- Original changelog modified as the intent of the repr change is now
  limited to the helper (as data was corrected during development of the
  es2015 branch).

- Done so while still retaining 2.7 support.

Fixes #31

- Should have just combined those earlier minor linting fixes...

- Also ensure that optimised mode is added to the CI workflow.

- Optimised mode is a thing.

Fixes #29

- Made use of the more fine-tuned specific flake8 per-file-ignores
  configuration option to limit what formatting issues should be ignored
  for the affected files.
- Also removed all remaining overspecified ignores and corrected the
  missing raw strings prefix for strings that have escape sequences that
  are not currently defined.

- Small refactoring to decode_vlqs into a generator function; also
  provide a standalone decode_vlq function to accompany the encode_vlq
  function.
- Correction and clarity to relevant documentation in that module.

- This is to ensure that the replacement for the generic cases across
  all parsers will not unintentionally be testing against a single
  implementation.

- Will need to ensure the test cases are generalised.
- Should introduce a more formal metaclass constructor for these tests.

- Include additional notable changes.

- Also included the trove classifier.

- Allow the specification of number of tokens, and document how that is
  used inside the relevant test case.

- Only do so if the current token is DIV and the previous token falls
  under one of the affected ambiguous token types.
- Also corrected the tracking of previous token to be more useful for
  error reporting, as part of the fix to what that lexer variable
  tracks.
- Based on hints at: https://stackoverflow.com/a/27120110/2904896

Fixes #25.

- This is done in a similar manner and reason as what the lexer tests
  went through.

- Also updated the language for the code sections.