Skip to content

docs(journal): 2026-05-23 日報 + journal/ subdirectory 新設 #425

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
hideyukiMORI merged 1 commit into from
May 22, 2026
Merged

docs(journal): 2026-05-23 日報 + journal/ subdirectory 新設 #425

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
115 changes: 115 additions & 0 deletions docs/journal/2026-05-23.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,115 @@
# 日報 2026-05-23

> セッション継続のため、作業範囲は **2026-05-22 後半 〜 2026-05-23** に跨る。`Date:` ヘッダは記録日 = 2026-05-23。

## 作業範囲

NeNe フレームワーク改修。FT13 (前セッションから継続) 〜 FT17 の 5 trial + 評価レポート (#401) 由来 7 件 + cross-repo handoff 1 件を完走。

## 完了タスク

### Field Trial (5 件)

| FT | テーマ | 採用 ADR | Trial Issue | Report |
| --- | --- | --- | --- | --- |
| FT13 | email-sending | **ADR-0006** Symfony Mailer | #374 | #375 |
| FT14 | security-headers | **ADR-0007** ResponseDecoration boundary | #383 | #384 |
| FT15 | request-id | (ADR-0007 generality 実証) | #389 | #390 |
| FT16 | agent-bearer-auth | **ADR-0008** Optional Bearer | #380 (cross-repo) | #396 |
| FT17 | schema-diff CLI | **ADR-0009** schema migrations | #417 | #418 |

### 評価レポート (#401) 由来 7 件

別 Claude による peer review (`REPORT_project_evaluation.md` 88/100、`REPORT_commercial_feasibility.md`) を保管し、指摘事項を 7 PR で完全消化:

- **#404** ADR README index 0003-0008 追加 → PR #410
- **#403** CI 強化 (Phan + format:check + checkout v4 + Docker buildx cache) → PR #411
- **#406** "debug" controller dead-code 削除 (2020 年初版以来の 6 年放置を git log で確認) → PR #412
- **#405** 型 tightening (`$REQUEST_JSON` parameterize + `ApiResponse::failure` array-shape) → PR #413
- **#408** Security tightening (`Request::getFile` defense-in-depth + 新 `RedirectGuard` fail-closed allowlist) → PR #414
- **#407** test 補強 (TransactionManager nested + DataModelBase fixture-based、+15 cases) → PR #415
- **#409** ADR-0009 schema migration story draft → PR #416 → **FT17 で実装まで完遂**

### Merged PR (24 件)

FT13: #375 / #379 / #381 / #382
FT14: #384 / #387 / #388
FT15: #390 / #393 / #394
FT16: #396 / #399 / #400 / #402 (#395 final HTTP smoke)
評価レポート対応: #401 (report 保管) / #410 / #411 / #412 / #413 / #414 / #415 / #416
FT17: #418 / #422 / #423 / #424

## 数値変化

- **ADR**: 5 本 → **9 本** (0006 / 0007 / 0008 / 0009 採用)
- **Unit tests**: 89 → **142** (+53, +60%)
- **CI strictness**: Phan + format:check が PR ごとに走るように (baseline drift 検知有効化)
- **CI 速度**: runtime-smoke 2m28s → 1m11s (Docker buildx + gha cache、~50% 短縮)
- **REPORT_commercial_feasibility.md 最大懸念** (migration 機構欠如): **解消** (operator-applied `composer schema:diff`)

## 構造的に重要だったもの

### **ADR-0007 が長年の宿題を消した**

FT7 (#311) で documented した「decoration trap at dispatcher boundary」を FT14 で `ResponseDecorator` + `HttpEmitter::emit()` + `View::execute()` の **2 hook 設計**で正面解決。FT15 で同 boundary が observability (request-id) でも動くことを **validation trial として実証**。これで future の cross-cutting concern (Server-Timing / OTel / 監査 fingerprint) は同 recipe で plug-in 可能。

### **cross-repo trial governance が確立**

FT16 で nene-mcp #380 を新 Issue 切らずに直接 trial Issue 化する pattern を実証 → `docs/field-trials/README.md` に **"Cross-repo Issue reuse" subsection** として正式項目化。今後 NENE2 / nene2-python からの handoff も同様運用。Phase 6 thesis (cross-project governance) の最も clean な実装事例。

### **評価レポート由来 7 issue を 1 セッションで全消化**

外部視点 → Issue → PR → merge のループが 1 セッションで一周。`composer check` が定義されてたのに CI で呼ばれてなかった、`actions/checkout@v6` が存在しないバージョンだった等、外部視点でなければ気付けなかった問題が複数 surface。

## 技術的な学び (memory 化済)

1. **PHP の `getenv($k) ?: 'default'` は `'0'` / `''` を falsy 扱いで壊れる** (FT15 F-1)
`$raw = getenv(...); $value = $raw === false ? 'default' : (string)$raw;` が正解。`coding-standards.md` に予防 note 追加。FT16 / FT17 では再発防止できた。

2. **Apache mod_php が `$_SERVER['HTTP_AUTHORIZATION']` を strip する** (FT16 F-1)
`apache_request_headers()` / `getallheaders()` まで降りる 4-layer fallback が必要。`agent-bearer-auth.md` に recipe 化。1 時間 hammer 中の最も重い debug。

3. **NeNe TODO REST に `GET /todo/item/id_{id}` が無かった** (FT16 F-2)
姉妹リポの catalog が assumption していた read-by-id 穴。`itemGetRest` 追加で解消。実 user 視点が無ければ気付けなかった。

4. **UNIQUE constraint 由来 index が STATISTICS 経由で false-positive warning** (FT17 F-5)
同 PR 内で `NON_UNIQUE = 1` (MySQL) / `unique = 0` (SQLite) フィルタで解消。SchemaDefinition の `unique` キーは CREATE TABLE 一括管理なので diff path から外す方針が確立。

## ハマったところ

- FT16 の Apache mod_php 問題で T-1 acceptance が 401 になり、debug に約 30 分。`$_SERVER['HTTP_AUTHORIZATION']` 単独に頼ってはいけないことが framework knowledge として固まったので投資対効果はプラス。
- FT15 で `getenv() ?:` の罠を 1 度踏んだが unit test が即 catch、5 分で fix。test-first 設計が効いた瞬間。

## Trial の性格分化

FT12 までの「real-app entity 立ち上げ」一辺倒だったのが、本セッションで **5 種の異なる trial 性格** に分化:

| trial | 性格 |
| --- | --- |
| FT13 | 完全空白 surface 立ち上げ + dep ADR |
| FT14 | structural pattern の正面解決 (長期宿題) |
| FT15 | 前 ADR の generality validation |
| FT16 | cross-repo handoff の handler trial |
| FT17 | 前 ADR の implementation trial |

trial methodology が成熟してきている。

## 翌日以降の候補

| Priority | 候補 | 規模 |
| --- | --- | --- |
| Real-app gap | session-storage-backend (Redis) | 中〜大 (ADR-class) |
| Defer chunk | DataMapperBase の PDO 依存 unit test 補強 (#407 残り) | 中 |
| Meta | ai-agent-journey (Phase 6 thesis 自己検証) | 長尺 |
| Observability | structured-logs (JSON formatter swap) | 小 |
| Architecture | constraint-changes ADR (FK / UNIQUE 追加 ALTER 経路) | ADR-class、需要次第 |

## 雑感

評価レポート (#401) が来たタイミングと、ADR-0007 → ADR-0008 → ADR-0009 が連鎖した流れが噛み合って、これまで個別だった structural debt が 1 セッションで複数 closed された。特に **FT7 F-6 が 7 trial 越しに resolved** された瞬間は trial methodology が effective に働いている強い証拠。

明日以降は real-app gap (Redis session 等) か meta trial (ai-agent-journey) のどちらかで、また性格を変えたい。

---

**Status (セッション終了時)**: main green、未 merged ブランチなし、全 trial / 全 issue close 済み。
31 changes: 31 additions & 0 deletions docs/journal/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,31 @@
# Journal

Day-by-day session summaries — what was worked on, what merged, what got learned, what's queued for next time. Different from:

- `docs/field-trials/` — trial-specific reports, scoped to one FT each.
- `docs/todo/current.md` — current state of the work, kept fresh; loses history when items move out of "Active".
- `docs/roadmap.md` — phase-level direction, slow-moving.
- `docs/releases.md` — versioned release notes for users of the framework.

The journal is the **working log** for maintainers and reviewing AI agents. It captures the texture of a session (what was hard, what surprised, what was deferred) that a list-of-merged-PRs would lose.

## File naming

`docs/journal/YYYY-MM-DD.md`, one file per session-day. If multiple sessions happen on the same date, append the file in place rather than creating a sibling.

## What belongs here

- Work completed (with PR / Issue references)
- ADRs accepted / drafted
- Notable bugs surfaced + fixed
- Technical traps documented (so future selves don't re-discover them)
- Cross-repo handoffs
- What's queued for the next session

## What does **not** belong here

- Code recipes — those go in `docs/development/`.
- Design decisions — those become ADRs.
- Active TODO — that lives in `docs/todo/current.md`.

A journal entry should be useful 6 months later as "what was the context when this PR was merged?" not as a substitute for the actual artifacts.
Loading