maintenance/namespaceDupes.php

   1 <?php
   2 # Copyright (C) 2005-2007 Brion Vibber <brion@pobox.com>
   3 # http://www.mediawiki.org/
   4 #
   5 # This program is free software; you can redistribute it and/or modify
   6 # it under the terms of the GNU General Public License as published by
   7 # the Free Software Foundation; either version 2 of the License, or
   8 # (at your option) any later version.
   9 #
  10 # This program is distributed in the hope that it will be useful,
  11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13 # GNU General Public License for more details.
  14 #
  15 # You should have received a copy of the GNU General Public License along
  16 # with this program; if not, write to the Free Software Foundation, Inc.,
  17 # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18 # http://www.gnu.org/copyleft/gpl.html
  19
  20 /**
  21  * @file
  22  * @ingroup Maintenance
  23  */
  24
  25 $options = array( 'fix', 'suffix', 'help' );
  26
  27 /** */
  28 require_once( 'commandLine.inc' );
  29
  30 if(isset( $options['help'] ) ) {
  31 print <<<ENDS
  32 usage: namespaceDupes.php [--fix] [--suffix=<text>] [--help]
  33     --help          : this help message
  34     --fix           : attempt to automatically fix errors
  35     --suffix=<text> : dupes will be renamed with correct namespace with <text>
  36                       appended after the article name.
  37     --prefix=<text> : Do an explicit check for the given title prefix
  38                       in place of the standard namespace list.
  39     --verbose       : Display output for checked namespaces without conflicts
  40
  41 ENDS;
  42 die;
  43 }
  44
  45 class NamespaceConflictChecker {
  46         function NamespaceConflictChecker( $db, $verbose=false ) {
  47                 $this->db = $db;
  48                 $this->verbose = $verbose;
  49         }
  50
  51         function checkAll( $fix, $suffix = '' ) {
  52                 global $wgContLang, $wgNamespaceAliases, $wgCanonicalNamespaceNames;
  53                 global $wgCapitalLinks;
  54
  55                 $spaces = array();
  56
  57                 // List interwikis first, so they'll be overridden
  58                 // by any conflicting local namespaces.
  59                 foreach( $this->getInterwikiList() as $prefix ) {
  60                         $name = $wgContLang->ucfirst( $prefix );
  61                         $spaces[$name] = 0;
  62                 }
  63
  64                 // Now pull in all canonical and alias namespaces...
  65                 foreach( $wgCanonicalNamespaceNames as $ns => $name ) {
  66                         // This includes $wgExtraNamespaces
  67                         if( $name !== '' ) {
  68                                 $spaces[$name] = $ns;
  69                         }
  70                 }
  71                 foreach( $wgContLang->getNamespaces() as $ns => $name ) {
  72                         if( $name !== '' ) {
  73                                 $spaces[$name] = $ns;
  74                         }
  75                 }
  76                 foreach( $wgNamespaceAliases as $name => $ns ) {
  77                         $spaces[$name] = $ns;
  78                 }
  79                 foreach( $wgContLang->namespaceAliases as $name => $ns ) {
  80                         $spaces[$name] = $ns;
  81                 }
  82
  83                 // We'll need to check for lowercase keys as well,
  84                 // since we're doing case-sensitive searches in the db.
  85                 foreach( $spaces as $name => $ns ) {
  86                         $moreNames = array();
  87                         $moreNames[] = $wgContLang->uc( $name );
  88                         $moreNames[] = $wgContLang->ucfirst( $wgContLang->lc( $name ) );
  89                         $moreNames[] = $wgContLang->ucwords( $name );
  90                         $moreNames[] = $wgContLang->ucwords( $wgContLang->lc( $name ) );
  91                         $moreNames[] = $wgContLang->ucwordbreaks( $name );
  92                         $moreNames[] = $wgContLang->ucwordbreaks( $wgContLang->lc( $name ) );
  93                         if( !$wgCapitalLinks ) {
  94                                 foreach( $moreNames as $altName ) {
  95                                         $moreNames[] = $wgContLang->lcfirst( $altName );
  96                                 }
  97                                 $moreNames[] = $wgContLang->lcfirst( $name );
  98                         }
  99                         foreach( array_unique( $moreNames ) as $altName ) {
 100                                 if( $altName !== $name ) {
 101                                         $spaces[$altName] = $ns;
 102                                 }
 103                         }
 104                 }
 105
 106                 ksort( $spaces );
 107                 asort( $spaces );
 108
 109                 $ok = true;
 110                 foreach( $spaces as $name => $ns ) {
 111                         $ok = $this->checkNamespace( $ns, $name, $fix, $suffix ) && $ok;
 112                 }
 113                 return $ok;
 114         }
 115
 116         private function getInterwikiList() {
 117                 $result = $this->db->select( 'interwiki', array( 'iw_prefix' ) );
 118                 while( $row = $this->db->fetchObject( $result ) ) {
 119                         $prefixes[] = $row->iw_prefix;
 120                 }
 121                 $this->db->freeResult( $result );
 122                 return $prefixes;
 123         }
 124
 125         function checkNamespace( $ns, $name, $fix, $suffix = '' ) {
 126                 if( $ns == 0 ) {
 127                         $header = "Checking interwiki prefix: \"$name\"\n";
 128                 } else {
 129                         $header = "Checking namespace $ns: \"$name\"\n";
 130                 }
 131
 132                 $conflicts = $this->getConflicts( $ns, $name );
 133                 $count = count( $conflicts );
 134                 if( $count == 0 ) {
 135                         if( $this->verbose ) {
 136                                 echo $header;
 137                                 echo "... no conflicts detected!\n";
 138                         }
 139                         return true;
 140                 }
 141
 142                 echo $header;
 143                 echo "... $count conflicts detected:\n";
 144                 $ok = true;
 145                 foreach( $conflicts as $row ) {
 146                         $resolvable = $this->reportConflict( $row, $suffix );
 147                         $ok = $ok && $resolvable;
 148                         if( $fix && ( $resolvable || $suffix != '' ) ) {
 149                                 $ok = $this->resolveConflict( $row, $resolvable, $suffix ) && $ok;
 150                         }
 151                 }
 152                 return $ok;
 153         }
 154
 155         /**
 156          * @todo: do this for reals
 157          */
 158         function checkPrefix( $key, $prefix, $fix, $suffix = '' ) {
 159                 echo "Checking prefix \"$prefix\" vs namespace $key\n";
 160                 return $this->checkNamespace( $key, $prefix, $fix, $suffix );
 161         }
 162
 163         function getConflicts( $ns, $name ) {
 164                 $page  = 'page';
 165                 $table = $this->db->tableName( $page );
 166
 167                 $prefix     = $this->db->strencode( $name );
 168                 $likeprefix = str_replace( '_', '\\_', $prefix);
 169                 $encNamespace = $this->db->addQuotes( $ns );
 170
 171                 $titleSql = "TRIM(LEADING '$prefix:' FROM {$page}_title)";
 172                 if( $ns == 0 ) {
 173                         // An interwiki; try an alternate encoding with '-' for ':'
 174                         $titleSql = "CONCAT('$prefix-',$titleSql)";
 175                 }
 176
 177                 $sql = "SELECT {$page}_id    AS id,
 178                                {$page}_title AS oldtitle,
 179                                $encNamespace AS namespace,
 180                                $titleSql     AS title
 181                           FROM {$table}
 182                          WHERE {$page}_namespace=0
 183                            AND {$page}_title LIKE '$likeprefix:%'";
 184
 185                 $result = $this->db->query( $sql, 'NamespaceConflictChecker::getConflicts' );
 186
 187                 $set = array();
 188                 while( $row = $this->db->fetchObject( $result ) ) {
 189                         $set[] = $row;
 190                 }
 191                 $this->db->freeResult( $result );
 192
 193                 return $set;
 194         }
 195
 196         function reportConflict( $row, $suffix ) {
 197                 $newTitle = Title::makeTitleSafe( $row->namespace, $row->title );
 198                 if( is_null($newTitle) || !$newTitle->canExist() ) {
 199                         // Title is also an illegal title...
 200                         // For the moment we'll let these slide to cleanupTitles or whoever.
 201                         printf( "... %d (0,\"%s\")\n",
 202                                 $row->id,
 203                                 $row->oldtitle );
 204                         echo "...  *** cannot resolve automatically; illegal title ***\n";
 205                         return false;
 206                 }
 207
 208                 printf( "... %d (0,\"%s\") -> (%d,\"%s\") [[%s]]\n",
 209                         $row->id,
 210                         $row->oldtitle,
 211                         $newTitle->getNamespace(),
 212                         $newTitle->getDBkey(),
 213                         $newTitle->getPrefixedText() );
 214
 215                 $id = $newTitle->getArticleId();
 216                 if( $id ) {
 217                         echo "...  *** cannot resolve automatically; page exists with ID $id ***\n";
 218                         return false;
 219                 } else {
 220                         return true;
 221                 }
 222         }
 223
 224         function resolveConflict( $row, $resolvable, $suffix ) {
 225                 if( !$resolvable ) {
 226                         echo "...  *** old title {$row->title}\n";
 227                         while( true ) {
 228                                 $row->title .= $suffix;
 229                                 echo "...  *** new title {$row->title}\n";
 230                                 $title = Title::makeTitleSafe( $row->namespace, $row->title );
 231                                 if ( ! $title ) {
 232                                         echo "... !!! invalid title\n";
 233                                         return false;
 234                                 }
 235                                 if ( $id = $title->getArticleId() ) {
 236                                         echo "...  *** page exists with ID $id ***\n";
 237                                 } else {
 238                                         break;
 239                                 }
 240                         }
 241                         echo "...  *** using suffixed form [[" . $title->getPrefixedText() . "]] ***\n";
 242                 }
 243                 $tables = array( 'page' );
 244                 foreach( $tables as $table ) {
 245                         $this->resolveConflictOn( $row, $table );
 246                 }
 247                 return true;
 248         }
 249
 250         function resolveConflictOn( $row, $table ) {
 251                 echo "... resolving on $table... ";
 252                 $newTitle = Title::makeTitleSafe( $row->namespace, $row->title );
 253                 $this->db->update( $table,
 254                         array(
 255                                 "{$table}_namespace" => $newTitle->getNamespace(),
 256                                 "{$table}_title"     => $newTitle->getDBkey(),
 257                         ),
 258                         array(
 259                                 "{$table}_namespace" => 0,
 260                                 "{$table}_title"     => $row->oldtitle,
 261                         ),
 262                         __METHOD__ );
 263                 echo "ok.\n";
 264                 return true;
 265         }
 266 }
 267
 268
 269
 270
 271 $wgTitle = Title::newFromText( 'Namespace title conflict cleanup script' );
 272
 273 $verbose = isset( $options['verbose'] );
 274 $fix = isset( $options['fix'] );
 275 $suffix = isset( $options['suffix'] ) ? $options['suffix'] : '';
 276 $prefix = isset( $options['prefix'] ) ? $options['prefix'] : '';
 277 $key = isset( $options['key'] ) ? intval( $options['key'] ) : 0;
 278
 279 $dbw = wfGetDB( DB_MASTER );
 280 $duper = new NamespaceConflictChecker( $dbw, $verbose );
 281
 282 if( $prefix ) {
 283         $retval = $duper->checkPrefix( $key, $prefix, $fix, $suffix );
 284 } else {
 285         $retval = $duper->checkAll( $fix, $suffix );
 286 }
 287
 288 if( $retval ) {
 289         echo "\nLooks good!\n";
 290         exit( 0 );
 291 } else {
 292         echo "\nOh noeees\n";
 293         exit( -1 );
 294 }
 295
 296