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          * Takes HTTP headers in a name => value format and converts them to the weird format
  44          * expected by stream().
  45          * @param string[] $headers
  46          * @return array[] [ $headers, $optHeaders ]
  47          * @since 1.34
  48          */
  49         public static function preprocessHeaders( $headers ) {
  50                 $rawHeaders = [];
  51                 $optHeaders = [];
  52                 foreach ( $headers as $name => $header ) {
  53                         $nameLower = strtolower( $name );
  54                         if ( in_array( $nameLower, [ 'range', 'if-modified-since' ], true ) ) {
  55                                 $optHeaders[$nameLower] = $header;
  56                         } else {
  57                                 $rawHeaders[] = "$name: $header";
  58                         }
  59                 }
  60                 return [ $rawHeaders, $optHeaders ];
  61         }
  62
  63         /**
  64          * @param string $path Local filesystem path to a file
  65          * @param array $params Options map, which includes:
  66          *   - obResetFunc : alternative callback to clear the output buffer
  67          *   - streamMimeFunc : alternative method to determine the content type from the path
  68          */
  69         public function __construct( $path, array $params = [] ) {
  70                 $this->path = $path;
  71                 $this->obResetFunc = isset( $params['obResetFunc'] )
  72                         ? $params['obResetFunc']
  73                         : [ __CLASS__, 'resetOutputBuffers' ];
  74                 $this->streamMimeFunc = isset( $params['streamMimeFunc'] )
  75                         ? $params['streamMimeFunc']
  76                         : [ __CLASS__, 'contentTypeFromPath' ];
  77         }
  78
  79         /**
  80          * Stream a file to the browser, adding all the headings and fun stuff.
  81          * Headers sent include: Content-type, Content-Length, Last-Modified,
  82          * and Content-Disposition.
  83          *
  84          * @param array $headers Any additional headers to send if the file exists
  85          * @param bool $sendErrors Send error messages if errors occur (like 404)
  86          * @param array $optHeaders HTTP request header map (e.g. "range") (use lowercase keys)
  87          * @param int $flags Bitfield of STREAM_* constants
  88          * @throws MWException
  89          * @return bool Success
  90          */
  91         public function stream(
  92                 $headers = [], $sendErrors = true, $optHeaders = [], $flags = 0
  93         ) {
  94                 // Don't stream it out as text/html if there was a PHP error
  95                 if ( ( ( $flags & self::STREAM_HEADLESS ) == 0 || $headers ) && headers_sent() ) {
  96                         echo "Headers already sent, terminating.\n";
  97                         return false;
  98                 }
  99
 100                 $headerFunc = ( $flags & self::STREAM_HEADLESS )
 101                         ? function ( $header ) {
 102                                 // no-op
 103                         }
 104                         : function ( $header ) {
 105                                 is_int( $header ) ? HttpStatus::header( $header ) : header( $header );
 106                         };
 107
 108                 Wikimedia\suppressWarnings();
 109                 $info = stat( $this->path );
 110                 Wikimedia\restoreWarnings();
 111
 112                 if ( !is_array( $info ) ) {
 113                         if ( $sendErrors ) {
 114                                 self::send404Message( $this->path, $flags );
 115                         }
 116                         return false;
 117                 }
 118
 119                 // Send Last-Modified HTTP header for client-side caching
 120                 $mtimeCT = new ConvertibleTimestamp( $info['mtime'] );
 121                 $headerFunc( 'Last-Modified: ' . $mtimeCT->getTimestamp( TS_RFC2822 ) );
 122
 123                 if ( ( $flags & self::STREAM_ALLOW_OB ) == 0 ) {
 124                         call_user_func( $this->obResetFunc );
 125                 }
 126
 127                 $type = call_user_func( $this->streamMimeFunc, $this->path );
 128                 if ( $type && $type != 'unknown/unknown' ) {
 129                         $headerFunc( "Content-type: $type" );
 130                 } else {
 131                         // Send a content type which is not known to Internet Explorer, to
 132                         // avoid triggering IE's content type detection. Sending a standard
 133                         // unknown content type here essentially gives IE license to apply
 134                         // whatever content type it likes.
 135                         $headerFunc( 'Content-type: application/x-wiki' );
 136                 }
 137
 138                 // Don't send if client has up to date cache
 139                 if ( isset( $optHeaders['if-modified-since'] ) ) {
 140                         $modsince = preg_replace( '/;.*$/', '', $optHeaders['if-modified-since'] );
 141                         if ( $mtimeCT->getTimestamp( TS_UNIX ) <= strtotime( $modsince ) ) {
 142                                 ini_set( 'zlib.output_compression', 0 );
 143                                 $headerFunc( 304 );
 144                                 return true; // ok
 145                         }
 146                 }
 147
 148                 // Send additional headers
 149                 foreach ( $headers as $header ) {
 150                         header( $header ); // always use header(); specifically requested
 151                 }
 152
 153                 if ( isset( $optHeaders['range'] ) ) {
 154                         $range = self::parseRange( $optHeaders['range'], $info['size'] );
 155                         if ( is_array( $range ) ) {
 156                                 $headerFunc( 206 );
 157                                 $headerFunc( 'Content-Length: ' . $range[2] );
 158                                 $headerFunc( "Content-Range: bytes {$range[0]}-{$range[1]}/{$info['size']}" );
 159                         } elseif ( $range === 'invalid' ) {
 160                                 if ( $sendErrors ) {
 161                                         $headerFunc( 416 );
 162                                         $headerFunc( 'Cache-Control: no-cache' );
 163                                         $headerFunc( 'Content-Type: text/html; charset=utf-8' );
 164                                         $headerFunc( 'Content-Range: bytes */' . $info['size'] );
 165                                 }
 166                                 return false;
 167                         } else { // unsupported Range request (e.g. multiple ranges)
 168                                 $range = null;
 169                                 $headerFunc( 'Content-Length: ' . $info['size'] );
 170                         }
 171                 } else {
 172                         $range = null;
 173                         $headerFunc( 'Content-Length: ' . $info['size'] );
 174                 }
 175
 176                 if ( is_array( $range ) ) {
 177                         $handle = fopen( $this->path, 'rb' );
 178                         if ( $handle ) {
 179                                 $ok = true;
 180                                 fseek( $handle, $range[0] );
 181                                 $remaining = $range[2];
 182                                 while ( $remaining > 0 && $ok ) {
 183                                         $bytes = min( $remaining, 8 * 1024 );
 184                                         $data = fread( $handle, $bytes );
 185                                         $remaining -= $bytes;
 186                                         $ok = ( $data !== false );
 187                                         print $data;
 188                                 }
 189                         } else {
 190                                 return false;
 191                         }
 192                 } else {
 193                         return readfile( $this->path ) !== false; // faster
 194                 }
 195
 196                 return true;
 197         }
 198
 199         /**
 200          * Send out a standard 404 message for a file
 201          *
 202          * @param string $fname Full name and path of the file to stream
 203          * @param int $flags Bitfield of STREAM_* constants
 204          * @since 1.24
 205          */
 206         public static function send404Message( $fname, $flags = 0 ) {
 207                 if ( ( $flags & self::STREAM_HEADLESS ) == 0 ) {
 208                         HttpStatus::header( 404 );
 209                         header( 'Cache-Control: no-cache' );
 210                         header( 'Content-Type: text/html; charset=utf-8' );
 211                 }
 212                 $encFile = htmlspecialchars( $fname );
 213                 $encScript = htmlspecialchars( $_SERVER['SCRIPT_NAME'] );
 214                 echo "<!DOCTYPE html><html><body>
 215                         <h1>File not found</h1>
 216                         <p>Although this PHP script ($encScript) exists, the file requested for output
 217                         ($encFile) does not.</p>
 218                         </body></html>
 219                         ";
 220         }
 221
 222         /**
 223          * Convert a Range header value to an absolute (start, end) range tuple
 224          *
 225          * @param string $range Range header value
 226          * @param int $size File size
 227          * @return array|string Returns error string on failure (start, end, length)
 228          * @since 1.24
 229          */
 230         public static function parseRange( $range, $size ) {
 231                 $m = [];
 232                 if ( preg_match( '#^bytes=(\d*)-(\d*)$#', $range, $m ) ) {
 233                         list( , $start, $end ) = $m;
 234                         if ( $start === '' && $end === '' ) {
 235                                 $absRange = [ 0, $size - 1 ];
 236                         } elseif ( $start === '' ) {
 237                                 $absRange = [ $size - $end, $size - 1 ];
 238                         } elseif ( $end === '' ) {
 239                                 $absRange = [ $start, $size - 1 ];
 240                         } else {
 241                                 $absRange = [ $start, $end ];
 242                         }
 243                         if ( $absRange[0] >= 0 && $absRange[1] >= $absRange[0] ) {
 244                                 if ( $absRange[0] < $size ) {
 245                                         $absRange[1] = min( $absRange[1], $size - 1 ); // stop at EOF
 246                                         $absRange[2] = $absRange[1] - $absRange[0] + 1;
 247                                         return $absRange;
 248                                 } elseif ( $absRange[0] == 0 && $size == 0 ) {
 249                                         return 'unrecognized'; // the whole file should just be sent
 250                                 }
 251                         }
 252                         return 'invalid';
 253                 }
 254                 return 'unrecognized';
 255         }
 256
 257         protected static function resetOutputBuffers() {
 258                 while ( ob_get_status() ) {
 259                         if ( !ob_end_clean() ) {
 260                                 // Could not remove output buffer handler; abort now
 261                                 // to avoid getting in some kind of infinite loop.
 262                                 break;
 263                         }
 264                 }
 265         }
 266
 267         /**
 268          * Determine the file type of a file based on the path
 269          *
 270          * @param string $filename Storage path or file system path
 271          * @return null|string
 272          */
 273         protected static function contentTypeFromPath( $filename ) {
 274                 $ext = strrchr( $filename, '.' );
 275                 $ext = $ext === false ? '' : strtolower( substr( $ext, 1 ) );
 276
 277                 switch ( $ext ) {
 278                         case 'gif':
 279                                 return 'image/gif';
 280                         case 'png':
 281                                 return 'image/png';
 282                         case 'jpg':
 283                                 return 'image/jpeg';
 284                         case 'jpeg':
 285                                 return 'image/jpeg';
 286                 }
 287
 288                 return 'unknown/unknown';
 289         }
 290 }