Quoter

Updated testing matrix. Now certified under Python 3.6 and latest
versions of PyPy and PyPy3.

Updated testing for Python 3.5.0 final and PyPy 2.6.1 (based on
CPython 2.7.10).

Minor fix to Markdown module. Turns out HTML stops at ``h6``, not
``h7``. Oops!

Quoters can now take a variable number of arguments. A ``sep``
kwarg determins how multiple pieces are combined. Defaults to
empty string, so rank concatenation. Full rationalization re
``Joiner`` awaits another day.

Minor testing cleanups.

Handful of minor cleanups.

All previously skipped tests are now re-enabled and passing.

Argument handling of ``Quoter`` creation is now hardened. If a
prefix or suffix is explicitly set but not the other, the unset
one defaults to an empty string, not to the other one that is set.

The ``pair`` keyword for compact setting of balanced quote
prefix/suffix pairs is returned to service (in earlier
incarndations, this was the ``chars`` kwarg).

Extended documentation. Added API Reference.

Major redesign of the named styles facility. Refactored style
management out of ``Quoter`` and its subclasses and into separate
``StyleSet`` class. Much cleaner and more reliable.

The ``style=`` kwarg to ``Quoter`` classes, however, goes away. It
was "a bridge too far," complicating everything and not adding
real value. Its function is now found in ``StyleSet._define()``.

While this is largely an "infrastructure" rather than functional
release, there were some minor functional improvements: Markdown
quoting extended with ``hr`` and ``h1`` through ``h7`` header
methods. XML quoting now has ``cdata`` and ``pcdata`` methods.
Both of these are part of an incremental march toward functional
completeness in their respective quoting realms.

Withdrew branch coverage testing metric. While a good idea in
theory, current tools do not provide sufficient actionable
intelligence to directly identify source of branch misses.