du: our implementation is slower than GNU's

[wilmerdooley]

Resolves #6845.

On firefox source code (lot of files and directories):

[sylvestre]

not sure we need this comment

[codspeed-hq]

Merging this PR will improve performance by 6.3%

⚡ 4 improved benchmarks
✅ 327 untouched benchmarks
⏩ 46 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	du_summarize_balanced_tree[(5, 4, 10)]	6.7 ms	6.2 ms	+8.28%
⚡	Simulation	du_max_depth_balanced_tree[(6, 4, 10)]	25.6 ms	23.9 ms	+7.06%
⚡	Simulation	du_deep_tree[(100, 3)]	1.1 ms	1 ms	+6.25%
⚡	Simulation	du_wide_tree[(5000, 500)]	9.3 ms	9 ms	+3.65%

Tip

Curious why this is faster? Comment @codspeedbot explain why this is faster on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing wilmerdooley:oss/issue-6845 (4d185a5) with main (3bd6665)}

Footnotes

1. 46 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. ↩

[github-actions]

GNU testsuite comparison:

Skip an intermittent issue tests/cut/bounded-memory (fails in this run but passes in the 'main' branch)
Skip an intermittent issue tests/tail/tail-n0f (fails in this run but passes in the 'main' branch)
Skipping an intermittent issue tests/misc/io-errors (passes in this run but fails in the 'main' branch)