typo3/sysext/linkvalidator/classes/class.tx_linkvalidator_processor.php

   1 <?php
   2 /***************************************************************
   3  *  Copyright notice
   4  *
   5  *  (c) 2010 - 2011 Michael Miousse (michael.miousse@infoglobe.ca)
   6  *  All rights reserved
   7  *
   8  *  This script is part of the TYPO3 project. The TYPO3 project is
   9  *  free software; you can redistribute it and/or modify
  10  *  it under the terms of the GNU General Public License as published by
  11  *  the Free Software Foundation; either version 2 of the License, or
  12  *  (at your option) any later version.
  13  *
  14  *  The GNU General Public License can be found at
  15  *  http://www.gnu.org/copyleft/gpl.html.
  16  *
  17  *  This script is distributed in the hope that it will be useful,
  18  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  19  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  20  *  GNU General Public License for more details.
  21  *
  22  *  This copyright notice MUST APPEAR in all copies of the script!
  23  ***************************************************************/
  24
  25 /**
  26  * This class provides Processing plugin implementation
  27  *
  28  * @author Michael Miousse <michael.miousse@infoglobe.ca>
  29  * @author Jochen Rieger <j.rieger@connecta.ag>
  30  * @package TYPO3
  31  * @subpackage linkvalidator
  32  */
  33
  34 $GLOBALS['LANG']->includeLLFile('EXT:linkvalidator/modfuncreport/locallang.xml');
  35
  36 class tx_linkvalidator_Processor {
  37
  38         /**
  39          * Array of tables and fields to search for broken links
  40          *
  41          * @var array
  42          */
  43         protected $searchFields = array();
  44
  45         /**
  46          * List of comma separated page uids (rootline downwards)
  47          *
  48          * @var string
  49          */
  50         protected $pidList = '';
  51
  52         /**
  53          * Array of tables and the number of external links they contain
  54          *
  55          * @var array
  56          */
  57         protected $linkCounts = array();
  58
  59         /**
  60          * Array of tables and the number of broken external links they contain
  61          *
  62          * @var array
  63          */
  64         protected $brokenLinkCounts = array();
  65
  66         /**
  67          * Array of tables and records containing broken links
  68          *
  69          * @var array
  70          */
  71         protected $recordsWithBrokenLinks = array();
  72
  73         /**
  74          * Array for hooks for own checks
  75          *
  76          * @var tx_linkvalidator_linktype_Abstract[]
  77          */
  78         protected $hookObjectsArr = array();
  79
  80         /**
  81          * Array with information about the current page
  82          *
  83          * @var array
  84          */
  85         protected $extPageInTreeInfo = array();
  86
  87         /**
  88          * Reference to the current element with table:uid, e.g. pages:85
  89          *
  90          * @var string
  91          */
  92         protected $recordReference = '';
  93
  94         /**
  95          * Linked page together with a possible anchor, e.g. 85#c105
  96          *
  97          * @var string
  98          */
  99         protected $pageWithAnchor = '';
 100
 101         /**
 102          * Fill hookObjectsArr with different link types and possible XClasses.
 103          */
 104         public function __construct() {
 105                         // Hook to handle own checks
 106                 if (is_array($GLOBALS['TYPO3_CONF_VARS']['EXTCONF']['linkvalidator']['checkLinks'])) {
 107                         foreach ($GLOBALS['TYPO3_CONF_VARS']['EXTCONF']['linkvalidator']['checkLinks'] as $key => $classRef) {
 108                                 $this->hookObjectsArr[$key] = t3lib_div::getUserObj($classRef);
 109                         }
 110                 }
 111         }
 112
 113         /**
 114          * Store all the needed configuration values in class variables
 115          *
 116          * @param array $searchField List of fields in which to search for links
 117          * @param string $pid List of comma separated page uids in which to search for links
 118          * @return      void
 119          */
 120         public function init(array $searchField, $pid) {
 121                 $this->searchFields = $searchField;
 122                 $this->pidList = $pid;
 123
 124                 foreach ($searchField as $tableName => $table) {
 125                         t3lib_div::loadTCA($tableName);
 126                 }
 127         }
 128
 129         /**
 130          * Find all supported broken links and store them in tx_linkvalidator_link
 131          *
 132          * @param array $checkOptions List of hook object to activate
 133          * @param boolean $considerHidden Defines whether to look into hidden fields or not
 134          * @return void
 135          */
 136         public function getLinkStatistics($checkOptions = array(), $considerHidden = FALSE) {
 137                 $results = array();
 138                 if (count($checkOptions) > 0) {
 139                         $checkKeys = array_keys($checkOptions);
 140                         $checkLinkTypeCondition = ' and link_type in (\'' . implode('\',\'', $checkKeys) . '\')';
 141
 142                         $GLOBALS['TYPO3_DB']->exec_DELETEquery('tx_linkvalidator_link',
 143                                                                         '(record_pid in (' . $this->pidList . ')'
 144                                                                         . ' or ( record_uid IN (' . $this->pidList . ') and table_name like \'pages\')) '
 145                                                                         . $checkLinkTypeCondition);
 146
 147                                 // Traverse all configured tables
 148                         foreach ($this->searchFields as $table => $fields) {
 149                                 if ($table === 'pages') {
 150                                         $where = 'deleted = 0 AND uid IN (' . $this->pidList . ')';
 151                                 } else {
 152                                         $where = 'deleted = 0 AND pid IN (' . $this->pidList . ')';
 153                                 }
 154                                 if (!$considerHidden) {
 155                                         $where .= t3lib_BEfunc::BEenableFields($table);
 156                                 }
 157                                         // If table is not configured, assume the extension is not installed and therefore no need to check it
 158                                 if (!is_array($GLOBALS['TCA'][$table])) continue;
 159
 160                                         // Re-init selectFields for table
 161                                 $selectFields = 'uid, pid';
 162                                 $selectFields .= ', ' . $GLOBALS['TCA'][$table]['ctrl']['label'] . ', ' . implode(', ', $fields);
 163
 164                                         // TODO: only select rows that have content in at least one of the relevant fields (via OR)
 165                                 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery($selectFields, $table, $where);
 166                                         // Get record rows of table
 167                                 while (($row = $GLOBALS['TYPO3_DB']->sql_fetch_assoc($res)) !== FALSE) {
 168                                                 // Analyse each record
 169                                         $this->analyzeRecord($results, $table, $fields, $row);
 170                                 }
 171                                 $GLOBALS['TYPO3_DB']->sql_free_result($res);
 172                         }
 173
 174                         foreach ($this->hookObjectsArr as $key => $hookObj) {
 175                                 if ((is_array($results[$key])) && empty($checkOptions) || (is_array($results[$key]) && $checkOptions[$key])) {
 176                                                 //  Check them
 177                                         foreach ($results[$key] as $entryKey => $entryValue) {
 178                                                 $table = $entryValue['table'];
 179                                                 $record = array();
 180                                                 $record['headline'] = $entryValue['row'][$GLOBALS['TCA'][$table]['ctrl']['label']];
 181                                                 $record['record_pid'] = $entryValue['row']['pid'];
 182                                                 $record['record_uid'] = $entryValue['uid'];
 183                                                 $record['table_name'] = $table;
 184                                                 $record['link_title'] = $entryValue['link_title'];
 185                                                 $record['field'] = $entryValue['field'];
 186                                                 $record['last_check'] = time();
 187
 188                                                 $this->recordReference = $entryValue['substr']['recordRef'];
 189
 190                                                 $this->pageWithAnchor = $entryValue['pageAndAnchor'];
 191
 192                                                 if (!empty($this->pageWithAnchor)) {
 193                                                                 // Page with anchor, e.g. 18#1580
 194                                                         $url = $this->pageWithAnchor;
 195                                                 } else {
 196                                                         $url = $entryValue['substr']['tokenValue'];
 197                                                 }
 198
 199                                                 $this->linkCounts[$table]++;
 200                                                 $checkURL = $hookObj->checkLink($url, $entryValue, $this);
 201                                                         // Broken link found
 202                                                 if (!$checkURL) {
 203                                                         $response = array();
 204                                                         $response['valid'] = FALSE;
 205                                                         $response['errorParams'] = $hookObj->getErrorParams();
 206                                                         $this->brokenLinkCounts[$table]++;
 207                                                         $record['link_type'] = $key;
 208                                                         $record['url'] = $url;
 209                                                         $record['url_response'] = serialize($response);
 210                                                         $GLOBALS['TYPO3_DB']->exec_INSERTquery('tx_linkvalidator_link', $record);
 211                                                 } elseif (t3lib_div::_GP('showalllinks')) {
 212                                                         $response = array();
 213                                                         $response['valid'] = TRUE;
 214                                                         $this->brokenLinkCounts[$table]++;
 215                                                         $record['url'] = $url;
 216                                                         $record['link_type'] = $key;
 217                                                         $record['url_response'] = serialize($response);
 218                                                         $GLOBALS['TYPO3_DB']->exec_INSERTquery('tx_linkvalidator_link', $record);
 219                                                 }
 220                                         }
 221                                 }
 222                         }
 223                 }
 224         }
 225
 226
 227         /**
 228          * Find all supported broken links for a specific record
 229          *
 230          * @param array $results Array of broken links
 231          * @param string $table Table name of the record
 232          * @param array $fields Array of fields to analyze
 233          * @param array $record Record to analyse
 234          * @return void
 235          */
 236         public function analyzeRecord(array &$results, $table, array $fields, array $record) {
 237
 238                         // Put together content of all relevant fields
 239                 $haystack = '';
 240                         /** @var t3lib_parsehtml $htmlParser */
 241                 $htmlParser = t3lib_div::makeInstance('t3lib_parsehtml');
 242
 243                 $idRecord = $record['uid'];
 244
 245                         // Get all references
 246                 foreach ($fields as $field) {
 247                         $haystack .= $record[$field] . ' --- ';
 248                         $conf = $GLOBALS['TCA'][$table]['columns'][$field]['config'];
 249
 250                         $valueField = $record[$field];
 251
 252                                 // Check if a TCA configured field has soft references defined (see TYPO3 Core API document)
 253                         if ($conf['softref'] && strlen($valueField)) {
 254                                         // Explode the list of soft references/parameters
 255                                 $softRefs = t3lib_BEfunc::explodeSoftRefParserList($conf['softref']);
 256                                         // Traverse soft references
 257                                 foreach ($softRefs as $spKey => $spParams) {
 258                                                 /** @var t3lib_softrefproc $softRefObj Create or get the soft reference object */
 259                                         $softRefObj = &t3lib_BEfunc::softRefParserObj($spKey);
 260
 261                                                 // If there is an object returned...
 262                                         if (is_object($softRefObj)) {
 263
 264                                                         // Do processing
 265                                                 $resultArray = $softRefObj->findRef($table, $field, $idRecord, $valueField, $spKey, $spParams);
 266                                                 if (!empty($resultArray['elements'])) {
 267
 268                                                         if ($spKey == 'typolink_tag') {
 269                                                                 $this->analyseTypoLinks($resultArray, $results, $htmlParser, $record, $field, $table);
 270                                                         } else {
 271                                                                 $this->analyseLinks($resultArray, $results, $record, $field, $table);
 272                                                         }
 273                                                 }
 274                                         }
 275                                 }
 276                         }
 277                 }
 278         }
 279
 280         /**
 281          * Find all supported broken links for a specific link list
 282          *
 283          * @param array $resultArray findRef parsed records
 284          * @param array $results Array of broken links
 285          * @param array $record UID of the current record
 286          * @param string $field The current field
 287          * @param string $table The current table
 288          * @return      void
 289          */
 290         protected function analyseLinks(array $resultArray, array &$results, array $record, $field, $table) {
 291                 foreach ($resultArray['elements'] as $element) {
 292                         $r = $element['subst'];
 293                         $type = '';
 294                         $idRecord = $record['uid'];
 295                         if (!empty($r)) {
 296                                         /** @var tx_linkvalidator_linktype_Abstract $hookObj */
 297                                 foreach ($this->hookObjectsArr as $keyArr => $hookObj) {
 298                                         $type = $hookObj->fetchType($r, $type, $keyArr);
 299                                                 // Store the type that was found
 300                                                 // This prevents overriding by internal validator
 301                                         if (!empty($type)) {
 302                                                 $r['type'] = $type;
 303                                         }
 304                                 }
 305                                 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r["tokenID"]]["substr"] = $r;
 306                                 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r["tokenID"]]["row"] = $record;
 307                                 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r["tokenID"]]["table"] = $table;
 308                                 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r["tokenID"]]["field"] = $field;
 309                                 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r["tokenID"]]["uid"] = $idRecord;
 310
 311                         }
 312                 }
 313         }
 314
 315         /**
 316          * Find all supported broken links for a specific typoLink
 317          *
 318          * @param array $resultArray findRef parsed records
 319          * @param array $results Array of broken links
 320          * @param t3lib_parsehtml $htmlParser Instance of html parser
 321          * @param array $record The current record
 322          * @param string $field The current field
 323          * @param string $table The current table
 324          * @return void
 325          */
 326         protected function analyseTypoLinks(array $resultArray, array &$results, $htmlParser, array $record, $field, $table) {
 327                 $currentR = array();
 328                 $linkTags = $htmlParser->splitIntoBlock('link', $resultArray['content']);
 329                 $idRecord = $record['uid'];
 330                 $type = '';
 331                 $title = '';
 332                 for ($i = 1; $i < count($linkTags); $i += 2) {
 333                         $referencedRecordType = '';
 334                         foreach ($resultArray['elements'] as $element) {
 335                                 $type = '';
 336                                 $r = $element['subst'];
 337
 338                                 if (!empty($r['tokenID'])) {
 339                                         if (substr_count($linkTags[$i], $r['tokenID'])) {
 340                                                         // Type of referenced record
 341                                                 if (strpos($r['recordRef'], 'pages') !== FALSE) {
 342                                                         $currentR = $r;
 343                                                                 // Contains number of the page
 344                                                         $referencedRecordType = $r['tokenValue'];
 345                                                         $wasPage = TRUE;
 346
 347                                                         // Append number of content element to the page saved in the last loop
 348                                                 } elseif ((strpos($r['recordRef'], 'tt_content') !== FALSE) && (isset($wasPage) && $wasPage === TRUE)) {
 349                                                         $referencedRecordType = $referencedRecordType . '#c' . $r['tokenValue'];
 350                                                         $wasPage = FALSE;
 351                                                 } else {
 352                                                         $currentR = $r;
 353                                                 }
 354                                                 $title = strip_tags($linkTags[$i]);
 355                                         }
 356                                 }
 357                         }
 358                                 /** @var tx_linkvalidator_linktype_Abstract $hookObj */
 359                         foreach ($this->hookObjectsArr as $keyArr => $hookObj) {
 360                                 $type = $hookObj->fetchType($currentR, $type, $keyArr);
 361                                         // Store the type that was found
 362                                         // This prevents overriding by internal validator
 363                                 if (!empty($type)) {
 364                                         $currentR['type'] = $type;
 365                                 }
 366                         }
 367
 368                         $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR["tokenID"]]["substr"] = $currentR;
 369                         $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR["tokenID"]]["row"] = $record;
 370                         $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR["tokenID"]]["table"] = $table;
 371                         $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR["tokenID"]]["field"] = $field;
 372                         $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR["tokenID"]]["uid"] = $idRecord;
 373                         $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR["tokenID"]]["link_title"] = $title;
 374                         $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR["tokenID"]]["pageAndAnchor"] = $referencedRecordType;
 375
 376                 }
 377         }
 378
 379         /**
 380          * Fill a marker array with the number of links found in a list of pages
 381          *
 382          * @param string $curPage Comma separated list of page uids
 383          * @return array Marker array with the number of links found
 384          */
 385         public function getLinkCounts($curPage) {
 386                 $markerArray = array();
 387
 388                 if (empty($this->pidList)) {
 389                         $this->pidList = $curPage;
 390                 }
 391
 392                 if (($res = $GLOBALS['TYPO3_DB']->exec_SELECTquery(
 393                         'count(uid) as nbBrokenLinks,link_type',
 394                         'tx_linkvalidator_link',
 395                         'record_pid in (' . $this->pidList . ')',
 396                         'link_type'
 397                 ))
 398                 ) {
 399                         while (($row = $GLOBALS['TYPO3_DB']->sql_fetch_assoc($res)) !== FALSE) {
 400                                 $markerArray[$row['link_type']] = $row['nbBrokenLinks'];
 401                                 $markerArray['brokenlinkCount'] += $row['nbBrokenLinks'];
 402                         }
 403                 }
 404                 $GLOBALS['TYPO3_DB']->sql_free_result($res);
 405                 return $markerArray;
 406         }
 407
 408         /**
 409          * Calls t3lib_tsfeBeUserAuth::extGetTreeList.
 410          * Although this duplicates the function t3lib_tsfeBeUserAuth::extGetTreeList
 411          * this is necessary to create the object that is used recursively by the original function.
 412          *
 413          * Generates a list of page uids from $id. List does not include $id itself.
 414          * The only pages excluded from the list are deleted pages.
 415          *
 416          * @param integer $id Start page id
 417          * @param integer $depth Depth to traverse down the page tree.
 418          * @param integer $begin is an optional integer that determines at which
 419          * @param string $permsClause Perms clause
 420          * @param boolean $considerHidden Whether to consider hidden pages or not
 421          * @return string Returns the list with a comma in the end (if any pages selected!)
 422          */
 423         public function extGetTreeList($id, $depth, $begin = 0, $permsClause, $considerHidden = FALSE) {
 424                 $depth = intval($depth);
 425                 $begin = intval($begin);
 426                 $id = intval($id);
 427                 $theList = '';
 428
 429                 if ($depth > 0) {
 430                         $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery(
 431                                 'uid,title,hidden,extendToSubpages',
 432                                 'pages',
 433                                 'pid=' . $id . ' AND deleted=0 AND ' . $permsClause
 434                         );
 435
 436                         while (($row = $GLOBALS['TYPO3_DB']->sql_fetch_assoc($res)) !== FALSE) {
 437                                 if ($begin <= 0 && ($row['hidden'] == 0 || $considerHidden == 1)) {
 438                                         $theList .= $row['uid'] . ',';
 439                                         $this->extPageInTreeInfo[] = array($row['uid'], htmlspecialchars($row['title'], $depth));
 440                                 }
 441                                 if ($depth > 1 && (!($row['hidden'] == 1 && $row['extendToSubpages'] == 1) || $considerHidden == 1)) {
 442                                         $theList .= $this->extGetTreeList($row['uid'], $depth - 1, $begin - 1, $permsClause, $considerHidden);
 443                                 }
 444                         }
 445                         $GLOBALS['TYPO3_DB']->sql_free_result($res);
 446                 }
 447                 return $theList;
 448         }
 449
 450         /**
 451          * @param array $pageInfo Array with uid, title, hidden, extendToSubpages from pages table
 452          * @return boolean TRUE if rootline contains a hidden page, FALSE if not
 453          */
 454         public function getRootLineIsHidden(array $pageInfo) {
 455                 $hidden = FALSE;
 456                 if ($pageInfo['extendToSubpages'] == 1 && $pageInfo['hidden'] == 1) {
 457                         $hidden = TRUE;
 458                 } else {
 459                         if ($pageInfo['pid'] > 0) {
 460                                 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery(
 461                                         'uid,title,hidden,extendToSubpages',
 462                                         'pages',
 463                                         'uid=' . $pageInfo['pid']
 464                                 );
 465
 466                                 while (($row = $GLOBALS['TYPO3_DB']->sql_fetch_assoc($res)) !== FALSE) {
 467                                         $hidden = $this->getRootLineIsHidden($row);
 468                                 }
 469                                 $GLOBALS['TYPO3_DB']->sql_free_result($res);
 470                         } else {
 471                                 $hidden = FALSE;
 472                         }
 473                 }
 474                 return $hidden;
 475
 476         }
 477
 478 }
 479
 480 if (defined('TYPO3_MODE') && isset($GLOBALS['TYPO3_CONF_VARS'][TYPO3_MODE]['XCLASS']['ext/linkvalidator/classes/class.tx_linkvalidator_processor.php'])) {
 481         include_once($GLOBALS['TYPO3_CONF_VARS'][TYPO3_MODE]['XCLASS']['ext/linkvalidator/classes/class.tx_linkvalidator_processor.php']);
 482 }
 483 ?>