This changes the copyright file to use the Debian machine readable
standard detailed here:
http://www.debian.org/doc/packaging-manuals/copyright-format/1.0/

Fixes #28

We now `pytest.mark` tests that are mainly related to packaging and
disable them by default. This is a first workaround for broken
year-checks. These tests blindly fail when the current year does not
equal the one mentioned in docs and version string.

In fact the tests should compare the year in docs/version string with
the year of "current" commit (with "current commit" being the one
representing the tested codebase). Another issue to be addressed.

All tests can be run by: `py.test -m ""`, the packaging-only tests can
be run by `py.test -m "packaging"`. `tox` got a new env called `pkg`.

Tests that check logger states explicitly, have to setup logging from
now on, as any existing setups are ripped-off automatically now.

In tests we now automatically remove any handlers registered with
`ulif.diceware` loggers.

When calling `logger.removeHandler(h)` with one of our custom
`NullHandler` instances, things crashed.

Handle FileNotFoundError. Closes #44

Using `IOError` for compatibility, from Python 3.3 `IOError` is alias of `OSError`
https://stackoverflow.com/a/28633573/6202405

Checking the errno of 'No such file or directory' so that only that exception will be catched.
https://docs.python.org/2/library/errno.html
https://stackoverflow.com/a/39414133



Signed-off-by: Bhavin Gandhi <bhavin7392@gmail.com>

Normal errors should be displayed by default, even if they are not
catastrophic.

It looks like py26 gets less and less support by more and more core
packages.

Yet we can pin old `pytest` and `setuptools` versions but in the long
run we will drop support for python2.6 and python 3.3 too.

Tell about using real dice with more than 6 faces.

The manpage test failed when tests were run in a long path. The help
output then led to an extra line break which made comaparisons quite
cumbersome.

This one ruined the 0.9.2 release.

Use version generated by docutils from manpage template in `docs/`.

The wordlist length must be a power of the number of dice sides used to
result in fair choices. This is not being relatively prime.