fix some benchmark (some of them were failing)

[sylvestre]

  /home/runner/work/coreutils/coreutils/target/codspeed/instrumentation/uu_numfmt/numfmt_bench: numfmt-error-invalid-suffix
  /home/runner/work/coreutils/coreutils/target/codspeed/instrumentation/uu_sort/sort_bench: sort-cannot-read

[github-actions]

GNU testsuite comparison:

Skipping an intermittent issue tests/misc/tee (passes in this run but fails in the 'main' branch)

[codspeed-hq]

CodSpeed Performance Report

Merging #8773 will degrade performances by 100%

_{Comparing sylvestre:improv-bench (d1cd999) with main (31d1f06)}

Summary

⚡ 9 improvements
❌ 23 regressions
✅ 3 untouched
🆕 9 new
⏩ 73 skipped¹

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

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
❌	ls_recursive_balanced_tree[(6, 4, 15)]	458.3 µs	55,889.2 µs	-99.18%
❌	ls_recursive_deep_tree[(200, 2)]	416.4 µs	2,766.2 µs	-84.95%
❌	ls_recursive_long_all_balanced_tree[(6, 4, 15)]	563.2 µs	141,000 µs	-99.6%
❌	ls_recursive_long_all_deep_tree[(100, 4)]	465 µs	3,456.4 µs	-86.55%
❌	ls_recursive_long_all_mixed_tree	1.8 ms	4 ms	-53.69%
❌	ls_recursive_long_all_wide_tree[(15000, 1500)]	96.4 ms	156 ms	-38.18%
❌	ls_recursive_mixed_tree	509.4 µs	1,367.4 µs	-62.75%
❌	ls_recursive_wide_tree[(10000, 1000)]	46.4 ms	55.2 ms	-15.93%
🆕	numfmt_from_si[10000]	N/A	33.1 ms	N/A
🆕	numfmt_large_numbers_si[10000]	N/A	34.5 ms	N/A
🆕	numfmt_padding[(10000, 50)]	N/A	48.8 ms	N/A
🆕	numfmt_round_modes[("down", 10000)]	N/A	34.6 ms	N/A
🆕	numfmt_round_modes[("towards-zero", 10000)]	N/A	34.6 ms	N/A
🆕	numfmt_round_modes[("up", 10000)]	N/A	32.8 ms	N/A
🆕	numfmt_to_iec[10000]	N/A	34.1 ms	N/A
🆕	numfmt_to_si[10000]	N/A	33.9 ms	N/A
🆕	numfmt_to_si_precision[10000]	N/A	35.5 ms	N/A
❌	sort_accented_data[500000]	263.6 µs	819,171 µs	-99.97%
❌	sort_ascii_only[500000]	264.5 µs	799,451.7 µs	-99.97%
❌	sort_case_insensitive[500000]	325.5 ms	416.8 ms	-21.91%
...	...	...	...	...

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

Footnotes

1. 73 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. ↩