Fix regression that prevents async fixtures from working in sync tests

[seifertm]

The function _preprocess_async_fixtures does nothing if the fixture is not an async function. The change simply prepares all async fixtures of a test regardless of whether the test is an async function or not.

Do you see any harm in doing so?

This should fix the issue encountered in #286.

[codecov-commenter]

Codecov Report

Merging #287 (8a86c9e) into master (07e9922) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #287   +/-   ##
=======================================
  Coverage   92.61%   92.61%           
=======================================
  Files           3        3           
  Lines         271      271           
  Branches       39       39           
=======================================
  Hits          251      251           
  Misses         12       12           
  Partials        8        8           

Impacted Files	Coverage Δ
plugin.py	92.48% <0.00%> (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 07e9922...8a86c9e. Read the comment docs.

[asvetlov]

LGTM!
Thanks for the fix, my fault.
There are too many combinations between sync and async code :)

[asvetlov]

Please let me merge and make a new release right now.
Sorry for the review delay.

[seifertm]

Sorry for the review delay.

@asvetlov No worries. If someone has time to look at these things within a week, I'm already pretty happy :)
You're always very responsive, anyway.

A new release sounds good!

[seifertm]

@asvetlov The tag that being built was v18.0.1, instead of v0.18.1. I thought that might cause issues and cancelled the running release workflow before the package landed on PyPI. I hope this was okay!