X-Git-Url: https://git.immae.eu/?a=blobdiff_plain;f=application%2Fbookmark%2FLinkUtils.php;h=faf5dbfd4fe24906bf980d8f4cc72e0472b7e008;hb=458b6b9918ec27154dd45416947bb93bedb97109;hp=35a5b290454b32870464c6457fea73059d613f4c;hpb=786f35f2700d83687ef494b98e3dace37d268e1c;p=github%2Fshaarli%2FShaarli.git

diff --git a/application/bookmark/LinkUtils.php b/application/bookmark/LinkUtils.php
index 35a5b290..faf5dbfd 100644
--- a/application/bookmark/LinkUtils.php
+++ b/application/bookmark/LinkUtils.php
@@ -1,64 +1,6 @@
 <?php
 
-use Shaarli\Bookmark\LinkDB;
-
-/**
- * Get cURL callback function for CURLOPT_WRITEFUNCTION
- *
- * @param string $charset     to extract from the downloaded page (reference)
- * @param string $title       to extract from the downloaded page (reference)
- * @param string $curlGetInfo Optionally overrides curl_getinfo function
- *
- * @return Closure
- */
-function get_curl_download_callback(&$charset, &$title, $curlGetInfo = 'curl_getinfo')
-{
-    $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);
-    };
-}
+use Shaarli\Bookmark\Bookmark;
 
 /**
  * Extract title from an HTML document.
@@ -84,7 +26,7 @@ function html_extract_title($html)
  */
 function header_extract_charset($header)
 {
-    preg_match('/charset="?([^; ]+)/i', $header, $match);
+    preg_match('/charset=["\']?([^; "\']+)/i', $header, $match);
     if (! empty($match[1])) {
         return strtolower(trim($match[1]));
     }
@@ -111,30 +53,42 @@ function html_extract_charset($html)
 }
 
 /**
- * Count private links in given linklist.
+ * Extract meta tag from HTML content in either:
+ *   - OpenGraph: <meta property="og:[tag]" ...>
+ *   - Meta tag: <meta name="[tag]" ...>
  *
- * @param array|Countable $links Linklist.
+ * @param string $tag  Name of the tag to retrieve.
+ * @param string $html HTML content where to look for charset.
  *
- * @return int Number of private links.
+ * @return bool|string Charset string if found, false otherwise.
  */
-function count_private($links)
+function html_extract_tag($tag, $html)
 {
-    $cpt = 0;
-    foreach ($links as $link) {
-        if ($link['private']) {
-            $cpt += 1;
-        }
+    $propertiesKey = ['property', 'name', 'itemprop'];
+    $properties = implode('|', $propertiesKey);
+    // We need a OR here to accept either 'property=og:noquote' or 'property="og:unrelated og:my-tag"'
+    $orCondition  = '["\']?(?:og:)?'. $tag .'["\']?|["\'][^\'"]*?(?:og:)?' . $tag . '[^\'"]*?[\'"]';
+    // Try to retrieve OpenGraph image.
+    $ogRegex = '#<meta[^>]+(?:'. $properties .')=(?:'. $orCondition .')[^>]*content=["\'](.*?)["\'].*?>#';
+    // If the attributes are not in the order property => content (e.g. Github)
+    // New regex to keep this readable... more or less.
+    $ogRegexReverse = '#<meta[^>]+content=["\'](.*?)["\'][^>]+(?:'. $properties .')=(?:'. $orCondition .').*?>#';
+
+    if (preg_match($ogRegex, $html, $matches) > 0
+        || preg_match($ogRegexReverse, $html, $matches) > 0
+    ) {
+        return $matches[1];
     }
 
-    return $cpt;
+    return false;
 }
 
 /**
- * In a string, converts URLs to clickable links.
+ * In a string, converts URLs to clickable bookmarks.
  *
  * @param string $text       input string.
  *
- * @return string returns $text with all links converted to HTML links.
+ * @return string returns $text with all bookmarks converted to HTML bookmarks.
  *
  * @see Function inspired from http://www.php.net/manual/en/function.preg-replace.php#85722
  */
@@ -162,7 +116,7 @@ function hashtag_autolink($description, $indexUrl = '')
      * \p{Mn} - any non marking space (accents, umlauts, etc)
      */
     $regex = '/(^|\s)#([\p{Pc}\p{N}\p{L}\p{Mn}]+)/mui';
-    $replacement = '$1<a href="'. $indexUrl .'?addtag=$2" title="Hashtag $2">#$2</a>';
+    $replacement = '$1<a href="'. $indexUrl .'./add-tag/$2" title="Hashtag $2">#$2</a>';
     return preg_replace($regex, $replacement, $description);
 }
 
@@ -202,7 +156,7 @@ function format_description($description, $indexUrl = '')
  */
 function link_small_hash($date, $id)
 {
-    return smallHash($date->format(LinkDB::LINK_DATE_FORMAT) . $id);
+    return smallHash($date->format(Bookmark::LINK_DATE_FORMAT) . $id);
 }
 
 /**