Replace invalid UTF-8 sequences with U+FFFD in edit summaries

Bug: T95353
Change-Id: Ib22b563e7bc4022754752632f0eebfad00e8eb01

diff --git a/includes/Sanitizer.php b/includes/Sanitizer.php
index 44e4e3e..6779189 100644 (file)
--- a/includes/Sanitizer.php
+++ b/includes/Sanitizer.php
@@ -1262,8 +1262,9 @@ class Sanitizer {
        static function escapeHtmlAllowEntities( $html ) {
                $html = Sanitizer::decodeCharReferences( $html );
                # It seems wise to escape ' as well as ", as a matter of course.  Can't
-               # hurt.
-               $html = htmlspecialchars( $html, ENT_QUOTES );
+               # hurt. Use ENT_SUBSTITUTE so that incorrectly truncated multibyte characters
+               # don't cause the entire string to disappear.
+               $html = htmlspecialchars( $html, ENT_QUOTES | ENT_SUBSTITUTE );
                return $html;
        }