Description:
------------
When you use mb_strcut with an offset and this offset happens to fall within a multibyte (utf-8) character mb_strcut moves the start point of the string as to not break a multibyte character. As an undesirable side effect it strips one or more bytes from the end of the cut string.

Reproduce code:
---------------
$crap = 'AåBäCöDü' /* in utf-8 */
echo mb_strcut($crap, 0, 100, 'UTF-8')."\n"; /* 'AåBäCöDü' as 100 is more than strlen */
echo mb_strcut($crap, 1, 100, 'UTF-8')."\n"; /* 'åBäCöDü' as 100 is more than strlen and we're offset 1 byte */
echo mb_strcut($crap, 2, 100, 'UTF-8')."\n"; /* 'åBäCöD' */

Expected result:
----------------
$crap = 'AåBäCöDü' /* in utf-8 */
echo mb_strcut($crap, 0, 100, 'UTF-8')."\n"; /* 'AåBäCöDü' as 100 is more than strlen */
echo mb_strcut($crap, 1, 100, 'UTF-8')."\n"; /* 'åBäCöDü' as 100 is more than strlen and we're offset 1 byte */
echo mb_strcut($crap, 2, 100, 'UTF-8')."\n"; /* either 'åBäCöDü' or 'BäCöDü' */