maintenance/cleanupImages.php

   1 <?php
   2 /*
   3  * Script to clean up broken, unparseable upload filenames.
   4  *
   5  * Usage: php cleanupImages.php [--fix]
   6  * Options:
   7  *   --fix  Actually clean up titles; otherwise just checks for them
   8  *
   9  * Copyright (C) 2005-2006 Brion Vibber <brion@pobox.com>
  10  * http://www.mediawiki.org/
  11  *
  12  * This program is free software; you can redistribute it and/or modify
  13  * it under the terms of the GNU General Public License as published by
  14  * the Free Software Foundation; either version 2 of the License, or
  15  * (at your option) any later version.
  16  *
  17  * This program is distributed in the hope that it will be useful,
  18  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  19  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  20  * GNU General Public License for more details.
  21  *
  22  * You should have received a copy of the GNU General Public License along
  23  * with this program; if not, write to the Free Software Foundation, Inc.,
  24  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  25  * http://www.gnu.org/copyleft/gpl.html
  26  *
  27  * @author Brion Vibber <brion at pobox.com>
  28  * @ingroup Maintenance
  29  */
  30
  31 require_once( dirname(__FILE__) . '/cleanupTable.inc' );
  32
  33 class ImageCleanup extends TableCleanup {
  34         protected $targetTable = 'image';
  35         public function __construct() {
  36                 parent::__construct();
  37                 $this->mDescription = "Script to clean up broken, unparseable upload filenames";
  38         }
  39
  40         protected function processPage( $row ) {
  41                 global $wgContLang;
  42
  43                 $source = $row->img_name;
  44                 if( $source == '' ) {
  45                         // Ye olde empty rows. Just kill them.
  46                         $this->killRow( $source );
  47                         return $this->progress( 1 );
  48                 }
  49
  50                 $cleaned = $source;
  51
  52                 // About half of old bad image names have percent-codes
  53                 $cleaned = rawurldecode( $cleaned );
  54
  55                 // We also have some HTML entities there
  56                 $cleaned = Sanitizer::decodeCharReferences( $cleaned );
  57
  58                 // Some are old latin-1
  59                 $cleaned = $wgContLang->checkTitleEncoding( $cleaned );
  60
  61                 // Many of remainder look like non-normalized unicode
  62                 $cleaned = UtfNormal::cleanUp( $cleaned );
  63
  64                 $title = Title::makeTitleSafe( NS_FILE, $cleaned );
  65
  66                 if( is_null( $title ) ) {
  67                         $this->output( "page $source ($cleaned) is illegal.\n" );
  68                         $safe = $this->buildSafeTitle( $cleaned );
  69                         if( $safe === false )
  70                                 return $this->progress( 0 );
  71                         $this->pokeFile( $source, $safe );
  72                         return $this->progress( 1 );
  73                 }
  74
  75                 if( $title->getDBkey() !== $source ) {
  76                         $munged = $title->getDBkey();
  77                         $this->output( "page $source ($munged) doesn't match self.\n" );
  78                         $this->pokeFile( $source, $munged );
  79                         return $this->progress( 1 );
  80                 }
  81
  82                 $this->progress( 0 );
  83         }
  84
  85         private function killRow( $name ) {
  86                 if( $this->dryrun ) {
  87                         $this->output( "DRY RUN: would delete bogus row '$name'\n" );
  88                 } else {
  89                         $this->output( "deleting bogus row '$name'\n" );
  90                         $db = wfGetDB( DB_MASTER );
  91                         $db->delete( 'image',
  92                                 array( 'img_name' => $name ),
  93                                 __METHOD__ );
  94                 }
  95         }
  96
  97         private function filePath( $name ) {
  98                 if ( !isset( $this->repo ) ) {
  99                         $this->repo = RepoGroup::singleton()->getLocalRepo();
 100                 }
 101                 return $this->repo->getRootDirectory() . '/' . $this->repo->getHashPath( $name ) . $name;
 102         }
 103
 104         private function imageExists( $name, $db ) {
 105                 return $db->selectField( 'image', '1', array( 'img_name' => $name ), __METHOD__ );
 106         }
 107
 108         private function pageExists( $name, $db ) {
 109                 return $db->selectField( 'page', '1', array( 'page_namespace' => NS_FILE, 'page_title' => $name ), __METHOD__ );
 110         }
 111
 112         private function pokeFile( $orig, $new ) {
 113                 $path = $this->filePath( $orig );
 114                 if( !file_exists( $path ) ) {
 115                         $this->output( "missing file: $path\n" );
 116                         return $this->killRow( $orig );
 117                 }
 118
 119                 $db = wfGetDB( DB_MASTER );
 120
 121                 /*
 122                  * To prevent key collisions in the update() statements below,
 123                  * if the target title exists in the image table, or if both the
 124                  * original and target titles exist in the page table, append
 125                  * increasing version numbers until the target title exists in
 126                  * neither.  (See also bug 16916.)
 127                  */
 128                 $version = 0;
 129                 $final = $new;
 130                 $conflict = ( $this->imageExists( $final, $db ) ||
 131                               ( $this->pageExists( $orig, $db ) && $this->pageExists( $final, $db ) ) );
 132
 133                 while( $conflict ) {
 134                         $this->output( "Rename conflicts with '$final'...\n" );
 135                         $version++;
 136                         $final = $this->appendTitle( $new, "_$version" );
 137                         $conflict = ( $this->imageExists( $final, $db ) || $this->pageExists( $final, $db ) );
 138                 }
 139
 140                 $finalPath = $this->filePath( $final );
 141
 142                 if( $this->dryrun ) {
 143                         $this->output( "DRY RUN: would rename $path to $finalPath\n" );
 144                 } else {
 145                         $this->output( "renaming $path to $finalPath\n" );
 146                         // XXX: should this use File::move()?  FIXME?
 147                         $db->begin();
 148                         $db->update( 'image',
 149                                 array( 'img_name' => $final ),
 150                                 array( 'img_name' => $orig ),
 151                                 __METHOD__ );
 152                         $db->update( 'oldimage',
 153                                 array( 'oi_name' => $final ),
 154                                 array( 'oi_name' => $orig ),
 155                                 __METHOD__ );
 156                         $db->update( 'page',
 157                                 array( 'page_title' => $final ),
 158                                 array( 'page_title' => $orig, 'page_namespace' => NS_FILE ),
 159                                 __METHOD__ );
 160                         $dir = dirname( $finalPath );
 161                         if( !file_exists( $dir ) ) {
 162                                 if( !wfMkdirParents( $dir ) ) {
 163                                         $this->log( "RENAME FAILED, COULD NOT CREATE $dir" );
 164                                         $db->rollback();
 165                                         return;
 166                                 }
 167                         }
 168                         if( rename( $path, $finalPath ) ) {
 169                                 $db->commit();
 170                         } else {
 171                                 $this->error( "RENAME FAILED" );
 172                                 $db->rollback();
 173                         }
 174                 }
 175         }
 176
 177         private function appendTitle( $name, $suffix ) {
 178                 return preg_replace( '/^(.*)(\..*?)$/',
 179                         "\\1$suffix\\2", $name );
 180         }
 181
 182         private function buildSafeTitle( $name ) {
 183                 global $wgLegalTitleChars;
 184                 $x = preg_replace_callback(
 185                         "/([^$wgLegalTitleChars]|~)/",
 186                         array( $this, 'hexChar' ),
 187                         $name );
 188
 189                 $test = Title::makeTitleSafe( NS_FILE, $x );
 190                 if( is_null( $test ) || $test->getDBkey() !== $x ) {
 191                         $this->error( "Unable to generate safe title from '$name', got '$x'" );
 192                         return false;
 193                 }
 194
 195                 return $x;
 196         }
 197 }
 198
 199 $maintClass = "ImageCleanup";
 200 require_once( DO_MAINTENANCE );