Files
danyang-match/.github/workflows/ci.yml
T
WorkBuddy f63e579a8d ci(e2e): 落地「活栈供给方式」决策 — Compose in CI,并修掉 6 处真实缺陷
决策:E2E 活栈在 runner 内用 docker compose 起全栈(postgres + redis + minio + backend),
与本地验证路径完全一致;未采纳「service containers + runner 上 mvn package 后 java -jar」——
那需要为后端另写一套 host 网络配置,等于制造第二条与本地不一致的路径。

本机实测(与 CI 同一序列):中间件三项 healthcheck 全 Healthy(pg_isready / redis-cli ping /
mc ready local)→ backend /api/v1/health/live 返回 200 → seed 幂等可执行(INSERT 0 0)
→ pytest tests/e2e → 49 passed。
(本机限制:Docker Hub 不可达且无 temurin 基础镜像,未复现 --build;改用本地既有镜像为底座、
挂载当日 16:24 构建的 jar,使被验证的代码版本与 CI 一致 —— 否则是拿 09-09 的旧代码验 09-18 的用例。)

修掉的 6 处缺陷:
1. minio/minio:latest 是滚动 tag → pin RELEASE.2025-09-07T16-13-09Z(否则 CI 拉到的镜像
   与本地验证时不同,表现为不可复现的偶发失败)
2. up 未显式列出 minio(靠 depends_on 隐式拉起),而 minio 的 healthcheck 正是 backend
   能否启动的前置条件
3. up --build 把构建与起栈耦合,失败日志混杂 → 拆为 build / up 两步
4. 就绪窗口 5min 偏紧(首次构建后仍需 Spring 启动 + Flyway 46 个迁移)→ 7.5min
5. 缺 down -v 前置清卷:seed 用 ON CONFLICT (username) DO NOTHING,残留卷会让账号口令
   停留在上一轮 → 登录 401,并伪装成业务失败
6. job 无 timeout(GH Actions 默认 360 分钟)→ timeout-minutes: 30

另:pytest -v → pytest tests/e2e -v。原以为从仓库根收集会误收 tests/*_test.py
(其在 import 期依赖 psycopg2 与 E2E_DB_PASSWORD),实测否定:pytest.ini 的
testpaths = tests/e2e 已限定范围,从根收集正是 49 个用例且无 error。显式传参作为兜底保留。

计划 Task 2.3 由「依赖基础设施决策」改为「已决策并落地」,仅余 Step 3(CI 上实测一次绿)。
2026-09-18 21:58:56 +08:00

255 lines
9.5 KiB
YAML
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
name: CI/CD Pipeline
on:
push:
branches: [main, develop]
pull_request:
branches: [main, develop]
jobs:
backend:
name: Backend Build & Test
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Set up JDK 21
uses: actions/setup-java@v4
with:
java-version: '21'
distribution: 'temurin'
cache: maven
working-directory: backend
- name: Build and Test
working-directory: backend
run: mvn verify -B
- name: Upload Test Results
if: always()
uses: actions/upload-artifact@v4
with:
name: backend-test-results
path: backend/target/surefire-reports/
admin-web:
name: Admin Web Build
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Set up Node.js
uses: actions/setup-node@v4
with:
node-version: '20'
cache: 'pnpm'
cache-dependency-path: admin-web/pnpm-lock.yaml
- name: Install Dependencies
working-directory: admin-web
run: pnpm install --frozen-lockfile
- name: Build
working-directory: admin-web
run: pnpm build
miniapp:
name: Mini Program Compile Check
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Set up Node.js
uses: actions/setup-node@v4
with:
node-version: '20'
- name: Install Dependencies
working-directory: miniapp
run: npm install
- name: TypeScript Compile Check
working-directory: miniapp
run: npx tsc --noEmit
migration:
name: Migration Validation
runs-on: ubuntu-latest
needs: [backend]
steps:
- uses: actions/checkout@v4
- name: Set up Python
uses: actions/setup-python@v5
with:
python-version: '3.11'
# Phase 23 KEEP-VERIFY:本步骤原注释称「验证 migration SQL 语法」,实际只检查文件非空,
# 属名不副实的假校验(SQL 语法与基线表完整性由后端测试 MigrationValidationTest 覆盖,
# 该测试随 backend job 的 `mvn verify` 运行)。现改为该测试未覆盖的**序列**校验。
- name: 迁移序列校验(命名 / 非空 / 版本号唯一 / 无缺号)
run: |
python3 - <<'PY'
import re, sys, pathlib, collections
d = pathlib.Path("backend/src/main/resources/db/migration")
files = sorted(d.glob("V*.sql"))
if not files:
sys.exit("ERROR: 未找到任何迁移文件")
vers, bad_name, empty = [], [], []
for f in files:
m = re.match(r"V(\d+)__", f.name)
if not m:
bad_name.append(f.name)
continue
if f.stat().st_size == 0:
empty.append(f.name)
vers.append(int(m.group(1)))
problems = []
if bad_name:
problems.append(f"迁移文件命名不合规(应为 V<序号>__<描述>.sql): {bad_name}")
if empty:
problems.append(f"空迁移文件: {empty}")
dups = sorted(v for v, c in collections.Counter(vers).items() if c > 1)
if dups:
problems.append(f"迁移版本号重复: {dups}")
gaps = [v for v in range(min(vers), max(vers) + 1) if v not in set(vers)]
if gaps:
problems.append(f"迁移版本号缺号: {gaps}")
if problems:
for p in problems:
print("ERROR: " + p, file=sys.stderr)
sys.exit(1)
print(f"OK: {len(files)} 个迁移文件,版本 {min(vers)}~{max(vers)} 连续且唯一")
PY
test-shell-guard:
name: Test Shell Guard (Phase 23 Gate)
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Set up Python
uses: actions/setup-python@v5
with:
python-version: '3.11'
# Phase 23 Gate:「不允许 skip/pass 空壳」。
# 机械阻断空壳回归 —— 详见 scripts/check-no-empty-tests.py。
- name: 禁止 skip/pass 空壳测试
run: python3 scripts/check-no-empty-tests.py tests
ui-contracts:
name: UI Contracts (source-structure gates)
runs-on: ubuntu-latest
# Phase 23 遗留(收尾计划 批 1 Task 1.1):14 个 UI 契约门禁此前**只在本地手工执行**,
# 没有 CI 把关 —— 即「改动破坏了源码结构契约也不会被发现」。
# 守卫是纯 node 内建模块、零 npm 依赖,故无需装依赖;统一执行器负责逐个跑 + 汇总 + 非零退出,
# 并校验守卫数量下限(防止守卫被删除/改名后仍然报绿)。
steps:
- uses: actions/checkout@v4
- name: Set up Node.js
uses: actions/setup-node@v4
with:
node-version: '20'
- name: 运行 UI 契约门禁
run: node scripts/run-ui-contracts.mjs
e2e:
name: E2E (live stack)
runs-on: ubuntu-latest
# 防挂死:GH Actions 默认 360 分钟,起栈失败若无人超时会把流水线拖住数小时。
timeout-minutes: 30
# Phase 23 遗留 D-23-D:计划要求 CI 串联 E2E。此前 E2E 从未真正执行过——
# 无 conftest、fixture 不存在、文件在 import 期就崩。本 job 起真实栈后执行。
needs: [backend, migration]
env:
E2E_BASE_URL: http://localhost:8085/api/v1
# CI 专用测试口令,仅存在于 runner 运行期,不入库、不用于任何真实环境
E2E_OPERATOR_PASSWORD: ci-e2e-only-password
steps:
- uses: actions/checkout@v4
- uses: actions/setup-python@v5
with:
python-version: '3.11'
- name: Install E2E dependencies
run: pip install -r tests/e2e/requirements.txt
- name: 生成 CI 测试用 .env(仅运行期存在,不入库)
run: |
cat > infrastructure/.env <<'ENV'
DB_PASSWORD=ci-e2e-db-password
REDIS_PASSWORD=ci-e2e-redis-password
MINIO_ACCESS_KEY=ci-e2e-minio
MINIO_SECRET_KEY=ci-e2e-minio-secret
JWT_ACCESS_SECRET=ci-e2e-access-secret-only-for-ci
JWT_REFRESH_SECRET=ci-e2e-refresh-secret-only-for-ci
ENCRYPTION_KEY=0123456789abcdef0123456789abcdef0123456789abcdef0123456789abcdef
WECHAT_APP_SECRET=
ENV
# 「活栈供给方式」决策(2026-09-18):**在 runner 内用 Compose 起全栈**。
# 备选是 service containers + 在 runner 上 `mvn package` 后 `java -jar`,但那需要为后端
# 另写一套 host 网络配置,等于制造第二条与本地不一致的路径;Compose 全栈则与本地验证
# 路径完全相同(2026-09-18 本机按同一序列实测:中间件三项 healthcheck 全 Healthy、
# backend /health/live 200、seed 幂等可执行、**pytest tests/e2e 49 passed**)。
- name: 校验 compose 配置(秒级快速失败,早于任何构建)
working-directory: infrastructure
run: docker compose config --quiet
# 保证库为全新:seed 用 `ON CONFLICT (username) DO NOTHING`,若卷是上一轮残留,
# 账号密码会停留在上一轮的 `E2E_OPERATOR_PASSWORD` 上 → 本轮登录 401(表现为各种业务失败)。
- name: 清理可能的残留栈与数据卷
working-directory: infrastructure
run: docker compose down -v --remove-orphans || true
# 构建与起栈分两步:构建失败时的日志不混杂在 up 输出里,定位更快。
- name: 构建 backend 镜像
working-directory: infrastructure
run: docker compose build backend
- name: 起真实栈(postgres + redis + minio + backend)
# 显式列出 minio:backend 的 depends_on 虽会隐式拉起它,但隐式依赖不利于阅读,
# 且 minio 的 healthcheck(`mc ready local`)是 backend 能否启动的前置条件。
working-directory: infrastructure
run: docker compose up -d postgres redis minio backend
- name: 等待 backend 就绪(超时即失败,不静默跳过)
# 窗口 90×5s=7.5min:首次构建后还需 Spring 启动 + Flyway 46 个迁移,5min 偏紧。
run: |
for i in $(seq 1 90); do
if curl -sf "$E2E_BASE_URL/health/live" > /dev/null; then
echo "backend READY after ${i} checks"; exit 0
fi
sleep 5
done
echo "::error::backend 未在 7.5 分钟内就绪"
docker compose -f infrastructure/docker-compose.yml logs --tail=200 backend
exit 1
- name: Seed E2E 运营账号(迁移不含 operator 种子,全新库无账号可登录)
working-directory: infrastructure
run: |
docker compose exec -T postgres psql -U postgres -d danyang_match \
-v pwd="$E2E_OPERATOR_PASSWORD" -f - < ../tests/e2e/sql/seed_e2e_operators.sql
- name: Run E2E
# 显式给出目录:pytest.ini 的 `testpaths = tests/e2e` 已限定范围,但显式传参可防
# 该配置被改动后悄悄扩大收集范围(仓库根另有 tests/*_test.py 不是 E2E 用例,
# 且它们在 import 期即依赖 psycopg2 与 E2E_DB_PASSWORD,会被收集成 error)。
run: pytest tests/e2e -v
- name: Teardown(无论成败都清理并留日志)
if: always()
run: |
docker compose -f infrastructure/docker-compose.yml logs backend || true
docker compose -f infrastructure/docker-compose.yml down -v || true