Duplicated f for Enzyme to handle closures

[gdalle]

Fix #339

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 0% with 71 lines in your changes missing coverage. Please review.

Project coverage is 90.61%. Comparing base (ff529cb) to head (337cb5c).

Files	Patch %	Lines
...ifferentiationInterfaceEnzymeExt/reverse_onearg.jl	0.00%	29 Missing ⚠️
...ifferentiationInterfaceEnzymeExt/forward_onearg.jl	0.00%	26 Missing ⚠️
...ifferentiationInterfaceEnzymeExt/reverse_twoarg.jl	0.00%	10 Missing ⚠️
...ifferentiationInterfaceEnzymeExt/forward_twoarg.jl	0.00%	6 Missing ⚠️

❗ There is a different number of reports uploaded between BASE (ff529cb) and HEAD (337cb5c). Click for more details.

HEAD has 68 uploads less than BASE

| Flag | BASE (ff529cb) | HEAD (337cb5c) | |------|------|------| ||99|31|

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #341      +/-   ##
==========================================
- Coverage   96.77%   90.61%   -6.16%     
==========================================
  Files         101       99       -2     
  Lines        4831     4836       +5     
==========================================
- Hits         4675     4382     -293     
- Misses        156      454     +298     

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

[wsmoses]

This has highly non-trivial implications -- especially if you only support multi arguments etc via closures.

This will likely result in significant perf downgrades for folks, please do not merge if you want to maintain performance for downstream users.

[gdalle]

Superseded by #375