Section 06: Verification

Status: Not Started Goal: All parser frontend benchmarks show measurable improvement compared to Section 01 baselines. All test suites green. Findings documented for future reference.

Context: This section re-measures all benchmarks established in Section 01 after all optimization work (Sections 02-05) is complete. The before/after comparison proves the plan delivered measurable value and identifies which optimizations had the most impact.

Depends on: All prior sections (01-05).

06.1 Benchmark Comparison

File(s): plans/parser-perf/baselines.json

Re-run the complete benchmark suite and compare against Section 01 baselines.

Run all benchmarks with --baseline parser-perf-v0 to compare against saved baselines:

cargo bench -p oric --bench lexer_core -- --baseline parser-perf-v0
cargo bench -p oric --bench lexer -- --baseline parser-perf-v0
cargo bench -p oric --bench parser -- --baseline parser-perf-v0

Record final metrics in baselines.json alongside original baselines:

{
  "final_results": {
    "date": "YYYY-MM-DD",
    "commit": "<hash>",
    "lexer_core_raw_throughput_mib_s": null,
    "lexer_cooked_throughput_mib_s": null,
    "parser_raw_throughput_mib_s": null,
    "parser_salsa_throughput_mib_s": null,
    "salsa_overhead_percent": null
  },
  "improvement": {
    "lexer_cooked_percent": null,
    "parser_raw_percent": null,
    "parser_salsa_percent": null,
    "salsa_overhead_delta": null
  }
}

Verify new workloads from Section 01.2 are included in the final run:
- parser/salsa_overhead/raw and parser/salsa_overhead/via_salsa (Salsa isolation)
- String-heavy workload (in lexer.rs)
- Expression-heavy workload (in parser.rs)
Identify which optimizations had the most impact:
- Cross-crate inlining (Sections 03.1, 04.1)
- Arena pre-allocation (Sections 03.2, 04.2)
- Cooker/parser micro-optimizations (Sections 03.3, 04.4)
- Incremental parsing (Section 05.2)
Verify no benchmark regressions: every benchmark must be equal or better than baseline.
/tpr-review passed — independent review found no critical or major issues (or all findings triaged)
/impl-hygiene-review passed — hygiene review clean. MUST run AFTER /tpr-review is clean.
Subsection close-out (06.1) — MANDATORY before starting the next subsection. Run /improve-tooling retrospectively on THIS subsection’s debugging journey (per .claude/skills/improve-tooling/SKILL.md “Per-Subsection Workflow”): which diagnostics/ scripts you ran, where you added dbg!/tracing calls, where output was hard to interpret, where test failures gave unhelpful messages, where you ran the same command sequence repeatedly. Forward-look: what tool/log/diagnostic would shorten the next regression in this code path by 10 minutes? Implement improvements NOW (zero deferral) and commit each via SEPARATE /commit-push using a valid conventional-commit type (build(diagnostics): ... — surfaced by section-06.1 retrospective — build/test/chore/ci/docs are valid; tools(...) is rejected by the lefthook commit-msg hook). Mandatory even when nothing felt painful. If genuinely no gaps, document briefly: “Retrospective 06.1: no tooling gaps”. Update this subsection’s status in section frontmatter to complete.
/sync-claude section-close doc sync — verify Claude artifacts across all section commits. Map changed crates to rules files, check CLAUDE.md, canon.md. Fix drift NOW.
Repo hygiene check — run diagnostics/repo-hygiene.sh --check and clean any detected temp files.

06.2 Regression Testing

Run the complete test suite to verify zero behavioral regressions.

timeout 150 ./test-all.sh — all tests pass
./clippy-all.sh — no warnings
./fmt-all.sh — no formatting changes needed
timeout 150 cargo st — all Ori spec tests pass
timeout 150 cargo t -p ori_lexer — all lexer tests pass
timeout 150 cargo t -p ori_parse — all parser tests pass
timeout 150 cargo t -p oric — all Salsa integration tests pass

Verify debug AND release builds:

timeout 150 cargo t
cargo b --release  # build first (no timeout on compilation)
timeout 150 cargo t --release  # test execution only, respects mandatory 150s timeout

/tpr-review passed — independent review found no critical or major issues (or all findings triaged)
/impl-hygiene-review passed — hygiene review clean. MUST run AFTER /tpr-review is clean.
Subsection close-out (06.2) — MANDATORY before starting the next subsection. Run /improve-tooling retrospectively on THIS subsection’s debugging journey (per .claude/skills/improve-tooling/SKILL.md “Per-Subsection Workflow”): which diagnostics/ scripts you ran, where you added dbg!/tracing calls, where output was hard to interpret, where test failures gave unhelpful messages, where you ran the same command sequence repeatedly. Forward-look: what tool/log/diagnostic would shorten the next regression in this code path by 10 minutes? Implement improvements NOW (zero deferral) and commit each via SEPARATE /commit-push using a valid conventional-commit type (build(diagnostics): ... — surfaced by section-06.2 retrospective — build/test/chore/ci/docs are valid; tools(...) is rejected by the lefthook commit-msg hook). Mandatory even when nothing felt painful. If genuinely no gaps, document briefly: “Retrospective 06.2: no tooling gaps”. Update this subsection’s status in section frontmatter to complete.
/sync-claude section-close doc sync — verify Claude artifacts across all section commits. Map changed crates to rules files, check CLAUDE.md, canon.md. Fix drift NOW.
Repo hygiene check — run diagnostics/repo-hygiene.sh --check and clean any detected temp files.

06.3 Documentation

Update project documentation to reflect changes and record findings.

06.R Third Party Review Findings

None.

06.N Completion Checklist

Exit Criteria: baselines.json contains populated before/after results showing measurable improvement. All test suites green in both debug and release modes. Documentation updated. parser_raw_throughput_mib_s final > parser_raw_throughput_mib_s baseline.