CI: Reduce combinations in cargo hack check

[pitdicker]

I'm not sure why but within the last week the combinations cargo hack check tries on our CI has increased 4x to 672. Cargo hack has seen a couple of new releases, so there may be a change or bugfix there.

Disabling the oldtime (unused) and wasmbind (not checked on linux anyway) features reduces it back to 192 combinations.

We can also now inform cargo hack that some features are mutually exclusive, or don't effect each other. The rkyv features are truly mutually exclusive. The arbitrary feature is far removed from serialization, and serde also doesn't seem to have anything in common with rkyv. So I think we can treat them as mutually exclusive.

This allows us to run cargo hack check for more meaningful combinations with the rkyv-* features while we only check 84 combinations. Room to grow again 😄.

This should bring our CI time back below 5 minutes instead of close to 20 minutes.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 91.78%. Comparing base (ffe2745) to head (226ddc2).
Report is 7 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1553      +/-   ##
==========================================
- Coverage   91.79%   91.78%   -0.01%     
==========================================
  Files          37       37              
  Lines       18175    18167       -8     
==========================================
- Hits        16683    16674       -9     
- Misses       1492     1493       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[pitdicker]

4m3s 😄.

[djc]

Good stuff!