Localisation updates from https://translatewiki.net.
[mediawiki.git] / maintenance / namespaceDupes.php
blobb97aaf47c67f4c951675e19e4cb79bc1b9f8c74a
1 <?php
2 /**
3 * Check for articles to fix after adding/deleting namespaces
5 * Copyright © 2005-2007 Brooke Vibber <bvibber@wikimedia.org>
6 * https://www.mediawiki.org/
8 * This program is free software; you can redistribute it and/or modify
9 * it under the terms of the GNU General Public License as published by
10 * the Free Software Foundation; either version 2 of the License, or
11 * (at your option) any later version.
13 * This program is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 * GNU General Public License for more details.
18 * You should have received a copy of the GNU General Public License along
19 * with this program; if not, write to the Free Software Foundation, Inc.,
20 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
21 * http://www.gnu.org/copyleft/gpl.html
23 * @file
24 * @ingroup Maintenance
27 // @codeCoverageIgnoreStart
28 require_once __DIR__ . '/Maintenance.php';
29 // @codeCoverageIgnoreEnd
31 use MediaWiki\Deferred\DeferredUpdates;
32 use MediaWiki\Deferred\LinksUpdate\LinksDeletionUpdate;
33 use MediaWiki\Linker\LinkTarget;
34 use MediaWiki\MainConfigNames;
35 use MediaWiki\Maintenance\Maintenance;
36 use MediaWiki\Title\Title;
37 use MediaWiki\Title\TitleValue;
38 use Wikimedia\Rdbms\IDBAccessObject;
39 use Wikimedia\Rdbms\IExpression;
40 use Wikimedia\Rdbms\IResultWrapper;
41 use Wikimedia\Rdbms\LikeValue;
43 /**
44 * Maintenance script that checks for articles to fix after
45 * adding/deleting namespaces.
47 * @ingroup Maintenance
49 class NamespaceDupes extends Maintenance {
51 /**
52 * Total number of pages that need fixing that are automatically resolveable
53 * @var int
55 private $resolvablePages = 0;
57 /**
58 * Total number of pages that need fixing
59 * @var int
61 private $totalPages = 0;
63 /**
64 * Total number of links that need fixing that are automatically resolveable
65 * @var int
67 private $resolvableLinks = 0;
69 /**
70 * Total number of erroneous links
71 * @var int
73 private $totalLinks = 0;
75 /**
76 * Total number of links deleted because they weren't automatically resolveable due to the
77 * target already existing
78 * @var int
80 private $deletedLinks = 0;
82 public function __construct() {
83 parent::__construct();
84 $this->addDescription( 'Find and fix pages affected by namespace addition/removal' );
85 $this->addOption( 'fix', 'Attempt to automatically fix errors and delete broken links' );
86 $this->addOption( 'merge', "Instead of renaming conflicts, do a history merge with " .
87 "the correct title" );
88 $this->addOption( 'add-suffix', "Dupes will be renamed with correct namespace with " .
89 "<text> appended after the article name", false, true );
90 $this->addOption( 'add-prefix', "Dupes will be renamed with correct namespace with " .
91 "<text> prepended before the article name", false, true );
92 $this->addOption( 'source-pseudo-namespace', "Move all pages with the given source " .
93 "prefix (with an implied colon following it). If --dest-namespace is not specified, " .
94 "the colon will be replaced with a hyphen.",
95 false, true );
96 $this->addOption( 'dest-namespace', "In combination with --source-pseudo-namespace, " .
97 "specify the namespace ID of the destination.", false, true );
98 $this->addOption( 'move-talk', "If this is specified, pages in the Talk namespace that " .
99 "begin with a conflicting prefix will be renamed, for example " .
100 "Talk:File:Foo -> File_Talk:Foo" );
103 public function execute() {
104 $options = [
105 'fix' => $this->hasOption( 'fix' ),
106 'merge' => $this->hasOption( 'merge' ),
107 'add-suffix' => $this->getOption( 'add-suffix', '' ),
108 'add-prefix' => $this->getOption( 'add-prefix', '' ),
109 'move-talk' => $this->hasOption( 'move-talk' ),
110 'source-pseudo-namespace' => $this->getOption( 'source-pseudo-namespace', '' ),
111 'dest-namespace' => intval( $this->getOption( 'dest-namespace', 0 ) )
114 if ( $options['source-pseudo-namespace'] !== '' ) {
115 $retval = $this->checkPrefix( $options );
116 } else {
117 $retval = $this->checkAll( $options );
120 if ( $retval ) {
121 $this->output( "\nLooks good!\n" );
122 } else {
123 $this->output( "\nOh noeees\n" );
128 * Check all namespaces
130 * @param array $options Associative array of validated command-line options
132 * @return bool
134 private function checkAll( $options ) {
135 $contLang = $this->getServiceContainer()->getContentLanguage();
136 $spaces = [];
138 // List interwikis first, so they'll be overridden
139 // by any conflicting local namespaces.
140 foreach ( $this->getInterwikiList() as $prefix ) {
141 $name = $contLang->ucfirst( $prefix );
142 $spaces[$name] = 0;
145 // Now pull in all canonical and alias namespaces...
146 foreach (
147 $this->getServiceContainer()->getNamespaceInfo()->getCanonicalNamespaces()
148 as $ns => $name
150 // This includes $wgExtraNamespaces
151 if ( $name !== '' ) {
152 $spaces[$name] = $ns;
155 foreach ( $contLang->getNamespaces() as $ns => $name ) {
156 if ( $name !== '' ) {
157 $spaces[$name] = $ns;
160 foreach ( $contLang->getNamespaceAliases() as $name => $ns ) {
161 $spaces[$name] = $ns;
164 // We'll need to check for lowercase keys as well,
165 // since we're doing case-sensitive searches in the db.
166 $capitalLinks = $this->getConfig()->get( MainConfigNames::CapitalLinks );
167 foreach ( $spaces as $name => $ns ) {
168 $moreNames = [];
169 $moreNames[] = $contLang->uc( $name );
170 $moreNames[] = $contLang->ucfirst( $contLang->lc( $name ) );
171 $moreNames[] = $contLang->ucwords( $name );
172 $moreNames[] = $contLang->ucwords( $contLang->lc( $name ) );
173 $moreNames[] = $contLang->ucwordbreaks( $name );
174 $moreNames[] = $contLang->ucwordbreaks( $contLang->lc( $name ) );
175 if ( !$capitalLinks ) {
176 foreach ( $moreNames as $altName ) {
177 $moreNames[] = $contLang->lcfirst( $altName );
179 $moreNames[] = $contLang->lcfirst( $name );
181 foreach ( array_unique( $moreNames ) as $altName ) {
182 if ( $altName !== $name ) {
183 $spaces[$altName] = $ns;
188 // Sort by namespace index, and if there are two with the same index,
189 // break the tie by sorting by name
190 $origSpaces = $spaces;
191 uksort( $spaces, static function ( $a, $b ) use ( $origSpaces ) {
192 return $origSpaces[$a] <=> $origSpaces[$b]
193 ?: $a <=> $b;
194 } );
196 $ok = true;
197 foreach ( $spaces as $name => $ns ) {
198 $ok = $this->checkNamespace( $ns, $name, $options ) && $ok;
201 $this->output(
202 "{$this->totalPages} pages to fix, " .
203 "{$this->resolvablePages} were resolvable.\n\n"
206 foreach ( $spaces as $name => $ns ) {
207 if ( $ns != 0 ) {
208 /* Fix up link destinations for non-interwiki links only.
210 * For example if a page has [[Foo:Bar]] and then a Foo namespace
211 * is introduced, pagelinks needs to be updated to have
212 * page_namespace = NS_FOO.
214 * If instead an interwiki prefix was introduced called "Foo",
215 * the link should instead be moved to the iwlinks table. If a new
216 * language is introduced called "Foo", or if there is a pagelink
217 * [[fr:Bar]] when interlanguage magic links are turned on, the
218 * link would have to be moved to the langlinks table. Let's put
219 * those cases in the too-hard basket for now. The consequences are
220 * not especially severe.
221 * @fixme Handle interwiki links, and pagelinks to Category:, File:
222 * which probably need reparsing.
225 $this->checkLinkTable( 'pagelinks', 'pl', $ns, $name, $options );
226 $this->checkLinkTable( 'templatelinks', 'tl', $ns, $name, $options );
228 // The redirect table has interwiki links randomly mixed in, we
229 // need to filter those out. For example [[w:Foo:Bar]] would
230 // have rd_interwiki=w and rd_namespace=0, which would match the
231 // query for a conflicting namespace "Foo" if filtering wasn't done.
232 $this->checkLinkTable( 'redirect', 'rd', $ns, $name, $options,
233 [ 'rd_interwiki' => '' ] );
237 $this->output(
238 "{$this->totalLinks} links to fix, " .
239 "{$this->resolvableLinks} were resolvable, " .
240 "{$this->deletedLinks} were deleted.\n"
243 return $ok;
247 * @return string[]
249 private function getInterwikiList() {
250 $result = $this->getServiceContainer()->getInterwikiLookup()->getAllPrefixes();
251 return array_column( $result, 'iw_prefix' );
255 * Check a given prefix and try to move it into the given destination namespace
257 * @param int $ns Destination namespace id
258 * @param string $name
259 * @param array $options Associative array of validated command-line options
260 * @return bool
262 private function checkNamespace( $ns, $name, $options ) {
263 $targets = $this->getTargetList( $ns, $name, $options );
264 $count = $targets->numRows();
265 $this->totalPages += $count;
266 if ( $count == 0 ) {
267 return true;
270 $dryRunNote = $options['fix'] ? '' : ' DRY RUN ONLY';
272 $ok = true;
273 foreach ( $targets as $row ) {
274 // Find the new title and determine the action to take
276 $newTitle = $this->getDestinationTitle(
277 $ns, $name, $row->page_namespace, $row->page_title );
278 $logStatus = false;
279 if ( !$newTitle ) {
280 if ( $options['add-prefix'] == '' && $options['add-suffix'] == '' ) {
281 $logStatus = 'invalid title and --add-prefix not specified';
282 $action = 'abort';
283 } else {
284 $action = 'alternate';
286 } elseif ( $newTitle->exists( IDBAccessObject::READ_LATEST ) ) {
287 if ( $options['merge'] ) {
288 if ( $this->canMerge( $row->page_id, $newTitle, $logStatus ) ) {
289 $action = 'merge';
290 } else {
291 $action = 'abort';
293 } elseif ( $options['add-prefix'] == '' && $options['add-suffix'] == '' ) {
294 $action = 'abort';
295 $logStatus = 'dest title exists and --add-prefix not specified';
296 } else {
297 $action = 'alternate';
299 } else {
300 $action = 'move';
301 $logStatus = 'no conflict';
303 if ( $action === 'alternate' ) {
304 [ $ns, $dbk ] = $this->getDestination( $ns, $name, $row->page_namespace,
305 $row->page_title );
306 $newTitle = $this->getAlternateTitle( $ns, $dbk, $options );
307 if ( !$newTitle ) {
308 $action = 'abort';
309 $logStatus = 'alternate title is invalid';
310 } elseif ( $newTitle->exists() ) {
311 $action = 'abort';
312 $logStatus = 'alternate title conflicts';
313 } else {
314 $action = 'move';
315 $logStatus = 'alternate';
319 // Take the action or log a dry run message
321 $logTitle = "id={$row->page_id} ns={$row->page_namespace} dbk={$row->page_title}";
322 $pageOK = true;
324 switch ( $action ) {
325 case 'abort':
326 $this->output( "$logTitle *** $logStatus\n" );
327 $pageOK = false;
328 break;
329 case 'move':
330 $this->output( "$logTitle -> " .
331 $newTitle->getPrefixedDBkey() . " ($logStatus)$dryRunNote\n" );
333 if ( $options['fix'] ) {
334 $pageOK = $this->movePage( $row->page_id, $newTitle );
336 break;
337 case 'merge':
338 $this->output( "$logTitle => " .
339 $newTitle->getPrefixedDBkey() . " (merge)$dryRunNote\n" );
341 if ( $options['fix'] ) {
342 $pageOK = $this->mergePage( $row, $newTitle );
344 break;
347 if ( $pageOK ) {
348 $this->resolvablePages++;
349 } else {
350 $ok = false;
354 return $ok;
358 * Check and repair the destination fields in a link table
359 * @param string $table The link table name
360 * @param string $fieldPrefix The field prefix in the link table
361 * @param int $ns Destination namespace id
362 * @param string $name
363 * @param array $options Associative array of validated command-line options
364 * @param array $extraConds Extra conditions for the SQL query
366 private function checkLinkTable( $table, $fieldPrefix, $ns, $name, $options,
367 $extraConds = []
369 $dbw = $this->getPrimaryDB();
371 $batchConds = [];
372 $fromField = "{$fieldPrefix}_from";
373 $batchSize = 100;
374 $sqb = $dbw->newSelectQueryBuilder()
375 ->select( $fromField )
376 ->where( $extraConds )
377 ->limit( $batchSize );
379 $linksMigration = $this->getServiceContainer()->getLinksMigration();
380 if ( isset( $linksMigration::$mapping[$table] ) ) {
381 $sqb->queryInfo( $linksMigration->getQueryInfo( $table ) );
382 [ $namespaceField, $titleField ] = $linksMigration->getTitleFields( $table );
383 $schemaMigrationStage = $linksMigration::$mapping[$table]['config'] === -1
384 ? MIGRATION_NEW
385 : $this->getConfig()->get( $linksMigration::$mapping[$table]['config'] );
386 $linkTargetLookup = $this->getServiceContainer()->getLinkTargetLookup();
387 $targetIdField = $linksMigration::$mapping[$table]['target_id'];
388 } else {
389 $sqb->table( $table );
390 $namespaceField = "{$fieldPrefix}_namespace";
391 $titleField = "{$fieldPrefix}_title";
392 $sqb->fields( [ $namespaceField, $titleField ] );
393 // Variables only used for links migration, init only
394 $schemaMigrationStage = -1;
395 $linkTargetLookup = null;
396 $targetIdField = '';
398 $sqb->andWhere( [
399 $namespaceField => 0,
400 $dbw->expr( $titleField, IExpression::LIKE, new LikeValue( "$name:", $dbw->anyString() ) ),
402 ->orderBy( [ $titleField, $fromField ] )
403 ->caller( __METHOD__ );
405 $updateRowsPerQuery = $this->getConfig()->get( MainConfigNames::UpdateRowsPerQuery );
406 while ( true ) {
407 $res = ( clone $sqb )
408 ->andWhere( $batchConds )
409 ->fetchResultSet();
410 if ( $res->numRows() == 0 ) {
411 break;
414 $rowsToDeleteIfStillExists = [];
416 foreach ( $res as $row ) {
417 $logTitle = "from={$row->$fromField} ns={$row->$namespaceField} " .
418 "dbk={$row->$titleField}";
419 $destTitle = $this->getDestinationTitle(
420 $ns, $name, $row->$namespaceField, $row->$titleField );
421 $this->totalLinks++;
422 if ( !$destTitle ) {
423 $this->output( "$table $logTitle *** INVALID\n" );
424 continue;
426 $this->resolvableLinks++;
427 if ( !$options['fix'] ) {
428 $this->output( "$table $logTitle -> " .
429 $destTitle->getPrefixedDBkey() . " DRY RUN\n" );
430 continue;
433 if ( isset( $linksMigration::$mapping[$table] ) ) {
434 $setValue = [];
435 if ( $schemaMigrationStage & SCHEMA_COMPAT_WRITE_NEW ) {
436 $setValue[$targetIdField] = $linkTargetLookup->acquireLinkTargetId( $destTitle, $dbw );
438 if ( $schemaMigrationStage & SCHEMA_COMPAT_WRITE_OLD ) {
439 $setValue["{$fieldPrefix}_namespace"] = $destTitle->getNamespace();
440 $setValue["{$fieldPrefix}_title"] = $destTitle->getDBkey();
442 $whereCondition = $linksMigration->getLinksConditions(
443 $table,
444 new TitleValue( 0, $row->$titleField )
446 $deleteCondition = $linksMigration->getLinksConditions(
447 $table,
448 new TitleValue( (int)$row->$namespaceField, $row->$titleField )
450 } else {
451 $setValue = [
452 $namespaceField => $destTitle->getNamespace(),
453 $titleField => $destTitle->getDBkey()
455 $whereCondition = [
456 $namespaceField => 0,
457 $titleField => $row->$titleField
459 $deleteCondition = [
460 $namespaceField => $row->$namespaceField,
461 $titleField => $row->$titleField,
465 $dbw->newUpdateQueryBuilder()
466 ->update( $table )
467 ->ignore()
468 ->set( $setValue )
469 ->where( [ $fromField => $row->$fromField ] )
470 ->andWhere( $whereCondition )
471 ->caller( __METHOD__ )
472 ->execute();
474 // In case there is a key conflict on UPDATE IGNORE the row needs deletion
475 $rowsToDeleteIfStillExists[] = array_merge( [ $fromField => $row->$fromField ], $deleteCondition );
477 $this->output( "$table $logTitle -> " .
478 $destTitle->getPrefixedDBkey() . "\n"
482 if ( $options['fix'] && count( $rowsToDeleteIfStillExists ) > 0 ) {
483 $affectedRows = 0;
484 $deleteBatches = array_chunk( $rowsToDeleteIfStillExists, $updateRowsPerQuery );
485 foreach ( $deleteBatches as $deleteBatch ) {
486 $dbw->newDeleteQueryBuilder()
487 ->deleteFrom( $table )
488 ->where( $dbw->factorConds( $deleteBatch ) )
489 ->caller( __METHOD__ )
490 ->execute();
491 $affectedRows += $dbw->affectedRows();
492 if ( count( $deleteBatches ) > 1 ) {
493 $this->waitForReplication();
497 $this->deletedLinks += $affectedRows;
498 $this->resolvableLinks -= $affectedRows;
501 $batchConds = [
502 $dbw->buildComparison( '>', [
503 // @phan-suppress-next-line PhanPossiblyUndeclaredVariable rows contains at least one item
504 $titleField => $row->$titleField,
505 // @phan-suppress-next-line PhanPossiblyUndeclaredVariable rows contains at least one item
506 $fromField => $row->$fromField,
510 $this->waitForReplication();
515 * Move the given pseudo-namespace, either replacing the colon with a hyphen
516 * (useful for pseudo-namespaces that conflict with interwiki links) or move
517 * them to another namespace if specified.
518 * @param array $options Associative array of validated command-line options
519 * @return bool
521 private function checkPrefix( $options ) {
522 $prefix = $options['source-pseudo-namespace'];
523 $ns = $options['dest-namespace'];
524 $this->output( "Checking prefix \"$prefix\" vs namespace $ns\n" );
526 return $this->checkNamespace( $ns, $prefix, $options );
530 * Find pages in main and talk namespaces that have a prefix of the new
531 * namespace so we know titles that will need migrating
533 * @param int $ns Destination namespace id
534 * @param string $name Prefix that is being made a namespace
535 * @param array $options Associative array of validated command-line options
537 * @return IResultWrapper
539 private function getTargetList( $ns, $name, $options ) {
540 $dbw = $this->getPrimaryDB();
542 if (
543 $options['move-talk'] &&
544 $this->getServiceContainer()->getNamespaceInfo()->isSubject( $ns )
546 $checkNamespaces = [ NS_MAIN, NS_TALK ];
547 } else {
548 $checkNamespaces = NS_MAIN;
551 return $dbw->newSelectQueryBuilder()
552 ->select( [ 'page_id', 'page_title', 'page_namespace' ] )
553 ->from( 'page' )
554 ->where( [
555 'page_namespace' => $checkNamespaces,
556 $dbw->expr( 'page_title', IExpression::LIKE, new LikeValue( "$name:", $dbw->anyString() ) ),
558 ->caller( __METHOD__ )->fetchResultSet();
562 * Get the preferred destination for a given target page.
563 * @param int $ns The destination namespace ID
564 * @param string $name The conflicting prefix
565 * @param int $sourceNs The source namespace
566 * @param string $sourceDbk The source DB key (i.e. page_title)
567 * @return array [ ns, dbkey ], not necessarily valid
569 private function getDestination( $ns, $name, $sourceNs, $sourceDbk ) {
570 $dbk = substr( $sourceDbk, strlen( "$name:" ) );
571 if ( $ns <= 0 ) {
572 // An interwiki or an illegal namespace like "Special" or "Media"
573 // try an alternate encoding with '-' for ':'
574 $dbk = "$name-" . $dbk;
575 $ns = 0;
577 $destNS = $ns;
578 $nsInfo = $this->getServiceContainer()->getNamespaceInfo();
579 if ( $sourceNs == NS_TALK && $nsInfo->isSubject( $ns ) ) {
580 // This is an associated talk page moved with the --move-talk feature.
581 $destNS = $nsInfo->getTalk( $destNS );
583 return [ $destNS, $dbk ];
587 * Get the preferred destination title for a given target page.
588 * @param int $ns The destination namespace ID
589 * @param string $name The conflicting prefix
590 * @param int $sourceNs The source namespace
591 * @param string $sourceDbk The source DB key (i.e. page_title)
592 * @return Title|false
594 private function getDestinationTitle( $ns, $name, $sourceNs, $sourceDbk ) {
595 [ $destNS, $dbk ] = $this->getDestination( $ns, $name, $sourceNs, $sourceDbk );
596 $newTitle = Title::makeTitleSafe( $destNS, $dbk );
597 if ( !$newTitle || !$newTitle->canExist() ) {
598 return false;
600 return $newTitle;
604 * Get an alternative title to move a page to. This is used if the
605 * preferred destination title already exists.
607 * @param int $ns The destination namespace ID
608 * @param string $dbk The source DB key (i.e. page_title)
609 * @param array $options Associative array of validated command-line options
610 * @return Title|false
612 private function getAlternateTitle( $ns, $dbk, $options ) {
613 $prefix = $options['add-prefix'];
614 $suffix = $options['add-suffix'];
615 if ( $prefix == '' && $suffix == '' ) {
616 return false;
618 $newDbk = $prefix . $dbk . $suffix;
619 return Title::makeTitleSafe( $ns, $newDbk );
623 * Move a page
625 * @param int $id The page_id
626 * @param LinkTarget $newLinkTarget The new title link target
627 * @return bool
629 private function movePage( $id, LinkTarget $newLinkTarget ) {
630 $dbw = $this->getPrimaryDB();
632 $dbw->newUpdateQueryBuilder()
633 ->update( 'page' )
634 ->set( [
635 "page_namespace" => $newLinkTarget->getNamespace(),
636 "page_title" => $newLinkTarget->getDBkey(),
638 ->where( [
639 "page_id" => $id,
641 ->caller( __METHOD__ )
642 ->execute();
644 // Update *_from_namespace in links tables
645 $fromNamespaceTables = [
646 [ 'templatelinks', 'tl', [ 'tl_target_id' ] ],
647 [ 'imagelinks', 'il', [ 'il_to' ] ]
649 if ( $this->getConfig()->get( MainConfigNames::PageLinksSchemaMigrationStage ) & SCHEMA_COMPAT_WRITE_OLD ) {
650 $fromNamespaceTables[] = [ 'pagelinks', 'pl', [ 'pl_namespace', 'pl_title' ] ];
651 } else {
652 $fromNamespaceTables[] = [ 'pagelinks', 'pl', [ 'pl_target_id' ] ];
654 $updateRowsPerQuery = $this->getConfig()->get( MainConfigNames::UpdateRowsPerQuery );
655 foreach ( $fromNamespaceTables as [ $table, $fieldPrefix, $additionalPrimaryKeyFields ] ) {
656 $fromField = "{$fieldPrefix}_from";
657 $fromNamespaceField = "{$fieldPrefix}_from_namespace";
659 $res = $dbw->newSelectQueryBuilder()
660 ->select( $additionalPrimaryKeyFields )
661 ->from( $table )
662 ->where( [ $fromField => $id ] )
663 ->andWhere( $dbw->expr( $fromNamespaceField, '!=', $newLinkTarget->getNamespace() ) )
664 ->caller( __METHOD__ )
665 ->fetchResultSet();
666 if ( !$res ) {
667 continue;
670 $updateConds = [];
671 foreach ( $res as $row ) {
672 $updateConds[] = array_merge( [ $fromField => $id ], (array)$row );
674 $updateBatches = array_chunk( $updateConds, $updateRowsPerQuery );
675 foreach ( $updateBatches as $updateBatch ) {
676 $this->beginTransactionRound( __METHOD__ );
677 $dbw->newUpdateQueryBuilder()
678 ->update( $table )
679 ->set( [ $fromNamespaceField => $newLinkTarget->getNamespace() ] )
680 ->where( $dbw->factorConds( $updateBatch ) )
681 ->caller( __METHOD__ )
682 ->execute();
683 $this->commitTransactionRound( __METHOD__ );
687 return true;
691 * Determine if we can merge a page.
692 * We check if an inaccessible revision would become the latest and
693 * deny the merge if so -- it's theoretically possible to update the
694 * latest revision, but opens a can of worms -- search engine updates,
695 * recentchanges review, etc.
697 * @param int $id The page_id
698 * @param LinkTarget $linkTarget The new link target
699 * @param string &$logStatus This is set to the log status message on failure @phan-output-reference
700 * @return bool
702 private function canMerge( $id, LinkTarget $linkTarget, &$logStatus ) {
703 $revisionLookup = $this->getServiceContainer()->getRevisionLookup();
704 $latestDest = $revisionLookup->getRevisionByTitle( $linkTarget, 0,
705 IDBAccessObject::READ_LATEST );
706 $latestSource = $revisionLookup->getRevisionByPageId( $id, 0,
707 IDBAccessObject::READ_LATEST );
708 if ( $latestSource->getTimestamp() > $latestDest->getTimestamp() ) {
709 $logStatus = 'cannot merge since source is later';
710 return false;
711 } else {
712 return true;
717 * Merge page histories
719 * @param stdClass $row Page row
720 * @param Title $newTitle
721 * @return bool
723 private function mergePage( $row, Title $newTitle ) {
724 $updateRowsPerQuery = $this->getConfig()->get( MainConfigNames::UpdateRowsPerQuery );
726 $id = $row->page_id;
728 // Construct the WikiPage object we will need later, while the
729 // page_id still exists. Note that this cannot use makeTitleSafe(),
730 // we are deliberately constructing an invalid title.
731 $sourceTitle = Title::makeTitle( $row->page_namespace, $row->page_title );
732 $sourceTitle->resetArticleID( $id );
733 $wikiPage = $this->getServiceContainer()->getWikiPageFactory()->newFromTitle( $sourceTitle );
734 $wikiPage->loadPageData( IDBAccessObject::READ_LATEST );
735 $destId = $newTitle->getArticleID();
737 $dbw = $this->getPrimaryDB();
738 $this->beginTransactionRound( __METHOD__ );
739 $revIds = $dbw->newSelectQueryBuilder()
740 ->select( 'rev_id' )
741 ->from( 'revision' )
742 ->where( [ 'rev_page' => $id ] )
743 ->caller( __METHOD__ )
744 ->fetchFieldValues();
745 $updateBatches = array_chunk( array_map( 'intval', $revIds ), $updateRowsPerQuery );
746 foreach ( $updateBatches as $updateBatch ) {
747 $dbw->newUpdateQueryBuilder()
748 ->update( 'revision' )
749 ->set( [ 'rev_page' => $destId ] )
750 ->where( [ 'rev_id' => $updateBatch ] )
751 ->caller( __METHOD__ )
752 ->execute();
753 if ( count( $updateBatches ) > 1 ) {
754 $this->commitTransactionRound( __METHOD__ );
755 $this->beginTransactionRound( __METHOD__ );
758 $dbw->newDeleteQueryBuilder()
759 ->deleteFrom( 'page' )
760 ->where( [ 'page_id' => $id ] )
761 ->caller( __METHOD__ )
762 ->execute();
763 $this->commitTransactionRound( __METHOD__ );
765 /* Call LinksDeletionUpdate to delete outgoing links from the old title,
766 * and update category counts.
768 * Calling external code with a fake broken Title is a fairly dubious
769 * idea. It's necessary because it's quite a lot of code to duplicate,
770 * but that also makes it fragile since it would be easy for someone to
771 * accidentally introduce an assumption of title validity to the code we
772 * are calling.
774 DeferredUpdates::addUpdate( new LinksDeletionUpdate( $wikiPage ) );
775 DeferredUpdates::doUpdates();
777 return true;
781 // @codeCoverageIgnoreStart
782 $maintClass = NamespaceDupes::class;
783 require_once RUN_MAINTENANCE_IF_MAIN;
784 // @codeCoverageIgnoreEnd