====
* Refactored code for Python 3 compatibility via 2to3 during installation.
* Dropped compatibility for Python 2 versions less than 2.7.
* Added dependency on `mock`_ for unit tests. This dependency is
satisfied by the standard library for Python 3.3 and newer.
* Renamed ``convutils.convutils`` to ``convutils.utils``; renamed
``convutils.convstructs`` to ``convutils.structs``.
* Added unit tests for ``convutils.utils``.
* Renamed ``ExcelTabNewlineDialect`` to ``SimpleTsvDialect``, and
changed its quoting style to no quoting.
* Refactored ``make_csv_reader`` and ``make_simple_tsv_dict_writer`` to
use the ``csv.excel`` dialect by default, to be more in line with the
standard library. Added new functions ``make_simple_tsv_reader`` and
``make_simple_tsv_dict_writer`` for the previous functionality.
* Renamed the ``headers`` parameter to ``header`` for
``make_csv_reader`` and ``make_simple_tsv_reader``.
* Changed the signatures of ``split_file_by_num_lines`` and
``split_file_by_parts``. The functions now accept a file handle
instead of a file name. The parameter ``has_header`` has been renamed
``header``. Added two new parameters, ``pad_file_names`` and
``num_lines_total``. If ``pad_file_names`` is ``True``, the numerical
portion of the output file names will be zero-padded. If
``num_lines_total`` is provided in addition to ``pad_file_names``,
``split_file_by_num_lines`` and ``split_file_by_parts`` will skip
counting the number of lines in the file, itself, which can save time.
* ``SortedTupleKeysDict`` and ``TwoWaySetDict`` now subclass
``collections.MutableMapping`` instead of ``dict`` directly due to the
suggestions on best-practices from Stack Overflow:
http://stackoverflow.com/questions/3387691/python-how-to-perfectly-override-a-dict
* Relocated ``sample_list_dict`` and ``sample_list_dict_low_mem`` to
``convutils.structs``.
* Added Sphinx-based documentation.
.. _mock: http://www.voidspace.org.uk/python/mock/