Unquoting improvements

[Vizonex]

What do these changes do?

How do I keep finding things to optimize??? I got rid of the list in the unquoter and I replaced it with a custom unicode writer I wrote. I'm gonna allow anybody to use it as they wish since it's got a feature where it's able to dynamically handle unicode characters but now we can get rid of needing to join a bunch of python objects all at one time.

Are there changes in behavior for the user?

There's a new technique in the Unquoter that allocates the same amount of memory as the string it's taking in to be examined and decoded. This should be an optimization and not a performance bottleneck now.

Related issue number

Checklist

• I think the code is well written
• Unit tests for the changes exist
• Documentation reflects the changes

[codspeed-hq]

CodSpeed Performance Report

Merging #1529 will degrade performances by 72.21%

_{Comparing Vizonex:Unquoting-Improvements (3e4fdbe) with master (3871d2c)}

Summary

❌ 7 regressions
✅ 94 untouched benchmarks

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

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
❌	test_unquoter_long_ascii	12.3 ms	44.3 ms	-72.21%
❌	test_unquoter_long_pct	15.4 ms	45.6 ms	-66.17%
❌	test_unquoter_short	227.3 µs	501.6 µs	-54.68%
❌	test_parse_query_uncached[long]	23.6 ms	39.6 ms	-40.43%
❌	test_parse_query_uncached[short]	1.9 ms	3.1 ms	-36.8%
❌	test_path_safe_uncached	287.7 µs	445.2 µs	-35.37%
❌	test_update_query_string	662 µs	720.2 µs	-8.09%

[Vizonex]

how is that a regression do I stand corrected?