |
||
| 440 | // Contains number of the page |
||
| 441 | $referencedRecordType = $r['tokenValue']; |
||
| 442 | $wasPage = true; |
||
| 443 | } elseif (strpos($r['recordRef'] ?? '', 'tt_content') !== false && (isset($wasPage) && $wasPage === true)) { |
||
| 444 | $referencedRecordType = $referencedRecordType . '#c' . $r['tokenValue']; |
||
| 445 | $wasPage = false; |
||
| 446 | } else { |
||
| 447 | $currentR = $r; |
||
| 448 | } |
||
| 449 | $title = strip_tags($linkTags[$i]); |
||
| 450 | } |
||
| 451 | foreach ($this->hookObjectsArr as $keyArr => $hookObj) { |
||
| 452 | $type = $hookObj->fetchType($currentR, $type, $keyArr); |
||
| 453 | // Store the type that was found |
||
| 454 | // This prevents overriding by internal validator |
||
| 455 | if (!empty($type)) { |
||
| 456 | $currentR['type'] = $type; |
||
| 457 | } |
||
| 458 | } |
||
| 459 | $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR['tokenID']]['substr'] = $currentR; |
||
| 460 | $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR['tokenID']]['row'] = $record; |
||
| 461 | $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR['tokenID']]['table'] = $table; |
||
| 462 | $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR['tokenID']]['field'] = $field; |
||
| 463 | $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR['tokenID']]['uid'] = $idRecord; |
||
| 464 | $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR['tokenID']]['link_title'] = $title; |
||
| 465 | $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR['tokenID']]['pageAndAnchor'] = $referencedRecordType; |
||
| 466 | } |
||
| 467 | } |
||
| 468 | |||
| 469 | /** |
||
| 470 | * Fill a marker array with the number of links found in a list of pages |
||
| 471 | * |
||
| 472 | * @return array array with the number of links found |
||
| 473 | */ |
||
| 474 | public function getLinkCounts() |
||
| 475 | { |
||
| 476 | return $this->brokenLinkRepository->getNumberOfBrokenLinksForRecordsOnPages($this->pids, $this->searchFields); |
||
| 477 | } |
||
| 478 | |||
| 479 | /** |
||
| 480 | * @return LanguageService |
||
| 481 | */ |
||
| 482 | protected function getLanguageService() |
||
| 485 | } |
||
| 486 | } |
||
| 487 |
Loading history...