X-Git-Url: https://git.immae.eu/?a=blobdiff_plain;f=application%2FLinkUtils.php;h=d56e019f4246f40d94f7a2eb5b8f715d10c7f1a1;hb=1004742f09b55ff781c13745781b9a7e90986faa;hp=e3d95d0860d872dd371173a3b45eecf5c558f4f3;hpb=286757ab29660916f6b66d1fc0ad3b53aef337c8;p=github%2Fshaarli%2FShaarli.git

diff --git a/application/LinkUtils.php b/application/LinkUtils.php
index e3d95d08..d56e019f 100644
--- a/application/LinkUtils.php
+++ b/application/LinkUtils.php
@@ -1,60 +1,90 @@
 <?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;
+    $isRedirected = 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, &$isRedirected) {
+        $responseCode = $curlGetInfo($ch, CURLINFO_RESPONSE_CODE);
+        if (!empty($responseCode) && in_array($responseCode, [301, 302])) {
+            $isRedirected = true;
+            return strlen($data);
+        }
+        if (!empty($responseCode) && $responseCode !== 200) {
+            return false;
+        }
+        // After a redirection, the content type will keep the previous request value
+        // until it finds the next content-type header.
+        if (! $isRedirected || strpos(strtolower($data), 'content-type') !== false) {
+            $contentType = $curlGetInfo($ch, CURLINFO_CONTENT_TYPE);
+        }
+        if (!empty($contentType) && strpos($contentType, 'text/html') === false) {
+            return false;
+        }
+        if (!empty($contentType) && empty($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;
@@ -171,7 +201,8 @@ function space2nbsp($text)
 
  * @return string formatted description.
  */
-function format_description($description, $redirector = '', $urlEncode = true, $indexUrl = '') {
+function format_description($description, $redirector = '', $urlEncode = true, $indexUrl = '')
+{
     return nl2br(space2nbsp(hashtag_autolink(text2clickable($description, $redirector, $urlEncode), $indexUrl)));
 }