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 $defaultParams = array(
  35                 'table' => 'image',
  36                 'conds' => array(),
  37                 'index' => 'img_name',
  38                 'callback' => 'processRow',
  39         );
  40
  41         public function __construct() {
  42                 parent::__construct();
  43                 $this->mDescription = "Script to clean up broken, unparseable upload filenames";
  44         }
  45
  46         protected function processRow( $row ) {
  47                 global $wgContLang;
  48
  49                 $source = $row->img_name;
  50                 if ( $source == '' ) {
  51                         // Ye olde empty rows. Just kill them.
  52                         $this->killRow( $source );
  53                         return $this->progress( 1 );
  54                 }
  55
  56                 $cleaned = $source;
  57
  58                 // About half of old bad image names have percent-codes
  59                 $cleaned = rawurldecode( $cleaned );
  60
  61                 // We also have some HTML entities there
  62                 $cleaned = Sanitizer::decodeCharReferences( $cleaned );
  63
  64                 // Some are old latin-1
  65                 $cleaned = $wgContLang->checkTitleEncoding( $cleaned );
  66
  67                 // Many of remainder look like non-normalized unicode
  68                 $cleaned = $wgContLang->normalize( $cleaned );
  69
  70                 $title = Title::makeTitleSafe( NS_FILE, $cleaned );
  71
  72                 if ( is_null( $title ) ) {
  73                         $this->output( "page $source ($cleaned) is illegal.\n" );
  74                         $safe = $this->buildSafeTitle( $cleaned );
  75                         if ( $safe === false )
  76                                 return $this->progress( 0 );
  77                         $this->pokeFile( $source, $safe );
  78                         return $this->progress( 1 );
  79                 }
  80
  81                 if ( $title->getDBkey() !== $source ) {
  82                         $munged = $title->getDBkey();
  83                         $this->output( "page $source ($munged) doesn't match self.\n" );
  84                         $this->pokeFile( $source, $munged );
  85                         return $this->progress( 1 );
  86                 }
  87
  88                 $this->progress( 0 );
  89         }
  90
  91         private function killRow( $name ) {
  92                 if ( $this->dryrun ) {
  93                         $this->output( "DRY RUN: would delete bogus row '$name'\n" );
  94                 } else {
  95                         $this->output( "deleting bogus row '$name'\n" );
  96                         $db = wfGetDB( DB_MASTER );
  97                         $db->delete( 'image',
  98                                 array( 'img_name' => $name ),
  99                                 __METHOD__ );
 100                 }
 101         }
 102
 103         private function filePath( $name ) {
 104                 if ( !isset( $this->repo ) ) {
 105                         $this->repo = RepoGroup::singleton()->getLocalRepo();
 106                 }
 107                 return $this->repo->getRootDirectory() . '/' . $this->repo->getHashPath( $name ) . $name;
 108         }
 109
 110         private function imageExists( $name, $db ) {
 111                 return $db->selectField( 'image', '1', array( 'img_name' => $name ), __METHOD__ );
 112         }
 113
 114         private function pageExists( $name, $db ) {
 115                 return $db->selectField( 'page', '1', array( 'page_namespace' => NS_FILE, 'page_title' => $name ), __METHOD__ );
 116         }
 117
 118         private function pokeFile( $orig, $new ) {
 119                 $path = $this->filePath( $orig );
 120                 if ( !file_exists( $path ) ) {
 121                         $this->output( "missing file: $path\n" );
 122                         return $this->killRow( $orig );
 123                 }
 124
 125                 $db = wfGetDB( DB_MASTER );
 126
 127                 /*
 128                  * To prevent key collisions in the update() statements below,
 129                  * if the target title exists in the image table, or if both the
 130                  * original and target titles exist in the page table, append
 131                  * increasing version numbers until the target title exists in
 132                  * neither.  (See also bug 16916.)
 133                  */
 134                 $version = 0;
 135                 $final = $new;
 136                 $conflict = ( $this->imageExists( $final, $db ) ||
 137                               ( $this->pageExists( $orig, $db ) && $this->pageExists( $final, $db ) ) );
 138
 139                 while ( $conflict ) {
 140                         $this->output( "Rename conflicts with '$final'...\n" );
 141                         $version++;
 142                         $final = $this->appendTitle( $new, "_$version" );
 143                         $conflict = ( $this->imageExists( $final, $db ) || $this->pageExists( $final, $db ) );
 144                 }
 145
 146                 $finalPath = $this->filePath( $final );
 147
 148                 if ( $this->dryrun ) {
 149                         $this->output( "DRY RUN: would rename $path to $finalPath\n" );
 150                 } else {
 151                         $this->output( "renaming $path to $finalPath\n" );
 152                         // XXX: should this use File::move()?  FIXME?
 153                         $db->begin();
 154                         $db->update( 'image',
 155                                 array( 'img_name' => $final ),
 156                                 array( 'img_name' => $orig ),
 157                                 __METHOD__ );
 158                         $db->update( 'oldimage',
 159                                 array( 'oi_name' => $final ),
 160                                 array( 'oi_name' => $orig ),
 161                                 __METHOD__ );
 162                         $db->update( 'page',
 163                                 array( 'page_title' => $final ),
 164                                 array( 'page_title' => $orig, 'page_namespace' => NS_FILE ),
 165                                 __METHOD__ );
 166                         $dir = dirname( $finalPath );
 167                         if ( !file_exists( $dir ) ) {
 168                                 if ( !wfMkdirParents( $dir ) ) {
 169                                         $this->log( "RENAME FAILED, COULD NOT CREATE $dir" );
 170                                         $db->rollback();
 171                                         return;
 172                                 }
 173                         }
 174                         if ( rename( $path, $finalPath ) ) {
 175                                 $db->commit();
 176                         } else {
 177                                 $this->error( "RENAME FAILED" );
 178                                 $db->rollback();
 179                         }
 180                 }
 181         }
 182
 183         private function appendTitle( $name, $suffix ) {
 184                 return preg_replace( '/^(.*)(\..*?)$/',
 185                         "\\1$suffix\\2", $name );
 186         }
 187
 188         private function buildSafeTitle( $name ) {
 189                 global $wgLegalTitleChars;
 190                 $x = preg_replace_callback(
 191                         "/([^$wgLegalTitleChars]|~)/",
 192                         array( $this, 'hexChar' ),
 193                         $name );
 194
 195                 $test = Title::makeTitleSafe( NS_FILE, $x );
 196                 if ( is_null( $test ) || $test->getDBkey() !== $x ) {
 197                         $this->error( "Unable to generate safe title from '$name', got '$x'" );
 198                         return false;
 199                 }
 200
 201                 return $x;
 202         }
 203 }
 204
 205 $maintClass = "ImageCleanup";
 206 require_once( DO_MAINTENANCE );