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 $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 ( function_exists( 'wfCustomExtractThumbParams' ) ) {
  29                 $params = wfCustomExtractThumbParams( $uri ); // overridden by configuration
  30         } else {
  31                 $params = wfExtractThumbParams( $uri ); // basic wiki URL param extracting
  32         }
  33         if ( $params === null ) { // not a valid thumb request
  34                 header( 'X-Debug: no regex match' ); // useful for debugging
  35                 require_once( $thgThumb404File ); // standard 404 message
  36                 return;
  37         }
  38
  39         # Do some basic checks on the filename...
  40         if ( preg_match( '/[\x80-\xff]/', $uri ) ) {
  41                 header( 'HTTP/1.0 400 Bad request' );
  42                 header( 'Content-Type: text/html' );
  43                 echo "<html><head><title>Bad request</title></head><body>" .
  44                         "The URI contained bytes with the high bit set, this is not allowed." .
  45                         "</body></html>";
  46                 return;
  47         } elseif ( strpos( $params['f'], '%20' ) !== false ) {
  48                 header( 'HTTP/1.0 404 Not found' );
  49                 header( 'Content-Type: text/html' );
  50                 header( 'X-Debug: filename contains a space' ); // useful for debugging
  51                 echo "<html><head><title>Not found</title></head><body>" .
  52                         "The URL contained spaces, we don't have any thumbnail files with spaces." .
  53                         "</body></html>";
  54                 return;
  55         }
  56
  57         wfStreamThumbViaCurl( $params, $uri );
  58 }
  59
  60 /**
  61  * Extract the required params for thumb.php from the thumbnail request URI.
  62  * At least 'width' and 'f' should be set if the result is an array.
  63  *
  64  * @param $uri String Thumbnail request URI
  65  * @return Array|null associative params array or null
  66  */
  67 function wfExtractThumbParams( $uri ) {
  68         global $thgThumbServer, $thgThumbFragment, $thgThumbHashFragment;
  69
  70         $thumbRegex = '!^(?:' . preg_quote( $thgThumbServer ) . ')?/' .
  71                 preg_quote( $thgThumbFragment ) . '(/archive|/temp|)/' .
  72                 $thgThumbHashFragment . '([^/]*)/' . '(page(\d*)-)*(\d*)px-([^/]*)$!';
  73
  74         # Is this a thumbnail?
  75         if ( preg_match( $thumbRegex, $uri, $matches ) ) {
  76                 list( $all, $archOrTemp, $filename, $pagefull, $pagenum, $size, $fn2 ) = $matches;
  77                 $params = array( 'f' => $filename, 'width' => $size );
  78                 if ( $pagenum ) {
  79                         $params['page'] = $pagenum;
  80                 }
  81                 if ( $archOrTemp == '/archive' ) {
  82                         $params['archived'] = 1;
  83                 } elseif ( $archOrTemp == '/temp' ) {
  84                         $params['temp'] = 1;
  85                 }
  86         } else {
  87                 $params = null;
  88         }
  89
  90         return $params;
  91 }
  92
  93 /**
  94  * cURL to thumb.php and stream back the resulting file or give an error message.
  95  *
  96  * @param $params Array Parameters to thumb.php
  97  * @param $uri String Thumbnail request URI
  98  * @return void
  99  */
 100 function wfStreamThumbViaCurl( array $params, $uri ) {
 101         global $thgThumbScriptPath, $thgThumbCurlProxy, $thgThumbCurlTimeout;
 102
 103         if ( !function_exists( 'curl_init' ) ) {
 104                 header( 'HTTP/1.0 404 Not found' );
 105                 header( 'Content-Type: text/html' );
 106                 header( 'X-Debug: cURL is not enabled' ); // useful for debugging
 107                 echo "<html><head><title>Not found</title></head><body>" .
 108                         "cURL is not enabled for PHP on this wiki. Unable to send request thumb.php." .
 109                         "</body></html>";
 110                 return;
 111         }
 112
 113         # Build up the request URL to use with CURL...
 114         $reqURL = "{$thgThumbScriptPath}?";
 115         $first = true;
 116         foreach ( $params as $name => $value ) {
 117                 if ( $first ) {
 118                         $first = false;
 119                 } else {
 120                         $reqURL .= '&';
 121                 }
 122                 // Note: value is already urlencoded
 123                 $reqURL .= "$name=$value";
 124         }
 125
 126         $ch = curl_init( $reqURL );
 127         if ( $thgThumbCurlProxy ) {
 128                 curl_setopt( $ch, CURLOPT_PROXY, $thgThumbCurlProxy );
 129         }
 130
 131         $headers = array(); // HTTP headers
 132         # Set certain headers...
 133         $headers[] = "X-Original-URI: " . str_replace( "\n", '', $uri );
 134         if ( function_exists( 'wfCustomThumbRequestHeaders' ) ) {
 135                 wfCustomThumbRequestHeaders( $headers ); // add on any custom headers (like XFF)
 136         }
 137         # Pass through some other headers...
 138         $passthrough = array( 'If-Modified-Since', 'Referer', 'User-Agent' );
 139         foreach ( $passthrough as $headerName ) {
 140                 $serverVarName = 'HTTP_' . str_replace( '-', '_', strtoupper( $headerName ) );
 141                 if ( !empty( $_SERVER[$serverVarName] ) ) {
 142                         $headers[] = $headerName . ': ' .
 143                                 str_replace( "\n", '', $_SERVER[$serverVarName] );
 144                 }
 145         }
 146
 147         curl_setopt( $ch, CURLOPT_HTTPHEADER, $headers );
 148         curl_setopt( $ch, CURLOPT_RETURNTRANSFER, true );
 149         curl_setopt( $ch, CURLOPT_TIMEOUT, $thgThumbCurlTimeout );
 150
 151         # Actually make the request
 152         $text = curl_exec( $ch );
 153
 154         # Send it on to the client
 155         $errno = curl_errno( $ch );
 156         $contentType = curl_getinfo( $ch, CURLINFO_CONTENT_TYPE );
 157         $httpCode = curl_getinfo( $ch, CURLINFO_HTTP_CODE );
 158         if ( $errno ) {
 159                 header( 'HTTP/1.1 500 Internal server error' );
 160                 header( 'Cache-Control: no-cache' );
 161                 list( $text, $contentType ) = wfCurlErrorText( $ch );
 162         } elseif ( $httpCode == 304 ) {
 163                 header( 'HTTP/1.1 304 Not modified' );
 164                 $contentType = '';
 165                 $text = '';
 166         } elseif ( strval( $text ) == '' ) {
 167                 header( 'HTTP/1.1 500 Internal server error' );
 168                 header( 'Cache-Control: no-cache' );
 169                 list( $text, $contentType ) = wfCurlEmptyText( $ch );
 170         } elseif ( $httpCode == 404 ) {
 171                 header( 'HTTP/1.1 404 Not found' );
 172                 header( 'Cache-Control: s-maxage=300, must-revalidate, max-age=0' );
 173         } elseif ( $httpCode != 200
 174                 || substr( $contentType, 0, 9 ) == 'text/html'
 175                 || substr( $text, 0, 5 ) == '<html' )
 176         {
 177                 # Error message, suppress cache
 178                 header( 'HTTP/1.1 500 Internal server error' );
 179                 header( 'Cache-Control: no-cache' );
 180         }
 181
 182         if ( !$contentType ) {
 183                 header( 'Content-Type:' );
 184         } else {
 185                 header( "Content-Type: $contentType" );
 186         }
 187
 188         print $text; // thumb data or error text
 189
 190         curl_close( $ch );
 191 }
 192
 193 /**
 194  * Get error message and content type for when the cURL response is empty.
 195  *
 196  * @param $ch cURL handle
 197  * @return Array (error html, content type)
 198  */
 199 function wfCurlErrorText( $ch ) {
 200         $contentType = 'text/html';
 201         $error = htmlspecialchars( curl_error( $ch ) );
 202         $text = <<<EOT
 203 <html>
 204 <head><title>Thumbnail error</title></head>
 205 <body>Error retrieving thumbnail from scaling server: $error</body>
 206 </html>
 207 EOT;
 208         return array( $text, $contentType );
 209 }
 210
 211 /**
 212  * Get error message and content type for when the cURL response is an error.
 213  *
 214  * @param $ch cURL handle
 215  * @return Array (error html, content type)
 216  */
 217 function wfCurlEmptyText( $ch ) {
 218         $contentType = 'text/html';
 219         $error = htmlspecialchars( curl_error( $ch ) );
 220         $text = <<<EOT
 221 <html>
 222 <head><title>Thumbnail error</title></head>
 223 <body>Error retrieving thumbnail from scaling server: empty response</body>
 224 </html>
 225 EOT;
 226         return array( $text, $contentType );
 227 }
 228
 229 # Entry point
 230 wfHandleThumb404();