diff --git a/ext/mbstring/mbstring.c b/ext/mbstring/mbstring.c
index a374539c2903b..5b8918ff6f9cf 100644
--- a/ext/mbstring/mbstring.c
+++ b/ext/mbstring/mbstring.c
@@ -3070,6 +3070,12 @@ static size_t count_demerits(struct candidate *array, size_t length, bool strict
 	uint32_t wchar_buf[128];
 	unsigned int finished = 0; /* For how many candidate encodings have we processed all the input? */
 
+	for (size_t i = 0; i < length; i++) {
+		if (array[i].in_len == 0) {
+			finished++;
+		}
+	}
+
 	while ((strict || length > 1) && finished < length) {
 		/* Iterate in reverse order to avoid moving candidates that can be eliminated. */
 		for (size_t i = length - 1; i != (size_t)-1; i--) {
diff --git a/ext/mbstring/tests/mb_detect_encoding.phpt b/ext/mbstring/tests/mb_detect_encoding.phpt
index b3c457738c1f8..2721ec3c83c6a 100644
--- a/ext/mbstring/tests/mb_detect_encoding.phpt
+++ b/ext/mbstring/tests/mb_detect_encoding.phpt
@@ -25,6 +25,13 @@ print("Bad ASCII (strict): " . mb_detect_encoding("\xDD\x92", ['ASCII', 'UTF-8']
 print("Bad ASCII/UTF-8, with more errors for ASCII (non-strict): " . mb_detect_encoding("\xD6\x8A\x8A", ['ASCII', 'UTF-8'], false) . "\n");
 print("Bad ASCII/UTF-8, with more errors for ASCII (strict): " . var_export(mb_detect_encoding("\xD6\x8A\x8A", ['ASCII', 'UTF-8'], true), true) . "\n");
 
+print("UTF-8 BOM (non-strict): " . mb_detect_encoding("\xEF\xBB\xBF", ["UTF-8", "ASCII"], false) . "\n");
+print("UTF-8 BOM (strict): " . mb_detect_encoding("\xEF\xBB\xBF", ["UTF-8", "ASCII"], true) . "\n");
+print("UTF-16BE BOM (non-strict): " . mb_detect_encoding("\xFE\xFF", ["UTF-8", "UTF-16BE", "UTF-16LE"], false) . "\n");
+print("UTF-16BE BOM (strict): " . mb_detect_encoding("\xFE\xFF", ["UTF-8", "UTF-16BE", "UTF-16LE"], true) . "\n");
+print("UTF-16LE BOM (non-strict): " . mb_detect_encoding("\xFF\xFE", ["UTF-8", "UTF-16BE", "UTF-16LE"], false) . "\n");
+print("UTF-16LE BOM (strict): " . mb_detect_encoding("\xFF\xFE", ["UTF-8", "UTF-16BE", "UTF-16LE"], true) . "\n");
+
 print("SJIS: " . mb_detect_encoding($sjis, 'SJIS', true) . "\n");
 print("JIS: " . mb_detect_encoding($jis, 'JIS', true) . "\n");
 print("EUC-JP (strict): " . mb_detect_encoding($euc_jp, 'UTF-8,EUC-JP,JIS', true) . "\n");
@@ -399,6 +406,12 @@ Bad ASCII (non-strict): UTF-8
 Bad ASCII (strict): UTF-8
 Bad ASCII/UTF-8, with more errors for ASCII (non-strict): UTF-8
 Bad ASCII/UTF-8, with more errors for ASCII (strict): false
+UTF-8 BOM (non-strict): UTF-8
+UTF-8 BOM (strict): UTF-8
+UTF-16BE BOM (non-strict): UTF-16BE
+UTF-16BE BOM (strict): UTF-16BE
+UTF-16LE BOM (non-strict): UTF-16LE
+UTF-16LE BOM (strict): UTF-16LE
 SJIS: SJIS
 JIS: JIS
 EUC-JP (strict): EUC-JP