* Added Feature #12447 Clearly separate Checks from Report in module
[Packages/TYPO3.CMS.git] / 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 seperated 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 array
77 */
78 protected $hookObjectsArr = array();
79
80 /**
81 * Array with information about the current page.
82 *
83 * @var array
84 */
85 protected $extPageInTreeInfo = array();
86
87 /**
88 * Fill hookObjectsArr with different link types and possible XClasses.
89 */
90 public function __construct() {
91 // Hook to handle own checks
92 if (is_array($GLOBALS['TYPO3_CONF_VARS']['EXTCONF']['linkvalidator']['checkLinks'])) {
93 foreach ($GLOBALS['TYPO3_CONF_VARS']['EXTCONF']['linkvalidator']['checkLinks'] as $key => $classRef) {
94 $this->hookObjectsArr[$key] = &t3lib_div::getUserObj($classRef);
95 }
96 }
97 }
98
99 /**
100 * Init Function: Here all the needed configuration values are stored in class variables.
101 *
102 * @param array $searchField: list of fields in which to search for links
103 * @param string $pid: list of comma separated page uids in which to search for links
104 * @return void
105 */
106 public function init($searchField, $pid) {
107 $this->searchFields = $searchField;
108 $this->pidList = $pid;
109 }
110
111 /**
112 * Find all supported broken links and store them in tx_linkvalidator_link.
113 *
114 * @param array $checkOptions: list of hook object to activate
115 * @param boolean $considerHidden: defines whether to look into hidden fields or not
116 * @return void
117 */
118 public function getLinkStatistics($checkOptions = array(), $considerHidden = FALSE) {
119 $results = array();
120 $checlLinkTypeCondition = '';
121 if(count($checkOptions) > 0) {
122 $checkKeys = array_keys($checkOptions);
123 $checlLinkTypeCondition = ' and link_type in (\'' . implode('\',\'',$checkKeys) . '\')';
124 }
125 $GLOBALS['TYPO3_DB']->exec_DELETEquery('tx_linkvalidator_link', '(record_pid in (' . $this->pidList . ') or ( record_uid IN (' . $this->pidList . ') and table_name like \'pages\')) ' . $checlLinkTypeCondition);
126
127 // let's traverse all configured tables
128 foreach ($this->searchFields as $table => $fields) {
129 if($table == 'pages'){
130 $where = 'deleted = 0 AND uid IN (' . $this->pidList . ')';
131 }
132 else{
133 $where = 'deleted = 0 AND pid IN (' . $this->pidList . ')';
134 }
135 if (!$considerHidden) {
136 $where .= t3lib_BEfunc::BEenableFields($table);
137 }
138 // if table is not configured, we assume the ext is not installed and therefore no need to check it
139 if (!is_array($GLOBALS['TCA'][$table])) continue;
140
141 // re-init selectFields for table
142 $selectFields = 'uid, pid';
143 $selectFields .= ', ' . $GLOBALS['TCA'][$table]['ctrl']['label'] . ', ' . implode(', ', $fields);
144
145 // TODO: only select rows that have content in at least one of the relevant fields (via OR)
146 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery($selectFields, $table, $where);
147 // Get record rows of table
148 while ($row = $GLOBALS['TYPO3_DB']->sql_fetch_assoc($res)) {
149 // Analyse each record
150 $this->analyzeRecord($results, $table, $fields, $row);
151 }
152 }
153
154 foreach ($this->hookObjectsArr as $key => $hookObj) {
155 if ((is_array($results[$key])) && empty($checkOptions) || (is_array($results[$key]) && $checkOptions[$key])) {
156 // check'em!
157 foreach ($results[$key] as $entryKey => $entryValue) {
158 $table = $entryValue['table'];
159 $record = array();
160 $record['headline'] = $entryValue['row'][$GLOBALS['TCA'][$table]['ctrl']['label']];
161 $record['record_pid'] = $entryValue['row']['pid'];
162 $record['record_uid'] = $entryValue['uid'];
163 $record['table_name'] = $table;
164 $record['link_title'] = $entryValue['link_title'];
165 $record['field'] = $entryValue['field'];
166 $record['last_check'] = time();
167
168 $this->recordReference = $entryValue['substr']['recordRef'];
169
170 $this->pageWithAnchor = $entryValue['pageAndAnchor'];
171
172 if (!empty($this->pageWithAnchor)) {
173 // page with anchor, e.g. 18#1580
174 $url = $this->pageWithAnchor;
175 } else {
176 $url = $entryValue['substr']['tokenValue'];
177 }
178
179 $this->linkCounts[$table]++;
180 $checkURL = $hookObj->checkLink($url, $entryValue, $this);
181 // broken link found!
182 if (!$checkURL) {
183 $response = array();
184 $response['valid'] = FALSE;
185 $response['errorParams'] = $hookObj->getErrorParams();
186 $this->brokenLinkCounts[$table]++;
187 $record['link_type'] = $key;
188 $record['url'] = $url;
189 $record['url_response'] = serialize($response);
190 $GLOBALS['TYPO3_DB']->exec_INSERTquery('tx_linkvalidator_link', $record);
191 } elseif (t3lib_div::_GP('showalllinks')) {
192 $response = array();
193 $response['valid'] = TRUE;
194 $this->brokenLinkCounts[$table]++;
195 $record['url'] = $url;
196 $record['link_type'] = $key;
197 $record['url_response'] = serialize($response);
198 $GLOBALS['TYPO3_DB']->exec_INSERTquery('tx_linkvalidator_link', $record);
199 }
200 }
201 }
202 }
203 }
204
205
206 /**
207 * Find all supported broken links for a specific record.
208 *
209 * @param array $results: array of broken links
210 * @param string $table: table name of the record
211 * @param array $fields: array of fields to analyze
212 * @param array $record: record to analyse
213 * @return void
214 */
215 public function analyzeRecord(&$results, $table, $fields, $record) {
216
217 // array to store urls from relevant field contents
218 $urls = array();
219
220 $referencedRecordType = '';
221 // last-parsed link element was a page.
222 $wasPage = TRUE;
223
224 // flag whether row contains a broken link in some field or not
225 $rowContainsBrokenLink = FALSE;
226
227 // put together content of all relevant fields
228 $haystack = '';
229 $htmlParser = t3lib_div::makeInstance('t3lib_parsehtml');
230
231 $idRecord = $record['uid'];
232
233 // get all references
234 foreach ($fields as $field) {
235 $haystack .= $record[$field] . ' --- ';
236 $conf = $GLOBALS['TCA'][$table]['columns'][$field]['config'];
237
238 $valueField = $record[$field];
239
240 // Check if a TCA configured field has softreferences defined (see TYPO3 Core API document)
241 if ($conf['softref'] && strlen($valueField)) {
242 // Explode the list of softreferences/parameters
243 $softRefs = t3lib_BEfunc::explodeSoftRefParserList($conf['softref']);
244 // Traverse soft references
245 foreach ($softRefs as $spKey => $spParams) {
246 // create / get object
247 $softRefObj = &t3lib_BEfunc::softRefParserObj($spKey);
248
249 // If there was an object returned...:
250 if (is_object($softRefObj)) {
251
252 // Do processing
253 $resultArray = $softRefObj->findRef($table, $field, $idRecord, $valueField, $spKey, $spParams);
254 if (!empty($resultArray['elements'])) {
255
256 if ($spKey == 'typolink_tag') {
257 $linkTags = $htmlParser->splitIntoBlock('link', $resultArray['content']);
258
259 for ($i = 1; $i < count($linkTags); $i += 2) {
260 $referencedRecordType = '';
261 foreach($resultArray['elements'] as $element) {
262 $type = '';
263 $r = $element['subst'];
264
265 if (!empty($r['tokenID'])) {
266 if (substr_count($linkTags[$i], $r['tokenID'])) {
267 // Type of referenced record
268 if (strpos($r['recordRef'], 'pages') !== FALSE) {
269 $currentR = $r;
270 // contains number of the page
271 $referencedRecordType = $r['tokenValue'];
272 $wasPage = TRUE;
273 }
274 // append number of content element to the page saved in the last loop
275 elseif ((strpos($r['recordRef'], 'tt_content') !== FALSE) && ($wasPage === TRUE)) {
276 $referencedRecordType = $referencedRecordType . '#c' . $r['tokenValue'];
277 $wasPage = FALSE;
278 } else {
279 $currentR = $r;
280 }
281 $title = strip_tags($linkTags[$i]);
282 }
283 }
284 }
285 foreach ($this->hookObjectsArr as $keyArr => $hookObj) {
286 $type = $hookObj->fetchType($currentR, $type, $keyArr);
287 }
288
289 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR["tokenID"]]["substr"] = $currentR;
290 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR["tokenID"]]["row"] = $record;
291 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR["tokenID"]]["table"] = $table;
292 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR["tokenID"]]["field"] = $field;
293 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR["tokenID"]]["uid"] = $idRecord;
294 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR["tokenID"]]["link_title"] = $title;
295 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $currentR["tokenID"]]["pageAndAnchor"] = $referencedRecordType;
296
297 }
298 } else {
299
300 foreach ($resultArray['elements'] as $element) {
301 $r = $element['subst'];
302 $title = '';
303 $type = '';
304
305 if (!empty($r)) {
306 // Parse string for special TYPO3 <link> tag:
307
308 foreach ($this->hookObjectsArr as $keyArr => $hookObj) {
309 $type = $hookObj->fetchType($r, $type, $keyArr);
310 }
311 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r["tokenID"]]["substr"] = $r;
312 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r["tokenID"]]["row"] = $record;
313 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r["tokenID"]]["table"] = $table;
314 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r["tokenID"]]["field"] = $field;
315 $results[$type][$table . ':' . $field . ':' . $idRecord . ':' . $r["tokenID"]]["uid"] = $idRecord;
316
317 }
318 }
319 }
320 }
321 }
322 }
323 }
324 }
325 }
326
327 /**
328 * Fill a markerarray with the number of links found in a list of pages.
329 *
330 * @param string $curPage: comma separated list of page uids
331 * @return array markerarray with the number of links found
332 */
333 public function getLinkCounts($curPage) {
334 $markerArray = array();
335 if (($res = $GLOBALS['TYPO3_DB']->exec_SELECTquery(
336 'count(uid) as nbBrokenLinks,link_type',
337 'tx_linkvalidator_link',
338 'record_pid in (' . $this->pidList . ')',
339 'link_type'
340 ))) {
341 while (($row = $GLOBALS['TYPO3_DB']->sql_fetch_assoc($res))) {
342 $markerArray[$row['link_type']] = $row['nbBrokenLinks'];
343 $markerArray['brokenlinkCount'] += $row['nbBrokenLinks'];
344 }
345 }
346 return $markerArray;
347 }
348
349 /**
350 * Calls t3lib_tsfeBeUserAuth::extGetTreeList.
351 * Although this duplicates the function t3lib_tsfeBeUserAuth::extGetTreeList
352 * this is necessary to create the object that is used recursively by the original function.
353 *
354 * Generates a list of page uids from $id. List does not include $id itself.
355 * The only pages excluded from the list are deleted pages.
356 *
357 * level in the tree to start collecting uids. Zero means
358 * 'start right away', 1 = 'next level and out'
359 *
360 * @param integer Start page id
361 * @param integer Depth to traverse down the page tree.
362 * @param integer $begin is an optional integer that determines at which
363 * @param string Perms clause
364 * @return string Returns the list with a comma in the end (if any pages selected!)
365 */
366 public function extGetTreeList($id, $depth, $begin = 0, $permsClause) {
367 $depth = intval($depth);
368 $begin = intval($begin);
369 $id = intval($id);
370 $theList = '';
371
372 if ($depth > 0) {
373 $res = $GLOBALS['TYPO3_DB']->exec_SELECTquery(
374 'uid,title',
375 'pages',
376 'pid=' . $id . ' AND deleted=0 AND ' . $permsClause
377 );
378 while ($row = $GLOBALS['TYPO3_DB']->sql_fetch_assoc($res)) {
379 if ($begin <= 0) {
380 $theList .= $row['uid'] . ',';
381 $this->extPageInTreeInfo[] = array($row['uid'], htmlspecialchars($row['title'], $depth));
382 }
383 if ($depth > 1) {
384 $theList .= $this->extGetTreeList($row['uid'], $depth - 1, $begin - 1, $permsClause);
385 }
386 }
387 }
388 return $theList;
389 }
390
391
392 }
393
394 if (defined('TYPO3_MODE') && isset($GLOBALS['TYPO3_CONF_VARS'][TYPO3_MODE]['XCLASS']['ext/linkvalidator/classes/class.tx_linkvalidator_processor.php'])) {
395 include_once($GLOBALS['TYPO3_CONF_VARS'][TYPO3_MODE]['XCLASS']['ext/linkvalidator/classes/class.tx_linkvalidator_processor.php']);
396 }
397 ?>