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