improve database connection handling, add docker-entrypoint.sh

improve database connection handling

improve database connection handling

minor update to index.ts

Fix GHCR publishing - use GITHUB_TOKEN v0.1.1

• workflow: switch from GH_TOKEN secret to GITHUB_TOKEN (auto-provided)
• The GITHUB_TOKEN is automatically available with packages:write scope

Add GitHub Actions release workflow and cleanup v0.5.1

• Add release.yml workflow for automated Docker builds on tag push
• Publishes to Docker Hub (tananetwork/deploy) and GHCR
• Remove test project artifacts from examples/ that were accidentally committed

PostgreSQL persistence and Docker packaging v0.5.0

-(db): add PostgreSQL persistence layer with Drizzle ORM schema and connection management -(docker): create production-ready Dockerfile with multi-stage build, Firecracker+kernel baked in -(docker): add .dockerignore to exclude node_modules, build artifacts, and ext4 images -(executor): implement DockerExecutor for local development fallback (uses node:22-alpine) -(jobs): integrate database layer for job persistence across restarts -(server): add database health checks and executor type reporting to /health endpoint

Docker image size: ~195MB (without rootfs), ~260MB with compressed rootfs baked in. Builds cross-platform (x86_64 and ARM64), runs on Linux with KVM only.

T4 integration for build artifact staging v0.4.0

-(t4-client): Add complete T4 client for two-stage upload system

• Upload files to /_pending/{deploymentId}/{path}
• Verify staged files with POST /_verify
• List pending deployments and files
• Health check and deployment cleanup

-(executor-firecracker): Extract build output files from VM via base64

• New extractFiles() method reads output directory from VM
• Files transferred as base64 through vsock for binary safety
• Skips node_modules, .map files, and hidden files
• Limited to 500 files per deployment

-(jobs): Auto-upload to T4 after successful build

• Uploads files to staging when address is provided
• Stores deploymentId, fileManifest, totalSize in job record
• Comprehensive error handling with log output

-(types): Extended BuildJob and ExecutionResult

• Added FileManifestEntry interface
• Added address, deploymentId, fileManifest, totalSize fields
• Added files?: Map<string, Buffer> to ExecutionResult

-(executor-docker): Added Docker-based executor as alternative

• Useful for development on macOS (no Firecracker/KVM)
• Automatic executor selection based on environment

-(scripts): Build scripts for rootfs variants

• build-rootfs.sh: Full Alpine+Bun image (700MB, 1GB RAM)
• build-rootfs-slim.sh: Slim Node-only image (200MB, 256MB RAM)

Flow: Build → Extract → Upload to /_pending → Verify → Blockchain confirms → Promoted to /content

add two-stage upload with _pending staging and verification

• storage: add _pending namespace for staging uploads before blockchain confirmation
• api: add /_verify endpoint to check content hash integrity
• api: add /_move endpoint to promote verified content to permanent storage
• this enables atomic deployments - content only becomes permanent after the blockchain transaction is confirmed

add website deployment support and fix block producer

• schema: add contract_assets table for tracking deployed website files
• block-producer: fix FK constraint by inserting contract before assets
• block-producer: fix transaction UPSERT to use metadata JSONB for timestamp/nonce instead of non-existent columns
• contracts: support website deployment transaction type with T4 verification
• api: add T4 content verification and promotion during block production
• migrations: add 0016_website_support.sql and 0017_contract_assets.sql