Release Process¶
This page documents the CI/CD pipeline: what each workflow does, how releases are built, and how to run the process manually.
CI workflows overview¶
The project uses 4 GitHub Actions workflows. All use only official GitHub
actions (actions/checkout, actions/setup-python, actions/upload-pages-artifact,
actions/deploy-pages). No third-party actions.
Budget target: ~175 minutes/month on the GitHub free tier.
| Workflow | File | Trigger |
|---|---|---|
| Build & Release | build.yml |
Manual dispatch only |
| Deploy Site | deploy-site.yml |
Push to main (platforms, emulators, provenance, wiki, scripts, database.json, mkdocs.yml), manual |
| PR Validation | validate.yml |
PR touching bios/**, platforms/** or emulators/** |
| Weekly Sync | watch.yml |
Cron Monday 06:00 UTC, manual dispatch |
build.yml - Build & Release¶
Releasing is deliberate. Pushing never cuts one: somebody decides the collection is worth publishing and dispatches the workflow.
Trigger. workflow_dispatch only, with an optional force_release flag to
bypass the rate limit.
Concurrency. Group build, queued rather than cancelled: a run that is
already uploading assets must finish.
Steps:
- Checkout, Python 3.12, install
pyyaml - Run
test_e2e - Rate limit check: skip if last release was less than 7 days ago (unless
force_releaseis set) - Restore large files from the
large-filesrelease into.cache/large/ - Refresh data directories (
refresh_data_dirs.py) - Build packs (
generate_pack.py --all --output-dir dist/) - Split any pack over 2 GB into
.zip.001,.zip.002, ... volumes. GitHub caps release assets at 2 GB, and the.001convention is what 7-Zip and PeaZip open directly (they reject.partNNnames as corrupt) - Create GitHub release with tag
v{YYYY.MM.DD}(appends.Nsuffix if a same-day release already exists) - Clean up old releases, keeping the 3 most recent plus
large-files
Release notes include file count, total size, per-pack sizes, the extract
path per platform, and the last 15 non-merge commits touching bios/ or
platforms/.
Pack variants. The workflow builds full packs only. Releases that also
carry *_Platform_BIOS_Pack.zip had a second --source platform run added by
hand, as in the manual process below. See
advanced usage.
deploy-site.yml - Deploy Documentation Site¶
Trigger. Push to main when any of these paths change: platforms/,
emulators/, provenance/, wiki/, scripts/generate_site.py,
scripts/generate_readme.py, scripts/verify.py, scripts/common.py,
database.json, mkdocs.yml. Also manual dispatch.
The list is the set of inputs the site is generated from. Adding a new input to
generate_site.py means adding its path here, or the site silently goes stale.
Steps:
- Checkout, Python 3.12
- Install
pyyaml,mkdocs-material>=9.7.5,<10,pymdown-extensions>=10.14 - Restore large files from the
large-filesrelease, refresh data directories - Run
generate_site.py(converts YAML data into MkDocs pages and rewritesmkdocs.yml) - Run
generate_readme.py(rebuilds README.md and CONTRIBUTING.md) mkdocs build --strictto produce the static site- Run
validate_site.pyon the rendered HTML (metadata, headings, image alternatives, duplicate ids, local links and fragments) - Require the committed README and CONTRIBUTING to match what the generator
just produced.
write_if_changed()compares content with the timestamp line stripped, so a run that only moves the clock leaves the files untouched and the check stays meaningful - Upload artifact, deploy to GitHub Pages
Data contracts are validated with scripts/validate_schemas.py before the site
is generated: database.json, the install and target manifests, the site API
envelopes and the stats file, plus the semantic invariants those schemas cannot
express (declared totals matching their lists, no destination both installed
and omitted).
The site is deployed via the github-pages environment using the official
actions/deploy-pages action. Pages deployments are queued rather than
cancelled (cancel-in-progress: false): cancelling one mid-flight leaves the
deployment stuck and the next runs time out waiting on it.
--strict turns MkDocs warnings into failures, so a broken internal link or a
dangling anchor fails the build instead of shipping. The validation: block in
mkdocs.yml is what promotes unrecognized links and missing anchors to
warnings in the first place.
The theme version is pinned on both sides: >=9.7.5 because that is the
release which caps mkdocs < 2 (MkDocs 2.0 ships without a license), <10
so a major theme release cannot change the site without a deliberate bump.
validate.yml - PR Validation¶
Trigger. Pull requests that modify bios/**, platforms/**,
emulators/**, schemas/**, scripts/**, tests/** or install.py.
Concurrency. Per-PR group, cancel in-progress.
Four parallel jobs:
validate-bios. Diffs the PR to find changed BIOS files, runs
validate_pr.py --markdown on each, and posts the validation report as a PR
comment (hash verification, database match status).
validate-configs. Runs python scripts/validate_schemas.py --source-only,
which validates every platform YAML against schemas/platform.schema.json and
every emulator profile against schemas/emulator.schema.json. Both schemas set
additionalProperties: false, so a typo in a field name fails the job instead
of being silently ignored.
run-tests. Runs python -m unittest discover tests -v. Must pass before
merge.
label-pr. Auto-labels the PR based on changed paths:
| Path pattern | Label |
|---|---|
bios/ |
bios |
bios/{Manufacturer}/ |
system:{manufacturer} |
platforms/ |
platform-config |
scripts/ |
automation |
watch.yml - Weekly Platform Sync¶
Trigger. Cron schedule every Monday at 06:00 UTC, or manual dispatch.
Flow:
- Scrape live upstream sources (System.dat, batocera-systems, es_bios.xml, etc.) and regenerate platform YAML configs
- Auto-fetch missing BIOS files
- Refresh data directories
- Run dedup
- Regenerate
database.json - Create or update a PR with labels
automatedandplatform-update
The PR contains all changes from the scrape cycle. A maintainer reviews and merges.
Large files management¶
Files larger than 50 MB are stored as assets on a permanent GitHub release
named large-files (to keep the git repository lightweight).
Examples: PS3UPDAT.PUP, PSVUPDAT.PUP, PSP2UPDAT.PUP, the DSi NAND images,
maclc3.zip, Firmware.19.0.0.zip (Switch), the QEMU EDK2 firmware, the ScummVM
data bundle, the EasyRPG soundfont, the Dolphin/Ishiiruka SD card images, and
the arcade sets over 100 MB. .gitignore is the authoritative list: every
bios/ path listed there is a release asset.
Storage. Listed in .gitignore so they stay out of git history. The
large-files release is excluded from cleanup (the build workflow only
deletes version-tagged releases).
Build-time restore. The build workflow downloads all assets from
large-files into .cache/large/ and copies them to their expected paths
before pack generation.
Upload. To add or update a large file:
Local cache. generate_pack.py calls fetch_large_file() which downloads
from the release and caches in .cache/large/ for subsequent runs.
Manual release process¶
When build.yml is disabled, build and release manually:
# Run the full pipeline (DB + verify + packs + manifests + integrity + docs)
python scripts/pipeline.py
# Or step by step:
python scripts/generate_db.py --force --bios-dir bios --output database.json
python scripts/verify.py --all
python scripts/generate_pack.py --all --output-dir dist/ # full packs
python scripts/generate_pack.py --all --source platform --output-dir dist/ # platform packs
python scripts/generate_pack.py --all --verify-packs --output-dir dist/
# Split anything over 2 GB (GitHub asset cap)
for f in dist/*.zip; do
[ "$(stat -c%s "$f")" -gt 2000000000 ] || continue
split --bytes=1900M --numeric-suffixes=1 --suffix-length=3 "$f" "$f." && rm "$f"
done
# Create the release
DATE=$(date +%Y.%m.%d)
gh release create "v${DATE}" dist/*.zip* \
--title "BIOS Pack v${DATE}" \
--notes "Release notes here" \
--latest
The two generate_pack.py runs are what puts both *_BIOS_Pack.zip and
*_Platform_BIOS_Pack.zip on the release. --all-variants builds all six
combinations instead, which is more than a release needs.
Run the pipeline online for a release: --offline skips the data directory
refresh and the MAME/FBNeo hash refresh, so the packs would ship stale data
directories.
The workflow carries no push trigger, so there is nothing to disable and no
guard to remove. Cutting a release means dispatching build.yml from the
Actions tab, or:
gh workflow run build.yml
gh workflow run build.yml -f force_release=true # within 7 days of the last
The rate limit refuses a second release inside seven days unless
force_release is set, which is what keeps a stray dispatch from publishing
twice in a day.