From: Kohei Yoshino <kohei.yoshino@gmail.com>
Date: Fri, 22 Mar 2019 15:42:55 +0000 (-0400)
Subject: Bug 1538047 - Plain text attachment cut off due to wrong charset detection (UTF-8... 
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=73aa0badaf5c4c9c3ff6598ad07e77dd40b25fef;p=thirdparty%2Fbugzilla.git

Bug 1538047 - Plain text attachment cut off due to wrong charset detection (UTF-8 as Windows-1252)
---

diff --git a/Bugzilla/Util.pm b/Bugzilla/Util.pm
index 128642f0b..a59ba5d02 100644
--- a/Bugzilla/Util.pm
+++ b/Bugzilla/Util.pm
@@ -913,6 +913,12 @@ sub detect_encoding {
     $encoding = $decoded_as if $decoded_as;
   }
 
+  # Encode::Detect sometimes mis-detects UTF-8 as Windows-1252
+  if ($encoding && $encoding eq 'cp1252') {
+    my $decoder = guess_encoding($data, ('utf8', 'cp1252'));
+    $encoding = $decoder->name if ref $decoder;
+  }
+
   return $encoding;
 }