문제: 빌드는 `dist`, `.generated/frontend-runtime`, `artifacts/release`·
`artifacts/quality`에 고정으로 쓰고, 다섯 Playwright 설정은 결과·report·
JUnit을 `artifacts/tests/<이름>/`에 고정으로 쓴다. 한 checkout에서 두
번째 e2e 실행이 시작하면 빌드 1단계가 첫 실행의 preview가 서빙 중인
`dist`를 지우고, Playwright가 `outputDir`을 비워 첫 실행의 trace를 지운다.
증거 파일도 늦게 끝난 쪽 것으로 섞인다(1105/12-04).
고른 방법: 변수 두 개로 나눈다(`scripts/lib/run-isolation.ts`).
- `TEST_RUN_ID`: Playwright 설정이 읽는다. 값이 있으면 그 실행의 결과·
report·JUnit과 서빙할 빌드를 `.tmp/runs/<id>/` 아래 같은 상대 경로에
둔다. 없으면 지금 경로 그대로라 게이트, CI upload, 증거 검사가 읽는
곳이 바뀌지 않는다.
- `RUN_OUTPUT_ROOT`: 설정이 web server에만 넘긴다. 빌드 단계 전부
(`build-frontend`, contract set, runtime config, SW asset 목록, SW 빌드,
build manifest)와 `vite.config.ts`가 읽어 `dist`·생성 모듈·증거·Vite
캐시를 그 아래에 쓴다. preview는 `--outDir`로 같은 `dist`를 서빙한다.
빌드 코드는 `TEST_RUN_ID`를 읽지 않는다. e2e worker가 저장소 복사본을
빌드할 때(fixture-builds, service-worker-builds) runner 환경을 물려받는데,
그 빌드는 복사본 자기 `dist`에 써야 하기 때문이다.
두 루트는 `.` 또는 `.tmp/` 아래만 받는다. git, Tailwind 소스 스캔(report
안 문자열이 CSS에 들어가지 않게), Vite watcher, vitest, 저장소 복사
helper가 모두 `.tmp`를 건너뛴다. service-worker-builds의 복사 제외 목록에
`.tmp`를 더했다(fixture-builds와 같아짐). build manifest의 경로는 빌드가
쓴 폴더 기준 상대 경로로 둔다.
`vite.config.ts`가 새로 가져오는 `scripts/lib/run-isolation.ts`는 저장소
일부를 임시 루트에 복사해 빌드하는 두 목록(초기 JS 측정
`scripts/scale/measure-feature-shape-growth.ts`,
`tests/unit/product-feature-selection-build.test.ts`)에도 더한다.
Storybook은 run ID가 있으면 `build:storybook`의 두 단계를 다른 `-o`로
돌린다. 같은 단계인지는 단위 테스트가 package script와 대조한다.
버린 대안: 운영 증거 경로(54개 스크립트, `config/ci/gates.json`)까지 run
ID로 옮기기. 게이트와 workflow upload가 고정 경로를 읽으므로 이번 범위
밖이다(아래 한계). 실행마다 저장소를 임시 폴더로 복사하기: 매번 수십
MB 복사와 node_modules 링크가 들고, 개발자가 보는 결과 위치가 사라진다.
위험·한계: 포트는 아직 고정이라 이 커밋만으로는 두 실행이 함께 돌지
못한다(다음 커밋). `ci:gate`와 증거 검사(`check:test-evidence`,
`verify:browser-capability-evidence`)는 기본 경로만 읽는다. run ID를 준
채 게이트를 돌리면 증거가 기본 경로에 새로 생기지 않아 게이트가 "not
freshly produced"로 실패한다(통과로 새지 않는다). Storybook 내부 캐시
(`node_modules/.cache`)는 여전히 공유한다. capability 설정
(`playwright.capabilities.config.ts`)은 test:browser-file-storage-removal이
capability와 함께 지운다. 설정 테스트는 그 파일이 있을 때만 그 설정을 읽고,
경로를 변수로 두어 타입 검사가 그 파일을 요구하지 않는다.
검증: desktop batch에서 확인.
node scripts/run-vitest.ts run tests/unit/run-isolation.test.ts tests/unit/playwright-run-isolation.test.ts tests/unit/run-output-root-build.test.ts
node scripts/run-vitest.ts run tests/unit/feature-shape-growth.test.ts tests/unit/product-feature-selection-build.test.ts
corepack pnpm test:browser-file-storage-removal
Refs: 1105/12-04
ENV_REGISTRY called VITE_BUILD_ID required with no default while every reader
fell back to local-build. Staging and production builds now fail without it;
local builds keep the placeholder, and the registry records that default.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
README said runtime settings live in public/config.json, which the build
replaces with config/runtime/<profile>.json, and named none of APP_PROFILE,
RUNTIME_*, VITE_BUILD_ID, RELEASE_ID or RELEASE_TARGET. A unit test now
requires README to name every variable the generator and admission read.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A /app/ build still fetched /config.json and /release-manifest.json and failed
at boot. The config URL default and a root-relative profile manifest URL now
follow VITE_ROUTER_BASE_PATH, and an explicit config URL outside the base fails
the build.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The provider sandbox never ran. bubblewrap 0.9.0 stops parsing an `--args`
file at the first non-option and never hands the remainder back, so the
command written into that file was silently dropped: bwrap printed its usage
text, exited 1, and the provider produced no evidence at all. The options
still travel in the args file — that is what keeps host paths and credentials
out of `/proc/<pid>/cmdline` — but the command now rides on real argv, and
`encodeProviderBwrapInput` refuses a `--` so the drop cannot come back.
The scope wrapper then could not exit. It read the supervisor's liveness pipe
through `fs`, which runs a blocking `read(2)` on a threadpool thread; the
supervisor holds that pipe open for the scope's whole life, so the read never
returned and closing the descriptor did not interrupt it. Once bubblewrap
finished the wrapper deadlocked in `process.exit`, the scope outlived the
provider, and a completed run was reported as a timeout kill. The channel is
now read through the event loop, so teardown is observable and terminal.
Creation modes were left to the ambient umask. `mkdir(mode)` and `open(mode)`
are requests the kernel subtracts the umask from, so a runner exporting a
restrictive umask produced directories it could not enter and handed `tar` a
file it could not re-open. Private modes are pinned instead of inherited.
Promotion cleanup deleted before it checked. Removals run through a pinned
descriptor, so a leaf substituted after validation had this promotion's exact
five destroyed first and the substitution reported afterwards, leaving a
half-emptied directory a retry could not tell from a completed one. The name
is re-bound to the inode before anything is removed, so the failure is total.
Separately, release coherence proved the artifacts agreed with each other but
never that they belonged where they were going: a build whose runtime document
said `APP_ENV: local`, `AUTH_MODE: demo` and a loopback API is coherent with
itself and passed every gate. `public/` is copied verbatim into `dist/`, so
that local document shipped with every build regardless of what the build was
for. Runtime configuration now comes from a declared profile, and FE-GATE-027
refuses to admit an artifact to an environment it does not match — including
refusing an undeclared destination, so nothing is admitted by omission.
`REQUEST_TIMEOUT_MS` and `VITE_ROUTER_BASE_PATH` were validated and then
dropped: the V3 executor ran every operation on its contract's own deadline,
and Vite emitted root-absolute assets for a sub-path deployment. The timeout is
now a ceiling that may tighten a contract but never loosen one, and one base
path feeds the router, the Service Worker scope and the asset base together.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>