The previous approach was too easily fooled (a rotate() sufficed).

* Use it->counter to determine when iteration is complete.  The
  previous approach was too complex.

* Strengthen an assertion and add a comment here or there.

* Change the centering by one to make it possible to test the module
  with BLOCKLEN's as low as two.  Testing small blocks makes end-point
  errors surface more readily.

BLOCKLEN-1, this assert-failed in a debug build, or went wild with a
NULL pointer in a release build.  Reported on c.l.py by Stefan Behnel.

which isn't being converted from reST yet).

Many of these tests are redundant, but this will ensure
that the mapping protocols all stay in sync.
Also, added a test for dictionary subclasses.

* Updated docs to reflected signed return values.
* Fixed a doubled word typo.

Rather than introduce new logic, took the approach of making the message
itself more general.

(Suggested by Michael Hudson.)

repr(66.6) == "66.6", so doubt that the claimed output has ever been seen.
Changed it to 66.25 everywhere, and manually verified that the new claimed
output is correct.

request.  Tim says that "correct 'fuzzy' comparison of floats cannot
be automated."  (The motivation behind adding the new option
was verifying interactive examples in Python's latex documentation;
several such examples use numbers that don't print consistently on
different platforms.)

* Have groupby() be careful about decreffing structure members.

the expected output to match corresponding number literals in the
actual output if their values are equal (to ten digits of precision).

the proper path directories for compiling extension modules.

Added a clarifying line to the docs.

fact that compiler.ast.Function now takes a "decorators" argument.

the htons() function.