This is an experiment to see if reducing string allocations with Cow has any performance improvement.
My local benchmarks show mixed results. My guess is that there is a small additional overhead due to branching when destructuring the Cow enum.
It might be worthwhile to further split these changes into individual branches to target benchmarks on specific changes instead of testing all of them at once.
This is an experiment to see if reducing string allocations with
Cowhas any performance improvement.My local benchmarks show mixed results. My guess is that there is a small additional overhead due to branching when destructuring the
Cowenum.It might be worthwhile to further split these changes into individual branches to target benchmarks on specific changes instead of testing all of them at once.