test: cover slash conversion policies

[hyfdev]

Summary

• exercise strict, fallible, and lossy slash conversion across str, String, Path, borrowed PathBuf, and consuming PathBuf receivers
• pin exact non-normalizing spelling and Cow ownership on Unix and Windows, including UNC spelling, mixed separators, repeated and trailing separators, and multibyte UTF-8
• cover invalid native encoding with strict panic, exact fallible recovery of the original bytes or wide units, and both separator-replacing and no-replacement lossy branches

This PR changes tests only. It does not change production code, dependencies, benchmarks, or benchmark inputs.

Validation

• SUGAR_PATH_EXPECT_CACHED_CURRENT_DIR=0 cargo +1.97.0 test --locked -p sugar_path
• SUGAR_PATH_EXPECT_CACHED_CURRENT_DIR=1 cargo +1.97.0 test --locked --workspace --all-features
• cargo +1.97.0 fmt --all -- --check
• cargo +1.97.0 clippy --locked --workspace --all-targets --all-features -- -D warnings
• targeted slash-policy integration tests
• changed integration tests checked for x86_64-pc-windows-msvc through a dependency-only harness, avoiding the repository dev-only mimalloc build that requires a local Windows SDK
• final adversarial review: PASS after adding explicit Windows invalid-Unicode lossy coverage with no native separator

Final coverage audit

• The audit found that invalid Unix and Windows try_into_slash cases asserted pointer and capacity identity even though storage reuse is an optimization rather than part of the returned value contract.
• Those assertions are removed. Exact native bytes or wide units remain required, while the valid-UTF-8 clean owned cases continue to protect the accepted storage-reuse path.

At the rebased head e4372a2, all eight repository Actions jobs pass. Native Windows confirms that the invalid-wide recovery test and valid receiver-policy matrix execute under both default and all features; Linux and macOS likewise execute their Unix policy cases.

External CodSpeed alone reports as_path/rolldown_corpus/string 226.7→255.8 ns (-11.4%) while explicitly warning that the runtime environments differ. This is the exact non-attributable result already observed on merged test-only PR #52; this PR's diff contains only tests and changes no production code, dependency, benchmark, or benchmark input.

[codspeed-hq]

Merging this PR will degrade performance by 11.4%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

❌ 1 regressed benchmark
✅ 227 untouched benchmarks
⏩ 38 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	string	226.7 ns	255.8 ns	-11.4%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing agent/slash-policy-matrix (e4372a2) with main (36ea1df)}

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. ↩

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.