cp: fix metadata handling on WASI

[DePasqualeOrg]

Correcting cp metadata handling on WASI requires extensive changes. I’m offering this as a possible reference implementation. It was developed over multiple rounds of review with Codex, and I don’t claim to fully understand everything. I’ll maintain this as a private patch for my own use until the issues are addressed here.

Changes:

• Timestamp preservation for files, symlinks, and recursive directories
• Pre---progress timestamp capture and stale-snapshot checks
• Correct --update skip behavior, including recursive copies
• Optional metadata error handling on WASI
• No-follow source safety on WASI

[codspeed-hq]

Merging this PR will improve performance by ×2.2

⚡ 7 improved benchmarks
✅ 354 untouched benchmarks
⏩ 50 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Memory	cp_recursive_wide_tree[(6000, 800)]	4,376.8 KB	182.5 KB	×24
⚡	Memory	cp_preserve_metadata[(5, 4, 10)]	1,357.3 KB	416.6 KB	×3.3
⚡	Memory	cp_recursive_balanced_tree[(5, 4, 10)]	1,438.5 KB	497.8 KB	×2.9
⚡	Simulation	cp_recursive_wide_tree[(6000, 800)]	236.1 ms	221.3 ms	+6.71%
⚡	Simulation	cp_preserve_metadata[(5, 4, 10)]	102.4 ms	97.4 ms	+5.15%
⚡	Simulation	cp_recursive_balanced_tree[(5, 4, 10)]	96.5 ms	91.9 ms	+5.09%
⚡	Simulation	cp_recursive_deep_tree[(120, 4)]	18.1 ms	17.3 ms	+4.59%

Tip

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

---

_{Comparing DePasqualeOrg:cp-wasi-file-symlink-timestamps (f65f68f) with main (dfb953d)²}

Footnotes

1. 50 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. ↩

2. No successful run was found on main (48a2013) during the generation of this report, so dfb953d was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

[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/pid-pipe (fails in this run but passes in the 'main' branch)
Skipping an intermittent issue tests/date/date-locale-hour (passes in this run but fails in the 'main' branch)