#13899: \A, \Z, and \B now correctly match the A, Z, and B literals when used inside character classes (e.g. [A]).  Patch by Matthew Barnett.

With '%', non-ascii worked because the '%' automatically got promoted to
unicode.  With format that doesn't happen, which led to encoding errors.  This
fix goes back to using %, and adds a test to make sure non-ascii string values
work in iterdump.

Although there is not a regression in Python2, we make the same update here to
keep the code bases in sync.

(The fix for issue 9750 introduced a regression in Python 3 by sorting the row
objects returned by fetchall.  But if a row_factory such as sqlite3.Row is
used, the rows may not be sortable (in Python3), which leads to an exception.
The sorting is still a nice idea, so the patch moves the sort into the sql.)

Fix and test by Peter Otten.

_Qdoffs when compiling with an SDK of 10.7 or later.  The OS X APIs they
wrap have long been deprecated and have now been removed with 10.7.
These modules were already empty for 64-bit builds and have been removed
in Python 3.  (Original patch by Ronald Oussoren.)

Amaury Forgeot d'Arc. Added tests for partial decoding of non-BMP
characters.

Patch by Martin Packman.

Patch from Tom Tromey.