feat(hooks-post-tool-linter): no-time-field-strict-greater rule (順位 47 / §A-2 P-2)

[aloekun]

Summary

• §A-2 Phase 5 dogfood P-2 (順位 47) — PR feat(check-ci-coderabbit): --list-findings モード追加 (Bundle a Sub-PR 1) #101 で実証された latent drift (parse_findings 系 3 関数で時刻フィールドの > vs >= 混在) を構造的に防止する custom lint rule を追加
• Bundle Z #B-α と同じ「決定論的防止層」哲学 (ADR-007 正規表現層)
• master 上 classifier (P-0 で有効化) が本 PR の post-pr-monitor で classification を起動 → §A-2 計測ログに dogfood 結果を記録予定

Changes

.claude/custom-lint-rules.toml

新ルール no-time-field-strict-greater 追加:

[[rules]]
id = "no-time-field-strict-greater"
pattern = '\b(created_at|submitted_at|updated_at|comment_event_time|event_time|comment_created_at|published_at|posted_at|commented_at)\s*>\s*[a-zA-Z_]'
severity = "warning"
extensions = ["rs"]

検出対象: 9 種の時刻フィールド (created_at / submitted_at / updated_at / comment_event_time / event_time / comment_created_at / published_at / posted_at / commented_at) の strict > 比較で、RHS が alphabetic identifier の場合。

Direction-specific 設計 (> のみ flag):

• < の strict 比較 (例: is_stale の created_at < threshold) は意図的な exclusive 境界として正しい用途があるため scope 外
• < 側の boundary 問題は観測されたら別 rule で対応

src/hooks-post-tool-linter/src/main.rs

11 unit tests + 3 helper 関数 で transition matrix を網羅:

Test	入力	期待
detects_created_at_gt_push_time	created_at > push_time	1 violation
detects_submitted_at_gt_since	submitted_at > since	1 violation
detects_updated_at_gt_threshold	updated_at > threshold	1 violation
detects_comment_event_time	comment_event_time > now	1 violation
skips_inclusive_comparison	created_at >= push_time	0 violations
skips_strict_less_than	created_at < threshold	0 violations
skips_le_inclusive	created_at <= cutoff	0 violations
skips_numeric_rhs	created_at > 0	0 violations
skips_doc_comment_with_inclusive	/// created_at >= push_time	0 violations
skips_unrelated_field	count > limit	0 violations
only_targets_rs	TOML file with created_at > push_time	0 violations

Helper 関数 (test fixture を runtime 構築):

• build_rs_source_with_op(field, op, rhs) -> String
• build_doc_comment_source(field, op, rhs) -> String
• build_toml_with_field(field, op, rhs) -> String

理由: literal created_at > push_time を test source 内に書くと 新ルール自身が test source を flag して self-trigger ループに陥る。format!() 経由で runtime 構築すると Edit-time lint の regex (literal adjacency 必須) を回避できる (= dogfood で実証された "lint rule self-violation" の structural pattern)。

.claude/hooks-post-tool-linter.exe

pnpm build:hooks-post-tool-linter で release build を再生成し配備

既存 codebase の lint 適用結果

• 0 violations (grep -rn '\b(created_at|...)_at\s*>\s*[a-zA-Z_]' で 0 件)
• 既存の created_at >= push_time (doc コメント) は >= のため対象外
• 既存の is_stale の created_at < threshold は < のため対象外 (Direction-specific 設計の効果)

§A-2 dogfood (P-2) 観測ポイント

本 PR の post-pr-monitor で classifier が CodeRabbit findings を分類する 2 回目の dogfood サイクル (P-1 では findings ゼロで起動せず)。lint rule 追加 PR は CR から style / regex 厳密性 / edge case 指摘が出やすく、findings が複数発生する想定。

計測対象は P-1 と同じ:

1. agreement rate: classifier の action 一致率 — 目標 ≥80%
2. classifier latency: finding あたり秒数 — 目標 ≤5s/件
3. fallback rate: human_review fallback 比率 — 目標 ≤20%
4. normalized_issue 言語制約違反率: 英語 8 文字以上連続検出 — 目標 ≤10%

Self-violation 知見の副産物

本 PR で 新 lint rule が test source に self-trigger するパターンを実体験。Edit-time lint hook (PostToolUse) は new_string (Edit の追加文字列) を scan するため、test fixture に literal pattern を書くと Edit が即 reject される。回避策として test fixture を format!() で runtime 構築する DSL pattern を確立。これは新 lint rule 追加時の standard pattern として再利用可能 (今後の T1 lint task で同様のテクニックが必要)。

Test plan

• cargo test -p hooks-post-tool-linter rs_time_field_strict_greater — 11/11 pass
• cargo test -p hooks-post-tool-linter — 75/75 pass、regression なし
• pre-push-review APPROVE (1 iter / 3m 4s)
• release build + deploy
• 既存 codebase で > strict 違反 0 件確認
• (post-merge) classifier 起動確認 + §A-2 計測ログ記録

関連

• 順位 47 — docs/todo5.md 詳細エントリ
• §A-2 P-2 — docs/local-llm-offload-analysis.md
• ADR-007 — custom lint rule の正規表現/AST 層線引き
• Bundle Z #B-α (ADR-036) — 決定論的防止層哲学

Summary by CodeRabbit

リリースノート

• Tests

  • タイムスタンプ比較に関連する新しいリントルールのテストカバレッジを追加しました。

• Chores

  • 開発ツールにカスタムリント規則を追加し、コード品質チェックを強化しました。

[coderabbitai]

📝 Walkthrough

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 75.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	プルリクエストのタイトルは、新しいカスタムリント規則 no-time-field-strict-greater の追加という主要な変更を明確に要約しており、変更セットの中心内容と完全に関連している。
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

🧹 Nitpick comments (1)

src/hooks-post-tool-linter/src/main.rs (1)

1360-1366: ⚡ Quick win

9 フィールドのうち 5 フィールドに検出テストがありません。

TOML の交代パターンには 9 フィールドが列挙されていますが、陽性検出テスト (violations.len() == 1 を assert するもの) は created_at、submitted_at、updated_at、comment_event_time の 4 フィールドのみです。以下の 5 フィールドには検出テストがなく、TOML の交代式への誤字があってもテストスイートでは捕捉できません。

• event_time
• comment_created_at
• published_at
• posted_at
• commented_at

🧪 追加テストの例 (残り 5 フィールド)

+    #[test]
+    fn rs_time_field_strict_greater_detects_event_time() {
+        let dir = tempfile::tempdir().unwrap();
+        let file = write_file(
+            dir.path(),
+            "parse.rs",
+            &build_rs_source_with_op("event_time", ">", "cutoff"),
+        );
+        let rules = compile_test_rules(vec![rs_time_field_strict_greater_rule()]);
+        let violations = run_custom_rules(file.to_str().unwrap(), &rules);
+        assert_eq!(violations.len(), 1);
+    }
+
+    #[test]
+    fn rs_time_field_strict_greater_detects_comment_created_at() {
+        let dir = tempfile::tempdir().unwrap();
+        let file = write_file(
+            dir.path(),
+            "parse.rs",
+            &build_rs_source_with_op("comment_created_at", ">", "since"),
+        );
+        let rules = compile_test_rules(vec![rs_time_field_strict_greater_rule()]);
+        let violations = run_custom_rules(file.to_str().unwrap(), &rules);
+        assert_eq!(violations.len(), 1);
+    }
+
+    #[test]
+    fn rs_time_field_strict_greater_detects_published_at() {
+        let dir = tempfile::tempdir().unwrap();
+        let file = write_file(
+            dir.path(),
+            "parse.rs",
+            &build_rs_source_with_op("published_at", ">", "threshold"),
+        );
+        let rules = compile_test_rules(vec![rs_time_field_strict_greater_rule()]);
+        let violations = run_custom_rules(file.to_str().unwrap(), &rules);
+        assert_eq!(violations.len(), 1);
+    }
+
+    #[test]
+    fn rs_time_field_strict_greater_detects_posted_at() {
+        let dir = tempfile::tempdir().unwrap();
+        let file = write_file(
+            dir.path(),
+            "parse.rs",
+            &build_rs_source_with_op("posted_at", ">", "now"),
+        );
+        let rules = compile_test_rules(vec![rs_time_field_strict_greater_rule()]);
+        let violations = run_custom_rules(file.to_str().unwrap(), &rules);
+        assert_eq!(violations.len(), 1);
+    }
+
+    #[test]
+    fn rs_time_field_strict_greater_detects_commented_at() {
+        let dir = tempfile::tempdir().unwrap();
+        let file = write_file(
+            dir.path(),
+            "parse.rs",
+            &build_rs_source_with_op("commented_at", ">", "boundary"),
+        );
+        let rules = compile_test_rules(vec![rs_time_field_strict_greater_rule()]);
+        let violations = run_custom_rules(file.to_str().unwrap(), &rules);
+        assert_eq!(violations.len(), 1);
+    }

Also applies to: 1380-1521

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/hooks-post-tool-linter/src/main.rs` around lines 1360 - 1366, The test
suite is missing positive-detection tests for five time-field alternatives
listed in the regex used by rs_time_field_strict_greater_rule (event_time,
comment_created_at, published_at, posted_at, commented_at); add unit tests
alongside the existing positive cases (the tests that assert violations.len() ==
1) to cover each of these five fields so any typo in the TOML pattern will be
caught. Locate rs_time_field_strict_greater_rule and the test block(s) that
currently assert violations.len() == 1 for created_at, submitted_at, updated_at,
comment_event_time and add similar tests that feed sample Rust input containing
each of event_time, comment_created_at, published_at, posted_at, and
commented_at (using the same test harness and expectations) to ensure each
alternative triggers a violation.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In `@src/hooks-post-tool-linter/src/main.rs`:
- Around line 1360-1366: The test suite is missing positive-detection tests for
five time-field alternatives listed in the regex used by
rs_time_field_strict_greater_rule (event_time, comment_created_at, published_at,
posted_at, commented_at); add unit tests alongside the existing positive cases
(the tests that assert violations.len() == 1) to cover each of these five fields
so any typo in the TOML pattern will be caught. Locate
rs_time_field_strict_greater_rule and the test block(s) that currently assert
violations.len() == 1 for created_at, submitted_at, updated_at,
comment_event_time and add similar tests that feed sample Rust input containing
each of event_time, comment_created_at, published_at, posted_at, and
commented_at (using the same test harness and expectations) to ensure each
alternative triggers a violation.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 1d4e3e97-31ae-466d-96fe-e93eb72cfea1

📥 Commits

Reviewing files that changed from the base of the PR and between be5ee6f and 39f0d95.

📒 Files selected for processing (2)

• .claude/custom-lint-rules.toml
• src/hooks-post-tool-linter/src/main.rs