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         /**
  51          * @suppress SecurityCheck-XSS Non html mime type
  52          */
  53         function onView() {
  54                 $this->getOutput()->disable();
  55                 $request = $this->getRequest();
  56                 $response = $request->response();
  57                 $config = $this->context->getConfig();
  58
  59                 if ( !$request->checkUrlExtension() ) {
  60                         return;
  61                 }
  62
  63                 if ( $this->getOutput()->checkLastModified( $this->page->getTouched() ) ) {
  64                         return; // Client cache fresh and headers sent, nothing more to do.
  65                 }
  66
  67                 $contentType = $this->getContentType();
  68
  69                 $maxage = $request->getInt( 'maxage', $config->get( 'SquidMaxage' ) );
  70                 $smaxage = $request->getIntOrNull( 'smaxage' );
  71                 if ( $smaxage === null ) {
  72                         if (
  73                                 $contentType == 'text/css' ||
  74                                 $contentType == 'application/json' ||
  75                                 $contentType == 'text/javascript'
  76                         ) {
  77                                 // CSS/JSON/JS raw content has its own CDN max age configuration.
  78                                 // Note: Title::getCdnUrls() includes action=raw for css/json/js
  79                                 // pages, so if using the canonical url, this will get HTCP purges.
  80                                 $smaxage = intval( $config->get( 'ForcedRawSMaxage' ) );
  81                         } else {
  82                                 // No CDN cache for anything else
  83                                 $smaxage = 0;
  84                         }
  85                 }
  86
  87                 // Set standard Vary headers so cache varies on cookies and such (T125283)
  88                 $response->header( $this->getOutput()->getVaryHeader() );
  89                 if ( $config->get( 'UseKeyHeader' ) ) {
  90                         $response->header( $this->getOutput()->getKeyHeader() );
  91                 }
  92
  93                 // Output may contain user-specific data;
  94                 // vary generated content for open sessions on private wikis
  95                 $privateCache = !User::isEveryoneAllowed( 'read' ) &&
  96                         ( $smaxage == 0 || MediaWiki\Session\SessionManager::getGlobalSession()->isPersistent() );
  97                 // Don't accidentally cache cookies if user is logged in (T55032)
  98                 $privateCache = $privateCache || $this->getUser()->isLoggedIn();
  99                 $mode = $privateCache ? 'private' : 'public';
 100                 $response->header(
 101                         'Cache-Control: ' . $mode . ', s-maxage=' . $smaxage . ', max-age=' . $maxage
 102                 );
 103
 104                 // In the event of user JS, don't allow loading a user JS/CSS/Json
 105                 // subpage that has no registered user associated with, as
 106                 // someone could register the account and take control of the
 107                 // JS/CSS/Json page.
 108                 $title = $this->getTitle();
 109                 if ( $title->isUserConfigPage() && $contentType !== 'text/x-wiki' ) {
 110                         // not using getRootText() as we want this to work
 111                         // even if subpages are disabled.
 112                         $rootPage = strtok( $title->getText(), '/' );
 113                         $userFromTitle = User::newFromName( $rootPage, 'usable' );
 114                         if ( !$userFromTitle || $userFromTitle->getId() === 0 ) {
 115                                 $elevated = $this->getUser()->isAllowed( 'editinterface' );
 116                                 $elevatedText = $elevated ? 'by elevated ' : '';
 117                                 $log = LoggerFactory::getInstance( "security" );
 118                                 $log->warning(
 119                                         "Unsafe JS/CSS/Json $elevatedText" . "load - {user} loaded {title} with {ctype}",
 120                                         [
 121                                                 'user' => $this->getUser()->getName(),
 122                                                 'title' => $title->getPrefixedDBKey(),
 123                                                 'ctype' => $contentType,
 124                                                 'elevated' => $elevated
 125                                         ]
 126                                 );
 127                                 $msg = wfMessage( 'unregistered-user-config' );
 128                                 throw new HttpError( 403, $msg );
 129                         }
 130                 }
 131
 132                 $response->header( 'Content-type: ' . $contentType . '; charset=UTF-8' );
 133
 134                 $text = $this->getRawText();
 135
 136                 // Don't return a 404 response for CSS or JavaScript;
 137                 // 404s aren't generally cached and it would create
 138                 // extra hits when user CSS/JS are on and the user doesn't
 139                 // have the pages.
 140                 if ( $text === false && $contentType == 'text/x-wiki' ) {
 141                         $response->statusHeader( 404 );
 142                 }
 143
 144                 // Avoid PHP 7.1 warning of passing $this by reference
 145                 $rawAction = $this;
 146                 if ( !Hooks::run( 'RawPageViewBeforeOutput', [ &$rawAction, &$text ] ) ) {
 147                         wfDebug( __METHOD__ . ": RawPageViewBeforeOutput hook broke raw page output.\n" );
 148                 }
 149
 150                 echo $text;
 151         }
 152
 153         /**
 154          * Get the text that should be returned, or false if the page or revision
 155          * was not found.
 156          *
 157          * @return string|bool
 158          */
 159         public function getRawText() {
 160                 global $wgParser;
 161
 162                 $text = false;
 163                 $title = $this->getTitle();
 164                 $request = $this->getRequest();
 165
 166                 // If it's a page in the MediaWiki namespace, we can just hit the message cache
 167                 if ( $request->getBool( 'usemsgcache' ) && $title->getNamespace() == NS_MEDIAWIKI ) {
 168                         // FIXME: The overhead and complexity of using MessageCache for serving
 169                         // source code is not worth the marginal gain in performance. This should
 170                         // instead use Revision::getRevisionText, which already has its own caching
 171                         // layer, which is good enough fine given action=raw only responds with
 172                         // a single page (no need for batch).
 173                         //
 174                         // Use of MessageCache:
 175                         // - is unsustainable (T193271),
 176                         // - can cause bugs due to "post-processing" (see MessageCache::get) not
 177                         //   intending to apply to program source code,
 178                         // - causes uncertaintly around whether or not localisation default
 179                         //   placeholders are, can, and should be used, or not.
 180                         $text = MessageCache::singleton()->get(
 181                                 $title->getDBkey(),
 182                                 // Yes, use the database.
 183                                 true,
 184                                 // Yes, use the content language.
 185                                 true,
 186                                 // Yes, the message key already contains the language in it ("/de", etc.)
 187                                 true
 188                         );
 189                         // If the local page doesn't exist, return a blank (not the default)
 190                         if ( $text === false ) {
 191                                 $text = '';
 192                         }
 193                 } else {
 194                         // Get it from the DB
 195                         $rev = Revision::newFromTitle( $title, $this->getOldId() );
 196                         if ( $rev ) {
 197                                 $lastmod = wfTimestamp( TS_RFC2822, $rev->getTimestamp() );
 198                                 $request->response()->header( "Last-modified: $lastmod" );
 199
 200                                 // Public-only due to cache headers
 201                                 $content = $rev->getContent();
 202
 203                                 if ( $content === null ) {
 204                                         // revision not found (or suppressed)
 205                                         $text = false;
 206                                 } elseif ( !$content instanceof TextContent ) {
 207                                         // non-text content
 208                                         wfHttpError( 415, "Unsupported Media Type", "The requested page uses the content model `"
 209                                                 . $content->getModel() . "` which is not supported via this interface." );
 210                                         die();
 211                                 } else {
 212                                         // want a section?
 213                                         $section = $request->getIntOrNull( 'section' );
 214                                         if ( $section !== null ) {
 215                                                 $content = $content->getSection( $section );
 216                                         }
 217
 218                                         if ( $content === null || $content === false ) {
 219                                                 // section not found (or section not supported, e.g. for JS, JSON, and CSS)
 220                                                 $text = false;
 221                                         } else {
 222                                                 $text = $content->getNativeData();
 223                                         }
 224                                 }
 225                         }
 226                 }
 227
 228                 if ( $text !== false && $text !== '' && $request->getRawVal( 'templates' ) === 'expand' ) {
 229                         $text = $wgParser->preprocess(
 230                                 $text,
 231                                 $title,
 232                                 ParserOptions::newFromContext( $this->getContext() )
 233                         );
 234                 }
 235
 236                 return $text;
 237         }
 238
 239         /**
 240          * Get the ID of the revision that should used to get the text.
 241          *
 242          * @return int
 243          */
 244         public function getOldId() {
 245                 $oldid = $this->getRequest()->getInt( 'oldid' );
 246                 switch ( $this->getRequest()->getText( 'direction' ) ) {
 247                         case 'next':
 248                                 # output next revision, or nothing if there isn't one
 249                                 $nextid = 0;
 250                                 if ( $oldid ) {
 251                                         $nextid = $this->getTitle()->getNextRevisionID( $oldid );
 252                                 }
 253                                 $oldid = $nextid ?: -1;
 254                                 break;
 255                         case 'prev':
 256                                 # output previous revision, or nothing if there isn't one
 257                                 if ( !$oldid ) {
 258                                         # get the current revision so we can get the penultimate one
 259                                         $oldid = $this->page->getLatest();
 260                                 }
 261                                 $previd = $this->getTitle()->getPreviousRevisionID( $oldid );
 262                                 $oldid = $previd ?: -1;
 263                                 break;
 264                         case 'cur':
 265                                 $oldid = 0;
 266                                 break;
 267                 }
 268
 269                 return $oldid;
 270         }
 271
 272         /**
 273          * Get the content type to use for the response
 274          *
 275          * @return string
 276          */
 277         public function getContentType() {
 278                 // Use getRawVal instead of getVal because we only
 279                 // need to match against known strings, there is no
 280                 // storing of localised content or other user input.
 281                 $ctype = $this->getRequest()->getRawVal( 'ctype' );
 282
 283                 if ( $ctype == '' ) {
 284                         // Legacy compatibilty
 285                         $gen = $this->getRequest()->getRawVal( 'gen' );
 286                         if ( $gen == 'js' ) {
 287                                 $ctype = 'text/javascript';
 288                         } elseif ( $gen == 'css' ) {
 289                                 $ctype = 'text/css';
 290                         }
 291                 }
 292
 293                 $allowedCTypes = [
 294                         'text/x-wiki',
 295                         'text/javascript',
 296                         'text/css',
 297                         // FIXME: Should we still allow Zope editing? External editing feature was dropped
 298                         'application/x-zope-edit',
 299                         'application/json'
 300                 ];
 301                 if ( $ctype == '' || !in_array( $ctype, $allowedCTypes ) ) {
 302                         $ctype = 'text/x-wiki';
 303                 }
 304
 305                 return $ctype;
 306         }
 307 }