Benchmark macro

[gavofyork]

Fixes #4861

[kianenigma]

Will it still be possible, given this, to manually build the setup? There is a situation in #4947 that you can see and it cannot be really tracked with this benchmarking setup. One of the parameters is

Enum Mode {
   BenchMode1,
   BenchMode2, 
}

and stepping over this is not possible.

[marcio-diaz]

Will it still be possible, given this, to manually build the setup? There is a situation in #4947 that you can see and it cannot be really tracked with this benchmarking setup. One of the parameters is

Enum Mode {
   BenchMode1,
   BenchMode2, 
}

and stepping over this is not possible.

I think you can make two benchmarks out of the enum and refactor the common parts.

[marcio-diaz]

Looks good to me, very ergonomic. Maybe we want @shawntabrizi or @gavofyork to take a look to my changes (balances/benchmarking mostly) before merging.

[gnunicorn]

So, the reports say that it consistently slightly improves on wasm but is sometimes a regression on native ... how do you want to proceed?

[marcio-diaz]

So, the reports say that it consistently slightly improves on wasm but is sometimes a regression on native ... how do you want to proceed?

Wrong PR :)? I'll merge once CI finish, other benchmarks depend on this.