Given a well-formed utf-8 string 'message' of messageLen bytes and a
desire to truncate to approximately abbrevLen bytes return the shortest
string greater or equal to abbrevLen that does not split a utf-8
sequence.
Signed-off-by: Caolán McNamara <caolan.mcnamara@collabora.com>
Change-Id: Ie623d8c1027bb4724485cab5b0bb6d3d1cd3d9ab