Files
Vijay Janapa Reddi e5d00632cf fix(ci): gate sibling preview deploys on validate workflows passing
Rolls the same gating pattern from staffml out to kits, slides,
instructors, site, and mlsysim, all of which had the same race —
preview-dev and validate-dev triggered independently on push to dev with
no `needs:` between them, so a deploy could ship even when validate
failed on the same SHA. (labs and tinytorch already gate via
`workflow_run` and were left alone.)

For each pair: validate-dev gains a workflow_call trigger and rekeys
concurrency to `head_ref || run_id`. preview-dev adds a `validate` job
that calls validate-dev via `uses:` and the existing build-and-deploy
job grows a `needs: [validate, ...]`. Build doesn't start until validate
finishes, so wall-clock per push regresses by roughly the validate
duration; can be upgraded to the staffml-style parallel-build split per
component if that becomes painful.
2026-05-01 13:25:27 -04:00

170 lines
5.7 KiB
YAML

name: '📦 Kits · 👁️ Preview (Dev)'
# =============================================================================
# Hardware Kits — Dev Preview Deploy
# =============================================================================
#
# Builds the Hardware Kits Quarto site with PDF and deploys to the dev
# preview site via SSH.
#
# Flow:
# 1. BUILD_PDF — Compile Hardware Kits PDF (reusable workflow)
# 2. BUILD_SITE — Quarto render + inject PDF into site
# 3. DEPLOY — Push to dev preview repo via SSH
#
# Triggers:
# - push: dev branch, kits/** paths
# - workflow_dispatch: manual
#
# Deploys to: harvard-edge.github.io/{DEV_REPO}/{DEV_KITS_PATH}/
# Secrets: SSH_DEPLOY_KEY
# Vars: KITS_ROOT, BOOK_TOOLS, DEV_REPO, DEV_REPO_URL, DEV_KITS_PATH
#
# Related:
# - kits-validate-dev.yml — Content & build validation
# - kits-publish-live.yml — Production deploy to mlsysbook.ai/kits/
# - kits-build-pdfs.yml — Reusable PDF build
#
# =============================================================================
on:
workflow_dispatch:
push:
branches: [dev]
paths:
- 'kits/**'
permissions:
contents: read
actions: read
concurrency:
group: kits-dev-deploy
cancel-in-progress: true
jobs:
# Run the full validate-dev workflow inline. Deploy depends on it so
# Preview can never deploy on a SHA where validate failed (the bug
# this gating closes is the screenshot pattern of "Preview ✅ but
# Validate ❌ on the same SHA"). Validate runs in parallel with
# build-pdf, so wall-clock impact is small.
validate:
name: '✅ Validate (Dev)'
uses: ./.github/workflows/kits-validate-dev.yml
# Build PDF first
build-pdf:
name: '📚 Build PDF'
uses: ./.github/workflows/kits-build-pdfs.yml
with:
ref: ${{ github.ref }}
build-and-deploy:
name: '📦 Build & Deploy Kits (Dev)'
runs-on: ubuntu-latest
needs: [validate, build-pdf]
steps:
- name: 📥 Checkout
uses: actions/checkout@v6
- name: 🔧 Setup Quarto
uses: quarto-dev/quarto-actions/setup@v2
- name: 🔨 Build Kits Site
working-directory: ${{ vars.KITS_ROOT }}
run: |
quarto render
touch _build/.nojekyll
- name: 📥 Download PDF Artifact
uses: actions/download-artifact@v8
continue-on-error: true
with:
name: Kits-PDF
path: ./pdf-artifacts/
- name: 📁 Inject PDF into site
run: |
echo "📁 Injecting PDF into built site..."
mkdir -p ${{ vars.KITS_ROOT }}/_build/assets/downloads
if [ -f ./pdf-artifacts/Hardware-Kits.pdf ]; then
cp ./pdf-artifacts/Hardware-Kits.pdf ${{ vars.KITS_ROOT }}/_build/assets/downloads/
echo "✅ Injected Hardware-Kits.pdf"
else
echo "⚠️ PDF not found"
fi
echo ""
echo "📦 Downloads folder contents:"
ls -la ${{ vars.KITS_ROOT }}/_build/assets/downloads/ || echo "No downloads folder"
- name: 🔗 Rewrite URLs for dev site
run: bash .github/scripts/rewrite-dev-urls.sh kits ${{ vars.KITS_ROOT }}/_build
- name: 🔧 Modify announcement for dev preview
run: |
echo "🔧 Modifying announcement banner for development preview..."
COMMIT_SHORT="${{ github.sha }}"
COMMIT_SHORT="${COMMIT_SHORT:0:8}"
python3 ${{ vars.BOOK_TOOLS }}/scripts/publish/modify_dev_announcement.py \
${{ vars.KITS_ROOT }}/_build \
--verbose \
--commit-hash "${{ github.sha }}" \
--commit-short "$COMMIT_SHORT"
- name: 🚀 Deploy to Dev Site via SSH
env:
SSH_DEPLOY_KEY: ${{ secrets.SSH_DEPLOY_KEY }}
run: |
echo "🔐 Starting ssh-agent..."
eval "$(ssh-agent -s)"
echo "$SSH_DEPLOY_KEY" | tr -d '\r' | ssh-add - > /dev/null
# Add github.com to known hosts
mkdir -p ~/.ssh
ssh-keyscan github.com >> ~/.ssh/known_hosts
echo "🔧 Configuring git..."
git config --global user.email "actions@github.com"
git config --global user.name "GitHub Actions"
echo "🔄 Cloning dev preview repository (${{ vars.DEV_REPO }})..."
git clone --depth=1 ${{ vars.DEV_REPO_URL }} target-repo
cd target-repo
echo "🧹 Cleaning ${{ vars.DEV_KITS_PATH }} directory..."
rm -rf ${{ vars.DEV_KITS_PATH }}
mkdir -p ${{ vars.DEV_KITS_PATH }}
echo "🚚 Copying Kits site content..."
cp -r "${{ github.workspace }}/${{ vars.KITS_ROOT }}/_build/." ${{ vars.DEV_KITS_PATH }}/
echo "🔍 Validating deployment content..."
if [ ! -f "${{ vars.DEV_KITS_PATH }}/index.html" ]; then
echo "❌ CRITICAL: ${{ vars.DEV_KITS_PATH }}/index.html is missing. Aborting deployment."
exit 1
fi
echo "📦 Contents of ${{ vars.DEV_KITS_PATH }}/:"
ls -la ${{ vars.DEV_KITS_PATH }}/ | head -10
echo "📦 Committing and pushing changes..."
git add .
git commit -m "📦 Deploy Kits dev to /${{ vars.DEV_KITS_PATH }}/ from ${{ github.sha }}" --allow-empty || echo "🟡 Nothing to commit"
# Retry push with rebase to handle concurrent deploys
for i in 1 2 3; do
if git push origin main 2>/dev/null; then
echo "✅ Push succeeded on attempt $i"
break
fi
echo "⚠️ Push failed (attempt $i/3), pulling with rebase..."
git pull --rebase origin main || true
done
echo "✅ Kits deployed to: https://harvard-edge.github.io/${{ vars.DEV_REPO }}/${{ vars.DEV_KITS_PATH }}/"