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