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 use Wikimedia\Rdbms\IMaintainableDatabase;
  28
  29 /**
  30  * Look for duplicate user table entries and optionally prune them.
  31  *
  32  * This is still used by our MysqlUpdater at:
  33  * includes/installer/MysqlUpdater.php
  34  *
  35  * @ingroup Maintenance
  36  */
  37 class UserDupes {
  38         /**
  39          * @var IMaintainableDatabase
  40          */
  41         private $db;
  42         private $reassigned;
  43         private $trimmed;
  44         private $failed;
  45         private $outputCallback;
  46
  47         /**
  48          * @param IMaintainableDatabase &$database
  49          * @param callback $outputCallback
  50          */
  51         public function __construct( &$database, $outputCallback ) {
  52                 $this->db = $database;
  53                 $this->outputCallback = $outputCallback;
  54         }
  55
  56         /**
  57          * Output some text via the output callback provided
  58          * @param string $str Text to print
  59          */
  60         private function out( $str ) {
  61                 call_user_func( $this->outputCallback, $str );
  62         }
  63
  64         /**
  65          * Check if this database's user table has already had a unique
  66          * user_name index applied.
  67          * @return bool
  68          */
  69         public function hasUniqueIndex() {
  70                 $info = $this->db->indexInfo( 'user', 'user_name', __METHOD__ );
  71                 if ( !$info ) {
  72                         $this->out( "WARNING: doesn't seem to have user_name index at all!\n" );
  73
  74                         return false;
  75                 }
  76
  77                 # Confusingly, 'Non_unique' is 0 for *unique* indexes,
  78                 # and 1 for *non-unique* indexes. Pass the crack, MySQL,
  79                 # it's obviously some good stuff!
  80                 return ( $info[0]->Non_unique == 0 );
  81         }
  82
  83         /**
  84          * Checks the database for duplicate user account records
  85          * and remove them in preparation for application of a unique
  86          * index on the user_name field. Returns true if the table is
  87          * clean or if duplicates have been resolved automatically.
  88          *
  89          * May return false if there are unresolvable problems.
  90          * Status information will be echo'd to stdout.
  91          *
  92          * @return bool
  93          */
  94         public function clearDupes() {
  95                 return $this->checkDupes( true );
  96         }
  97
  98         /**
  99          * Checks the database for duplicate user account records
 100          * in preparation for application of a unique index on the
 101          * user_name field. Returns true if the table is clean or
 102          * if duplicates can be resolved automatically.
 103          *
 104          * Returns false if there are duplicates and resolution was
 105          * not requested. (If doing resolution, edits may be reassigned.)
 106          * Status information will be echo'd to stdout.
 107          *
 108          * @param bool $doDelete Pass true to actually remove things
 109          *   from the database; false to just check.
 110          * @return bool
 111          */
 112         private function checkDupes( $doDelete = false ) {
 113                 if ( $this->hasUniqueIndex() ) {
 114                         echo wfWikiID() . " already has a unique index on its user table.\n";
 115
 116                         return true;
 117                 }
 118
 119                 $this->lock();
 120
 121                 $this->out( "Checking for duplicate accounts...\n" );
 122                 $dupes = $this->getDupes();
 123                 $count = count( $dupes );
 124
 125                 $this->out( "Found $count accounts with duplicate records on " . wfWikiID() . ".\n" );
 126                 $this->trimmed = 0;
 127                 $this->reassigned = 0;
 128                 $this->failed = 0;
 129                 foreach ( $dupes as $name ) {
 130                         $this->examine( $name, $doDelete );
 131                 }
 132
 133                 $this->unlock();
 134
 135                 $this->out( "\n" );
 136
 137                 if ( $this->reassigned > 0 ) {
 138                         if ( $doDelete ) {
 139                                 $this->out( "$this->reassigned duplicate accounts had edits "
 140                                         . "reassigned to a canonical record id.\n" );
 141                         } else {
 142                                 $this->out( "$this->reassigned duplicate accounts need to have edits reassigned.\n" );
 143                         }
 144                 }
 145
 146                 if ( $this->trimmed > 0 ) {
 147                         if ( $doDelete ) {
 148                                 $this->out( "$this->trimmed duplicate user records were deleted from "
 149                                         . wfWikiID() . ".\n" );
 150                         } else {
 151                                 $this->out( "$this->trimmed duplicate user accounts were found on "
 152                                         . wfWikiID() . " which can be removed safely.\n" );
 153                         }
 154                 }
 155
 156                 if ( $this->failed > 0 ) {
 157                         $this->out( "Something terribly awry; $this->failed duplicate accounts were not removed.\n" );
 158
 159                         return false;
 160                 }
 161
 162                 if ( $this->trimmed == 0 || $doDelete ) {
 163                         $this->out( "It is now safe to apply the unique index on user_name.\n" );
 164
 165                         return true;
 166                 } else {
 167                         $this->out( "Run this script again with the --fix option to automatically delete them.\n" );
 168
 169                         return false;
 170                 }
 171         }
 172
 173         /**
 174          * We don't want anybody to mess with our stuff...
 175          */
 176         private function lock() {
 177                 $set = [ 'user', 'revision' ];
 178                 $names = array_map( [ $this, 'lockTable' ], $set );
 179                 $tables = implode( ',', $names );
 180
 181                 $this->db->query( "LOCK TABLES $tables", __METHOD__ );
 182         }
 183
 184         private function lockTable( $table ) {
 185                 return $this->db->tableName( $table ) . ' WRITE';
 186         }
 187
 188         /**
 189          * @private
 190          */
 191         private function unlock() {
 192                 $this->db->query( "UNLOCK TABLES", __METHOD__ );
 193         }
 194
 195         /**
 196          * Grab usernames for which multiple records are present in the database.
 197          * @return array
 198          */
 199         private function getDupes() {
 200                 $user = $this->db->tableName( 'user' );
 201                 $result = $this->db->query(
 202                         "SELECT user_name,COUNT(*) AS n
 203                                 FROM $user
 204                         GROUP BY user_name
 205                                 HAVING n > 1", __METHOD__ );
 206
 207                 $list = [];
 208                 foreach ( $result as $row ) {
 209                         $list[] = $row->user_name;
 210                 }
 211
 212                 return $list;
 213         }
 214
 215         /**
 216          * Examine user records for the given name. Try to see which record
 217          * will be the one that actually gets used, then check remaining records
 218          * for edits. If the dupes have no edits, we can safely remove them.
 219          * @param string $name
 220          * @param bool $doDelete
 221          */
 222         private function examine( $name, $doDelete ) {
 223                 $result = $this->db->select( 'user',
 224                         [ 'user_id' ],
 225                         [ 'user_name' => $name ],
 226                         __METHOD__ );
 227
 228                 $firstRow = $this->db->fetchObject( $result );
 229                 $firstId = $firstRow->user_id;
 230                 $this->out( "Record that will be used for '$name' is user_id=$firstId\n" );
 231
 232                 foreach ( $result as $row ) {
 233                         $dupeId = $row->user_id;
 234                         $this->out( "... dupe id $dupeId: " );
 235                         $edits = $this->editCount( $dupeId );
 236                         if ( $edits > 0 ) {
 237                                 $this->reassigned++;
 238                                 $this->out( "has $edits edits! " );
 239                                 if ( $doDelete ) {
 240                                         $this->reassignEdits( $dupeId, $firstId );
 241                                         $newEdits = $this->editCount( $dupeId );
 242                                         if ( $newEdits == 0 ) {
 243                                                 $this->out( "confirmed cleaned. " );
 244                                         } else {
 245                                                 $this->failed++;
 246                                                 $this->out( "WARNING! $newEdits remaining edits for $dupeId; NOT deleting user.\n" );
 247                                                 continue;
 248                                         }
 249                                 } else {
 250                                         $this->out( "(will need to reassign edits on fix)" );
 251                                 }
 252                         } else {
 253                                 $this->out( "ok, no edits. " );
 254                         }
 255                         $this->trimmed++;
 256                         if ( $doDelete ) {
 257                                 $this->trimAccount( $dupeId );
 258                         }
 259                         $this->out( "\n" );
 260                 }
 261         }
 262
 263         /**
 264          * Count the number of edits attributed to this user.
 265          * Does not currently check log table or other things
 266          * where it might show up...
 267          * @param int $userid
 268          * @return int
 269          */
 270         private function editCount( $userid ) {
 271                 return intval( $this->db->selectField(
 272                         'revision',
 273                         'COUNT(*)',
 274                         [ 'rev_user' => $userid ],
 275                         __METHOD__ ) );
 276         }
 277
 278         /**
 279          * @param int $from
 280          * @param int $to
 281          */
 282         private function reassignEdits( $from, $to ) {
 283                 $this->out( 'reassigning... ' );
 284                 $this->db->update( 'revision',
 285                         [ 'rev_user' => $to ],
 286                         [ 'rev_user' => $from ],
 287                         __METHOD__ );
 288                 $this->out( "ok. " );
 289         }
 290
 291         /**
 292          * Remove a user account line.
 293          * @param int $userid
 294          */
 295         private function trimAccount( $userid ) {
 296                 $this->out( "deleting..." );
 297                 $this->db->delete( 'user', [ 'user_id' => $userid ], __METHOD__ );
 298                 $this->out( " ok" );
 299         }
 300 }