fix: resolve schedule data staleness in production builds

- Replace window.location.reload() with router.refresh() in ScheduleGenerator, MatchEditor, RecalculateEloButton for proper Next.js cache invalidation
- Add revalidatePath() call after schedule generation in POST handler
- Add ownerId to tournament creation in cucumber tests for proper permission checks
- Assign tournament_admin role via Prisma after user creation in cucumber tests
- Fix TypeScript type annotations in hooks.ts (tournament id map)
- Update page reload to use networkidle in common-steps.ts
- Clear .next/ cache before cucumber tests in justfile
- Add .turbo to clean target
- Add comprehensive debug logging to schedule API route
- Document findings in docs/TROUBLESHOOTING_SCHEDULE_GENERATION.md

justfile

@@ -85,17 +85,23 @@ test-acceptance-postgres:
 
 # Run Cucumber e2e tests with SQLite
 test-cucumber-sqlite:
+    @echo "Clearing Next.js cache..."
+    rm -rf .next/
     @echo "Running Cucumber e2e tests with SQLite..."
     DATABASE_PROVIDER=sqlite DATABASE_URL=file:./prisma/ci.db npm run test:acceptance:cucumber
 
 # Run Cucumber e2e tests with PostgreSQL (uses .env.development)
 test-cucumber-postgres:
+    @echo "Clearing Next.js cache..."
+    rm -rf .next/
     @echo "Running Cucumber e2e tests with PostgreSQL..."
     npm run test:acceptance:cucumber
 
 # Run Cucumber e2e tests with PostgreSQL against production build
 # This is more reliable than dev server (no HMR, faster API responses)
 test-cucumber-postgres-prod:
+    @echo "Clearing Next.js cache..."
+    rm -rf .next/
     @echo "Building application for production..."
     bun run build
     @echo "Starting production server in background..."
@@ -211,7 +217,7 @@ help:
 # Clean up project (remove node_modules, build artifacts)
 clean:
     @echo "Cleaning project..."
-    rm -rf node_modules .next dist
+    rm -rf node_modules .next dist .turbo
 
 # Generate Prisma client
 prisma-generate: