X-Git-Url: https://git.immae.eu/?a=blobdiff_plain;f=application%2FHttpUtils.php;h=e9282506188f4d19969e9a49ac27664171c79938;hb=bede8e1b633d720d6a7d1b05ba367811f3ac2b87;hp=ec54dcd4faa78a92bd3fb6533b7c4f9a6b608776;hpb=877491b4ad0a6a9119e667901cef40cc56116901;p=github%2Fshaarli%2FShaarli.git

diff --git a/application/HttpUtils.php b/application/HttpUtils.php
index ec54dcd4..e9282506 100644
--- a/application/HttpUtils.php
+++ b/application/HttpUtils.php
@@ -1,11 +1,13 @@
 <?php
 /**
  * GET an HTTP URL to retrieve its content
- * Uses the cURL library or a fallback method 
+ * Uses the cURL library or a fallback method
  *
- * @param string $url      URL to get (http://...)
- * @param int    $timeout  network timeout (in seconds)
- * @param int    $maxBytes maximum downloaded bytes (default: 4 MiB)
+ * @param string          $url               URL to get (http://...)
+ * @param int             $timeout           network timeout (in seconds)
+ * @param int             $maxBytes          maximum downloaded bytes (default: 4 MiB)
+ * @param callable|string $curlWriteFunction Optional callback called during the download (cURL CURLOPT_WRITEFUNCTION).
+ *                                           Can be used to add download conditions on the headers (response code, content type, etc.).
  *
  * @return array HTTP response headers, downloaded content
  *
@@ -29,7 +31,7 @@
  * @see http://stackoverflow.com/q/9183178
  * @see http://stackoverflow.com/q/1462720
  */
-function get_http_response($url, $timeout = 30, $maxBytes = 4194304)
+function get_http_response($url, $timeout = 30, $maxBytes = 4194304, $curlWriteFunction = null)
 {
     $urlObj = new Url($url);
     $cleanUrl = $urlObj->idnToAscii();
@@ -75,6 +77,10 @@ function get_http_response($url, $timeout = 30, $maxBytes = 4194304)
     curl_setopt($ch, CURLOPT_TIMEOUT,           $timeout);
     curl_setopt($ch, CURLOPT_USERAGENT,         $userAgent);
 
+    if (is_callable($curlWriteFunction)) {
+        curl_setopt($ch, CURLOPT_WRITEFUNCTION, $curlWriteFunction);
+    }
+
     // Max download size management
     curl_setopt($ch, CURLOPT_BUFFERSIZE,        1024*16);
     curl_setopt($ch, CURLOPT_NOPROGRESS,        false);
@@ -302,6 +308,13 @@ function server_url($server)
                 $port = $server['HTTP_X_FORWARDED_PORT'];
             }
 
+            // This is a workaround for proxies that don't forward the scheme properly.
+            // Connecting over port 443 has to be in HTTPS.
+            // See https://github.com/shaarli/Shaarli/issues/1022
+            if ($port == '443') {
+                $scheme = 'https';
+            }
+
             if (($scheme == 'http' && $port != '80')
                 || ($scheme == 'https' && $port != '443')
             ) {
@@ -402,6 +415,37 @@ function getIpAddressFromProxy($server, $trustedIps)
     return array_pop($ips);
 }
 
+
+/**
+ * Return an identifier based on the advertised client IP address(es)
+ *
+ * This aims at preventing session hijacking from users behind the same proxy
+ * by relying on HTTP headers.
+ *
+ * See:
+ * - https://secure.php.net/manual/en/reserved.variables.server.php
+ * - https://stackoverflow.com/questions/3003145/how-to-get-the-client-ip-address-in-php
+ * - https://stackoverflow.com/questions/12233406/preventing-session-hijacking
+ * - https://stackoverflow.com/questions/21354859/trusting-x-forwarded-for-to-identify-a-visitor
+ *
+ * @param array $server The $_SERVER array
+ *
+ * @return string An identifier based on client IP address information
+ */
+function client_ip_id($server)
+{
+    $ip = $server['REMOTE_ADDR'];
+
+    if (isset($server['HTTP_X_FORWARDED_FOR'])) {
+        $ip = $ip . '_' . $server['HTTP_X_FORWARDED_FOR'];
+    }
+    if (isset($server['HTTP_CLIENT_IP'])) {
+        $ip = $ip . '_' . $server['HTTP_CLIENT_IP'];
+    }
+    return $ip;
+}
+
+
 /**
  * Returns true if Shaarli's currently browsed in HTTPS.
  * Supports reverse proxies (if the headers are correctly set).