benchmarks: prepare args outside benchmark

[xtqqczze]

Currently, we construct uucore::Args within the measured section of benchmarks. That adds allocation and conversion overhead and will skew the results.

[github-actions]

GNU testsuite comparison:

GNU test failed: tests/tail/retry. tests/tail/retry is passing on 'main'. Maybe you have to rebase?

[codspeed-hq]

CodSpeed Performance Report

Merging this PR will degrade performance by 26.29%

_{Comparing xtqqczze:get_benchmark_args (1936e7b) with main (0f7a7c4)}

Summary

⚡ 8 improved benchmarks
❌ 3 regressed benchmarks
✅ 273 untouched benchmarks
⏩ 38 skipped benchmarks¹

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

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Memory	sort_ascii_utf8_locale	5.6 MB	6.1 MB	-8.02%
❌	Memory	sort_reverse_utf8_locale	2.6 MB	3.5 MB	-26.29%
⚡	Memory	sort_numeric[500000]	48.7 MB	44.8 MB	+8.72%
⚡	Memory	sort_key_field[500000]	32.8 MB	28.9 MB	+13.45%
⚡	Memory	cksum_sm3	66.5 KB	64 KB	+3.85%
⚡	Memory	cksum_sha2	70.7 KB	60.8 KB	+16.4%
❌	Memory	cksum_md5	64 KB	66.5 KB	-3.76%
⚡	Memory	cksum_sha512	59.2 KB	56.6 KB	+4.55%
⚡	Memory	cksum_default	58.5 KB	56.3 KB	+3.92%
⚡	Memory	cksum_blake2b	59.1 KB	55.4 KB	+6.71%
⚡	Simulation	cp_large_file[16]	385.9 µs	366.5 µs	+5.3%

Footnotes

1. 38 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[xtqqczze]

@sylvestre Please could you confirm whether these changes are viable to apply in full?

[sylvestre]

i guess ? :)

[xtqqczze]

Need to refactor to handle the following case:

coreutils/src/uu/cp/benches/cp_bench.rs

Lines 27 to 33 in 456252f

	bencher.bench(|| {
	fs_utils::remove_path(&dest);
	let mut full_args = Vec::with_capacity(args.len() + 2);
	full_args.extend_from_slice(args);
	full_args.push(source_str);
	full_args.push(dest_str);

[oech3]

Should we wait working about #10486 until your PR was merged?
Any chance for splitting PR?