Extract the title/charset during page download, and check content type

Use CURLOPT_WRITEFUNCTION to check the response code and content type (only allow HTML).
Also extract the title and charset during downloading chunk of data, and stop it when everything has been extracted.

Closes #579

2 files changed, 65 insertions, 38 deletions

diff --git a/application/HttpUtils.php b/application/HttpUtils.php
index 00835966..2edf5ce2 100644
--- a/application/HttpUtils.php
+++ b/application/HttpUtils.php
@@ -3,9 +3,11 @@
  * GET an HTTP URL to retrieve its content
  * 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);
     curl_setopt($ch, CURLOPT_NOPROGRESS,        false);
diff --git a/application/LinkUtils.php b/application/LinkUtils.php
index 976474de..c0dd32a6 100644
--- a/application/LinkUtils.php
+++ b/application/LinkUtils.php
@@ -1,60 +1,81 @@
 <?php
 
 /**
- * Extract title from an HTML document.
+ * Get cURL callback function for CURLOPT_WRITEFUNCTION
  *
- * @param string $html HTML content where to look for a title.
+ * @param string $charset     to extract from the downloaded page (reference)
+ * @param string $title       to extract from the downloaded page (reference)
+ * @param string $curlGetInfo Optionnaly overrides curl_getinfo function
  *
- * @return bool|string Extracted title if found, false otherwise.
+ * @return Closure
  */
-function html_extract_title($html)
+function get_curl_download_callback(&$charset, &$title, $curlGetInfo = 'curl_getinfo')
 {
-    if (preg_match('!<title.*?>(.*?)</title>!is', $html, $matches)) {
-        return trim(str_replace("\n", '', $matches[1]));
-    }
-    return false;
+    /**
+     * cURL callback function for CURLOPT_WRITEFUNCTION (called during the download).
+     *
+     * While downloading the remote page, we check that the HTTP code is 200 and content type is 'html/text'
+     * Then we extract the title and the charset and stop the download when it's done.
+     *
+     * @param resource $ch   cURL resource
+     * @param string   $data chunk of data being downloaded
+     *
+     * @return int|bool length of $data or false if we need to stop the download
+     */
+    return function(&$ch, $data) use ($curlGetInfo, &$charset, &$title) {
+        $responseCode = $curlGetInfo($ch, CURLINFO_RESPONSE_CODE);
+        if (!empty($responseCode) && $responseCode != 200) {
+            return false;
+        }
+        $contentType = $curlGetInfo($ch, CURLINFO_CONTENT_TYPE);
+        if (!empty($contentType) && strpos($contentType, 'text/html') === false) {
+            return false;
+        }
+        if (empty($charset)) {
+            $charset = header_extract_charset($contentType);
+        }
+        if (empty($charset)) {
+            $charset = html_extract_charset($data);
+        }
+        if (empty($title)) {
+            $title = html_extract_title($data);
+        }
+        // We got everything we want, stop the download.
+        if (!empty($responseCode) && !empty($contentType) && !empty($charset) && !empty($title)) {
+            return false;
+        }
+
+        return strlen($data);
+    };
 }
 
 /**
- * Determine charset from downloaded page.
- * Priority:
- *   1. HTTP headers (Content type).
- *   2. HTML content page (tag <meta charset>).
- *   3. Use a default charset (default: UTF-8).
+ * Extract title from an HTML document.
  *
- * @param array  $headers           HTTP headers array.
- * @param string $htmlContent       HTML content where to look for charset.
- * @param string $defaultCharset    Default charset to apply if other methods failed.
+ * @param string $html HTML content where to look for a title.
  *
- * @return string Determined charset.
+ * @return bool|string Extracted title if found, false otherwise.
  */
-function get_charset($headers, $htmlContent, $defaultCharset = 'utf-8')
+function html_extract_title($html)
 {
-    if ($charset = headers_extract_charset($headers)) {
-        return $charset;
-    }
-
-    if ($charset = html_extract_charset($htmlContent)) {
-        return $charset;
+    if (preg_match('!<title.*?>(.*?)</title>!is', $html, $matches)) {
+        return trim(str_replace("\n", '', $matches[1]));
     }
-
-    return $defaultCharset;
+    return false;
 }
 
 /**
- * Extract charset from HTTP headers if it's defined.
+ * Extract charset from HTTP header if it's defined.
  *
- * @param array $headers HTTP headers array.
+ * @param string $header HTTP header Content-Type line.
  *
  * @return bool|string Charset string if found (lowercase), false otherwise.
  */
-function headers_extract_charset($headers)
+function header_extract_charset($header)
 {
-    if (! empty($headers['Content-Type']) && strpos($headers['Content-Type'], 'charset=') !== false) {
-        preg_match('/charset="?([^; ]+)/i', $headers['Content-Type'], $match);
-        if (! empty($match[1])) {
-            return strtolower(trim($match[1]));
-        }
+    preg_match('/charset="?([^; ]+)/i', $header, $match);
+    if (! empty($match[1])) {
+        return strtolower(trim($match[1]));
     }
 
     return false;