Case insensitivity no longer locale specific
All case insensitive matching for function, class and constant names is
now performed in a locale independent manner according to ASCII rules.
This improves support for languages using the Latin alphabet with unusual
collating rules, such as Turkish and Azeri.
This may cause issues for code that uses case insensitive matches for
non-ASCII characters in multibyte character sets (including UTF-8), such
as accented characters in many European languages. If you have a
non-English, non-ASCII code base, then you will need to test that you are
not inadvertently relying on this behaviour before deploying PHP 5.5 to
production systems.