extensions/time: make to_not_compact_s 2x faster

[unleashed]

No description provided.

[davidor]

I don't think we need to optimize to_not_compact_s.

This case is very different from #81 where to_compact_s was optimized.

to_compact_s is called multiple times per job and using profiling tools we could show that it appears as the second method that takes more CPU time.

to_not_compact_s is only called from

apisonator/lib/3scale/backend/stats/aggregator.rb

Line 55 in a99715d

.to_not_compact_s

which is called at most once per job and it does not appear in the list of methods that take a significant amount of CPU time.

[unleashed]

We don't need to do it, but we have it now. :)

[davidor]

If I were a new contributor and saw a benchmark for that method I'd think that it's one of the important ones, but it's not. I don't see any reason to merge that benchmark.

[unleashed]

I agree, don't merge the benchmark! :)

[unleashed]

@davidor benchmark removed from this PR

[davidor]

bors r+

[bors]

Build succeeded

• ci/circleci