3 * Rebuild recent changes from scratch. This takes several hours,
4 * depending on the database size and server configuration.
6 * This program is free software; you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License as published by
8 * the Free Software Foundation; either version 2 of the License, or
9 * (at your option) any later version.
11 * This program is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 * GNU General Public License for more details.
16 * You should have received a copy of the GNU General Public License along
17 * with this program; if not, write to the Free Software Foundation, Inc.,
18 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
19 * http://www.gnu.org/copyleft/gpl.html
22 * @ingroup Maintenance
26 require_once __DIR__
. '/Maintenance.php';
27 use MediaWiki\MediaWikiServices
;
30 * Maintenance script that rebuilds recent changes from scratch.
32 * @ingroup Maintenance
34 class RebuildRecentchanges
extends Maintenance
{
35 /** @var int UNIX timestamp */
37 /** @var int UNIX timestamp */
40 public function __construct() {
41 parent
::__construct();
42 $this->addDescription( 'Rebuild recent changes' );
46 "Only rebuild rows in requested time range (in YYYYMMDDHHMMSS format)",
52 "Only rebuild rows in requested time range (in YYYYMMDDHHMMSS format)",
56 $this->setBatchSize( 200 );
59 public function execute() {
61 ( $this->hasOption( 'from' ) && !$this->hasOption( 'to' ) ) ||
62 ( !$this->hasOption( 'from' ) && $this->hasOption( 'to' ) )
64 $this->fatalError( "Both 'from' and 'to' must be given, or neither" );
67 $this->rebuildRecentChangesTablePass1();
68 $this->rebuildRecentChangesTablePass2();
69 $this->rebuildRecentChangesTablePass3();
70 $this->rebuildRecentChangesTablePass4();
71 $this->rebuildRecentChangesTablePass5();
72 if ( !( $this->hasOption( 'from' ) && $this->hasOption( 'to' ) ) ) {
75 $this->output( "Done.\n" );
79 * Rebuild pass 1: Insert `recentchanges` entries for page revisions.
81 private function rebuildRecentChangesTablePass1() {
82 $dbw = $this->getDB( DB_MASTER
);
83 $commentStore = CommentStore
::getStore();
85 if ( $this->hasOption( 'from' ) && $this->hasOption( 'to' ) ) {
86 $this->cutoffFrom
= wfTimestamp( TS_UNIX
, $this->getOption( 'from' ) );
87 $this->cutoffTo
= wfTimestamp( TS_UNIX
, $this->getOption( 'to' ) );
89 $sec = $this->cutoffTo
- $this->cutoffFrom
;
90 $days = $sec / 24 / 3600;
91 $this->output( "Rebuilding range of $sec seconds ($days days)\n" );
95 $days = $wgRCMaxAge / 24 / 3600;
96 $this->output( "Rebuilding \$wgRCMaxAge=$wgRCMaxAge seconds ($days days)\n" );
98 $this->cutoffFrom
= time() - $wgRCMaxAge;
99 $this->cutoffTo
= time();
102 $this->output( "Clearing recentchanges table for time range...\n" );
103 $rcids = $dbw->selectFieldValues(
107 'rc_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom
) ),
108 'rc_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo
) )
111 foreach ( array_chunk( $rcids, $this->getBatchSize() ) as $rcidBatch ) {
112 $dbw->delete( 'recentchanges', [ 'rc_id' => $rcidBatch ], __METHOD__
);
113 wfGetLBFactory()->waitForReplication();
116 $this->output( "Loading from page and revision tables...\n" );
118 $commentQuery = $commentStore->getJoin( 'rev_comment' );
119 $actorQuery = ActorMigration
::newMigration()->getJoin( 'rev_user' );
121 [ 'revision', 'page' ] +
$commentQuery['tables'] +
$actorQuery['tables'],
131 ] +
$commentQuery['fields'] +
$actorQuery['fields'],
133 'rev_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom
) ),
134 'rev_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo
) )
137 [ 'ORDER BY' => 'rev_timestamp DESC' ],
139 'page' => [ 'JOIN', 'rev_page=page_id' ],
140 ] +
$commentQuery['joins'] +
$actorQuery['joins']
143 $this->output( "Inserting from page and revision tables...\n" );
145 $actorMigration = ActorMigration
::newMigration();
146 foreach ( $res as $row ) {
147 $comment = $commentStore->getComment( 'rev_comment', $row );
148 $user = User
::newFromAnyId( $row->rev_user
, $row->rev_user_text
, $row->rev_actor
);
152 'rc_timestamp' => $row->rev_timestamp
,
153 'rc_namespace' => $row->page_namespace
,
154 'rc_title' => $row->page_title
,
155 'rc_minor' => $row->rev_minor_edit
,
157 'rc_new' => $row->page_is_new
,
158 'rc_cur_id' => $row->page_id
,
159 'rc_this_oldid' => $row->rev_id
,
160 'rc_last_oldid' => 0, // is this ok?
161 'rc_type' => $row->page_is_new ? RC_NEW
: RC_EDIT
,
162 'rc_source' => $row->page_is_new ? RecentChange
::SRC_NEW
: RecentChange
::SRC_EDIT
,
163 'rc_deleted' => $row->rev_deleted
164 ] +
$commentStore->insert( $dbw, 'rc_comment', $comment )
165 +
$actorMigration->getInsertValues( $dbw, 'rc_user', $user ),
168 if ( ( ++
$inserted %
$this->getBatchSize() ) == 0 ) {
169 wfGetLBFactory()->waitForReplication();
175 * Rebuild pass 2: Enhance entries for page revisions with references to the previous revision
176 * (rc_last_oldid, rc_new etc.) and size differences (rc_old_len, rc_new_len).
178 private function rebuildRecentChangesTablePass2() {
179 $dbw = $this->getDB( DB_MASTER
);
181 $this->output( "Updating links and size differences...\n" );
183 # Fill in the rc_last_oldid field, which points to the previous edit
186 [ 'rc_cur_id', 'rc_this_oldid', 'rc_timestamp' ],
188 "rc_timestamp > " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom
) ),
189 "rc_timestamp < " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo
) )
192 [ 'ORDER BY' => 'rc_cur_id,rc_timestamp' ]
199 foreach ( $res as $obj ) {
202 if ( $obj->rc_cur_id
!= $lastCurId ) {
203 # Switch! Look up the previous last edit, if any
204 $lastCurId = intval( $obj->rc_cur_id
);
205 $emit = $obj->rc_timestamp
;
207 $row = $dbw->selectRow(
209 [ 'rev_id', 'rev_len' ],
210 [ 'rev_page' => $lastCurId, "rev_timestamp < " . $dbw->addQuotes( $emit ) ],
212 [ 'ORDER BY' => 'rev_timestamp DESC' ]
215 $lastOldId = intval( $row->rev_id
);
216 # Grab the last text size if available
217 $lastSize = !is_null( $row->rev_len
) ?
intval( $row->rev_len
) : null;
222 $new = 1; // probably true
226 if ( $lastCurId == 0 ) {
227 $this->output( "Uhhh, something wrong? No curid\n" );
229 # Grab the entry's text size
230 $size = (int)$dbw->selectField(
233 [ 'rev_id' => $obj->rc_this_oldid
],
240 'rc_last_oldid' => $lastOldId,
242 'rc_type' => $new ? RC_NEW
: RC_EDIT
,
243 'rc_source' => $new === 1 ? RecentChange
::SRC_NEW
: RecentChange
::SRC_EDIT
,
244 'rc_old_len' => $lastSize,
245 'rc_new_len' => $size,
248 'rc_cur_id' => $lastCurId,
249 'rc_this_oldid' => $obj->rc_this_oldid
,
250 'rc_timestamp' => $obj->rc_timestamp
// index usage
255 $lastOldId = intval( $obj->rc_this_oldid
);
258 if ( ( ++
$updated %
$this->getBatchSize() ) == 0 ) {
259 wfGetLBFactory()->waitForReplication();
266 * Rebuild pass 3: Insert `recentchanges` entries for action logs.
268 private function rebuildRecentChangesTablePass3() {
269 global $wgLogTypes, $wgLogRestrictions;
271 $dbw = $this->getDB( DB_MASTER
);
272 $commentStore = CommentStore
::getStore();
274 $this->output( "Loading from user, page, and logging tables...\n" );
276 $commentQuery = $commentStore->getJoin( 'log_comment' );
277 $actorQuery = ActorMigration
::newMigration()->getJoin( 'log_user' );
279 [ 'logging', 'page' ] +
$commentQuery['tables'] +
$actorQuery['tables'],
290 ] +
$commentQuery['fields'] +
$actorQuery['fields'],
292 'log_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom
) ),
293 'log_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo
) ),
294 // Some logs don't go in RC since they are private.
295 // @FIXME: core/extensions also have spammy logs that don't go in RC.
296 'log_type' => array_diff( $wgLogTypes, array_keys( $wgLogRestrictions ) ),
299 [ 'ORDER BY' => 'log_timestamp DESC' ],
302 [ 'LEFT JOIN', [ 'log_namespace=page_namespace', 'log_title=page_title' ] ]
303 ] +
$commentQuery['joins'] +
$actorQuery['joins']
306 $field = $dbw->fieldInfo( 'recentchanges', 'rc_cur_id' );
309 $actorMigration = ActorMigration
::newMigration();
310 foreach ( $res as $row ) {
311 $comment = $commentStore->getComment( 'log_comment', $row );
312 $user = User
::newFromAnyId( $row->log_user
, $row->log_user_text
, $row->log_actor
);
316 'rc_timestamp' => $row->log_timestamp
,
317 'rc_namespace' => $row->log_namespace
,
318 'rc_title' => $row->log_title
,
323 'rc_this_oldid' => 0,
324 'rc_last_oldid' => 0,
326 'rc_source' => RecentChange
::SRC_LOG
,
327 'rc_cur_id' => $field->isNullable()
329 : (int)$row->page_id
, // NULL => 0,
330 'rc_log_type' => $row->log_type
,
331 'rc_log_action' => $row->log_action
,
332 'rc_logid' => $row->log_id
,
333 'rc_params' => $row->log_params
,
334 'rc_deleted' => $row->log_deleted
335 ] +
$commentStore->insert( $dbw, 'rc_comment', $comment )
336 +
$actorMigration->getInsertValues( $dbw, 'rc_user', $user ),
340 if ( ( ++
$inserted %
$this->getBatchSize() ) == 0 ) {
341 wfGetLBFactory()->waitForReplication();
347 * Rebuild pass 4: Mark bot and autopatrolled entries.
349 private function rebuildRecentChangesTablePass4() {
350 global $wgUseRCPatrol, $wgMiserMode;
352 $dbw = $this->getDB( DB_MASTER
);
354 $userQuery = User
::getQueryInfo();
356 # @FIXME: recognize other bot account groups (not the same as users with 'bot' rights)
357 # @NOTE: users with 'bot' rights choose when edits are bot edits or not. That information
358 # may be lost at this point (aside from joining on the patrol log table entries).
359 $botgroups = [ 'bot' ];
360 $autopatrolgroups = $wgUseRCPatrol ? User
::getGroupsWithPermission( 'autopatrol' ) : [];
362 # Flag our recent bot edits
364 $this->output( "Flagging bot account edits...\n" );
366 # Find all users that are bots
368 array_merge( [ 'user_groups' ], $userQuery['tables'] ),
369 $userQuery['fields'],
370 [ 'ug_group' => $botgroups ],
373 [ 'user_group' => [ 'JOIN', 'user_id = ug_user' ] ] +
$userQuery['joins']
377 foreach ( $res as $obj ) {
378 $botusers[] = User
::newFromRow( $obj );
381 # Fill in the rc_bot field
383 $actorQuery = ActorMigration
::newMigration()->getWhere( $dbw, 'rc_user', $botusers, false );
385 foreach ( $actorQuery['orconds'] as $cond ) {
386 $rcids = array_merge( $rcids, $dbw->selectFieldValues(
387 [ 'recentchanges' ] +
$actorQuery['tables'],
390 "rc_timestamp > " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom
) ),
391 "rc_timestamp < " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo
) ),
399 $rcids = array_values( array_unique( $rcids ) );
401 foreach ( array_chunk( $rcids, $this->getBatchSize() ) as $rcidBatch ) {
405 [ 'rc_id' => $rcidBatch ],
408 wfGetLBFactory()->waitForReplication();
413 # Flag our recent autopatrolled edits
414 if ( !$wgMiserMode && $autopatrolgroups ) {
417 $this->output( "Flagging auto-patrolled edits...\n" );
419 # Find all users in RC with autopatrol rights
421 array_merge( [ 'user_groups' ], $userQuery['tables'] ),
422 $userQuery['fields'],
423 [ 'ug_group' => $autopatrolgroups ],
426 [ 'user_group' => [ 'JOIN', 'user_id = ug_user' ] ] +
$userQuery['joins']
429 foreach ( $res as $obj ) {
430 $patrolusers[] = User
::newFromRow( $obj );
433 # Fill in the rc_patrolled field
434 if ( $patrolusers ) {
435 $actorQuery = ActorMigration
::newMigration()->getWhere( $dbw, 'rc_user', $patrolusers, false );
436 foreach ( $actorQuery['orconds'] as $cond ) {
439 [ 'rc_patrolled' => 1 ],
442 'rc_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom
) ),
443 'rc_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo
) ),
447 wfGetLBFactory()->waitForReplication();
454 * Rebuild pass 5: Delete duplicate entries where we generate both a page revision and a log entry
455 * for a single action (upload only, at the moment, but potentially also move, protect, ...).
457 private function rebuildRecentChangesTablePass5() {
458 $dbw = wfGetDB( DB_MASTER
);
460 $this->output( "Removing duplicate revision and logging entries...\n" );
463 [ 'logging', 'log_search' ],
464 [ 'ls_value', 'ls_log_id' ],
466 'ls_log_id = log_id',
467 'ls_field' => 'associated_rev_id',
468 'log_type' => 'upload',
469 'log_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom
) ),
470 'log_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo
) ),
476 foreach ( $res as $obj ) {
477 $rev_id = $obj->ls_value
;
478 $log_id = $obj->ls_log_id
;
480 // Mark the logging row as having an associated rev id
483 /*SET*/ [ 'rc_this_oldid' => $rev_id ],
484 /*WHERE*/ [ 'rc_logid' => $log_id ],
488 // Delete the revision row
491 /*WHERE*/ [ 'rc_this_oldid' => $rev_id, 'rc_logid' => 0 ],
495 if ( ( ++
$updates %
$this->getBatchSize() ) == 0 ) {
496 wfGetLBFactory()->waitForReplication();
502 * Purge cached feeds in $wanCache
504 private function purgeFeeds() {
505 global $wgFeedClasses;
507 $this->output( "Deleting feed timestamps.\n" );
509 $wanCache = MediaWikiServices
::getInstance()->getMainWANObjectCache();
510 foreach ( $wgFeedClasses as $feed => $className ) {
511 $wanCache->delete( $wanCache->makeKey( 'rcfeed', $feed, 'timestamp' ) ); # Good enough for now.
516 $maintClass = RebuildRecentchanges
::class;
517 require_once RUN_MAINTENANCE_IF_MAIN
;