From 2cb87aca330bf773da47c1a4c8101a00e83b4e7a Mon Sep 17 00:00:00 2001
From: Hypolite Petovan <hypolite@mrpetovan.com>
Date: Tue, 22 Sep 2020 20:35:08 -0400
Subject: [PATCH] Improve charset detection in Util\ParseUrl

- Pages with charset meta tag weren't properly decoded
---
 src/Util/ParseUrl.php | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)

diff --git a/src/Util/ParseUrl.php b/src/Util/ParseUrl.php
index 01ad79d4f1..bb3ebbc10b 100644
--- a/src/Util/ParseUrl.php
+++ b/src/Util/ParseUrl.php
@@ -201,9 +201,18 @@ class ParseUrl
 			}
 		}
 
-		// Fetch the first mentioned charset. Can be in body or header
 		$charset = '';
-		if (preg_match('/charset=(.*?)[\'"\s\n]/', $header, $matches)) {
+		// Look for a charset, first in headers
+		// Expected form: Content-Type: text/html; charset=ISO-8859-4
+		if (preg_match('/charset=(.+?)\s/', $header, $matches)) {
+			$charset = trim(trim(trim(array_pop($matches)), ';,'));
+		}
+
+		// Then in body that gets precedence
+		// Expected forms:
+		// - <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+		// - <meta charset="utf-8">
+		if (preg_match('/charset=["\']?([^\'"]*?)[\'"]/', $body, $matches)) {
 			$charset = trim(trim(trim(array_pop($matches)), ';,'));
 		}