Fix link regex (#922)

* Fix link regex

* add some tests

---------

Co-authored-by: Matthias Pfefferle <pfefferle@users.noreply.github.com>

Author: mattwiebe

activitypub.php

@@ -34,7 +34,7 @@
 \defined( 'ACTIVITYPUB_MAX_IMAGE_ATTACHMENTS' ) || \define( 'ACTIVITYPUB_MAX_IMAGE_ATTACHMENTS', 3 );
 \defined( 'ACTIVITYPUB_HASHTAGS_REGEXP' ) || \define( 'ACTIVITYPUB_HASHTAGS_REGEXP', '(?:(?<=\s)|(?<=<p>)|(?<=<br>)|^)#([A-Za-z0-9_]+)(?:(?=\s|[[:punct:]]|$))' );
 \defined( 'ACTIVITYPUB_USERNAME_REGEXP' ) || \define( 'ACTIVITYPUB_USERNAME_REGEXP', '(?:([A-Za-z0-9\._-]+)@((?:[A-Za-z0-9_-]+\.)+[A-Za-z]+))' );
-\defined( 'ACTIVITYPUB_URL_REGEXP' ) || \define( 'ACTIVITYPUB_URL_REGEXP', '(www.|http:|https:)+[^\s]+[\w\/]' );
+\defined( 'ACTIVITYPUB_URL_REGEXP' ) || \define( 'ACTIVITYPUB_URL_REGEXP', '(https?:|www\.)\S+[\w\/]' );
 \defined( 'ACTIVITYPUB_CUSTOM_POST_CONTENT' ) || \define( 'ACTIVITYPUB_CUSTOM_POST_CONTENT', "<h2>[ap_title]</h2>\n\n[ap_content]\n\n[ap_hashtags]\n\n[ap_shortlink]" );
 \defined( 'ACTIVITYPUB_AUTHORIZED_FETCH' ) || \define( 'ACTIVITYPUB_AUTHORIZED_FETCH', false );
 \defined( 'ACTIVITYPUB_DISABLE_REWRITES' ) || \define( 'ACTIVITYPUB_DISABLE_REWRITES', false );

includes/class-link.php

@@ -62,7 +62,7 @@ public static function replace_with_links( $result ) {
 			$result[0] = 'https://' . $result[0];
 		}
 		$parsed_url = \wp_parse_url( html_entity_decode( $result[0] ) );
-		if ( ! $parsed_url ) {
+		if ( ! $parsed_url || empty( $parsed_url['host'] ) ) {
 			return $result[0];
 		}
 

tests/test-class-activitypub-link.php

@@ -37,6 +37,8 @@ public function the_content_provider() {
 			array( 'hello https://www.test.de test', 'hello <a href="https://www.test.de" target="_blank" rel="nofollow noopener noreferrer" translate="no"><span class="invisible">https://www.</span><span class="">test.de</span><span class="invisible"></span></a> test' ),
 			array( 'hello www.test.de test', 'hello <a href="https://www.test.de" target="_blank" rel="nofollow noopener noreferrer" translate="no"><span class="invisible">https://www.</span><span class="">test.de</span><span class="invisible"></span></a> test' ),
 			array( 'hello https://test:test@test.de test', 'hello <a href="https://test:test@test.de" target="_blank" rel="nofollow noopener noreferrer" translate="no"><span class="invisible">https://test:test@</span><span class="">test.de</span><span class="invisible"></span></a> test' ),
+			array( 'wwwfoocom', 'wwwfoocom' ),
+			array( 'www.foo.com', '<a href="https://www.foo.com" target="_blank" rel="nofollow noopener noreferrer" translate="no"><span class="invisible">https://www.</span><span class="">foo.com</span><span class="invisible"></span></a>' ),
 			array( $code, $code ),
 			array( $style, $style ),
 			array( $textarea, $textarea ),