Oops, I missed one "cast a char to int and use it as a subscript" case.

Casting a signed char with a negative value to int will preserve the
value, so it'll still be a negative subscript.  Cast to guchar instead,
to make sure 0x80 through 0xFF are treated as 128 to 255, not -128 to
-1.

Change-Id: I1f0b33ba3686e963d45317b45465ff335431d17f
Reviewed-on: https://code.wireshark.org/review/4742
Reviewed-by: Guy Harris <guy@alum.mit.edu>

1 files changed, 1 insertions, 1 deletions

diff --git a/epan/strutil.c b/epan/strutil.c
index 148eaf417c..de52320568 100644
--- a/epan/strutil.c
+++ b/epan/strutil.c
@@ -647,7 +647,7 @@ hex_str_to_bytes_encoding(const gchar *hex_str, GByteArray *bytes, const gchar *
             ++end;
 
             /* check for separator and peek at next char to make sure we should keep going */
-            if (sep > 0 && *end == sep && str_to_nibble[(int)*(end+1)] > -1) {
+            if (sep > 0 && *end == sep && str_to_nibble[(guchar)*(end+1)] > -1) {
                 /* yes, it's the right sep and followed by more hex, so skip the sep */
                 ++end;
             } else if (sep != 0 && *end) {