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                 $dbDomain = WikiMap::getCurrentWikiDbDomain()->getId();
 114                 if ( $this->hasUniqueIndex() ) {
 115                         echo "$dbDomain 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 $dbDomain.\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                                 $this->out(
 150                                         "$this->trimmed duplicate user records were deleted from $dbDomain.\n" );
 151                         } else {
 152                                 $this->out(
 153                                         "$this->trimmed duplicate user accounts were found on $dbDomain " .
 154                                         "which can be removed safely.\n"
 155                                 );
 156                         }
 157                 }
 158
 159                 if ( $this->failed > 0 ) {
 160                         $this->out( "Something terribly awry; $this->failed duplicate accounts were not removed.\n" );
 161
 162                         return false;
 163                 }
 164
 165                 if ( $this->trimmed == 0 || $doDelete ) {
 166                         $this->out( "It is now safe to apply the unique index on user_name.\n" );
 167
 168                         return true;
 169                 } else {
 170                         $this->out( "Run this script again with the --fix option to automatically delete them.\n" );
 171
 172                         return false;
 173                 }
 174         }
 175
 176         /**
 177          * We don't want anybody to mess with our stuff...
 178          */
 179         private function lock() {
 180                 $set = [ 'user', 'revision' ];
 181                 $names = array_map( [ $this, 'lockTable' ], $set );
 182                 $tables = implode( ',', $names );
 183
 184                 $this->db->query( "LOCK TABLES $tables", __METHOD__ );
 185         }
 186
 187         private function lockTable( $table ) {
 188                 return $this->db->tableName( $table ) . ' WRITE';
 189         }
 190
 191         /**
 192          * @private
 193          */
 194         private function unlock() {
 195                 $this->db->query( "UNLOCK TABLES", __METHOD__ );
 196         }
 197
 198         /**
 199          * Grab usernames for which multiple records are present in the database.
 200          * @return array
 201          */
 202         private function getDupes() {
 203                 $user = $this->db->tableName( 'user' );
 204                 $result = $this->db->query(
 205                         "SELECT user_name,COUNT(*) AS n
 206                                 FROM $user
 207                         GROUP BY user_name
 208                                 HAVING n > 1", __METHOD__ );
 209
 210                 $list = [];
 211                 foreach ( $result as $row ) {
 212                         $list[] = $row->user_name;
 213                 }
 214
 215                 return $list;
 216         }
 217
 218         /**
 219          * Examine user records for the given name. Try to see which record
 220          * will be the one that actually gets used, then check remaining records
 221          * for edits. If the dupes have no edits, we can safely remove them.
 222          * @param string $name
 223          * @param bool $doDelete
 224          */
 225         private function examine( $name, $doDelete ) {
 226                 $result = $this->db->select( 'user',
 227                         [ 'user_id' ],
 228                         [ 'user_name' => $name ],
 229                         __METHOD__ );
 230
 231                 $firstRow = $this->db->fetchObject( $result );
 232                 $firstId = $firstRow->user_id;
 233                 $this->out( "Record that will be used for '$name' is user_id=$firstId\n" );
 234
 235                 foreach ( $result as $row ) {
 236                         $dupeId = $row->user_id;
 237                         $this->out( "... dupe id $dupeId: " );
 238                         $edits = $this->editCount( $dupeId );
 239                         if ( $edits > 0 ) {
 240                                 $this->reassigned++;
 241                                 $this->out( "has $edits edits! " );
 242                                 if ( $doDelete ) {
 243                                         $this->reassignEdits( $dupeId, $firstId );
 244                                         $newEdits = $this->editCount( $dupeId );
 245                                         if ( $newEdits == 0 ) {
 246                                                 $this->out( "confirmed cleaned. " );
 247                                         } else {
 248                                                 $this->failed++;
 249                                                 $this->out( "WARNING! $newEdits remaining edits for $dupeId; NOT deleting user.\n" );
 250                                                 continue;
 251                                         }
 252                                 } else {
 253                                         $this->out( "(will need to reassign edits on fix)" );
 254                                 }
 255                         } else {
 256                                 $this->out( "ok, no edits. " );
 257                         }
 258                         $this->trimmed++;
 259                         if ( $doDelete ) {
 260                                 $this->trimAccount( $dupeId );
 261                         }
 262                         $this->out( "\n" );
 263                 }
 264         }
 265
 266         /**
 267          * Count the number of edits attributed to this user.
 268          * Does not currently check log table or other things
 269          * where it might show up...
 270          * @param int $userid
 271          * @return int
 272          */
 273         private function editCount( $userid ) {
 274                 return intval( $this->db->selectField(
 275                         'revision',
 276                         'COUNT(*)',
 277                         [ 'rev_user' => $userid ],
 278                         __METHOD__ ) );
 279         }
 280
 281         /**
 282          * @param int $from
 283          * @param int $to
 284          */
 285         private function reassignEdits( $from, $to ) {
 286                 $this->out( 'reassigning... ' );
 287                 $this->db->update( 'revision',
 288                         [ 'rev_user' => $to ],
 289                         [ 'rev_user' => $from ],
 290                         __METHOD__ );
 291                 $this->out( "ok. " );
 292         }
 293
 294         /**
 295          * Remove a user account line.
 296          * @param int $userid
 297          */
 298         private function trimAccount( $userid ) {
 299                 $this->out( "deleting..." );
 300                 $this->db->delete( 'user', [ 'user_id' => $userid ], __METHOD__ );
 301                 $this->out( " ok" );
 302         }
 303 }