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';
28 use MediaWiki\MediaWikiServices
;
29 use Wikimedia\Rdbms\ILBFactory
;
32 * Maintenance script that rebuilds recent changes from scratch.
34 * @ingroup Maintenance
36 class RebuildRecentchanges
extends Maintenance
{
37 /** @var int UNIX timestamp */
39 /** @var int UNIX timestamp */
42 public function __construct() {
43 parent
::__construct();
44 $this->addDescription( 'Rebuild recent changes' );
48 "Only rebuild rows in requested time range (in YYYYMMDDHHMMSS format)",
54 "Only rebuild rows in requested time range (in YYYYMMDDHHMMSS format)",
58 $this->setBatchSize( 200 );
61 public function execute() {
63 ( $this->hasOption( 'from' ) && !$this->hasOption( 'to' ) ) ||
64 ( !$this->hasOption( 'from' ) && $this->hasOption( 'to' ) )
66 $this->fatalError( "Both 'from' and 'to' must be given, or neither" );
69 $lbFactory = MediaWikiServices
::getInstance()->getDBLoadBalancerFactory();
70 $this->rebuildRecentChangesTablePass1( $lbFactory );
71 $this->rebuildRecentChangesTablePass2( $lbFactory );
72 $this->rebuildRecentChangesTablePass3( $lbFactory );
73 $this->rebuildRecentChangesTablePass4( $lbFactory );
74 $this->rebuildRecentChangesTablePass5( $lbFactory );
75 if ( !( $this->hasOption( 'from' ) && $this->hasOption( 'to' ) ) ) {
78 $this->output( "Done.\n" );
82 * Rebuild pass 1: Insert `recentchanges` entries for page revisions.
84 private function rebuildRecentChangesTablePass1( ILBFactory
$lbFactory ) {
85 $dbw = $this->getDB( DB_MASTER
);
86 $commentStore = CommentStore
::getStore();
88 if ( $this->hasOption( 'from' ) && $this->hasOption( 'to' ) ) {
89 $this->cutoffFrom
= wfTimestamp( TS_UNIX
, $this->getOption( 'from' ) );
90 $this->cutoffTo
= wfTimestamp( TS_UNIX
, $this->getOption( 'to' ) );
92 $sec = $this->cutoffTo
- $this->cutoffFrom
;
93 $days = $sec / 24 / 3600;
94 $this->output( "Rebuilding range of $sec seconds ($days days)\n" );
98 $days = $wgRCMaxAge / 24 / 3600;
99 $this->output( "Rebuilding \$wgRCMaxAge=$wgRCMaxAge seconds ($days days)\n" );
101 $this->cutoffFrom
= time() - $wgRCMaxAge;
102 $this->cutoffTo
= time();
105 $this->output( "Clearing recentchanges table for time range...\n" );
106 $rcids = $dbw->selectFieldValues(
110 'rc_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom
) ),
111 'rc_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo
) )
114 foreach ( array_chunk( $rcids, $this->getBatchSize() ) as $rcidBatch ) {
115 $dbw->delete( 'recentchanges', [ 'rc_id' => $rcidBatch ], __METHOD__
);
116 $lbFactory->waitForReplication();
119 $this->output( "Loading from page and revision tables...\n" );
121 $commentQuery = $commentStore->getJoin( 'rev_comment' );
122 $actorQuery = ActorMigration
::newMigration()->getJoin( 'rev_user' );
124 [ 'revision', 'page' ] +
$commentQuery['tables'] +
$actorQuery['tables'],
134 ] +
$commentQuery['fields'] +
$actorQuery['fields'],
136 'rev_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom
) ),
137 'rev_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo
) )
140 [ 'ORDER BY' => 'rev_timestamp DESC' ],
142 'page' => [ 'JOIN', 'rev_page=page_id' ],
143 ] +
$commentQuery['joins'] +
$actorQuery['joins']
146 $this->output( "Inserting from page and revision tables...\n" );
148 $actorMigration = ActorMigration
::newMigration();
149 foreach ( $res as $row ) {
150 $comment = $commentStore->getComment( 'rev_comment', $row );
151 $user = User
::newFromAnyId( $row->rev_user
, $row->rev_user_text
, $row->rev_actor
);
155 'rc_timestamp' => $row->rev_timestamp
,
156 'rc_namespace' => $row->page_namespace
,
157 'rc_title' => $row->page_title
,
158 'rc_minor' => $row->rev_minor_edit
,
160 'rc_new' => $row->page_is_new
,
161 'rc_cur_id' => $row->page_id
,
162 'rc_this_oldid' => $row->rev_id
,
163 'rc_last_oldid' => 0, // is this ok?
164 'rc_type' => $row->page_is_new ? RC_NEW
: RC_EDIT
,
165 'rc_source' => $row->page_is_new ? RecentChange
::SRC_NEW
: RecentChange
::SRC_EDIT
,
166 'rc_deleted' => $row->rev_deleted
167 ] +
$commentStore->insert( $dbw, 'rc_comment', $comment )
168 +
$actorMigration->getInsertValues( $dbw, 'rc_user', $user ),
172 $rcid = $dbw->insertId();
175 [ 'ct_rc_id' => $rcid ],
176 [ 'ct_rev_id' => $row->rev_id
],
180 if ( ( ++
$inserted %
$this->getBatchSize() ) == 0 ) {
181 $lbFactory->waitForReplication();
187 * Rebuild pass 2: Enhance entries for page revisions with references to the previous revision
188 * (rc_last_oldid, rc_new etc.) and size differences (rc_old_len, rc_new_len).
190 private function rebuildRecentChangesTablePass2( ILBFactory
$lbFactory ) {
191 $dbw = $this->getDB( DB_MASTER
);
193 $this->output( "Updating links and size differences...\n" );
195 # Fill in the rc_last_oldid field, which points to the previous edit
198 [ 'rc_cur_id', 'rc_this_oldid', 'rc_timestamp' ],
200 "rc_timestamp > " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom
) ),
201 "rc_timestamp < " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo
) )
204 [ 'ORDER BY' => 'rc_cur_id,rc_timestamp' ]
211 foreach ( $res as $obj ) {
214 if ( $obj->rc_cur_id
!= $lastCurId ) {
215 # Switch! Look up the previous last edit, if any
216 $lastCurId = intval( $obj->rc_cur_id
);
217 $emit = $obj->rc_timestamp
;
219 $row = $dbw->selectRow(
221 [ 'rev_id', 'rev_len' ],
222 [ 'rev_page' => $lastCurId, "rev_timestamp < " . $dbw->addQuotes( $emit ) ],
224 [ 'ORDER BY' => 'rev_timestamp DESC' ]
227 $lastOldId = intval( $row->rev_id
);
228 # Grab the last text size if available
229 $lastSize = !is_null( $row->rev_len
) ?
intval( $row->rev_len
) : null;
234 $new = 1; // probably true
238 if ( $lastCurId == 0 ) {
239 $this->output( "Uhhh, something wrong? No curid\n" );
241 # Grab the entry's text size
242 $size = (int)$dbw->selectField(
245 [ 'rev_id' => $obj->rc_this_oldid
],
252 'rc_last_oldid' => $lastOldId,
254 'rc_type' => $new ? RC_NEW
: RC_EDIT
,
255 'rc_source' => $new === 1 ? RecentChange
::SRC_NEW
: RecentChange
::SRC_EDIT
,
256 'rc_old_len' => $lastSize,
257 'rc_new_len' => $size,
260 'rc_cur_id' => $lastCurId,
261 'rc_this_oldid' => $obj->rc_this_oldid
,
262 'rc_timestamp' => $obj->rc_timestamp
// index usage
267 $lastOldId = intval( $obj->rc_this_oldid
);
270 if ( ( ++
$updated %
$this->getBatchSize() ) == 0 ) {
271 $lbFactory->waitForReplication();
278 * Rebuild pass 3: Insert `recentchanges` entries for action logs.
280 private function rebuildRecentChangesTablePass3( ILBFactory
$lbFactory ) {
281 global $wgLogTypes, $wgLogRestrictions;
283 $dbw = $this->getDB( DB_MASTER
);
284 $commentStore = CommentStore
::getStore();
286 $this->output( "Loading from user, page, and logging tables...\n" );
288 $commentQuery = $commentStore->getJoin( 'log_comment' );
289 $actorQuery = ActorMigration
::newMigration()->getJoin( 'log_user' );
291 [ 'logging', 'page' ] +
$commentQuery['tables'] +
$actorQuery['tables'],
302 ] +
$commentQuery['fields'] +
$actorQuery['fields'],
304 'log_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom
) ),
305 'log_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo
) ),
306 // Some logs don't go in RC since they are private.
307 // @FIXME: core/extensions also have spammy logs that don't go in RC.
308 'log_type' => array_diff( $wgLogTypes, array_keys( $wgLogRestrictions ) ),
311 [ 'ORDER BY' => 'log_timestamp DESC' ],
314 [ 'LEFT JOIN', [ 'log_namespace=page_namespace', 'log_title=page_title' ] ]
315 ] +
$commentQuery['joins'] +
$actorQuery['joins']
318 $field = $dbw->fieldInfo( 'recentchanges', 'rc_cur_id' );
321 $actorMigration = ActorMigration
::newMigration();
322 foreach ( $res as $row ) {
323 $comment = $commentStore->getComment( 'log_comment', $row );
324 $user = User
::newFromAnyId( $row->log_user
, $row->log_user_text
, $row->log_actor
);
328 'rc_timestamp' => $row->log_timestamp
,
329 'rc_namespace' => $row->log_namespace
,
330 'rc_title' => $row->log_title
,
333 'rc_patrolled' => $row->log_type
== 'upload' ?
0 : 2,
335 'rc_this_oldid' => 0,
336 'rc_last_oldid' => 0,
338 'rc_source' => RecentChange
::SRC_LOG
,
339 'rc_cur_id' => $field->isNullable()
341 : (int)$row->page_id
, // NULL => 0,
342 'rc_log_type' => $row->log_type
,
343 'rc_log_action' => $row->log_action
,
344 'rc_logid' => $row->log_id
,
345 'rc_params' => $row->log_params
,
346 'rc_deleted' => $row->log_deleted
347 ] +
$commentStore->insert( $dbw, 'rc_comment', $comment )
348 +
$actorMigration->getInsertValues( $dbw, 'rc_user', $user ),
352 $rcid = $dbw->insertId();
355 [ 'ct_rc_id' => $rcid ],
356 [ 'ct_log_id' => $row->log_id
],
360 if ( ( ++
$inserted %
$this->getBatchSize() ) == 0 ) {
361 $lbFactory->waitForReplication();
367 * Rebuild pass 4: Mark bot and autopatrolled entries.
369 private function rebuildRecentChangesTablePass4( ILBFactory
$lbFactory ) {
370 global $wgUseRCPatrol, $wgMiserMode;
372 $dbw = $this->getDB( DB_MASTER
);
374 $userQuery = User
::getQueryInfo();
376 # @FIXME: recognize other bot account groups (not the same as users with 'bot' rights)
377 # @NOTE: users with 'bot' rights choose when edits are bot edits or not. That information
378 # may be lost at this point (aside from joining on the patrol log table entries).
379 $botgroups = [ 'bot' ];
380 $autopatrolgroups = $wgUseRCPatrol ? User
::getGroupsWithPermission( 'autopatrol' ) : [];
382 # Flag our recent bot edits
384 $this->output( "Flagging bot account edits...\n" );
386 # Find all users that are bots
388 array_merge( [ 'user_groups' ], $userQuery['tables'] ),
389 $userQuery['fields'],
390 [ 'ug_group' => $botgroups ],
393 [ 'user_groups' => [ 'JOIN', 'user_id = ug_user' ] ] +
$userQuery['joins']
397 foreach ( $res as $obj ) {
398 $botusers[] = User
::newFromRow( $obj );
401 # Fill in the rc_bot field
403 $actorQuery = ActorMigration
::newMigration()->getWhere( $dbw, 'rc_user', $botusers, false );
405 foreach ( $actorQuery['orconds'] as $cond ) {
406 $rcids = array_merge( $rcids, $dbw->selectFieldValues(
407 [ 'recentchanges' ] +
$actorQuery['tables'],
410 "rc_timestamp > " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom
) ),
411 "rc_timestamp < " . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo
) ),
419 $rcids = array_values( array_unique( $rcids ) );
421 foreach ( array_chunk( $rcids, $this->getBatchSize() ) as $rcidBatch ) {
425 [ 'rc_id' => $rcidBatch ],
428 $lbFactory->waitForReplication();
433 # Flag our recent autopatrolled edits
434 if ( !$wgMiserMode && $autopatrolgroups ) {
437 $this->output( "Flagging auto-patrolled edits...\n" );
439 # Find all users in RC with autopatrol rights
441 array_merge( [ 'user_groups' ], $userQuery['tables'] ),
442 $userQuery['fields'],
443 [ 'ug_group' => $autopatrolgroups ],
446 [ 'user_groups' => [ 'JOIN', 'user_id = ug_user' ] ] +
$userQuery['joins']
449 foreach ( $res as $obj ) {
450 $patrolusers[] = User
::newFromRow( $obj );
453 # Fill in the rc_patrolled field
454 if ( $patrolusers ) {
455 $actorQuery = ActorMigration
::newMigration()->getWhere( $dbw, 'rc_user', $patrolusers, false );
456 foreach ( $actorQuery['orconds'] as $cond ) {
459 [ 'rc_patrolled' => 2 ],
462 'rc_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom
) ),
463 'rc_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo
) ),
468 $lbFactory->waitForReplication();
475 * Rebuild pass 5: Delete duplicate entries where we generate both a page revision and a log entry
476 * for a single action (upload only, at the moment, but potentially also move, protect, ...).
478 private function rebuildRecentChangesTablePass5( ILBFactory
$lbFactory ) {
479 $dbw = wfGetDB( DB_MASTER
);
481 $this->output( "Removing duplicate revision and logging entries...\n" );
484 [ 'logging', 'log_search' ],
485 [ 'ls_value', 'ls_log_id' ],
487 'ls_log_id = log_id',
488 'ls_field' => 'associated_rev_id',
489 'log_type' => 'upload',
490 'log_timestamp > ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffFrom
) ),
491 'log_timestamp < ' . $dbw->addQuotes( $dbw->timestamp( $this->cutoffTo
) ),
497 foreach ( $res as $obj ) {
498 $rev_id = $obj->ls_value
;
499 $log_id = $obj->ls_log_id
;
501 // Mark the logging row as having an associated rev id
504 /*SET*/ [ 'rc_this_oldid' => $rev_id ],
505 /*WHERE*/ [ 'rc_logid' => $log_id ],
509 // Delete the revision row
512 /*WHERE*/ [ 'rc_this_oldid' => $rev_id, 'rc_logid' => 0 ],
516 if ( ( ++
$updates %
$this->getBatchSize() ) == 0 ) {
517 $lbFactory->waitForReplication();
523 * Purge cached feeds in $wanCache
525 private function purgeFeeds() {
526 global $wgFeedClasses;
528 $this->output( "Deleting feed timestamps.\n" );
530 $wanCache = MediaWikiServices
::getInstance()->getMainWANObjectCache();
531 foreach ( $wgFeedClasses as $feed => $className ) {
532 $wanCache->delete( $wanCache->makeKey( 'rcfeed', $feed, 'timestamp' ) ); # Good enough for now.
537 $maintClass = RebuildRecentchanges
::class;
538 require_once RUN_MAINTENANCE_IF_MAIN
;