maintenance/cleanupImages.php

   1 <?php
   2 /**
   3  * Clean up broken, unparseable upload filenames.
   4  *
   5  * Copyright © 2005-2006 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  * @author Brion Vibber <brion at pobox.com>
  25  * @ingroup Maintenance
  26  */
  27
  28 use MediaWiki\MediaWikiServices;
  29
  30 require_once __DIR__ . '/cleanupTable.inc';
  31
  32 /**
  33  * Maintenance script to clean up broken, unparseable upload filenames.
  34  *
  35  * @ingroup Maintenance
  36  */
  37 class CleanupImages extends TableCleanup {
  38         protected $defaultParams = [
  39                 'table' => 'image',
  40                 'conds' => [],
  41                 'index' => 'img_name',
  42                 'callback' => 'processRow',
  43         ];
  44
  45         public function __construct() {
  46                 parent::__construct();
  47                 $this->addDescription( 'Script to clean up broken, unparseable upload filenames' );
  48         }
  49
  50         protected function processRow( $row ) {
  51                 $source = $row->img_name;
  52                 if ( $source == '' ) {
  53                         // Ye olde empty rows. Just kill them.
  54                         $this->killRow( $source );
  55
  56                         return $this->progress( 1 );
  57                 }
  58
  59                 $cleaned = $source;
  60
  61                 // About half of old bad image names have percent-codes
  62                 $cleaned = rawurldecode( $cleaned );
  63
  64                 // We also have some HTML entities there
  65                 $cleaned = Sanitizer::decodeCharReferences( $cleaned );
  66
  67                 $contLang = MediaWikiServices::getInstance()->getContentLanguage();
  68
  69                 // Some are old latin-1
  70                 $cleaned = $contLang->checkTitleEncoding( $cleaned );
  71
  72                 // Many of remainder look like non-normalized unicode
  73                 $cleaned = $contLang->normalize( $cleaned );
  74
  75                 $title = Title::makeTitleSafe( NS_FILE, $cleaned );
  76
  77                 if ( is_null( $title ) ) {
  78                         $this->output( "page $source ($cleaned) is illegal.\n" );
  79                         $safe = $this->buildSafeTitle( $cleaned );
  80                         if ( $safe === false ) {
  81                                 return $this->progress( 0 );
  82                         }
  83                         $this->pokeFile( $source, $safe );
  84
  85                         return $this->progress( 1 );
  86                 }
  87
  88                 if ( $title->getDBkey() !== $source ) {
  89                         $munged = $title->getDBkey();
  90                         $this->output( "page $source ($munged) doesn't match self.\n" );
  91                         $this->pokeFile( $source, $munged );
  92
  93                         return $this->progress( 1 );
  94                 }
  95
  96                 return $this->progress( 0 );
  97         }
  98
  99         /**
 100          * @param string $name
 101          */
 102         private function killRow( $name ) {
 103                 if ( $this->dryrun ) {
 104                         $this->output( "DRY RUN: would delete bogus row '$name'\n" );
 105                 } else {
 106                         $this->output( "deleting bogus row '$name'\n" );
 107                         $db = $this->getDB( DB_MASTER );
 108                         $db->delete( 'image',
 109                                 [ 'img_name' => $name ],
 110                                 __METHOD__ );
 111                 }
 112         }
 113
 114         private function filePath( $name ) {
 115                 if ( !isset( $this->repo ) ) {
 116                         $this->repo = RepoGroup::singleton()->getLocalRepo();
 117                 }
 118
 119                 return $this->repo->getRootDirectory() . '/' . $this->repo->getHashPath( $name ) . $name;
 120         }
 121
 122         private function imageExists( $name, $db ) {
 123                 return $db->selectField( 'image', '1', [ 'img_name' => $name ], __METHOD__ );
 124         }
 125
 126         private function pageExists( $name, $db ) {
 127                 return $db->selectField(
 128                         'page',
 129                         '1',
 130                         [ 'page_namespace' => NS_FILE, 'page_title' => $name ],
 131                         __METHOD__
 132                 );
 133         }
 134
 135         private function pokeFile( $orig, $new ) {
 136                 $path = $this->filePath( $orig );
 137                 if ( !file_exists( $path ) ) {
 138                         $this->output( "missing file: $path\n" );
 139                         $this->killRow( $orig );
 140
 141                         return;
 142                 }
 143
 144                 $db = $this->getDB( DB_MASTER );
 145
 146                 /*
 147                  * To prevent key collisions in the update() statements below,
 148                  * if the target title exists in the image table, or if both the
 149                  * original and target titles exist in the page table, append
 150                  * increasing version numbers until the target title exists in
 151                  * neither.  (See also T18916.)
 152                  */
 153                 $version = 0;
 154                 $final = $new;
 155                 $conflict = ( $this->imageExists( $final, $db ) ||
 156                         ( $this->pageExists( $orig, $db ) && $this->pageExists( $final, $db ) ) );
 157
 158                 while ( $conflict ) {
 159                         $this->output( "Rename conflicts with '$final'...\n" );
 160                         $version++;
 161                         $final = $this->appendTitle( $new, "_$version" );
 162                         $conflict = ( $this->imageExists( $final, $db ) || $this->pageExists( $final, $db ) );
 163                 }
 164
 165                 $finalPath = $this->filePath( $final );
 166
 167                 if ( $this->dryrun ) {
 168                         $this->output( "DRY RUN: would rename $path to $finalPath\n" );
 169                 } else {
 170                         $this->output( "renaming $path to $finalPath\n" );
 171                         // @todo FIXME: Should this use File::move()?
 172                         $this->beginTransaction( $db, __METHOD__ );
 173                         $db->update( 'image',
 174                                 [ 'img_name' => $final ],
 175                                 [ 'img_name' => $orig ],
 176                                 __METHOD__ );
 177                         $db->update( 'oldimage',
 178                                 [ 'oi_name' => $final ],
 179                                 [ 'oi_name' => $orig ],
 180                                 __METHOD__ );
 181                         $db->update( 'page',
 182                                 [ 'page_title' => $final ],
 183                                 [ 'page_title' => $orig, 'page_namespace' => NS_FILE ],
 184                                 __METHOD__ );
 185                         $dir = dirname( $finalPath );
 186                         if ( !file_exists( $dir ) ) {
 187                                 if ( !wfMkdirParents( $dir, null, __METHOD__ ) ) {
 188                                         $this->output( "RENAME FAILED, COULD NOT CREATE $dir" );
 189                                         $this->rollbackTransaction( $db, __METHOD__ );
 190
 191                                         return;
 192                                 }
 193                         }
 194                         if ( rename( $path, $finalPath ) ) {
 195                                 $this->commitTransaction( $db, __METHOD__ );
 196                         } else {
 197                                 $this->error( "RENAME FAILED" );
 198                                 $this->rollbackTransaction( $db, __METHOD__ );
 199                         }
 200                 }
 201         }
 202
 203         private function appendTitle( $name, $suffix ) {
 204                 return preg_replace( '/^(.*)(\..*?)$/',
 205                         "\\1$suffix\\2", $name );
 206         }
 207
 208         private function buildSafeTitle( $name ) {
 209                 $x = preg_replace_callback(
 210                         '/([^' . Title::legalChars() . ']|~)/',
 211                         [ $this, 'hexChar' ],
 212                         $name );
 213
 214                 $test = Title::makeTitleSafe( NS_FILE, $x );
 215                 if ( is_null( $test ) || $test->getDBkey() !== $x ) {
 216                         $this->error( "Unable to generate safe title from '$name', got '$x'" );
 217
 218                         return false;
 219                 }
 220
 221                 return $x;
 222         }
 223 }
 224
 225 $maintClass = CleanupImages::class;
 226 require_once RUN_MAINTENANCE_IF_MAIN;