android.yml's debug upload and release.yml's minstrel-apk pair went through the bvandeusen fork mirrors, with comments saying stock actions refuse this hostname, that the pair had to be matched on the bundled @actions/artifact major, and that download v7 was off-limits for node24. None of that holds on gitea/runner 3.x: the runner edits the GHES refusal out of the action bundles, every download major v4-v8 reads every upload major v4-v7 (Scribe spike #3843, CI-runner run 6312), and every CI image carries Node 24. The mirror pair itself was last verified at tag run 6286. Same artifact names, paths and if-no-files-found. ci-requirements.md drops the pairing table and keeps what is still true: @v3 is invisible. Scribe snippet #2271, milestone 395. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DwoKYuw3qJmUUYsJeNherB
7.0 KiB
CI Requirements — Minstrel
Copy of the template. Family-wide policy:
ci-runners.mdin the FabledRulebook.
Runtime images
Minstrel's four workflows consume two CI images:
git.fabledsword.com/bvandeusen/ci-go:1.26
git.fabledsword.com/bvandeusen/ci-android:36
ci-go:1.26— Go server tests (.gitea/workflows/test-go.yml), web SPA tests (.gitea/workflows/test-web.yml), and the release container build (release.yml'simage-releasejob).ci-android:36— native Kotlin/Compose client: ktlint + detekt + unit tests + debug APK (.gitea/workflows/android.yml), and the signed release APK (release.yml'sandroid-releasejob).
ci-flutter is no longer consumed, and ci-flutter can now be retired.
The M8 rewrite replaced the Flutter client with the native Android app and
flutter.yml was removed; ci-android took its place. flutter_client/
itself was deleted on 2026-08-16, which was the condition CI-Runner was
waiting on before dropping the image — nothing in this repo needs a Flutter
toolchain any more.
Note this does not mean the flutter-ci runner label goes: the Android
jobs still schedule on it while pulling ci-android:36, per the label/image
split below. The label is a scheduling handle, not a toolchain assertion.
Image deps used
From ci-go:1.26
- Go (1.26 toolchain) —
go vet,go test -race,go build,go mod. - Node + npm —
npm ciandnpm test/npm run checkintest-web.yml. - golangci-lint — lint pass in
test-go.yml. - docker CLI — bridge-IP discovery of the per-job Postgres service container in
test-go.ymlintegration job (via the runner's shared/var/run/docker.sock). - docker buildx — release container build + push in
release.yml. - curl — release-asset polling / upload in
release.yml.
From ci-android:36
- JDK 25 — Gradle launcher + Android build. Requires Gradle 9.1.0+ in
android/gradle/wrapper; older Gradle rejects JDK 25 with an opaque"25.0.3"error. The workflows also setJAVA_TOOL_OPTIONS=--enable-native-access=ALL-UNNAMEDto silence Gradle's launcher-JVM restricted-method warning. - Android SDK + cmdline-tools + build-tools 36.0.0 — APK assembly + signing.
No NDK: the native client has no C/C++ sources (this is why it isn't on
ci-flutter). - ktlint + detekt —
./gradlew ktlintCheckand./gradlew detektinandroid.yml. Image pins trackandroid/gradle/libs.versions.tomlso local and CI checks agree. - git —
actions/checkout@v4baseline (and any shell git operations). - base64 + curl — keystore decode + release-asset upload in
release.yml'sandroid-releasejob.
Per-job tool installs
None.
Notes
-
Label/image split. Workflows keep
runs-on: go-ci/runs-on: flutter-cias the scheduling label per theci-runners.md"label = scheduling handle, image =container.image" pattern. The labels are intentional handles, not toolchain assertions — which is why the Android jobs still schedule onflutter-ciwhile pullingci-android:36. Switch them toandroid-ciif that runner label is ever registered; nothing breaks either way. -
Integration-job docker-socket dependency.
test-go.yml's integration job uses the runner's shared docker socket (/var/run/docker.sock) to bridge-IP-discover the per-job Postgres service container by name + network intersection — the dev compose'sminstrel-postgres-*containers are explicitly skipped as belt-and-suspenders. Depends onact_runner.valid_volumeswhitelisting the socket; if that ever stops auto-mounting, integration tests fail at thedocker inspectstep. -
Go toolchain pin.
go.modis ongo 1.25.0becausegolang.org/x/crypto v0.51.0declares 1.25 as its minimum.ci-go:1.26satisfies this with headroom. Futurex/cryptobumps that move the Go floor should be paired with an image-tag bump in this file + the workflows. -
In-app update channel —
needs:, not polling.release.yml'simage-releasejob declaresneeds: [android-release], so on tag pushes the signed APK is guaranteed present before the image build starts — no polling window, no race. (The old cross-workflow polling againstflutter.ymlis gone with that workflow.) On non-tagmainpushesandroid-releaseis skipped andimage-releaseinstead pulls the most recent release's APK and reconstructs its exactversionName, so:latestnever ships without an update channel. It degrades to an emptyclient/— never a wrong version — if no release, asset, or tag commit-count can be resolved. -
Cache server reachability.
test-web.ymldoes NOT usecache: 'npm'onactions/setup-node— the Gitea Actions cache server isn't reachable from this runner's container network andsetup-nodewas burning ~4m41s on ETIMEDOUT before failing open. With the migration toci-go:1.26,setup-nodeis removed entirely (Node is in the image). The cache concern reappears if a future change re-introduces a network-dependent action. -
Artifacts — stock
actions/upload-artifact@v7andactions/download-artifact@v8; never@v3.uses: actions/upload-artifact@v7 uses: actions/download-artifact@v8Stock works on this forge since the runner moved to gitea/runner 3.x, which edits the actions' client-side
isGhes()refusal out of their bundles. Proven on 2026-09-10 for upload v4–v7 and download v4–v8 (Scribe spike #3843). Until then this repo pinned SHA mirrors of the Forgejo project's forks, because upstream threw on the hostname before it opened a connection (Scribe 2255).@v3is still broken: it reports success, and Gitea serves artifacts back only through the v4 API (content_encoding = application/zip), so a v3 upload is stored but invisible to every retrieval path. That is how 72 unreachable artifacts accumulated on this repo (Scribe 2270).Pairing no longer needs managing. This entry used to pin upload v5 against download v6 so both bundled
@actions/artifact^4.0.0, warning that a mismatch acrossrelease.yml's producer/consumer pair would list empty. Tested, and not true on this instance: every download major v4–v8 read the artifacts of every upload major v4–v7, by name and by pattern (CI-runner run 6312). The only real protocol break is v3 → v4. node24 is no longer a concern either — every CI-runner image carries Node 24 and the runner runs actions with the image'snode.Upload steps set
if-no-files-found: errorrather than the defaultwarn, so an upload that matches nothing fails its own job instead of failing the consumer later.Retrieval:
GET /api/v1/repos/{owner}/{repo}/actions/runs/{run_id}/artifactsfor the id (global run id, not the repo-scoped run number), then…/actions/artifacts/{id}/zip. The workstation has nounzip— usepython3 -m zipfile -e. -
Friction asks. None pending. The two images cover everything Minstrel needs.