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