maintenance/orphans.php

   1 <?php
   2 /**
   3  * Look for 'orphan' revisions hooked to pages which don't exist and
   4  * 'childless' pages with no revisions.
   5  * Then, kill the poor widows and orphans.
   6  * Man this is depressing.
   7  *
   8  * Copyright © 2005 Brion Vibber <brion@pobox.com>
   9  * https://www.mediawiki.org/
  10  *
  11  * This program is free software; you can redistribute it and/or modify
  12  * it under the terms of the GNU General Public License as published by
  13  * the Free Software Foundation; either version 2 of the License, or
  14  * (at your option) any later version.
  15  *
  16  * This program is distributed in the hope that it will be useful,
  17  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  18  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  19  * GNU General Public License for more details.
  20  *
  21  * You should have received a copy of the GNU General Public License along
  22  * with this program; if not, write to the Free Software Foundation, Inc.,
  23  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  24  * http://www.gnu.org/copyleft/gpl.html
  25  *
  26  * @file
  27  * @author <brion@pobox.com>
  28  * @ingroup Maintenance
  29  */
  30
  31 require_once __DIR__ . '/Maintenance.php';
  32
  33 use MediaWiki\MediaWikiServices;
  34 use Wikimedia\Rdbms\IMaintainableDatabase;
  35
  36 /**
  37  * Maintenance script that looks for 'orphan' revisions hooked to pages which
  38  * don't exist and 'childless' pages with no revisions.
  39  *
  40  * @ingroup Maintenance
  41  */
  42 class Orphans extends Maintenance {
  43         public function __construct() {
  44                 parent::__construct();
  45                 $this->addDescription( "Look for 'orphan' revisions hooked to pages which don't exist\n" .
  46                         "and 'childless' pages with no revisions\n" .
  47                         "Then, kill the poor widows and orphans\n" .
  48                         "Man this is depressing"
  49                 );
  50                 $this->addOption( 'fix', 'Actually fix broken entries' );
  51         }
  52
  53         public function execute() {
  54                 $this->checkOrphans( $this->hasOption( 'fix' ) );
  55                 $this->checkSeparation( $this->hasOption( 'fix' ) );
  56                 # Does not work yet, do not use
  57                 # $this->checkWidows( $this->hasOption( 'fix' ) );
  58         }
  59
  60         /**
  61          * Lock the appropriate tables for the script
  62          * @param IMaintainableDatabase $db
  63          * @param string[] $extraTable The name of any extra tables to lock (eg: text)
  64          */
  65         private function lockTables( $db, $extraTable = [] ) {
  66                 $tbls = [ 'page', 'revision', 'redirect' ];
  67                 if ( $extraTable ) {
  68                         $tbls = array_merge( $tbls, $extraTable );
  69                 }
  70                 $db->lockTables( [], $tbls, __METHOD__, false );
  71         }
  72
  73         /**
  74          * Check for orphan revisions
  75          * @param bool $fix Whether to fix broken revisions when found
  76          */
  77         private function checkOrphans( $fix ) {
  78                 $dbw = $this->getDB( DB_MASTER );
  79                 $commentStore = CommentStore::getStore();
  80
  81                 if ( $fix ) {
  82                         $this->lockTables( $dbw );
  83                 }
  84
  85                 $commentQuery = $commentStore->getJoin( 'rev_comment' );
  86                 $actorQuery = ActorMigration::newMigration()->getJoin( 'rev_user' );
  87
  88                 $this->output( "Checking for orphan revision table entries... "
  89                         . "(this may take a while on a large wiki)\n" );
  90                 $result = $dbw->select(
  91                         [ 'revision', 'page' ] + $commentQuery['tables'] + $actorQuery['tables'],
  92                         [ 'rev_id', 'rev_page', 'rev_timestamp' ] + $commentQuery['fields'] + $actorQuery['fields'],
  93                         [ 'page_id' => null ],
  94                         __METHOD__,
  95                         [],
  96                         [ 'page' => [ 'LEFT JOIN', [ 'rev_page=page_id' ] ] ] + $commentQuery['joins']
  97                                 + $actorQuery['joins']
  98                 );
  99                 $orphans = $result->numRows();
 100                 if ( $orphans > 0 ) {
 101                         $this->output( "$orphans orphan revisions...\n" );
 102                         $this->output( sprintf(
 103                                 "%10s %10s %14s %20s %s\n",
 104                                 'rev_id', 'rev_page', 'rev_timestamp', 'rev_user_text', 'rev_comment'
 105                         ) );
 106
 107                         $contLang = MediaWikiServices::getInstance()->getContentLanguage();
 108                         foreach ( $result as $row ) {
 109                                 $comment = $commentStore->getComment( 'rev_comment', $row )->text;
 110                                 if ( $comment !== '' ) {
 111                                         $comment = '(' . $contLang->truncateForVisual( $comment, 40 ) . ')';
 112                                 }
 113                                 $rev_user_text = $contLang->truncateForVisual( $row->rev_user_text, 20 );
 114                                 # pad $rev_user_text to 20 characters.  Note that this may
 115                                 # yield poor results if $rev_user_text contains combining
 116                                 # or half-width characters.  Alas.
 117                                 if ( mb_strlen( $rev_user_text ) < 20 ) {
 118                                         $rev_user_text = str_repeat( ' ', 20 - mb_strlen( $rev_user_text ) );
 119                                 }
 120                                 $this->output( sprintf( "%10d %10d %14s %s %s\n",
 121                                         $row->rev_id,
 122                                         $row->rev_page,
 123                                         $row->rev_timestamp,
 124                                         $rev_user_text,
 125                                         $comment ) );
 126                                 if ( $fix ) {
 127                                         $dbw->delete( 'revision', [ 'rev_id' => $row->rev_id ] );
 128                                 }
 129                         }
 130                         if ( !$fix ) {
 131                                 $this->output( "Run again with --fix to remove these entries automatically.\n" );
 132                         }
 133                 } else {
 134                         $this->output( "No orphans! Yay!\n" );
 135                 }
 136
 137                 if ( $fix ) {
 138                         $dbw->unlockTables( __METHOD__ );
 139                 }
 140         }
 141
 142         /**
 143          * @param bool $fix
 144          * @todo DON'T USE THIS YET! It will remove entries which have children,
 145          *       but which aren't properly attached (eg if page_latest is bogus
 146          *       but valid revisions do exist)
 147          */
 148         private function checkWidows( $fix ) {
 149                 $dbw = $this->getDB( DB_MASTER );
 150                 $page = $dbw->tableName( 'page' );
 151                 $revision = $dbw->tableName( 'revision' );
 152
 153                 if ( $fix ) {
 154                         $this->lockTables( $dbw );
 155                 }
 156
 157                 $this->output( "\nChecking for childless page table entries... "
 158                         . "(this may take a while on a large wiki)\n" );
 159                 $result = $dbw->query( "
 160                         SELECT *
 161                         FROM $page LEFT OUTER JOIN $revision ON page_latest=rev_id
 162                         WHERE rev_id IS NULL
 163                 " );
 164                 $widows = $result->numRows();
 165                 if ( $widows > 0 ) {
 166                         $this->output( "$widows childless pages...\n" );
 167                         $this->output( sprintf( "%10s %11s %2s %s\n", 'page_id', 'page_latest', 'ns', 'page_title' ) );
 168                         foreach ( $result as $row ) {
 169                                 printf( "%10d %11d %2d %s\n",
 170                                         $row->page_id,
 171                                         $row->page_latest,
 172                                         $row->page_namespace,
 173                                         $row->page_title );
 174                                 if ( $fix ) {
 175                                         $dbw->delete( 'page', [ 'page_id' => $row->page_id ] );
 176                                 }
 177                         }
 178                         if ( !$fix ) {
 179                                 $this->output( "Run again with --fix to remove these entries automatically.\n" );
 180                         }
 181                 } else {
 182                         $this->output( "No childless pages! Yay!\n" );
 183                 }
 184
 185                 if ( $fix ) {
 186                         $dbw->unlockTables( __METHOD__ );
 187                 }
 188         }
 189
 190         /**
 191          * Check for pages where page_latest is wrong
 192          * @param bool $fix Whether to fix broken entries
 193          */
 194         private function checkSeparation( $fix ) {
 195                 $dbw = $this->getDB( DB_MASTER );
 196                 $page = $dbw->tableName( 'page' );
 197                 $revision = $dbw->tableName( 'revision' );
 198
 199                 if ( $fix ) {
 200                         $this->lockTables( $dbw, [ 'user', 'text' ] );
 201                 }
 202
 203                 $this->output( "\nChecking for pages whose page_latest links are incorrect... "
 204                         . "(this may take a while on a large wiki)\n" );
 205                 $result = $dbw->query( "
 206                         SELECT *
 207                         FROM $page LEFT OUTER JOIN $revision ON page_latest=rev_id
 208                 " );
 209                 $found = 0;
 210                 foreach ( $result as $row ) {
 211                         $result2 = $dbw->query( "
 212                                 SELECT MAX(rev_timestamp) as max_timestamp
 213                                 FROM $revision
 214                                 WHERE rev_page=" . (int)( $row->page_id )
 215                         );
 216                         $row2 = $dbw->fetchObject( $result2 );
 217                         if ( $row2 ) {
 218                                 if ( $row->rev_timestamp != $row2->max_timestamp ) {
 219                                         if ( $found == 0 ) {
 220                                                 $this->output( sprintf( "%10s %10s %14s %14s\n",
 221                                                         'page_id', 'rev_id', 'timestamp', 'max timestamp' ) );
 222                                         }
 223                                         ++$found;
 224                                         $this->output( sprintf( "%10d %10d %14s %14s\n",
 225                                                 $row->page_id,
 226                                                 $row->page_latest,
 227                                                 $row->rev_timestamp,
 228                                                 $row2->max_timestamp ) );
 229                                         if ( $fix ) {
 230                                                 # ...
 231                                                 $maxId = $dbw->selectField(
 232                                                         'revision',
 233                                                         'rev_id',
 234                                                         [
 235                                                                 'rev_page' => $row->page_id,
 236                                                                 'rev_timestamp' => $row2->max_timestamp ] );
 237                                                 $this->output( "... updating to revision $maxId\n" );
 238                                                 $maxRev = Revision::newFromId( $maxId );
 239                                                 $title = Title::makeTitle( $row->page_namespace, $row->page_title );
 240                                                 $article = WikiPage::factory( $title );
 241                                                 $article->updateRevisionOn( $dbw, $maxRev );
 242                                         }
 243                                 }
 244                         } else {
 245                                 $this->output( "wtf\n" );
 246                         }
 247                 }
 248
 249                 if ( $found ) {
 250                         $this->output( "Found $found pages with incorrect latest revision.\n" );
 251                 } else {
 252                         $this->output( "No pages with incorrect latest revision. Yay!\n" );
 253                 }
 254                 if ( !$fix && $found > 0 ) {
 255                         $this->output( "Run again with --fix to remove these entries automatically.\n" );
 256                 }
 257
 258                 if ( $fix ) {
 259                         $dbw->unlockTables( __METHOD__ );
 260                 }
 261         }
 262 }
 263
 264 $maintClass = Orphans::class;
 265 require_once RUN_MAINTENANCE_IF_MAIN;