Merge pull request #11531 from annando/display-polls

[friendica.git] / src / Content / PageInfo.php

diff --git a/src/Content/PageInfo.php b/src/Content/PageInfo.php
index 3cfab3769d28d86b737cd77af437614ec37a717f..41ecb3d7fa35a0ec207747a7c174fe9707e51b39 100644 (file)
--- a/src/Content/PageInfo.php
+++ b/src/Content/PageInfo.php
@@ -1,6 +1,6 @@
 <?php
 /**
- * @copyright Copyright (C) 2020, Friendica
+ * @copyright Copyright (C) 2010-2022, the Friendica project
  *
  * @license GNU AGPL version 3 or any later version
  *
@@ -73,7 +73,7 @@ class PageInfo
                        // Additional link attachments are prepended before the existing [attachment] tag
                        $body = substr_replace($body, "\n[bookmark=" . $data['url'] . ']' . $linkTitle . "[/bookmark]\n", $existingAttachmentPos, 0);
                } else {
-                       $footer = PageInfo::getFooterFromData($data, $no_photos);
+                       $footer = self::getFooterFromData($data, $no_photos);
                        $body = self::stripTrailingUrlFromBody($body, $data['url']);
                        $body .= "\n" . $footer;
                }
@@ -155,7 +155,7 @@ class PageInfo
                                if (empty($data['text'])) {
                                        $data['text'] = $data['title'];
                                }
-               
+
                                if (empty($data['text'])) {
                                        $data['text'] = $data['url'];
                                }
@@ -246,17 +246,22 @@ class PageInfo
         * @param bool   $searchNakedUrls Whether we should pick a naked URL (outside of BBCode tags) as a last resort
         * @return string|null
         */
-       protected static function getRelevantUrlFromBody(string $body, bool $searchNakedUrls = false)
+       public static function getRelevantUrlFromBody(string $body, bool $searchNakedUrls = false)
        {
                $URLSearchString = 'https?://[^\[\]]*';
 
                // Fix for Mastodon where the mentions are in a different format
                $body = preg_replace("~\[url=($URLSearchString)]([#!@])(.*?)\[/url]~is", '$2[url=$1]$3[/url]', $body);
 
-               preg_match("~(?<![!#@])\[url]($URLSearchString)\[/url]$~is", $body, $matches);
+               // Remove all hashtags and mentions
+               $body = preg_replace("/([#@!])\[url\=(.*?)\](.*?)\[\/url\]/ism", '', $body);
+
+               // Search for pure links
+               preg_match("/\[url\](https?:.*?)\[\/url\]/ism", $body, $matches);
 
                if (!$matches) {
-                       preg_match("~(?<![!#@])\[url=($URLSearchString)].*\[/url]$~is", $body, $matches);
+                       // Search for links with descriptions
+                       preg_match("/\[url\=(https?:.*?)\].*?\[\/url\]/ism", $body, $matches);
                }
 
                if (!$matches && $searchNakedUrls) {
@@ -292,7 +297,8 @@ class PageInfo
                        }
 
                        // Stripping link labels that include a shortened version of the URL
-                       if (strpos($url, trim($match[1], '.…')) !== false) {
+                       $trimMatch = trim($match[1], '.…');
+                       if (!empty($trimMatch) && strpos($url, $trimMatch) !== false) {
                                return '';
                        }