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 $outputCallback;
  46
  47         /**
  48          * UserDupes constructor.
  49          * @param IMaintainableDatabase &$database
  50          * @param callback $outputCallback
  51          */
  52         public function __construct( &$database, $outputCallback ) {
  53                 $this->db = $database;
  54                 $this->outputCallback = $outputCallback;
  55         }
  56
  57         /**
  58          * Output some text via the output callback provided
  59          * @param string $str Text to print
  60          */
  61         private function out( $str ) {
  62                 call_user_func( $this->outputCallback, $str );
  63         }
  64
  65         /**
  66          * Check if this database's user table has already had a unique
  67          * user_name index applied.
  68          * @return bool
  69          */
  70         public function hasUniqueIndex() {
  71                 $info = $this->db->indexInfo( 'user', 'user_name', __METHOD__ );
  72                 if ( !$info ) {
  73                         $this->out( "WARNING: doesn't seem to have user_name index at all!\n" );
  74
  75                         return false;
  76                 }
  77
  78                 # Confusingly, 'Non_unique' is 0 for *unique* indexes,
  79                 # and 1 for *non-unique* indexes. Pass the crack, MySQL,
  80                 # it's obviously some good stuff!
  81                 return ( $info[0]->Non_unique == 0 );
  82         }
  83
  84         /**
  85          * Checks the database for duplicate user account records
  86          * and remove them in preparation for application of a unique
  87          * index on the user_name field. Returns true if the table is
  88          * clean or if duplicates have been resolved automatically.
  89          *
  90          * May return false if there are unresolvable problems.
  91          * Status information will be echo'd to stdout.
  92          *
  93          * @return bool
  94          */
  95         public function clearDupes() {
  96                 return $this->checkDupes( true );
  97         }
  98
  99         /**
 100          * Checks the database for duplicate user account records
 101          * in preparation for application of a unique index on the
 102          * user_name field. Returns true if the table is clean or
 103          * if duplicates can be resolved automatically.
 104          *
 105          * Returns false if there are duplicates and resolution was
 106          * not requested. (If doing resolution, edits may be reassigned.)
 107          * Status information will be echo'd to stdout.
 108          *
 109          * @param bool $doDelete Pass true to actually remove things
 110          *   from the database; false to just check.
 111          * @return bool
 112          */
 113         private function checkDupes( $doDelete = false ) {
 114                 if ( $this->hasUniqueIndex() ) {
 115                         echo wfWikiID() . " already has a unique index on its user table.\n";
 116
 117                         return true;
 118                 }
 119
 120                 $this->lock();
 121
 122                 $this->out( "Checking for duplicate accounts...\n" );
 123                 $dupes = $this->getDupes();
 124                 $count = count( $dupes );
 125
 126                 $this->out( "Found $count accounts with duplicate records on " . wfWikiID() . ".\n" );
 127                 $this->trimmed = 0;
 128                 $this->reassigned = 0;
 129                 $this->failed = 0;
 130                 foreach ( $dupes as $name ) {
 131                         $this->examine( $name, $doDelete );
 132                 }
 133
 134                 $this->unlock();
 135
 136                 $this->out( "\n" );
 137
 138                 if ( $this->reassigned > 0 ) {
 139                         if ( $doDelete ) {
 140                                 $this->out( "$this->reassigned duplicate accounts had edits "
 141                                         . "reassigned to a canonical record id.\n" );
 142                         } else {
 143                                 $this->out( "$this->reassigned duplicate accounts need to have edits reassigned.\n" );
 144                         }
 145                 }
 146
 147                 if ( $this->trimmed > 0 ) {
 148                         if ( $doDelete ) {
 149                                 $this->out( "$this->trimmed duplicate user records were deleted from "
 150                                         . wfWikiID() . ".\n" );
 151                         } else {
 152                                 $this->out( "$this->trimmed duplicate user accounts were found on "
 153                                         . wfWikiID() . " which can be removed safely.\n" );
 154                         }
 155                 }
 156
 157                 if ( $this->failed > 0 ) {
 158                         $this->out( "Something terribly awry; $this->failed duplicate accounts were not removed.\n" );
 159
 160                         return false;
 161                 }
 162
 163                 if ( $this->trimmed == 0 || $doDelete ) {
 164                         $this->out( "It is now safe to apply the unique index on user_name.\n" );
 165
 166                         return true;
 167                 } else {
 168                         $this->out( "Run this script again with the --fix option to automatically delete them.\n" );
 169
 170                         return false;
 171                 }
 172         }
 173
 174         /**
 175          * We don't want anybody to mess with our stuff...
 176          */
 177         private function lock() {
 178                 $set = [ 'user', 'revision' ];
 179                 $names = array_map( [ $this, 'lockTable' ], $set );
 180                 $tables = implode( ',', $names );
 181
 182                 $this->db->query( "LOCK TABLES $tables", __METHOD__ );
 183         }
 184
 185         private function lockTable( $table ) {
 186                 return $this->db->tableName( $table ) . ' WRITE';
 187         }
 188
 189         /**
 190          * @private
 191          */
 192         private function unlock() {
 193                 $this->db->query( "UNLOCK TABLES", __METHOD__ );
 194         }
 195
 196         /**
 197          * Grab usernames for which multiple records are present in the database.
 198          * @return array
 199          */
 200         private function getDupes() {
 201                 $user = $this->db->tableName( 'user' );
 202                 $result = $this->db->query(
 203                         "SELECT user_name,COUNT(*) AS n
 204                                 FROM $user
 205                         GROUP BY user_name
 206                                 HAVING n > 1", __METHOD__ );
 207
 208                 $list = [];
 209                 foreach ( $result as $row ) {
 210                         $list[] = $row->user_name;
 211                 }
 212
 213                 return $list;
 214         }
 215
 216         /**
 217          * Examine user records for the given name. Try to see which record
 218          * will be the one that actually gets used, then check remaining records
 219          * for edits. If the dupes have no edits, we can safely remove them.
 220          * @param string $name
 221          * @param bool $doDelete
 222          */
 223         private function examine( $name, $doDelete ) {
 224                 $result = $this->db->select( 'user',
 225                         [ 'user_id' ],
 226                         [ 'user_name' => $name ],
 227                         __METHOD__ );
 228
 229                 $firstRow = $this->db->fetchObject( $result );
 230                 $firstId = $firstRow->user_id;
 231                 $this->out( "Record that will be used for '$name' is user_id=$firstId\n" );
 232
 233                 foreach ( $result as $row ) {
 234                         $dupeId = $row->user_id;
 235                         $this->out( "... dupe id $dupeId: " );
 236                         $edits = $this->editCount( $dupeId );
 237                         if ( $edits > 0 ) {
 238                                 $this->reassigned++;
 239                                 $this->out( "has $edits edits! " );
 240                                 if ( $doDelete ) {
 241                                         $this->reassignEdits( $dupeId, $firstId );
 242                                         $newEdits = $this->editCount( $dupeId );
 243                                         if ( $newEdits == 0 ) {
 244                                                 $this->out( "confirmed cleaned. " );
 245                                         } else {
 246                                                 $this->failed++;
 247                                                 $this->out( "WARNING! $newEdits remaining edits for $dupeId; NOT deleting user.\n" );
 248                                                 continue;
 249                                         }
 250                                 } else {
 251                                         $this->out( "(will need to reassign edits on fix)" );
 252                                 }
 253                         } else {
 254                                 $this->out( "ok, no edits. " );
 255                         }
 256                         $this->trimmed++;
 257                         if ( $doDelete ) {
 258                                 $this->trimAccount( $dupeId );
 259                         }
 260                         $this->out( "\n" );
 261                 }
 262         }
 263
 264         /**
 265          * Count the number of edits attributed to this user.
 266          * Does not currently check log table or other things
 267          * where it might show up...
 268          * @param int $userid
 269          * @return int
 270          */
 271         private function editCount( $userid ) {
 272                 return intval( $this->db->selectField(
 273                         'revision',
 274                         'COUNT(*)',
 275                         [ 'rev_user' => $userid ],
 276                         __METHOD__ ) );
 277         }
 278
 279         /**
 280          * @param int $from
 281          * @param int $to
 282          */
 283         private function reassignEdits( $from, $to ) {
 284                 $this->out( 'reassigning... ' );
 285                 $this->db->update( 'revision',
 286                         [ 'rev_user' => $to ],
 287                         [ 'rev_user' => $from ],
 288                         __METHOD__ );
 289                 $this->out( "ok. " );
 290         }
 291
 292         /**
 293          * Remove a user account line.
 294          * @param int $userid
 295          */
 296         private function trimAccount( $userid ) {
 297                 $this->out( "deleting..." );
 298                 $this->db->delete( 'user', [ 'user_id' => $userid ], __METHOD__ );
 299                 $this->out( " ok" );
 300         }
 301 }