Feat default surface routing + fix startup hang (#4)
Some checks failed
CI/CD / detect-changes (pull_request) Successful in 4s
CI/CD / test-frontend (pull_request) Successful in 4s
CI/CD / test-go-pg (pull_request) Failing after 2m18s
CI/CD / test-sqlite (pull_request) Successful in 2m40s
CI/CD / build-and-deploy (pull_request) Has been skipped

Default surface routing (v0.2.1):
- `/` redirects to configurable default surface with fallback chain:
  configured → first enabled extension surface → /admin
- Auto-sets default_surface on first extension surface install
- Admin settings UI: Default Surface dropdown
- disabledRedirect() and extension surface disabled redirect now go to
  /admin instead of / to prevent redirect loops

Fix Gin route param conflict causing backend startup hang:
- Team package settings routes used `:pkgId` while sibling routes used
  `:id` — Gin's radix tree entered infinite loop on conflicting param
  names. Unified to `:id`.

Docker entrypoint hardening:
- Health check timeout 10s → 60s
- Stale backend process cleanup on restart
- Crash detection (exit early if backend dies)

Roadmap direction updates:
- v0.3.0: Notes surface (Obsidian-style) replaces Editor surface
- v0.4.0: MVP (was v0.5.0) — chat moved to post-MVP
- No built-in auto-install — explicit install only

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
2026-03-26 19:38:02 +00:00
parent a36be83aaf
commit 70678cb352
9 changed files with 169 additions and 40 deletions

View File

@@ -22,19 +22,29 @@ fi
echo "🔀 Starting Switchboard Core backend..."
# Kill any stale backend from a previous entrypoint run
pkill -f /usr/local/bin/switchboard 2>/dev/null || true
sleep 0.2
# Launch Go backend in background (from /app so migrations are found)
cd /app
/usr/local/bin/switchboard &
BACKEND_PID=$!
# Wait for backend to be ready (max 10s)
for i in $(seq 1 20); do
# Wait for backend to be ready (max 60s)
for i in $(seq 1 120); do
if wget -q --spider http://localhost:${PORT:-8080}${BASE_PATH}/health 2>/dev/null; then
echo "✅ Backend ready (PID ${BACKEND_PID})"
exit 0
fi
# Check backend didn't crash
if ! kill -0 ${BACKEND_PID} 2>/dev/null; then
echo "❌ Backend process died"
exit 1
fi
sleep 0.5
done
echo "❌ Backend failed to start within 10s"
echo "❌ Backend failed to start within 60s"
kill ${BACKEND_PID} 2>/dev/null || true
exit 1