add comprehensions benchmark by carljm · Pull Request #265 · python/pyperformance

LGTM.

before merging I would like to see a comparison with one of your PRs which implements comprehensions inlining against main.

That seems backwards. Whether an optimization works on a particular benchmark determines the value of the optimization, not the benchmark.

The question is "does this benchmark represent real-world code?", or the weaker form "does this benchmark make the benchmark suite more closely represent real-world code?"

Answering those question objectively is difficult without a huge amount of real-word data, so we need to apply judgement.

Given that list comprehensions are clearly under represented in the rest of benchmarks, I think that this does improve the benchmark suite overall.

@kumaraditya303 do you have reason to think otherwise?