Do not implement ZipEq::fold (and change test)

[kinto-b]

Relates to #755.

I can't really think of a better way to implement this, but it still yields a regression :(

cargo bench --bench specializations "zip_eq/fold"

zip_eq/fold             time:   [509.89 ns 510.95 ns 512.02 ns]
                        change: [+4.8744% +5.1665% +5.4970%] (p = 0.00 < 0.05)
                        Performance has regressed.

[Philippe-Cholet]

To be honest, I did specialize the folds that felt the most promising to me but I thought this would be a win.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 94.42%. Comparing base (6814180) to head (2e6c570).
Report is 48 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #915      +/-   ##
==========================================
+ Coverage   94.38%   94.42%   +0.03%     
==========================================
  Files          48       48              
  Lines        6665     6837     +172     
==========================================
+ Hits         6291     6456     +165     
- Misses        374      381       +7     

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

[Philippe-Cholet]

After changes, if you could squash commits.

[Philippe-Cholet]

Thanks.

[kinto-b]

@Philippe-Cholet pleasure, thank you :)