thumb-handler.php

   1 <?php
   2
   3 # Valid web server entry point
   4 define( 'THUMB_HANDLER', true );
   5
   6 # Load thumb-handler configuration. We don't want to use
   7 # WebStart.php or the like as it would kill performance.
   8 $configPath = dirname( __FILE__ ) . "/thumb.config.php";
   9 if ( !file_exists( $configPath ) ) {
  10         die( "Thumb-handler.php is not enabled for this wiki.\n" );
  11 }
  12 require( $configPath );
  13
  14 function wfHandleThumb404() {
  15         global $thgThumbCallbacks, $thgThumb404File;
  16
  17         # lighttpd puts the original request in REQUEST_URI, while
  18         # sjs sets that to the 404 handler, and puts the original
  19         # request in REDIRECT_URL.
  20         if ( isset( $_SERVER['REDIRECT_URL'] ) ) {
  21                 # The URL is un-encoded, so put it back how it was.
  22                 $uri = str_replace( "%2F", "/", urlencode( $_SERVER['REDIRECT_URL'] ) );
  23         } else {
  24                 $uri = $_SERVER['REQUEST_URI'];
  25         }
  26
  27         # Extract thumb.php params from the URI...
  28         if ( isset( $thgThumbCallbacks['extractParams'] )
  29                 && is_callable( $thgThumbCallbacks['extractParams'] ) )
  30         {
  31                 # Overridden by configuration
  32                 $params = call_user_func_array( $thgThumbCallbacks['extractParams'], array( $uri ) );
  33         } else {
  34                 $params = wfExtractThumbParams( $uri ); // basic wiki URL param extracting
  35         }
  36         if ( $params === null ) { // not a valid thumb request
  37                 header( 'X-Debug: no regex match' ); // useful for debugging
  38                 require_once( $thgThumb404File ); // standard 404 message
  39                 return;
  40         }
  41
  42         # Do some basic checks on the filename...
  43         if ( preg_match( '/[\x80-\xff]/', $uri ) ) {
  44                 header( 'HTTP/1.0 400 Bad request' );
  45                 header( 'Content-Type: text/html' );
  46                 print "<html><head><title>Bad request</title></head><body>" .
  47                         "The URI contained bytes with the high bit set, this is not allowed." .
  48                         "</body></html>";
  49                 return;
  50         } elseif ( strpos( $params['f'], '%20' ) !== false ) {
  51                 header( 'HTTP/1.0 404 Not found' );
  52                 header( 'Content-Type: text/html' );
  53                 header( 'X-Debug: filename contains a space' ); // useful for debugging
  54                 print "<html><head><title>Not found</title></head><body>" .
  55                         "The URL contained spaces, we don't have any thumbnail files with spaces." .
  56                         "</body></html>";
  57                 return;
  58         }
  59
  60         # Check any backend caches for the thumbnail...
  61         if ( isset( $thgThumbCallbacks['checkCache'] )
  62                 && is_callable( $thgThumbCallbacks['checkCache'] ) )
  63         {
  64                 if ( call_user_func_array( $thgThumbCallbacks['checkCache'], array( $uri, $params ) ) ) {
  65                         return; // file streamed from backend thumb cache
  66                 }
  67         }
  68
  69         wfStreamThumbViaCurl( $params, $uri );
  70 }
  71
  72 /**
  73  * Extract the required params for thumb.php from the thumbnail request URI.
  74  * At least 'width' and 'f' should be set if the result is an array.
  75  *
  76  * @param $uri String Thumbnail request URI
  77  * @return Array|null associative params array or null
  78  */
  79 function wfExtractThumbParams( $uri ) {
  80         global $thgThumbServer, $thgThumbFragment, $thgThumbHashFragment;
  81
  82         $thumbRegex = '!^(?:' . preg_quote( $thgThumbServer ) . ')?/' .
  83                 preg_quote( $thgThumbFragment ) . '(/archive|/temp|)/' .
  84                 $thgThumbHashFragment . '([^/]*)/(page(\d*)-)*(\d*)px-[^/]*$!';
  85
  86         # Is this a thumbnail?
  87         if ( preg_match( $thumbRegex, $uri, $matches ) ) {
  88                 list( $all, $archOrTemp, $filename, $pagefull, $pagenum, $size ) = $matches;
  89                 $params = array( 'f' => $filename, 'width' => $size );
  90                 if ( $pagenum ) {
  91                         $params['page'] = $pagenum;
  92                 }
  93                 if ( $archOrTemp == '/archive' ) {
  94                         $params['archived'] = 1;
  95                 } elseif ( $archOrTemp == '/temp' ) {
  96                         $params['temp'] = 1;
  97                 }
  98         } else {
  99                 $params = null;
 100         }
 101
 102         return $params;
 103 }
 104
 105 /**
 106  * cURL to thumb.php and stream back the resulting file or give an error message.
 107  *
 108  * @param $params Array Parameters to thumb.php
 109  * @param $uri String Thumbnail request URI
 110  * @return void
 111  */
 112 function wfStreamThumbViaCurl( array $params, $uri ) {
 113         global $thgThumbCallbacks, $thgThumbScriptPath, $thgThumbCurlProxy, $thgThumbCurlTimeout;
 114
 115         if ( !function_exists( 'curl_init' ) ) {
 116                 header( 'HTTP/1.0 404 Not found' );
 117                 header( 'Content-Type: text/html' );
 118                 header( 'X-Debug: cURL is not enabled' ); // useful for debugging
 119                 print "<html><head><title>Not found</title></head><body>" .
 120                         "cURL is not enabled for PHP on this wiki. Unable to send request thumb.php." .
 121                         "</body></html>";
 122                 return;
 123         }
 124
 125         # Build up the request URL to use with CURL...
 126         $reqURL = "{$thgThumbScriptPath}?";
 127         $first = true;
 128         foreach ( $params as $name => $value ) {
 129                 if ( $first ) {
 130                         $first = false;
 131                 } else {
 132                         $reqURL .= '&';
 133                 }
 134                 $reqURL .= "$name=$value"; // Note: value is already urlencoded
 135         }
 136
 137         # Set relevant HTTP headers...
 138         $headers = array();
 139         $headers[] = "X-Original-URI: " . str_replace( "\n", '', $uri );
 140         if ( isset( $thgThumbCallbacks['curlHeaders'] )
 141                 && is_callable( $thgThumbCallbacks['curlHeaders'] ) )
 142         {
 143                 # Add on any custom headers (like XFF)
 144                 call_user_func_array( $thgThumbCallbacks['curlHeaders'], array( &$headers ) );
 145         }
 146
 147         # Pass through some other headers...
 148         $passThrough = array( 'If-Modified-Since', 'Referer', 'User-Agent' );
 149         foreach ( $passThrough as $headerName ) {
 150                 $serverVarName = 'HTTP_' . str_replace( '-', '_', strtoupper( $headerName ) );
 151                 if ( !empty( $_SERVER[$serverVarName] ) ) {
 152                         $headers[] = $headerName . ': ' .
 153                                 str_replace( "\n", '', $_SERVER[$serverVarName] );
 154                 }
 155         }
 156
 157         $ch = curl_init( $reqURL );
 158         if ( $thgThumbCurlProxy ) {
 159                 curl_setopt( $ch, CURLOPT_PROXY, $thgThumbCurlProxy );
 160         }
 161
 162         curl_setopt( $ch, CURLOPT_HTTPHEADER, $headers );
 163         curl_setopt( $ch, CURLOPT_RETURNTRANSFER, true );
 164         curl_setopt( $ch, CURLOPT_TIMEOUT, $thgThumbCurlTimeout );
 165
 166         # Actually make the request
 167         $text = curl_exec( $ch );
 168
 169         # Send it on to the client
 170         $errno = curl_errno( $ch );
 171         $contentType = curl_getinfo( $ch, CURLINFO_CONTENT_TYPE );
 172         $httpCode = curl_getinfo( $ch, CURLINFO_HTTP_CODE );
 173         if ( $errno ) {
 174                 header( 'HTTP/1.1 500 Internal server error' );
 175                 header( 'Cache-Control: no-cache' );
 176                 list( $text, $contentType ) = wfCurlErrorText( $ch );
 177         } elseif ( $httpCode == 304 ) {
 178                 header( 'HTTP/1.1 304 Not modified' );
 179                 $contentType = '';
 180                 $text = '';
 181         } elseif ( strval( $text ) == '' ) {
 182                 header( 'HTTP/1.1 500 Internal server error' );
 183                 header( 'Cache-Control: no-cache' );
 184                 list( $text, $contentType ) = wfCurlEmptyText( $ch );
 185         } elseif ( $httpCode == 404 ) {
 186                 header( 'HTTP/1.1 404 Not found' );
 187                 header( 'Cache-Control: s-maxage=300, must-revalidate, max-age=0' );
 188         } elseif ( $httpCode != 200
 189                 || substr( $contentType, 0, 9 ) == 'text/html'
 190                 || substr( $text, 0, 5 ) == '<html' )
 191         {
 192                 # Error message, suppress cache
 193                 header( 'HTTP/1.1 500 Internal server error' );
 194                 header( 'Cache-Control: no-cache' );
 195         } else {
 196                 # OK thumbnail; save to any backend caches...
 197                 if ( isset( $thgThumbCallbacks['fillCache'] )
 198                         && is_callable( $thgThumbCallbacks['fillCache'] ) )
 199                 {
 200                         call_user_func_array( $thgThumbCallbacks['fillCache'], array( $uri, $text ) );
 201                 }
 202         }
 203
 204         if ( !$contentType ) {
 205                 header( 'Content-Type:' );
 206         } else {
 207                 header( "Content-Type: $contentType" );
 208         }
 209
 210         print $text; // thumb data or error text
 211
 212         curl_close( $ch );
 213 }
 214
 215 /**
 216  * Get error message and content type for when the cURL response is empty.
 217  *
 218  * @param $ch cURL handle
 219  * @return Array (error html, content type)
 220  */
 221 function wfCurlErrorText( $ch ) {
 222         $contentType = 'text/html';
 223         $error = htmlspecialchars( curl_error( $ch ) );
 224         $text = <<<EOT
 225 <html>
 226 <head><title>Thumbnail error</title></head>
 227 <body>Error retrieving thumbnail from scaling server: $error</body>
 228 </html>
 229 EOT;
 230         return array( $text, $contentType );
 231 }
 232
 233 /**
 234  * Get error message and content type for when the cURL response is an error.
 235  *
 236  * @param $ch cURL handle
 237  * @return Array (error html, content type)
 238  */
 239 function wfCurlEmptyText( $ch ) {
 240         $contentType = 'text/html';
 241         $error = htmlspecialchars( curl_error( $ch ) );
 242         $text = <<<EOT
 243 <html>
 244 <head><title>Thumbnail error</title></head>
 245 <body>Error retrieving thumbnail from scaling server: empty response</body>
 246 </html>
 247 EOT;
 248         return array( $text, $contentType );
 249 }
 250
 251 # Entry point
 252 wfHandleThumb404();