locale - Add two new API functions

locale - Add two new API functions

* Add wcrtombin() and mbintowcr(), powerful buffer-to-buffer functions
  to translate UTF8, which can also perform escaping and de-escaping,
  allow legacy parsing, and do checks with varying levels of strictness.

  NOTE: If no locale or "C" locale, these functions are 1:1 (do not do
	UTF8 processing).  A UTF8 locale must be used to access these
	functions.

* Add wcrtoutf8() and utf8towcr().  These work exactly the same as
  the above functions but are locale-agnostic.  They always operate in
  UTF8 and locale initialization is not necessary.

* Escaping and de-escaping for 8-bit-clean round-trips with two of the
  modes.

* Can construct UTF8B from unescaped 8-bit-clean data, can deconstruct
  UTF8B (restore the original clean data), and can also be flagged
  generically to take already-escaped UTF8B or wchars as input.

* Can be used to perform strict checks.

* Very powerful incremental buffer validation and processing capabilities.
  Far more visibility to the caller rather than hiding state in a tracking
  state structure.  No tracking state structure.

* Does not do idiotic nul buffer termination.  Caller passes explicitly
  sized data, can opt to include or not include a terminator (depending
  on what is desired), and the byte buffers can contain any 8-bit code
  including 0x00.