Add Constants.encode as a utility for quick encoding in UTF-8

Add Constants.encode as a utility for quick encoding in UTF-8

We often need to convert a string into a UTF-8 encoding, so that
we can use this string as a path filter in a TreeWalk or in some
other suitable place where we assume a standard UTF-8 encoding is
being used.  As we have already done the lookup for the CHARSET
we can reuse that same CHARSET reference during future encoding
calls, while allowing the CharSet implementation to cache and
reuse the actual encoder instance.

Whenever possible we try to avoid copying the result as most of
the time the returned ByteBuffer's internal array matches the
result array we need to return to our caller.

Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Signed-off-by: Robin Rosenberg <robin.rosenberg@dewire.com>