maintenance/userDupes.inc

   1 <?php
   2 /**
   3  * Helper class for update.php.
   4  *
   5  * Copyright © 2005 Brion Vibber <brion@pobox.com>
   6  * https://www.mediawiki.org/
   7  *
   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.
  12  *
  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.
  17  *
  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
  22  *
  23  * @file
  24  * @ingroup Maintenance
  25  */
  26
  27 /**
  28  * Look for duplicate user table entries and optionally prune them.
  29  *
  30  * This is still used by our MysqlUpdater at:
  31  * includes/installer/MysqlUpdater.php
  32  *
  33  * @ingroup Maintenance
  34  */
  35 class UserDupes {
  36         private $db;
  37         private $reassigned;
  38         private $trimmed;
  39         private $failed;
  40         private $outputCallback;
  41
  42         function __construct( &$database, $outputCallback ) {
  43                 $this->db = $database;
  44                 $this->outputCallback = $outputCallback;
  45         }
  46
  47         /**
  48          * Output some text via the output callback provided
  49          * @param string $str Text to print
  50          */
  51         private function out( $str ) {
  52                 call_user_func( $this->outputCallback, $str );
  53         }
  54
  55         /**
  56          * Check if this database's user table has already had a unique
  57          * user_name index applied.
  58          * @return bool
  59          */
  60         function hasUniqueIndex() {
  61                 $info = $this->db->indexInfo( 'user', 'user_name', __METHOD__ );
  62                 if ( !$info ) {
  63                         $this->out( "WARNING: doesn't seem to have user_name index at all!\n" );
  64
  65                         return false;
  66                 }
  67
  68                 # Confusingly, 'Non_unique' is 0 for *unique* indexes,
  69                 # and 1 for *non-unique* indexes. Pass the crack, MySQL,
  70                 # it's obviously some good stuff!
  71                 return ( $info[0]->Non_unique == 0 );
  72         }
  73
  74         /**
  75          * Checks the database for duplicate user account records
  76          * and remove them in preparation for application of a unique
  77          * index on the user_name field. Returns true if the table is
  78          * clean or if duplicates have been resolved automatically.
  79          *
  80          * May return false if there are unresolvable problems.
  81          * Status information will be echo'd to stdout.
  82          *
  83          * @return bool
  84          */
  85         function clearDupes() {
  86                 return $this->checkDupes( true );
  87         }
  88
  89         /**
  90          * Checks the database for duplicate user account records
  91          * in preparation for application of a unique index on the
  92          * user_name field. Returns true if the table is clean or
  93          * if duplicates can be resolved automatically.
  94          *
  95          * Returns false if there are duplicates and resolution was
  96          * not requested. (If doing resolution, edits may be reassigned.)
  97          * Status information will be echo'd to stdout.
  98          *
  99          * @param bool $doDelete Pass true to actually remove things
 100          *   from the database; false to just check.
 101          * @return bool
 102          */
 103         function checkDupes( $doDelete = false ) {
 104                 if ( $this->hasUniqueIndex() ) {
 105                         echo wfWikiID() . " already has a unique index on its user table.\n";
 106
 107                         return true;
 108                 }
 109
 110                 $this->lock();
 111
 112                 $this->out( "Checking for duplicate accounts...\n" );
 113                 $dupes = $this->getDupes();
 114                 $count = count( $dupes );
 115
 116                 $this->out( "Found $count accounts with duplicate records on " . wfWikiID() . ".\n" );
 117                 $this->trimmed = 0;
 118                 $this->reassigned = 0;
 119                 $this->failed = 0;
 120                 foreach ( $dupes as $name ) {
 121                         $this->examine( $name, $doDelete );
 122                 }
 123
 124                 $this->unlock();
 125
 126                 $this->out( "\n" );
 127
 128                 if ( $this->reassigned > 0 ) {
 129                         if ( $doDelete ) {
 130                                 $this->out( "$this->reassigned duplicate accounts had edits "
 131                                         . "reassigned to a canonical record id.\n" );
 132                         } else {
 133                                 $this->out( "$this->reassigned duplicate accounts need to have edits reassigned.\n" );
 134                         }
 135                 }
 136
 137                 if ( $this->trimmed > 0 ) {
 138                         if ( $doDelete ) {
 139                                 $this->out( "$this->trimmed duplicate user records were deleted from "
 140                                         . wfWikiID() . ".\n" );
 141                         } else {
 142                                 $this->out( "$this->trimmed duplicate user accounts were found on "
 143                                         . wfWikiID() . " which can be removed safely.\n" );
 144                         }
 145                 }
 146
 147                 if ( $this->failed > 0 ) {
 148                         $this->out( "Something terribly awry; $this->failed duplicate accounts were not removed.\n" );
 149
 150                         return false;
 151                 }
 152
 153                 if ( $this->trimmed == 0 || $doDelete ) {
 154                         $this->out( "It is now safe to apply the unique index on user_name.\n" );
 155
 156                         return true;
 157                 } else {
 158                         $this->out( "Run this script again with the --fix option to automatically delete them.\n" );
 159
 160                         return false;
 161                 }
 162         }
 163
 164         /**
 165          * We don't want anybody to mess with our stuff...
 166          * @private
 167          */
 168         function lock() {
 169                 $set = [ 'user', 'revision' ];
 170                 $names = array_map( [ $this, 'lockTable' ], $set );
 171                 $tables = implode( ',', $names );
 172
 173                 $this->db->query( "LOCK TABLES $tables", __METHOD__ );
 174         }
 175
 176         function lockTable( $table ) {
 177                 return $this->db->tableName( $table ) . ' WRITE';
 178         }
 179
 180         /**
 181          * @private
 182          */
 183         function unlock() {
 184                 $this->db->query( "UNLOCK TABLES", __METHOD__ );
 185         }
 186
 187         /**
 188          * Grab usernames for which multiple records are present in the database.
 189          * @return array
 190          * @private
 191          */
 192         function getDupes() {
 193                 $user = $this->db->tableName( 'user' );
 194                 $result = $this->db->query(
 195                         "SELECT user_name,COUNT(*) AS n
 196                                 FROM $user
 197                         GROUP BY user_name
 198                                 HAVING n > 1", __METHOD__ );
 199
 200                 $list = [];
 201                 foreach ( $result as $row ) {
 202                         $list[] = $row->user_name;
 203                 }
 204
 205                 return $list;
 206         }
 207
 208         /**
 209          * Examine user records for the given name. Try to see which record
 210          * will be the one that actually gets used, then check remaining records
 211          * for edits. If the dupes have no edits, we can safely remove them.
 212          * @param string $name
 213          * @param bool $doDelete
 214          * @private
 215          */
 216         function examine( $name, $doDelete ) {
 217                 $result = $this->db->select( 'user',
 218                         [ 'user_id' ],
 219                         [ 'user_name' => $name ],
 220                         __METHOD__ );
 221
 222                 $firstRow = $this->db->fetchObject( $result );
 223                 $firstId = $firstRow->user_id;
 224                 $this->out( "Record that will be used for '$name' is user_id=$firstId\n" );
 225
 226                 foreach ( $result as $row ) {
 227                         $dupeId = $row->user_id;
 228                         $this->out( "... dupe id $dupeId: " );
 229                         $edits = $this->editCount( $dupeId );
 230                         if ( $edits > 0 ) {
 231                                 $this->reassigned++;
 232                                 $this->out( "has $edits edits! " );
 233                                 if ( $doDelete ) {
 234                                         $this->reassignEdits( $dupeId, $firstId );
 235                                         $newEdits = $this->editCount( $dupeId );
 236                                         if ( $newEdits == 0 ) {
 237                                                 $this->out( "confirmed cleaned. " );
 238                                         } else {
 239                                                 $this->failed++;
 240                                                 $this->out( "WARNING! $newEdits remaining edits for $dupeId; NOT deleting user.\n" );
 241                                                 continue;
 242                                         }
 243                                 } else {
 244                                         $this->out( "(will need to reassign edits on fix)" );
 245                                 }
 246                         } else {
 247                                 $this->out( "ok, no edits. " );
 248                         }
 249                         $this->trimmed++;
 250                         if ( $doDelete ) {
 251                                 $this->trimAccount( $dupeId );
 252                         }
 253                         $this->out( "\n" );
 254                 }
 255         }
 256
 257         /**
 258          * Count the number of edits attributed to this user.
 259          * Does not currently check log table or other things
 260          * where it might show up...
 261          * @param int $userid
 262          * @return int
 263          * @private
 264          */
 265         function editCount( $userid ) {
 266                 return intval( $this->db->selectField(
 267                         'revision',
 268                         'COUNT(*)',
 269                         [ 'rev_user' => $userid ],
 270                         __METHOD__ ) );
 271         }
 272
 273         /**
 274          * @param int $from
 275          * @param int $to
 276          * @private
 277          */
 278         function reassignEdits( $from, $to ) {
 279                 $this->out( 'reassigning... ' );
 280                 $this->db->update( 'revision',
 281                         [ 'rev_user' => $to ],
 282                         [ 'rev_user' => $from ],
 283                         __METHOD__ );
 284                 $this->out( "ok. " );
 285         }
 286
 287         /**
 288          * Remove a user account line.
 289          * @param int $userid
 290          * @private
 291          */
 292         function trimAccount( $userid ) {
 293                 $this->out( "deleting..." );
 294                 $this->db->delete( 'user', [ 'user_id' => $userid ], __METHOD__ );
 295                 $this->out( " ok" );
 296         }
 297 }