test: a registry of known breakages that tests must catch
Validate / provenance (push) Successful in 46s
Validate / hacs (push) Failing after 10s
Validate / hassfest (push) Failing after 12s
Validate / process-gate (push) Failing after 35s
Validate / frontend (push) Successful in 6m11s
Validate / backend (push) Failing after 9m24s
Validate / golden (push) Failing after 10m10s
Validate / performance_smoke (push) Failing after 12m46s
Validate / smoke (push) Failing after 30m15s

Five times in this project a green test meant nothing was checked. The
continuity smoke stayed green after the entire mechanism it guards was cut out.
The golden scene created to protect doorway light was empty — 1,177 warm pixels
against 107,119, all of them icons. The shadow smoke passed while no shadow was
drawn. Each time the test had been written alongside the code, went green at
once, and nobody ever asked whether it could go red.

The gate makes that question routine. Each mutant is a few lines of patch that
reproduce a known breakage, plus the name of the test that must fail on it. A
worktree is patched, the bundle rebuilt, the guard run — and a guard that stays
green fails the gate. Six mutants cover the holes documented in #85; the anchors
are exact strings from today's source, so the registry cannot silently drift —
a unit test that runs with the ordinary suite refuses a stale anchor.

The full run rebuilds the bundle per mutant, so it lives in its own workflow,
before a stable release and on a weekly schedule, not in Validate. The rules for
new tests are written at the top of docs/TESTING.md, and the sixth of them is
the cheapest: an assertion that reads back the property the code just set is
not written at all.

Issue: #85
User-Visible: no
This commit is contained in:
Matysh
2026-08-14 02:05:53 +03:00
parent 0e69c4a183
commit 328ed7afc0
3 changed files with 399 additions and 0 deletions
+61
View File
@@ -0,0 +1,61 @@
name: Mutation gate
# Реестр известных поломок (issue #85): каждый мутант ломает продуктовый код
# известным способом, и объявленный тест ОБЯЗАН на этом покраснеть. Тест,
# оставшийся зелёным на сломанном коде, ничего не защищает — он лишь выглядит
# защитой, и это хуже его отсутствия.
#
# Прогон дорогой: пересборка бандла на каждого мутанта. Поэтому он не входит в
# Validate и не идёт на каждый push. Его место — перед стабильным релизом
# (PROCESS.md §8) и раз в неделю по расписанию, чтобы дрейф тестов не копился
# до релиза. Дешёвая половина — «якоря патчей живы, guard-файлы существуют» —
# идёт с обычными юнитами: test/mutation-gate.test.mjs.
on:
workflow_dispatch:
schedule:
# Понедельник, 05:20 UTC — до начала рабочего дня владельца.
- cron: '20 5 * * 1'
permissions:
contents: read
concurrency:
group: mutation-gate
cancel-in-progress: true
jobs:
mutants:
runs-on: ubuntu-latest
# Шесть мутантов × (сборка + браузерный смок) — это десятки минут, и это
# нормально: гейт предрелизный. Час — потолок против зависшего Chromium.
timeout-minutes: 60
steps:
- uses: actions/checkout@v4
with:
ref: dev
fetch-depth: 0
- uses: actions/setup-node@v4
with:
node-version: 22
cache: npm
- run: npm ci
- name: Кэш браузеров Playwright
id: pw
uses: actions/cache@v4
with:
path: ~/.cache/ms-playwright
key: playwright-${{ runner.os }}-${{ hashFiles('package-lock.json') }}
- name: Установить Chromium
if: steps.pw.outputs.cache-hit != 'true'
run: npx playwright install --with-deps chromium
- name: Реестр применим к текущему коду
run: node scripts/mutation-gate.mjs --check
- name: Каждый тест ловит свою поломку
run: node scripts/mutation-gate.mjs