Testing¶
文件角色：implementation flow 投影；圖與步驟不得另建產品規則或參數 authority。
| Implementation authority | 產品 canon／流程 | 全域索引 |
|---|---|---|
| testing.md | 程式架構 · 升版 | 流程.md §2 |
測試層與 runner¶
graph TB
subgraph 測試層
Unit[Unit<br/>Angular unit-test builder + Vitest]
Int[Integration<br/>Vitest node environment + libp2p memory]
E2E[E2E<br/>Playwright]
Det[Determinism<br/>Cross-browser/OS harness]
end
Unit --> Int
Int --> E2E
Unit --> Det
E2E --> Det
classDef local fill:#e8f3ec,stroke:#3f8f5f,stroke-width:1.4px,color:#173525;
classDef consensus fill:#fdf3df,stroke:#c08a2d,stroke-width:1.4px,color:#3d2c0d;
classDef fail fill:#fae9e7,stroke:#b4544a,stroke-width:1.4px,color:#471d18;
class Unit,Int local
class E2E consensus
class Det failCI 觸發流程¶
flowchart TD
Event{觸發事件}
Event -->|master push / PR| CI[validate + unit + integration + E2E<br/>+ canonical assets gate]
Event -->|nightly schedule / 手動| Night[fuzz + performance<br/>+ canonical assets perf]
Event -->|workflow_dispatch| Release[determinism + perf + deployment assets<br/>+ production build + E2E]
CI --> Result{結果}
Night --> Result
Release --> Result
Result -->|綠| Pass[繼續]
Result -->|紅| Block[阻擋]
Result -->|效能退化落在警告區間| Warn[警告但通過<br/>依正文門檻判定]
classDef local fill:#e8f3ec,stroke:#3f8f5f,stroke-width:1.4px,color:#173525;
classDef consensus fill:#fdf3df,stroke:#c08a2d,stroke-width:1.4px,color:#3d2c0d;
classDef fail fill:#fae9e7,stroke:#b4544a,stroke-width:1.4px,color:#471d18;
class Pass local
class Warn consensus
class Block fail確定性測試流程¶
sequenceDiagram
participant Test as Test Runner
participant Browser1 as Chromium
participant Browser2 as Firefox
participant Browser3 as WebKit
participant Fixture as 期望 Checksum
Test->>Browser1: 跑固定 seed 與 frame 數的 sim
Browser1->>Test: checksum_chromium
Test->>Browser2: 跑相同 sim
Browser2->>Test: checksum_firefox
Test->>Browser3: 跑相同 sim
Browser3->>Test: checksum_webkit
Test->>Fixture: 比對
alt 全部一致
Fixture->>Test: ✅ PASS
else 任一不同
Fixture->>Test: ❌ FAIL + diff trace
Test->>Test: 阻擋 release
end8 Peer Mesh 壓測¶
graph TB
Setup[啟動完整 in-memory peer 集合]
Setup --> Mesh[建立 full mesh]
Mesh --> Assert[驗證全部 pairwise connections]
Assert -->|全部建立| Pass[PASS]
Assert -->|任一缺席| Fail[FAIL]
Note["此 in-process memory proof 不模擬<br/>packet loss、latency 或賽事流量"]
classDef local fill:#e8f3ec,stroke:#3f8f5f,stroke-width:1.4px,color:#173525;
classDef fail fill:#fae9e7,stroke:#b4544a,stroke-width:1.4px,color:#471d18;
class Pass local
class Fail failFuzz 流程¶
flowchart TD
Start[fuzz runner 啟動] --> Gen[fast-check 隨機產生 input]
Gen --> Run[呼叫 sanitizer / protocol parser]
Run --> Result{結果}
Result -->|throw| Bug[找到 BUG]
Result -->|crash| Bug
Result -->|hang > timeout| Bug
Result -->|ok| Continue[繼續]
Bug --> Shrink[fast-check shrink<br/>找最小 repro]
Shrink --> Save[存到 corpus]
Save --> Report[輸出本機可重現報告<br/>由維護者審閱]
Continue --> Loop{達到 budget？}
Loop -->|否| Gen
Loop -->|是| Done[結束]
classDef local fill:#e8f3ec,stroke:#3f8f5f,stroke-width:1.4px,color:#173525;
classDef fail fill:#fae9e7,stroke:#b4544a,stroke-width:1.4px,color:#471d18;
class Done local
class Bug failCoverage 門檻流程¶
graph LR
Run[pnpm test:coverage] --> Core[coverage-core gate]
Run --> Pages[coverage-pages gate]
Run --> Bootstrap[coverage-bootstrap-adapters gate]
Core --> Check{三個 gate 全數通過?}
Pages --> Check
Bootstrap --> Check
Check -->|否| Fail[CI Fail]
Check -->|是| Pass[Coverage gate Pass]
Pass --> Report[保留 CI coverage artifact]
classDef local fill:#e8f3ec,stroke:#3f8f5f,stroke-width:1.4px,color:#173525;
classDef fail fill:#fae9e7,stroke:#b4544a,stroke-width:1.4px,color:#471d18;
class Pass local
class Fail fail私密漏洞重現與公開 regression test¶
sequenceDiagram
participant Reporter
participant Private as 私密通報管道
participant Maintainer
participant Isolated as 隔離重現環境
participant PR as 公開 PR
participant CI as 一般 CI
Reporter->>Private: 私下提交漏洞與重現資料
Private->>Maintainer: 通知並協調補充資訊
Maintainer->>Isolated: 在私密隔離環境驗證
alt repro 成功
Isolated->>Maintainer: 確認漏洞並修復
Maintainer->>Private: 私下回覆結果與揭露時程
Maintainer->>PR: 修復或協調揭露後提交去敏 regression test
PR->>CI: 執行一般測試
CI->>PR: 回報 regression 結果
else repro 失敗
Isolated->>Maintainer: 無法重現
Maintainer->>Private: 私下請求補充細節
end測試環境隔離¶
graph TB
Host[CI Runner]
subgraph Test 1
IDB1[Memory store／手寫 IDB port fake]
WRTC1["@libp2p/memory transport"]
Time1[fake timers]
end
subgraph Test 2
IDB2[Memory store／手寫 IDB port fake]
WRTC2["@libp2p/memory transport"]
Time2[fake timers]
end
Host --> Test1
Host --> Test2
Note["每個 worker 獨立<br/>避免狀態洩漏"]
classDef chain fill:#e7eefb,stroke:#3f6bb0,stroke-width:1.4px,color:#152848;
class Test1,Test2 chainPerf Baseline 比對¶
flowchart LR
PR[新 PR] --> Bench[跑 Perf 基準]
Bench --> Compare{對比 baseline}
Compare -->|低於警告門檻| Green[✅ 通過]
Compare -->|警告區間| Yellow[⚠️ 警告但通過]
Compare -->|超過阻擋門檻| Red[❌ 阻擋 merge]
Red --> Investigate[調查退化原因]
Yellow --> Note[寫入 PR comment]
Green --> Update[可選：更新 baseline]
Thresholds["讀取正文的門檻與校準規則"]
classDef local fill:#e8f3ec,stroke:#3f8f5f,stroke-width:1.4px,color:#173525;
classDef consensus fill:#fdf3df,stroke:#c08a2d,stroke-width:1.4px,color:#3d2c0d;
classDef fail fill:#fae9e7,stroke:#b4544a,stroke-width:1.4px,color:#471d18;
class Green local
class Yellow consensus
class Red fail測試資料生命週期¶
graph TB
Source[測試資料來源] --> Type{類型}
Type --> Fixtures[Fixtures<br/>git-tracked]
Type --> Generated[Generated<br/>at-runtime]
Type --> Recorded[Recorded<br/>從真實比賽]
Fixtures --> Use[測試使用]
Generated --> Use
Recorded --> Use
Use --> Update{需更新？}
Update -->|fixture 過期| RegFix[重新產出]
Update -->|new feature| AddFix[新增]
RegFix --> Source
AddFix --> Source
classDef local fill:#e8f3ec,stroke:#3f8f5f,stroke-width:1.4px,color:#173525;
classDef consensus fill:#fdf3df,stroke:#c08a2d,stroke-width:1.4px,color:#3d2c0d;
classDef chain fill:#e7eefb,stroke:#3f6bb0,stroke-width:1.4px,color:#152848;
class Recorded local
class Generated consensus
class Fixtures chain