If you retrieve user input with multibyte characters you should additionally normalize the string using utf8_normalize_nfc() before you work with it. This is necessary to make sure that equal characters can only occur in one particular binary representation. For example the character Å can be represented either as U+00C5 (LATIN CAPITAL LETTER A WITH RING ABOVE) or as U+212B (ANGSTROM SIGN). phpBB uses Normalization Form Canonical Composition (NFC) for all text. So the correct version of the above example would look like this:
-$_REQUEST['multibyte_string'] = 'Käse';
-
-// normalize multibyte strings
-echo utf8_normalize_nfc($request->variable('multibyte_string', '', true));
-// ASCII strings do not need to be normalized
-echo $request->variable('multibyte_string', '');
-Case insensitive comparison of strings is no longer possible with strtolower or strtoupper as some characters have multiple lower case or multiple upper case forms depending on their position in a word. The utf8_strtolower and the utf8_strtoupper functions suffer from the same problem so they can only be used to display upper/lower case versions of a string but they cannot be used for case insensitive comparisons either. So instead you should use case folding which gives you a case insensitive version of the string which can be used for case insensitive comparisons. An NFC normalized string can be case folded using utf8_case_fold_nfc().