separated_list doc and test update

[SaltyKitkat]

• related issue: separated_list1 by multipspace0 sep failed #1691 (comment)
• related pr: move the loop check of separated list to cover both parsers #1756

[Geal]

this looks good, thanks

[codspeed-hq]

CodSpeed Performance Report

Merging #1784 will degrade performances by 11.95%

_{Comparing SaltyKitkat:separated_list_doc_and_test_update (587ba95) with main (5455747)}

Summary

❌ 2 regressions
✅ 22 untouched benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Benchmark	main	SaltyKitkat:separated_list_doc_and_test_update	Change
❌	float str	1.8 µs	2 µs	-10.02%
❌	number	215 ns	244.2 ns	-11.95%

[gdennie]

I haven't studied the code so perhaps it resolves other concerns or bugs but 11.95% seems like a big hit for functionality that seems to be available elsewise. Nonetheless, great. 😉

In a fantasy world we could have a marking trait, NonConsuming, that attaches to parsers derived from non-consuming subparsers that would be given the trait. That may mean an extended algebra of trait composition. Of course, for clarity we require restating the trait but it would be compiler checked. Then the trait would then make this constraint on the combinators compiler visible.

[Geal]

This is a flaky benchmark so do not read too much into it