chore: extend path performance baselines

[hyfdev]

Summary

• add separate normalization rows for early and late dirty markers around valid and invalid native encoding
• add absolute invalid-encoding and direct string-receiver relative rows
• require new timed benchmarks to land in a behavior-neutral baseline PR before their implementation PR

Why

This PR establishes stable CodSpeed benchmark identities before any implementation change. Draft #47 will use this branch as its base so each new case has both BASE and HEAD data instead of appearing only as new.

Validation

• cargo fmt --all --check
• cargo bench --locked --no-run

[codspeed-hq]

Merging this PR will not alter performance

✅ 210 untouched benchmarks
🆕 18 new benchmarks
⏩ 36 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
🆕	Simulation	dirty_early	N/A	2.9 µs	N/A
🆕	Simulation	dirty_late	N/A	3.1 µs	N/A
🆕	Simulation	module_to_cwd	N/A	2.1 µs	N/A
🆕	Simulation	dirty_before_non_utf8	N/A	2.5 µs	N/A
🆕	Simulation	non_utf8_before_dirty_late	N/A	2.7 µs	N/A
🆕	Simulation	non_utf8_absolute_target	N/A	8.4 µs	N/A
🆕	Simulation	non_utf8_lexically_clean	N/A	2.5 µs	N/A
🆕	Simulation	borrowed_native/non_utf8_lossy	N/A	2.3 µs	N/A
🆕	Simulation	borrowed_receiver/non_utf8_strict	N/A	441.1 ns	N/A
🆕	Memory	dirty_early	N/A	70 B	N/A
🆕	Memory	dirty_late	N/A	70 B	N/A
🆕	Memory	module_to_cwd	N/A	0 B	N/A
🆕	Memory	dirty_before_non_utf8	N/A	68 B	N/A
🆕	Memory	non_utf8_before_dirty_late	N/A	81 B	N/A
🆕	Memory	non_utf8_absolute_target	N/A	149 B	N/A
🆕	Memory	non_utf8_lexically_clean	N/A	66 B	N/A
🆕	Memory	borrowed_native/non_utf8_lossy	N/A	132 B	N/A
🆕	Memory	borrowed_receiver/non_utf8_strict	N/A	0 B	N/A

---

_{Comparing agent/safe-path-hot-path-baselines (3bd4208) with main (7599911)}

Footnotes

1. 36 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]

Pull request overview

This PR extends SugarPath’s performance baseline suite by introducing new, stable Criterion/CodSpeed benchmark identities (notably for non-UTF-8/native-encoding and “dirty marker position” cases) and by documenting a baseline-first protocol so future optimization PRs (e.g., #47) can get BASE vs HEAD deltas instead of “new benchmark” rows.

Changes:

• Add a dedicated hot_path_baselines benchmark binary to isolate new baseline families under fat LTO.
• Rename/reshape existing non-UTF-8 benchmark groups/leaf names to keep non_utf8 visible at the leaf level.
• Update benchmarking and project performance-strategy documentation to codify baseline protocol and non-UTF-8 performance scope.

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
Cargo.toml	Registers the new hot_path_baselines Criterion benchmark binary.
benchmarks/README.md	Documents non-UTF-8 naming visibility and the “baseline PR before implementation PR” workflow.
benches/to_slash.rs	Renames the non-UTF-8 to-slash benchmark group/IDs to be more explicit.
benches/normalize.rs	Renames the non-UTF-8 normalize benchmark group/ID to a non_utf8-explicit identity.
benches/hot_path_baselines.rs	Adds new benchmark rows for classifier-position, non-UTF-8 normalization variants, and additional relative baselines.
.agents/docs/README.md	Updates the PCR map to point optimization/non-UTF-8 tradeoff discussions to performance-strategy guidance.
.agents/docs/performance-strategy.md	Adds/expands the documented performance strategy and baseline protocol rationale.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.