Ticket #4801: Improved handling of invalid UTF-8 in the prompt, part 2/2

Do not filter out invalid UTF-8 components, instead let them pass through
as-is. This will result in mc converting them to a replacement symbol.

Signed-off-by: Egmont Koblinger <[email protected]>

Author: egmontkob

lib/terminal.c

@@ -237,17 +237,9 @@ strip_ctrl_codes (char *s)
                 r++;
         }
         else
-        {
-            // FIXME Improve the handling of invalid UTF-8, insert a replacement symbol (#4801)
-            const char *n = str_cget_next_char_safe (r);
-
-            if (str_isprint (r))
-            {
-                memmove (w, r, n - r);
-                w += n - r;
-            }
-            r = n;
-        }
+            // Copy byte by byte, thereby letting mc use its standard mechanism to denote invalid
+            // UTF-8 sequences. See #4801.
+            *(w++) = *(r++);
     }
 
     *w = '\0';

tests/lib/terminal.c

@@ -94,9 +94,8 @@ START_TEST (test_strip_ctrl_codes2)
     memcpy (s, s_orig, sizeof (s_orig));
 
     char *actual = strip_ctrl_codes (s);
-    const char *expected = "\342\235\244 bcdfghjklm";
 
-    ck_assert_str_eq (actual, expected);
+    ck_assert_str_eq (actual, s_orig);
     g_free (s);
 }
 END_TEST