diff --git a/core/js/public/comments.js b/core/js/public/comments.js
index 9811528e4c..f4825f0864 100644
--- a/core/js/public/comments.js
+++ b/core/js/public/comments.js
@@ -21,7 +21,7 @@
 		 * The downside: anything not ascii is excluded. Not sure how common it is in areas using different
 		 * alphabets… the upside: fake domains with similar looking characters won't be formatted as links
 		 */
-		urlRegex: /(\s|^)(https?:\/\/)?((?:[-A-Z0-9+_]*\.)+[-A-Z]+(?:\/[-A-Z0-9+&@#%?=~_|!:,.;()]*)*)(\s|$)/ig,
+		urlRegex: /(\s|^)(https?:\/\/)?((?:[-A-Z0-9+_]+\.)+[-A-Z]+(?:\/[-A-Z0-9+&@#%?=~_|!:,.;()]*)*)(\s|$)/ig,
 
 		plainToRich: function(content) {
 			content = this.formatLinksRich(content);
diff --git a/core/js/tests/specs/public/commentsSpec.js b/core/js/tests/specs/public/commentsSpec.js
index 57fd7264d2..77b00369ca 100644
--- a/core/js/tests/specs/public/commentsSpec.js
+++ b/core/js/tests/specs/public/commentsSpec.js
@@ -38,6 +38,7 @@ describe('OCP.Comments tests', function() {
 			{input: 'hi noreply@nextcloud.com', expected: 'hi noreply@nextcloud.com'},
 			{input: 'hi <noreply@nextcloud.com>', expected: 'hi <noreply@nextcloud.com>'},
 			{input: 'FirebaseInstanceId.getInstance().deleteInstanceId()', expected: 'FirebaseInstanceId.getInstance().deleteInstanceId()'},
+			{input: 'I mean...it', expected: 'I mean...it'},
 		];
 	}