includes/libs/filebackend/HTTPFileStreamer.php

   1 <?php
   2 /**
   3  * Functions related to the output of file content.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  */
  22 use Wikimedia\Timestamp\ConvertibleTimestamp;
  23
  24 /**
  25  * Functions related to the output of file content
  26  *
  27  * @since 1.28
  28  */
  29 class HTTPFileStreamer {
  30         /** @var string */
  31         protected $path;
  32         /** @var callable */
  33         protected $obResetFunc;
  34         /** @var callable */
  35         protected $streamMimeFunc;
  36
  37         // Do not send any HTTP headers unless requested by caller (e.g. body only)
  38         const STREAM_HEADLESS = 1;
  39         // Do not try to tear down any PHP output buffers
  40         const STREAM_ALLOW_OB = 2;
  41
  42         /**
  43          * @param string $path Local filesystem path to a file
  44          * @param array $params Options map, which includes:
  45          *   - obResetFunc : alternative callback to clear the output buffer
  46          *   - streamMimeFunc : alternative method to determine the content type from the path
  47          */
  48         public function __construct( $path, array $params = [] ) {
  49                 $this->path = $path;
  50                 $this->obResetFunc = isset( $params['obResetFunc'] )
  51                         ? $params['obResetFunc']
  52                         : [ __CLASS__, 'resetOutputBuffers' ];
  53                 $this->streamMimeFunc = isset( $params['streamMimeFunc'] )
  54                         ? $params['streamMimeFunc']
  55                         : [ __CLASS__, 'contentTypeFromPath' ];
  56         }
  57
  58         /**
  59          * Stream a file to the browser, adding all the headings and fun stuff.
  60          * Headers sent include: Content-type, Content-Length, Last-Modified,
  61          * and Content-Disposition.
  62          *
  63          * @param array $headers Any additional headers to send if the file exists
  64          * @param bool $sendErrors Send error messages if errors occur (like 404)
  65          * @param array $optHeaders HTTP request header map (e.g. "range") (use lowercase keys)
  66          * @param int $flags Bitfield of STREAM_* constants
  67          * @throws MWException
  68          * @return bool Success
  69          */
  70         public function stream(
  71                 $headers = [], $sendErrors = true, $optHeaders = [], $flags = 0
  72         ) {
  73                 // Don't stream it out as text/html if there was a PHP error
  74                 if ( ( ( $flags & self::STREAM_HEADLESS ) == 0 || $headers ) && headers_sent() ) {
  75                         echo "Headers already sent, terminating.\n";
  76                         return false;
  77                 }
  78
  79                 $headerFunc = ( $flags & self::STREAM_HEADLESS )
  80                         ? function ( $header ) {
  81                                 // no-op
  82                         }
  83                         : function ( $header ) {
  84                                 is_int( $header ) ? HttpStatus::header( $header ) : header( $header );
  85                         };
  86
  87                 MediaWiki\suppressWarnings();
  88                 $info = stat( $this->path );
  89                 MediaWiki\restoreWarnings();
  90
  91                 if ( !is_array( $info ) ) {
  92                         if ( $sendErrors ) {
  93                                 self::send404Message( $this->path, $flags );
  94                         }
  95                         return false;
  96                 }
  97
  98                 // Send Last-Modified HTTP header for client-side caching
  99                 $mtimeCT = new ConvertibleTimestamp( $info['mtime'] );
 100                 $headerFunc( 'Last-Modified: ' . $mtimeCT->getTimestamp( TS_RFC2822 ) );
 101
 102                 if ( ( $flags & self::STREAM_ALLOW_OB ) == 0 ) {
 103                         call_user_func( $this->obResetFunc );
 104                 }
 105
 106                 $type = call_user_func( $this->streamMimeFunc, $this->path );
 107                 if ( $type && $type != 'unknown/unknown' ) {
 108                         $headerFunc( "Content-type: $type" );
 109                 } else {
 110                         // Send a content type which is not known to Internet Explorer, to
 111                         // avoid triggering IE's content type detection. Sending a standard
 112                         // unknown content type here essentially gives IE license to apply
 113                         // whatever content type it likes.
 114                         $headerFunc( 'Content-type: application/x-wiki' );
 115                 }
 116
 117                 // Don't send if client has up to date cache
 118                 if ( isset( $optHeaders['if-modified-since'] ) ) {
 119                         $modsince = preg_replace( '/;.*$/', '', $optHeaders['if-modified-since'] );
 120                         if ( $mtimeCT->getTimestamp( TS_UNIX ) <= strtotime( $modsince ) ) {
 121                                 ini_set( 'zlib.output_compression', 0 );
 122                                 $headerFunc( 304 );
 123                                 return true; // ok
 124                         }
 125                 }
 126
 127                 // Send additional headers
 128                 foreach ( $headers as $header ) {
 129                         header( $header ); // always use header(); specifically requested
 130                 }
 131
 132                 if ( isset( $optHeaders['range'] ) ) {
 133                         $range = self::parseRange( $optHeaders['range'], $info['size'] );
 134                         if ( is_array( $range ) ) {
 135                                 $headerFunc( 206 );
 136                                 $headerFunc( 'Content-Length: ' . $range[2] );
 137                                 $headerFunc( "Content-Range: bytes {$range[0]}-{$range[1]}/{$info['size']}" );
 138                         } elseif ( $range === 'invalid' ) {
 139                                 if ( $sendErrors ) {
 140                                         $headerFunc( 416 );
 141                                         $headerFunc( 'Cache-Control: no-cache' );
 142                                         $headerFunc( 'Content-Type: text/html; charset=utf-8' );
 143                                         $headerFunc( 'Content-Range: bytes */' . $info['size'] );
 144                                 }
 145                                 return false;
 146                         } else { // unsupported Range request (e.g. multiple ranges)
 147                                 $range = null;
 148                                 $headerFunc( 'Content-Length: ' . $info['size'] );
 149                         }
 150                 } else {
 151                         $range = null;
 152                         $headerFunc( 'Content-Length: ' . $info['size'] );
 153                 }
 154
 155                 if ( is_array( $range ) ) {
 156                         $handle = fopen( $this->path, 'rb' );
 157                         if ( $handle ) {
 158                                 $ok = true;
 159                                 fseek( $handle, $range[0] );
 160                                 $remaining = $range[2];
 161                                 while ( $remaining > 0 && $ok ) {
 162                                         $bytes = min( $remaining, 8 * 1024 );
 163                                         $data = fread( $handle, $bytes );
 164                                         $remaining -= $bytes;
 165                                         $ok = ( $data !== false );
 166                                         print $data;
 167                                 }
 168                         } else {
 169                                 return false;
 170                         }
 171                 } else {
 172                         return readfile( $this->path ) !== false; // faster
 173                 }
 174
 175                 return true;
 176         }
 177
 178         /**
 179          * Send out a standard 404 message for a file
 180          *
 181          * @param string $fname Full name and path of the file to stream
 182          * @param int $flags Bitfield of STREAM_* constants
 183          * @since 1.24
 184          */
 185         public static function send404Message( $fname, $flags = 0 ) {
 186                 if ( ( $flags & self::STREAM_HEADLESS ) == 0 ) {
 187                         HttpStatus::header( 404 );
 188                         header( 'Cache-Control: no-cache' );
 189                         header( 'Content-Type: text/html; charset=utf-8' );
 190                 }
 191                 $encFile = htmlspecialchars( $fname );
 192                 $encScript = htmlspecialchars( $_SERVER['SCRIPT_NAME'] );
 193                 echo "<!DOCTYPE html><html><body>
 194                         <h1>File not found</h1>
 195                         <p>Although this PHP script ($encScript) exists, the file requested for output
 196                         ($encFile) does not.</p>
 197                         </body></html>
 198                         ";
 199         }
 200
 201         /**
 202          * Convert a Range header value to an absolute (start, end) range tuple
 203          *
 204          * @param string $range Range header value
 205          * @param int $size File size
 206          * @return array|string Returns error string on failure (start, end, length)
 207          * @since 1.24
 208          */
 209         public static function parseRange( $range, $size ) {
 210                 $m = [];
 211                 if ( preg_match( '#^bytes=(\d*)-(\d*)$#', $range, $m ) ) {
 212                         list( , $start, $end ) = $m;
 213                         if ( $start === '' && $end === '' ) {
 214                                 $absRange = [ 0, $size - 1 ];
 215                         } elseif ( $start === '' ) {
 216                                 $absRange = [ $size - $end, $size - 1 ];
 217                         } elseif ( $end === '' ) {
 218                                 $absRange = [ $start, $size - 1 ];
 219                         } else {
 220                                 $absRange = [ $start, $end ];
 221                         }
 222                         if ( $absRange[0] >= 0 && $absRange[1] >= $absRange[0] ) {
 223                                 if ( $absRange[0] < $size ) {
 224                                         $absRange[1] = min( $absRange[1], $size - 1 ); // stop at EOF
 225                                         $absRange[2] = $absRange[1] - $absRange[0] + 1;
 226                                         return $absRange;
 227                                 } elseif ( $absRange[0] == 0 && $size == 0 ) {
 228                                         return 'unrecognized'; // the whole file should just be sent
 229                                 }
 230                         }
 231                         return 'invalid';
 232                 }
 233                 return 'unrecognized';
 234         }
 235
 236         protected static function resetOutputBuffers() {
 237                 while ( ob_get_status() ) {
 238                         if ( !ob_end_clean() ) {
 239                                 // Could not remove output buffer handler; abort now
 240                                 // to avoid getting in some kind of infinite loop.
 241                                 break;
 242                         }
 243                 }
 244         }
 245
 246         /**
 247          * Determine the file type of a file based on the path
 248          *
 249          * @param string $filename Storage path or file system path
 250          * @return null|string
 251          */
 252         protected static function contentTypeFromPath( $filename ) {
 253                 $ext = strrchr( $filename, '.' );
 254                 $ext = $ext === false ? '' : strtolower( substr( $ext, 1 ) );
 255
 256                 switch ( $ext ) {
 257                         case 'gif':
 258                                 return 'image/gif';
 259                         case 'png':
 260                                 return 'image/png';
 261                         case 'jpg':
 262                                 return 'image/jpeg';
 263                         case 'jpeg':
 264                                 return 'image/jpeg';
 265                 }
 266
 267                 return 'unknown/unknown';
 268         }
 269 }