maintenance/namespaceDupes.php

   1 <?php
   2 /**
   3  * Check for articles to fix after adding/deleting namespaces
   4  *
   5  * Copyright © 2005-2007 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 require_once __DIR__ . '/Maintenance.php';
  28
  29 /**
  30  * Maintenance script that checks for articles to fix after
  31  * adding/deleting namespaces.
  32  *
  33  * @ingroup Maintenance
  34  */
  35 class NamespaceConflictChecker extends Maintenance {
  36
  37         /**
  38          * @var DatabaseBase
  39          */
  40         protected $db;
  41
  42         public function __construct() {
  43                 parent::__construct();
  44                 $this->mDescription = "";
  45                 $this->addOption( 'fix', 'Attempt to automatically fix errors' );
  46                 $this->addOption( 'suffix', "Dupes will be renamed with correct namespace with " .
  47                         "<text> appended after the article name", false, true );
  48                 $this->addOption( 'prefix', "Do an explicit check for the given title prefix " .
  49                         "appended after the article name", false, true );
  50         }
  51
  52         public function execute() {
  53                 $this->db = wfGetDB( DB_MASTER );
  54
  55                 $fix = $this->hasOption( 'fix' );
  56                 $suffix = $this->getOption( 'suffix', '' );
  57                 $prefix = $this->getOption( 'prefix', '' );
  58                 $key = intval( $this->getOption( 'key', 0 ) );
  59
  60                 if ( $prefix ) {
  61                         $retval = $this->checkPrefix( $key, $prefix, $fix, $suffix );
  62                 } else {
  63                         $retval = $this->checkAll( $fix, $suffix );
  64                 }
  65
  66                 if ( $retval ) {
  67                         $this->output( "\nLooks good!\n" );
  68                 } else {
  69                         $this->output( "\nOh noeees\n" );
  70                 }
  71         }
  72
  73         /**
  74          * @todo Document
  75          * @param bool $fix Whether or not to fix broken entries
  76          * @param string $suffix Suffix to append to renamed articles
  77          *
  78          * @return bool
  79          */
  80         private function checkAll( $fix, $suffix = '' ) {
  81                 global $wgContLang, $wgNamespaceAliases, $wgCapitalLinks;
  82
  83                 $spaces = array();
  84
  85                 // List interwikis first, so they'll be overridden
  86                 // by any conflicting local namespaces.
  87                 foreach ( $this->getInterwikiList() as $prefix ) {
  88                         $name = $wgContLang->ucfirst( $prefix );
  89                         $spaces[$name] = 0;
  90                 }
  91
  92                 // Now pull in all canonical and alias namespaces...
  93                 foreach ( MWNamespace::getCanonicalNamespaces() as $ns => $name ) {
  94                         // This includes $wgExtraNamespaces
  95                         if ( $name !== '' ) {
  96                                 $spaces[$name] = $ns;
  97                         }
  98                 }
  99                 foreach ( $wgContLang->getNamespaces() as $ns => $name ) {
 100                         if ( $name !== '' ) {
 101                                 $spaces[$name] = $ns;
 102                         }
 103                 }
 104                 foreach ( $wgNamespaceAliases as $name => $ns ) {
 105                         $spaces[$name] = $ns;
 106                 }
 107                 foreach ( $wgContLang->getNamespaceAliases() as $name => $ns ) {
 108                         $spaces[$name] = $ns;
 109                 }
 110
 111                 // We'll need to check for lowercase keys as well,
 112                 // since we're doing case-sensitive searches in the db.
 113                 foreach ( $spaces as $name => $ns ) {
 114                         $moreNames = array();
 115                         $moreNames[] = $wgContLang->uc( $name );
 116                         $moreNames[] = $wgContLang->ucfirst( $wgContLang->lc( $name ) );
 117                         $moreNames[] = $wgContLang->ucwords( $name );
 118                         $moreNames[] = $wgContLang->ucwords( $wgContLang->lc( $name ) );
 119                         $moreNames[] = $wgContLang->ucwordbreaks( $name );
 120                         $moreNames[] = $wgContLang->ucwordbreaks( $wgContLang->lc( $name ) );
 121                         if ( !$wgCapitalLinks ) {
 122                                 foreach ( $moreNames as $altName ) {
 123                                         $moreNames[] = $wgContLang->lcfirst( $altName );
 124                                 }
 125                                 $moreNames[] = $wgContLang->lcfirst( $name );
 126                         }
 127                         foreach ( array_unique( $moreNames ) as $altName ) {
 128                                 if ( $altName !== $name ) {
 129                                         $spaces[$altName] = $ns;
 130                                 }
 131                         }
 132                 }
 133
 134                 ksort( $spaces );
 135                 asort( $spaces );
 136
 137                 $ok = true;
 138                 foreach ( $spaces as $name => $ns ) {
 139                         $ok = $this->checkNamespace( $ns, $name, $fix, $suffix ) && $ok;
 140                 }
 141
 142                 return $ok;
 143         }
 144
 145         /**
 146          * Get the interwiki list
 147          *
 148          * @return array
 149          */
 150         private function getInterwikiList() {
 151                 $result = Interwiki::getAllPrefixes();
 152                 $prefixes = array();
 153                 foreach ( $result as $row ) {
 154                         $prefixes[] = $row['iw_prefix'];
 155                 }
 156
 157                 return $prefixes;
 158         }
 159
 160         /**
 161          * @todo Document
 162          * @param int $ns A namespace id
 163          * @param string $name
 164          * @param bool $fix Whether to fix broken entries
 165          * @param string $suffix Suffix to append to renamed articles
 166          * @return bool
 167          */
 168         private function checkNamespace( $ns, $name, $fix, $suffix = '' ) {
 169                 $conflicts = $this->getConflicts( $ns, $name );
 170                 $count = count( $conflicts );
 171                 if ( $count == 0 ) {
 172                         return true;
 173                 }
 174
 175                 $resolveableCount = 0;
 176
 177                 $ok = true;
 178                 foreach ( $conflicts as $row ) {
 179                         $resolvable = $this->reportConflict( $row, $suffix );
 180                         $ok = $ok && $resolvable;
 181
 182                         if ( $resolvable ) {
 183                                 $resolveableCount++;
 184                         }
 185
 186                         if ( $fix && ( $resolvable || $suffix != '' ) ) {
 187                                 $ok = $this->resolveConflict( $row, $resolvable, $suffix ) && $ok;
 188                         }
 189                 }
 190
 191                 $this->output( "{$count} conflicts. {$resolveableCount} are resolveable." );
 192
 193                 return $ok;
 194         }
 195
 196         /**
 197          * @todo Do this for real
 198          * @param int $key
 199          * @param string $prefix
 200          * @param bool $fix
 201          * @param string $suffix
 202          * @return bool
 203          */
 204         private function checkPrefix( $key, $prefix, $fix, $suffix = '' ) {
 205                 $this->output( "Checking prefix \"$prefix\" vs namespace $key\n" );
 206
 207                 return $this->checkNamespace( $key, $prefix, $fix, $suffix );
 208         }
 209
 210         /**
 211          * Find pages in mainspace that have a prefix of the new namespace
 212          * so we know titles that will need migrating
 213          *
 214          * @param int $ns Namespace id (id for new namespace?)
 215          * @param string $name Prefix that is being made a namespace
 216          *
 217          * @return array
 218          */
 219         private function getConflicts( $ns, $name ) {
 220                 $titleSql = "TRIM(LEADING {$this->db->addQuotes( "$name:" )} FROM page_title)";
 221                 if ( $ns == 0 ) {
 222                         // An interwiki; try an alternate encoding with '-' for ':'
 223                         $titleSql = $this->db->buildConcat( array(
 224                                 $this->db->addQuotes( "$name-" ),
 225                                 $titleSql,
 226                         ) );
 227                 }
 228
 229                 return iterator_to_array( $this->db->select( 'page',
 230                         array(
 231                                 'id' => 'page_id',
 232                                 'oldtitle' => 'page_title',
 233                                 'namespace' => $this->db->addQuotes( $ns ) . ' + page_namespace',
 234                                 'title' => $titleSql,
 235                                 'oldnamespace' => 'page_namespace',
 236                         ),
 237                         array(
 238                                 'page_namespace' => array( 0, 1 ),
 239                                 'page_title' . $this->db->buildLike( "$name:", $this->db->anyString() ),
 240                         ),
 241                         __METHOD__
 242                 ) );
 243         }
 244
 245         /**
 246          * Report any conflicts we find
 247          *
 248          * @param stdClass $row
 249          * @param string $suffix
 250          * @return bool
 251          */
 252         private function reportConflict( $row, $suffix ) {
 253                 $newTitle = Title::makeTitleSafe( $row->namespace, $row->title );
 254                 if ( is_null( $newTitle ) || !$newTitle->canExist() ) {
 255                         // Title is also an illegal title...
 256                         // For the moment we'll let these slide to cleanupTitles or whoever.
 257                         $this->output( sprintf( "... %d (%d,\"%s\")\n",
 258                                 $row->id,
 259                                 $row->oldnamespace,
 260                                 $row->oldtitle ) );
 261                         $this->output( "...  *** cannot resolve automatically; illegal title ***\n" );
 262
 263                         return false;
 264                 }
 265
 266                 $this->output( sprintf( "... %d (%d,\"%s\") -> (%d,\"%s\") [[%s]]\n",
 267                         $row->id,
 268                         $row->oldnamespace,
 269                         $row->oldtitle,
 270                         $newTitle->getNamespace(),
 271                         $newTitle->getDBkey(),
 272                         $newTitle->getPrefixedText() )
 273                 );
 274
 275                 $id = $newTitle->getArticleID();
 276                 if ( $id ) {
 277                         $this->output( "...  *** cannot resolve automatically; page exists with ID $id ***\n" );
 278
 279                         return false;
 280                 } else {
 281                         return true;
 282                 }
 283         }
 284
 285         /**
 286          * Resolve any conflicts
 287          *
 288          * @param stClass $row Row from the page table to fix
 289          * @param bool $resolvable
 290          * @param string $suffix Suffix to append to the fixed page
 291          * @return bool
 292          */
 293         private function resolveConflict( $row, $resolvable, $suffix ) {
 294                 if ( !$resolvable ) {
 295                         $this->output( "...  *** old title {$row->title}\n" );
 296                         while ( true ) {
 297                                 $row->title .= $suffix;
 298                                 $this->output( "...  *** new title {$row->title}\n" );
 299                                 $title = Title::makeTitleSafe( $row->namespace, $row->title );
 300                                 if ( !$title ) {
 301                                         $this->output( "... !!! invalid title\n" );
 302
 303                                         return false;
 304                                 }
 305                                 $id = $title->getArticleID();
 306                                 if ( $id ) {
 307                                         $this->output( "...  *** page exists with ID $id ***\n" );
 308                                 } else {
 309                                         break;
 310                                 }
 311                         }
 312                         $this->output( "...  *** using suffixed form [[" . $title->getPrefixedText() . "]] ***\n" );
 313                 }
 314                 $this->resolveConflictOn( $row, 'page', 'page' );
 315
 316                 return true;
 317         }
 318
 319         /**
 320          * Resolve a given conflict
 321          *
 322          * @param stdClass $row Row from the old broken entry
 323          * @param string $table Table to update
 324          * @param string $prefix Prefix for column name, like page or ar
 325          * @return bool
 326          */
 327         private function resolveConflictOn( $row, $table, $prefix ) {
 328                 $this->output( "... resolving on $table... " );
 329                 $newTitle = Title::makeTitleSafe( $row->namespace, $row->title );
 330                 $this->db->update( $table,
 331                         array(
 332                                 "{$prefix}_namespace" => $newTitle->getNamespace(),
 333                                 "{$prefix}_title" => $newTitle->getDBkey(),
 334                         ),
 335                         array(
 336                                 // "{$prefix}_namespace" => 0,
 337                                 // "{$prefix}_title" => $row->oldtitle,
 338                                 "{$prefix}_id" => $row->id,
 339                         ),
 340                         __METHOD__ );
 341                 $this->output( "ok.\n" );
 342
 343                 return true;
 344         }
 345 }
 346
 347 $maintClass = "NamespaceConflictChecker";
 348 require_once RUN_MAINTENANCE_IF_MAIN;