Ecoxipy

- _Improved:_ Performance – more duck-typing and less `isinstance`, using
`collections.queue` instead of `list` for children handling.
- _Fixed:_ Text handling of `ecoxipy.etree_output.ETreeOutput`.
- _Added:_ Performance tests for `ecoxipy.etree_output.ETreeOutput`.

This is a major release introducing new capabilities and Python 3 support.
- _Added:_ Support for Python 3.
- _Added:_ Use `|` on `ecoxipy.MarkupBuilder` to create comments.
- _Added:_ Use slicing on `ecoxipy.MarkupBuilder` to create documents
or processing instructions.
- _Added:_ The module `ecoxipy.parsing` contains SAX to ECoXiPy parsing
facilities.
- _Changed:_ Renamed module `ecoxipy.element_output` to
`ecoxipy.pyxom.output` and moved the XML representation classes to their
own module `ecoxipy.pyxom` naming them _PyXOM_ - **Pythonic XML Object
Model** as well as adding new functionality. DOM creation was removed from
those classes.
- _Changed:_ All XML data is internally handled as Unicode, a
`ecoxipy.MarkupBuilder` instance converts byte strings from an encoding
given on creation (defaults to UTF-8).
- _Changed:_ XML parsing is now handled by `ecoxipy.MarkupBuilder` instead
of the `ecoxipy.Output` implementations.
- _Changed:_ Text node creation is now handled by `ecoxipy.MarkupBuilder`
instead of the `ecoxipy.Output` implementations.
- _Improved:_ Unpacking of content while processing is done recursively on
iterable and callable content.

- _Added:_ Use `&` on `ecoxipy.MarkupBuilder` to create text nodes.
- _Improvement:_ Better performance of `ecoxipy.string_output.StringOutput`.

Initial release.