Change isFOO_utf8_safe macros to return matched byte length

Or 0 when the character isn't of type FOO.  This allows these macros to
be used as booleans, as previously; or to give you how many bytes there
are in the matched UTF-8 character.

This was always trivially the case for ASCII-range characters, as the
former boolean 0,1 gave you the correct length if they matched.

The previous commit extended this to return the length for above-Latin1
characters.

This commit is the final piece.  Latin1 characters that aren't ASCII
always are two bytes.  So just multiply the return by 2, yielding 0 if
no match or 2 bytes if matched.

Author: khwilliamson

handy.h

@@ -2263,8 +2263,10 @@ END_EXTERN_C
           : (UTF8_IS_ABOVE_LATIN1_START(*(p))                               \
              ? above_latin1                                                 \
              : ((LIKELY((e) - (p) > 1 && UTF8_IS_CONTINUATION(*((p)+1))))   \
-                ? generic_isCC_(EIGHT_BIT_UTF8_TO_NATIVE(*(p), *((p)+1 )),  \
-                                classnum)                                   \
+                  /* Multiply by 2 to return byte length of matched         \
+                   * character */                                           \
+                ? 2 * generic_isCC_(EIGHT_BIT_UTF8_TO_NATIVE(*(p),*((p)+1)),\
+                                    classnum)                               \
                 : (force_out_malformed_utf8_message_((U8 *) (p), (U8 *) (e),\
                                                      0, MALFORMED_UTF8_DIE),\
                    0))))