includes/actions/RawAction.php

   1 <?php
   2 /**
   3  * Raw page text accessor
   4  *
   5  * Copyright © 2004 Gabriel Wicke <wicke@wikidev.net>
   6  * http://wikidev.net/
   7  *
   8  * Based on HistoryAction and SpecialExport
   9  *
  10  * This program is free software; you can redistribute it and/or modify
  11  * it under the terms of the GNU General Public License as published by
  12  * the Free Software Foundation; either version 2 of the License, or
  13  * (at your option) any later version.
  14  *
  15  * This program is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  18  * GNU General Public License for more details.
  19  *
  20  * You should have received a copy of the GNU General Public License along
  21  * with this program; if not, write to the Free Software Foundation, Inc.,
  22  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  23  * http://www.gnu.org/copyleft/gpl.html
  24  *
  25  * @author Gabriel Wicke <wicke@wikidev.net>
  26  * @file
  27  */
  28
  29 use MediaWiki\Logger\LoggerFactory;
  30
  31 /**
  32  * A simple method to retrieve the plain source of an article,
  33  * using "action=raw" in the GET request string.
  34  *
  35  * @ingroup Actions
  36  */
  37 class RawAction extends FormlessAction {
  38         public function getName() {
  39                 return 'raw';
  40         }
  41
  42         public function requiresWrite() {
  43                 return false;
  44         }
  45
  46         public function requiresUnblock() {
  47                 return false;
  48         }
  49
  50         function onView() {
  51                 $this->getOutput()->disable();
  52                 $request = $this->getRequest();
  53                 $response = $request->response();
  54                 $config = $this->context->getConfig();
  55
  56                 if ( !$request->checkUrlExtension() ) {
  57                         return;
  58                 }
  59
  60                 if ( $this->getOutput()->checkLastModified( $this->page->getTouched() ) ) {
  61                         return; // Client cache fresh and headers sent, nothing more to do.
  62                 }
  63
  64                 $contentType = $this->getContentType();
  65
  66                 $maxage = $request->getInt( 'maxage', $config->get( 'SquidMaxage' ) );
  67                 $smaxage = $request->getIntOrNull( 'smaxage' );
  68                 if ( $smaxage === null ) {
  69                         if (
  70                                 $contentType == 'text/css' ||
  71                                 $contentType == 'application/json' ||
  72                                 $contentType == 'text/javascript'
  73                         ) {
  74                                 // CSS/JSON/JS raw content has its own CDN max age configuration.
  75                                 // Note: Title::getCdnUrls() includes action=raw for css/json/js
  76                                 // pages, so if using the canonical url, this will get HTCP purges.
  77                                 $smaxage = intval( $config->get( 'ForcedRawSMaxage' ) );
  78                         } else {
  79                                 // No CDN cache for anything else
  80                                 $smaxage = 0;
  81                         }
  82                 }
  83
  84                 // Set standard Vary headers so cache varies on cookies and such (T125283)
  85                 $response->header( $this->getOutput()->getVaryHeader() );
  86                 if ( $config->get( 'UseKeyHeader' ) ) {
  87                         $response->header( $this->getOutput()->getKeyHeader() );
  88                 }
  89
  90                 // Output may contain user-specific data;
  91                 // vary generated content for open sessions on private wikis
  92                 $privateCache = !User::isEveryoneAllowed( 'read' ) &&
  93                         ( $smaxage == 0 || MediaWiki\Session\SessionManager::getGlobalSession()->isPersistent() );
  94                 // Don't accidentally cache cookies if user is logged in (T55032)
  95                 $privateCache = $privateCache || $this->getUser()->isLoggedIn();
  96                 $mode = $privateCache ? 'private' : 'public';
  97                 $response->header(
  98                         'Cache-Control: ' . $mode . ', s-maxage=' . $smaxage . ', max-age=' . $maxage
  99                 );
 100
 101                 // In the event of user JS, don't allow loading a user JS/CSS/Json
 102                 // subpage that has no registered user associated with, as
 103                 // someone could register the account and take control of the
 104                 // JS/CSS/Json page.
 105                 $title = $this->getTitle();
 106                 if ( $title->isUserConfigPage() && $contentType !== 'text/x-wiki' ) {
 107                         // not using getRootText() as we want this to work
 108                         // even if subpages are disabled.
 109                         $rootPage = strtok( $title->getText(), '/' );
 110                         $userFromTitle = User::newFromName( $rootPage, 'usable' );
 111                         if ( !$userFromTitle || $userFromTitle->getId() === 0 ) {
 112                                 $log = LoggerFactory::getInstance( "security" );
 113                                 $log->warning(
 114                                         "Unsafe JS/CSS/Json load - {user} loaded {title} with {ctype}",
 115                                         [
 116                                                 'user' => $this->getUser()->getName(),
 117                                                 'title' => $title->getPrefixedDBKey(),
 118                                                 'ctype' => $contentType,
 119                                         ]
 120                                 );
 121                                 $msg = wfMessage( 'unregistered-user-config' );
 122                                 throw new HttpError( 403, $msg );
 123                         }
 124                 }
 125
 126                 // Don't allow loading non-protected pages as javascript.
 127                 // In future we may further restrict this to only CONTENT_MODEL_JAVASCRIPT
 128                 // in NS_MEDIAWIKI or NS_USER, as well as including other config types,
 129                 // but for now be more permissive. Allowing protected pages outside of
 130                 // NS_USER and NS_MEDIAWIKI in particular should be considered a temporary
 131                 // allowance.
 132                 if (
 133                         $contentType === 'text/javascript' &&
 134                         !$title->isUserJsConfigPage() &&
 135                         !$title->inNamespace( NS_MEDIAWIKI ) &&
 136                         !in_array( 'sysop', $title->getRestrictions( 'edit' ) ) &&
 137                         !in_array( 'editprotected', $title->getRestrictions( 'edit' ) )
 138                 ) {
 139
 140                         $log = LoggerFactory::getInstance( "security" );
 141                         $log->info( "Blocked loading unprotected JS {title} for {user}",
 142                                 [
 143                                         'user' => $this->getUser()->getName(),
 144                                         'title' => $title->getPrefixedDBKey(),
 145                                 ]
 146                         );
 147                         throw new HttpError( 403, wfMessage( 'unprotected-js' ) );
 148                 }
 149
 150                 $response->header( 'Content-type: ' . $contentType . '; charset=UTF-8' );
 151
 152                 $text = $this->getRawText();
 153
 154                 // Don't return a 404 response for CSS or JavaScript;
 155                 // 404s aren't generally cached and it would create
 156                 // extra hits when user CSS/JS are on and the user doesn't
 157                 // have the pages.
 158                 if ( $text === false && $contentType == 'text/x-wiki' ) {
 159                         $response->statusHeader( 404 );
 160                 }
 161
 162                 // Avoid PHP 7.1 warning of passing $this by reference
 163                 $rawAction = $this;
 164                 if ( !Hooks::run( 'RawPageViewBeforeOutput', [ &$rawAction, &$text ] ) ) {
 165                         wfDebug( __METHOD__ . ": RawPageViewBeforeOutput hook broke raw page output.\n" );
 166                 }
 167
 168                 echo $text;
 169         }
 170
 171         /**
 172          * Get the text that should be returned, or false if the page or revision
 173          * was not found.
 174          *
 175          * @return string|bool
 176          */
 177         public function getRawText() {
 178                 global $wgParser;
 179
 180                 $text = false;
 181                 $title = $this->getTitle();
 182                 $request = $this->getRequest();
 183
 184                 // If it's a MediaWiki message we can just hit the message cache
 185                 if ( $request->getBool( 'usemsgcache' ) && $title->getNamespace() == NS_MEDIAWIKI ) {
 186                         // The first "true" is to use the database, the second is to use
 187                         // the content langue and the last one is to specify the message
 188                         // key already contains the language in it ("/de", etc.).
 189                         $text = MessageCache::singleton()->get( $title->getDBkey(), true, true, true );
 190                         // If the message doesn't exist, return a blank
 191                         if ( $text === false ) {
 192                                 $text = '';
 193                         }
 194                 } else {
 195                         // Get it from the DB
 196                         $rev = Revision::newFromTitle( $title, $this->getOldId() );
 197                         if ( $rev ) {
 198                                 $lastmod = wfTimestamp( TS_RFC2822, $rev->getTimestamp() );
 199                                 $request->response()->header( "Last-modified: $lastmod" );
 200
 201                                 // Public-only due to cache headers
 202                                 $content = $rev->getContent();
 203
 204                                 if ( $content === null ) {
 205                                         // revision not found (or suppressed)
 206                                         $text = false;
 207                                 } elseif ( !$content instanceof TextContent ) {
 208                                         // non-text content
 209                                         wfHttpError( 415, "Unsupported Media Type", "The requested page uses the content model `"
 210                                                 . $content->getModel() . "` which is not supported via this interface." );
 211                                         die();
 212                                 } else {
 213                                         // want a section?
 214                                         $section = $request->getIntOrNull( 'section' );
 215                                         if ( $section !== null ) {
 216                                                 $content = $content->getSection( $section );
 217                                         }
 218
 219                                         if ( $content === null || $content === false ) {
 220                                                 // section not found (or section not supported, e.g. for JS, JSON, and CSS)
 221                                                 $text = false;
 222                                         } else {
 223                                                 $text = $content->getNativeData();
 224                                         }
 225                                 }
 226                         }
 227                 }
 228
 229                 if ( $text !== false && $text !== '' && $request->getRawVal( 'templates' ) === 'expand' ) {
 230                         $text = $wgParser->preprocess(
 231                                 $text,
 232                                 $title,
 233                                 ParserOptions::newFromContext( $this->getContext() )
 234                         );
 235                 }
 236
 237                 return $text;
 238         }
 239
 240         /**
 241          * Get the ID of the revision that should used to get the text.
 242          *
 243          * @return int
 244          */
 245         public function getOldId() {
 246                 $oldid = $this->getRequest()->getInt( 'oldid' );
 247                 switch ( $this->getRequest()->getText( 'direction' ) ) {
 248                         case 'next':
 249                                 # output next revision, or nothing if there isn't one
 250                                 $nextid = 0;
 251                                 if ( $oldid ) {
 252                                         $nextid = $this->getTitle()->getNextRevisionID( $oldid );
 253                                 }
 254                                 $oldid = $nextid ?: -1;
 255                                 break;
 256                         case 'prev':
 257                                 # output previous revision, or nothing if there isn't one
 258                                 if ( !$oldid ) {
 259                                         # get the current revision so we can get the penultimate one
 260                                         $oldid = $this->page->getLatest();
 261                                 }
 262                                 $previd = $this->getTitle()->getPreviousRevisionID( $oldid );
 263                                 $oldid = $previd ?: -1;
 264                                 break;
 265                         case 'cur':
 266                                 $oldid = 0;
 267                                 break;
 268                 }
 269
 270                 return $oldid;
 271         }
 272
 273         /**
 274          * Get the content type to use for the response
 275          *
 276          * @return string
 277          */
 278         public function getContentType() {
 279                 // Use getRawVal instead of getVal because we only
 280                 // need to match against known strings, there is no
 281                 // storing of localised content or other user input.
 282                 $ctype = $this->getRequest()->getRawVal( 'ctype' );
 283
 284                 if ( $ctype == '' ) {
 285                         // Legacy compatibilty
 286                         $gen = $this->getRequest()->getRawVal( 'gen' );
 287                         if ( $gen == 'js' ) {
 288                                 $ctype = 'text/javascript';
 289                         } elseif ( $gen == 'css' ) {
 290                                 $ctype = 'text/css';
 291                         }
 292                 }
 293
 294                 $allowedCTypes = [
 295                         'text/x-wiki',
 296                         'text/javascript',
 297                         'text/css',
 298                         // FIXME: Should we still allow Zope editing? External editing feature was dropped
 299                         'application/x-zope-edit',
 300                         'application/json'
 301                 ];
 302                 if ( $ctype == '' || !in_array( $ctype, $allowedCTypes ) ) {
 303                         $ctype = 'text/x-wiki';
 304                 }
 305
 306                 return $ctype;
 307         }
 308 }