sort: fix extended test_sort::test_buffer_sizes failures

[xtqqczze]

Fixes the test failures observed in #11081 by reverting 4438adf and b61a38e

[codspeed-hq]

Merging this PR will degrade performance by 11.61%

⚡ 4 improved benchmarks
❌ 17 regressed benchmarks
✅ 267 untouched benchmarks
⏩ 42 skipped benchmarks¹

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

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	sort_ascii_utf8_locale	15.8 ms	17.7 ms	-10.93%
⚡	Simulation	sort_long_line[10000]	514.7 µs	432.1 µs	+19.12%
❌	Simulation	sort_ascii_c_locale	16 ms	17.5 ms	-8.7%
❌	Memory	sort_ascii_utf8_locale	4.6 MB	5.2 MB	-10.75%
❌	Memory	sort_numeric[500000]	43.4 MB	45.2 MB	-3.99%
❌	Memory	sort_mixed_data[500000]	20.5 MB	22.3 MB	-8.1%
❌	Memory	sort_general_numeric[200000]	22.8 MB	23.8 MB	-3.91%
❌	Memory	sort_mixed_utf8_locale	2.3 MB	2.6 MB	-11.61%
❌	Memory	sort_accented_data[500000]	20.5 MB	21.9 MB	-6.52%
⚡	Memory	sort_long_line[10000]	139 KB	58.8 KB	×2.4
❌	Memory	sort_dictionary_order[500000]	20.6 MB	22.4 MB	-8.07%
⚡	Memory	sort_german_c_locale	1.8 MB	1.7 MB	+4.04%
❌	Memory	sort_unique_locale[500000]	32 MB	33.4 MB	-4.28%
❌	Memory	sort_unique_utf8_locale	3.5 MB	3.8 MB	-8.08%
⚡	Memory	sort_german_de_locale	1.8 MB	1.7 MB	+4.04%
❌	Memory	sort_reverse_locale[500000]	20.5 MB	21.9 MB	-6.52%
❌	Memory	sort_mixed_c_locale	2.3 MB	2.6 MB	-11.61%
❌	Memory	sort_ascii_only[500000]	20.5 MB	22 MB	-6.79%
❌	Memory	sort_reverse_utf8_locale	2.3 MB	2.6 MB	-11.61%
❌	Memory	sort_key_field[500000]	28.1 MB	29.3 MB	-4.02%
...	...	...	...	...	...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

---

_{Comparing xtqqczze:sort-alloc-size (5e269ff) with main (79fc0c7)}

Footnotes

1. 42 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]

The test failures are real, modifying test_sort::test_buffer_sizes to iterate over 1..50_000 bisects the failure to commit b61a38e

Opened #11081

[github-actions]

GNU testsuite comparison:

GNU test failed: tests/date/resolution. tests/date/resolution is passing on 'main'. Maybe you have to rebase?
Congrats! The gnu test tests/cut/bounded-memory is no longer failing!
Congrats! The gnu test tests/pr/bounded-memory is now passing!

[github-actions]

GNU testsuite comparison:

Congrats! The gnu test tests/cut/bounded-memory is no longer failing!
Congrats! The gnu test tests/date/date-locale-hour is no longer failing!
Skip an intermittent issue tests/pr/bounded-memory (was skipped on 'main', now failing)

[github-actions]

GNU testsuite comparison:

Congrats! The gnu test tests/cut/bounded-memory is no longer failing!
Note: The gnu test tests/expand/bounded-memory is now being skipped but was previously passing.
Congrats! The gnu test tests/pr/bounded-memory is now passing!

[github-actions]

GNU testsuite comparison:

Congrats! The gnu test tests/cut/bounded-memory is no longer failing!
Congrats! The gnu test tests/date/date-locale-hour is no longer failing!
Note: The gnu test tests/expand/bounded-memory is now being skipped but was previously passing.
Note: The gnu test tests/unexpand/bounded-memory is now being skipped but was previously passing.
Congrats! The gnu test tests/pr/bounded-memory is now passing!

[xtqqczze]

@sylvestre please review so we can get into 0.7.0. I will rebase to remove 5e269ff as the files are too big,

[oech3]

Do you really wan to merge this with current size? There are 2 Simulation perf regression.

[xtqqczze]

The main priority is to fix the test failures observed in #11081

[oech3]

Are big fixtures new test cases we were lacking?