Fix test_str::test_conversion_types

[alexjdw]

• Updated the test for test_str.NewFormat with @transforms and a few additional bits.
• Above change ensures that the test now accurately reflects that the .format() function should provide an exact duplicate of the Python output and Python errors.
• @tranforms was very helpful to me, but having to hunt through the file wasn't fun, so I added a detailed docstring using the existing comments.
• There was a restriction on float_exp in @transforms; however, the option was available and it was causing python output not to match up correctly. I removed the restriction.
• fixed: 'n' format code was grouped with 'g' format code, but has slightly different behavior in one case.
• fixed: both 'n' and 'g' had issues selecting the right floating point precision.
• fixed: .format() would crash with a single argument as it was expecting a Tuple.

Str.format() with new-style formatting arguments a la https://docs.python.org/3.4/library/string.html#format-string-syntax now passes muster.

Replaces PR #804

PR Checklist:

• [ x ] All new features have been tested
• [ x ] All new features have been documented
• [ x ] I have read the CONTRIBUTING.md file
• [ x ] I will abide by the code of conduct

[martica]

The 3.4 tests are failing because of a difference in error messages, but if you sync your PR to head, the 3.4 tests won't run anymore because we've removed 3.4 support.

[alexjdw]

I think that worked, thanks again for all of your help!

[martica]

Nice. All the tests are passing. Let me know if any of my comments are too cryptic.