Specify UTF-8 encoding when reading TOML files.

[domdfcoding]

This avoids a UnicodeDecodeError when reading pyproject.toml files containing unicode characters on Windows. These characters may be present in a project's metadata, such as its description or an author's name.

Furthermore, the TOML specification requires files to be encoded in UTF-8.

[domdfcoding]

I can't directly test this on Windows because of xolox/python-humanfriendly#44, but the issue I was seeing goes away when using my branch.

[domdfcoding]

I have had a go at making the test suite pass on Windows in my fork.

[codecov]

Codecov Report

Merging #4 (0f57bd7) into master (c365c59) will not change coverage.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master       #4   +/-   ##
=======================================
  Coverage   99.74%   99.74%           
=======================================
  Files           9        9           
  Lines         782      782           
  Branches      180      180           
=======================================
  Hits          780      780           
  Misses          1        1           
  Partials        1        1           

Impacted Files	Coverage Δ
src/check_wheel_contents/config.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c365c59...0f57bd7. Read the comment docs.

[jwodder]

@domdfcoding Thank you very much! Please finish up the PR so I can merge it. I suppose I should add Windows to the CI matrix for future tests.

As for property-manager -> humanfriendly -> pyreadline causing problems, I believe I can replace it with functools.cached_property (or cached-property before Python 3.8).

[domdfcoding]

I think this is ready to go.

I have opened #22 to discuss Windows tests further.