maintenance/namespaceDupes.php

   1 <?php
   2 /**
   3  * Check for articles to fix after adding/deleting namespaces
   4  *
   5  * Copyright (C) 2005-2007 Brion Vibber <brion@pobox.com>
   6  * http://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  * @ingroup Maintenance
  24  */
  25
  26 require_once( dirname( __FILE__ ) . '/Maintenance.php' );
  27
  28 class NamespaceConflictChecker extends Maintenance {
  29
  30         /**
  31          * @var DatabaseBase
  32          */
  33         protected $db;
  34
  35         public function __construct() {
  36                 parent::__construct();
  37                 $this->mDescription = "";
  38                 $this->addOption( 'fix', 'Attempt to automatically fix errors' );
  39                 $this->addOption( 'suffix', "Dupes will be renamed with correct namespace with " .
  40                                                                         "<text> appended after the article name", false, true );
  41                 $this->addOption( 'prefix', "Do an explicit check for the given title prefix " .
  42                                                                         "appended after the article name", false, true );
  43         }
  44
  45         public function execute() {
  46                 global $wgTitle;
  47
  48                 $this->db = wfGetDB( DB_MASTER );
  49                 $wgTitle = Title::newFromText( 'Namespace title conflict cleanup script' );
  50
  51                 $fix = $this->hasOption( 'fix' );
  52                 $suffix = $this->getOption( 'suffix', '' );
  53                 $prefix = $this->getOption( 'prefix', '' );
  54                 $key = intval( $this->getOption( 'key', 0 ) );
  55
  56                 if ( $prefix ) {
  57                         $retval = $this->checkPrefix( $key, $prefix, $fix, $suffix );
  58                 } else {
  59                         $retval = $this->checkAll( $fix, $suffix );
  60                 }
  61
  62                 if ( $retval ) {
  63                         $this->output( "\nLooks good!\n" );
  64                 } else {
  65                         $this->output( "\nOh noeees\n" );
  66                 }
  67         }
  68
  69         /**
  70          * @todo Document
  71          * @param $fix Boolean: whether or not to fix broken entries
  72          * @param $suffix String: suffix to append to renamed articles
  73          *
  74          * @return bool
  75          */
  76         private function checkAll( $fix, $suffix = '' ) {
  77                 global $wgContLang, $wgNamespaceAliases, $wgCapitalLinks;
  78
  79                 $spaces = array();
  80
  81                 // List interwikis first, so they'll be overridden
  82                 // by any conflicting local namespaces.
  83                 foreach ( $this->getInterwikiList() as $prefix ) {
  84                         $name = $wgContLang->ucfirst( $prefix );
  85                         $spaces[$name] = 0;
  86                 }
  87
  88                 // Now pull in all canonical and alias namespaces...
  89                 foreach ( MWNamespace::getCanonicalNamespaces() as $ns => $name ) {
  90                         // This includes $wgExtraNamespaces
  91                         if ( $name !== '' ) {
  92                                 $spaces[$name] = $ns;
  93                         }
  94                 }
  95                 foreach ( $wgContLang->getNamespaces() as $ns => $name ) {
  96                         if ( $name !== '' ) {
  97                                 $spaces[$name] = $ns;
  98                         }
  99                 }
 100                 foreach ( $wgNamespaceAliases as $name => $ns ) {
 101                         $spaces[$name] = $ns;
 102                 }
 103                 foreach ( $wgContLang->getNamespaceAliases() as $name => $ns ) {
 104                         $spaces[$name] = $ns;
 105                 }
 106
 107                 // We'll need to check for lowercase keys as well,
 108                 // since we're doing case-sensitive searches in the db.
 109                 foreach ( $spaces as $name => $ns ) {
 110                         $moreNames = array();
 111                         $moreNames[] = $wgContLang->uc( $name );
 112                         $moreNames[] = $wgContLang->ucfirst( $wgContLang->lc( $name ) );
 113                         $moreNames[] = $wgContLang->ucwords( $name );
 114                         $moreNames[] = $wgContLang->ucwords( $wgContLang->lc( $name ) );
 115                         $moreNames[] = $wgContLang->ucwordbreaks( $name );
 116                         $moreNames[] = $wgContLang->ucwordbreaks( $wgContLang->lc( $name ) );
 117                         if ( !$wgCapitalLinks ) {
 118                                 foreach ( $moreNames as $altName ) {
 119                                         $moreNames[] = $wgContLang->lcfirst( $altName );
 120                                 }
 121                                 $moreNames[] = $wgContLang->lcfirst( $name );
 122                         }
 123                         foreach ( array_unique( $moreNames ) as $altName ) {
 124                                 if ( $altName !== $name ) {
 125                                         $spaces[$altName] = $ns;
 126                                 }
 127                         }
 128                 }
 129
 130                 ksort( $spaces );
 131                 asort( $spaces );
 132
 133                 $ok = true;
 134                 foreach ( $spaces as $name => $ns ) {
 135                         $ok = $this->checkNamespace( $ns, $name, $fix, $suffix ) && $ok;
 136                 }
 137                 return $ok;
 138         }
 139
 140         /**
 141          * Get the interwiki list
 142          *
 143          * @todo Needs to respect interwiki cache!
 144          * @return Array
 145          */
 146         private function getInterwikiList() {
 147                 $result = $this->db->select( 'interwiki', array( 'iw_prefix' ) );
 148                 $prefixes = array();
 149                 foreach ( $result as $row ) {
 150                         $prefixes[] = $row->iw_prefix;
 151                 }
 152                 return $prefixes;
 153         }
 154
 155         /**
 156          * @todo Document
 157          * @param $ns Integer: a namespace id
 158          * @param $name String
 159          * @param $fix Boolean: whether to fix broken entries
 160          * @param $suffix String: suffix to append to renamed articles
 161          * @return bool
 162          */
 163         private function checkNamespace( $ns, $name, $fix, $suffix = '' ) {
 164                 $conflicts = $this->getConflicts( $ns, $name );
 165                 $count = count( $conflicts );
 166                 if ( $count == 0 ) {
 167                         return true;
 168                 }
 169
 170                 $ok = true;
 171                 foreach ( $conflicts as $row ) {
 172                         $resolvable = $this->reportConflict( $row, $suffix );
 173                         $ok = $ok && $resolvable;
 174                         if ( $fix && ( $resolvable || $suffix != '' ) ) {
 175                                 $ok = $this->resolveConflict( $row, $resolvable, $suffix ) && $ok;
 176                         }
 177                 }
 178                 return $ok;
 179         }
 180
 181         /**
 182          * @todo: do this for reals
 183          * @param $key
 184          * @param $prefix
 185          * @param $fix
 186          * @param $suffix string
 187          * @return bool
 188          */
 189         private function checkPrefix( $key, $prefix, $fix, $suffix = '' ) {
 190                 $this->output( "Checking prefix \"$prefix\" vs namespace $key\n" );
 191                 return $this->checkNamespace( $key, $prefix, $fix, $suffix );
 192         }
 193
 194         /**
 195          * Find pages in mainspace that have a prefix of the new namespace
 196          * so we know titles that will need migrating
 197          *
 198          * @param $ns Integer: namespace id (id for new namespace?)
 199          * @param $name String: prefix that is being made a namespace
 200          *
 201          * @return array
 202          */
 203         private function getConflicts( $ns, $name ) {
 204                 $page  = 'page';
 205                 $table = $this->db->tableName( $page );
 206
 207                 $prefix     = $this->db->strencode( $name );
 208                 $encNamespace = $this->db->addQuotes( $ns );
 209
 210                 $titleSql = "TRIM(LEADING '$prefix:' FROM {$page}_title)";
 211                 if ( $ns == 0 ) {
 212                         // An interwiki; try an alternate encoding with '-' for ':'
 213                         $titleSql = $this->db->buildConcat( array( "'$prefix-'", $titleSql ) );
 214                 }
 215
 216                 $sql = "SELECT {$page}_id    AS id,
 217                                            {$page}_title AS oldtitle,
 218                                            $encNamespace + {$page}_namespace AS namespace,
 219                                    $titleSql     AS title,
 220                                    {$page}_namespace AS oldnamespace
 221                                   FROM {$table}
 222                                  WHERE ( {$page}_namespace=0 OR {$page}_namespace=1 )
 223                                    AND {$page}_title " . $this->db->buildLike( $name . ':', $this->db->anyString() );
 224
 225                 $result = $this->db->query( $sql, __METHOD__ );
 226
 227                 $set = array();
 228                 foreach ( $result as $row ) {
 229                         $set[] = $row;
 230                 }
 231                 return $set;
 232         }
 233
 234         /**
 235          * Report any conflicts we find
 236          *
 237          * @return bool
 238          */
 239         private function reportConflict( $row, $suffix ) {
 240                 $newTitle = Title::makeTitleSafe( $row->namespace, $row->title );
 241                 if ( is_null( $newTitle ) || !$newTitle->canExist() ) {
 242                         // Title is also an illegal title...
 243                         // For the moment we'll let these slide to cleanupTitles or whoever.
 244                         $this->output( sprintf( "... %d (%d,\"%s\")\n",
 245                                 $row->id,
 246                                 $row->oldnamespace,
 247                                 $row->oldtitle ) );
 248                         $this->output( "...  *** cannot resolve automatically; illegal title ***\n" );
 249                         return false;
 250                 }
 251
 252                 $this->output( sprintf( "... %d (%d,\"%s\") -> (%d,\"%s\") [[%s]]\n",
 253                         $row->id,
 254                         $row->oldnamespace,
 255                         $row->oldtitle,
 256                         $newTitle->getNamespace(),
 257                         $newTitle->getDBkey(),
 258                         $newTitle->getPrefixedText() ) );
 259
 260                 $id = $newTitle->getArticleID();
 261                 if ( $id ) {
 262                         $this->output( "...  *** cannot resolve automatically; page exists with ID $id ***\n" );
 263                         return false;
 264                 } else {
 265                         return true;
 266                 }
 267         }
 268
 269         /**
 270          * Resolve any conflicts
 271          *
 272          * @param $row Object: row from the page table to fix
 273          * @param $resolvable Boolean
 274          * @param $suffix String: suffix to append to the fixed page
 275          * @return bool
 276          */
 277         private function resolveConflict( $row, $resolvable, $suffix ) {
 278                 if ( !$resolvable ) {
 279                         $this->output( "...  *** old title {$row->title}\n" );
 280                         while ( true ) {
 281                                 $row->title .= $suffix;
 282                                 $this->output( "...  *** new title {$row->title}\n" );
 283                                 $title = Title::makeTitleSafe( $row->namespace, $row->title );
 284                                 if ( !$title ) {
 285                                         $this->output( "... !!! invalid title\n" );
 286                                         return false;
 287                                 }
 288                                 $id = $title->getArticleID();
 289                                 if ( $id ) {
 290                                         $this->output( "...  *** page exists with ID $id ***\n" );
 291                                 } else {
 292                                         break;
 293                                 }
 294                         }
 295                         $this->output( "...  *** using suffixed form [[" . $title->getPrefixedText() . "]] ***\n" );
 296                 }
 297                 $this->resolveConflictOn( $row, 'page', 'page' );
 298                 return true;
 299         }
 300
 301         /**
 302          * Resolve a given conflict
 303          *
 304          * @param $row Object: row from the old broken entry
 305          * @param $table String: table to update
 306          * @param $prefix String: prefix for column name, like page or ar
 307          * @return bool
 308          */
 309         private function resolveConflictOn( $row, $table, $prefix ) {
 310                 $this->output( "... resolving on $table... " );
 311                 $newTitle = Title::makeTitleSafe( $row->namespace, $row->title );
 312                 $this->db->update( $table,
 313                         array(
 314                                 "{$prefix}_namespace" => $newTitle->getNamespace(),
 315                                 "{$prefix}_title"     => $newTitle->getDBkey(),
 316                         ),
 317                         array(
 318                                 // "{$prefix}_namespace" => 0,
 319                                 // "{$prefix}_title"     => $row->oldtitle,
 320                                 "{$prefix}_id"           => $row->id,
 321                         ),
 322                         __METHOD__ );
 323                 $this->output( "ok.\n" );
 324                 return true;
 325         }
 326 }
 327
 328 $maintClass = "NamespaceConflictChecker";
 329 require_once( RUN_MAINTENANCE_IF_MAIN );