Add JOB_TIMEOUT constant to fail stuck jobs after 5 minutes, and
LOCK_EXPIRY to reclaim abandoned locks after 10 minutes. Introduce
force_unlock_all to recover orphaned jobs at startup. Fix retry limit
off-by-one error and update deduplication to include locked jobs.
Integrate ts-rs for Rust-to-TypeScript type generation, build course
search page with filters, pagination, and expandable detail rows,
and refactor theme toggle into a reactive store with view transition
animation.
Switch to Bun for 2-5x faster frontend builds, implement cargo-chef for
reliable Rust dependency caching, and add Biome for fast code
formatting.
Build system improvements:
- Replace pnpm with Bun for frontend package management
- Add cargo-chef to Dockerfile for better Rust build layer caching
- Update all commands to use bun instead of pnpm
Developer experience:
- Add comprehensive Justfile commands (format, format-check, db)
- Implement automated PostgreSQL Docker setup with random port
allocation
- Add stricter checks (--deny warnings on clippy, --all-features flag)
Code quality:
- Add Biome formatter for 10-100x faster TypeScript/JavaScript
formatting
- Add GitHub Actions CI/CD workflow for automated checks
- Update .dockerignore with comprehensive exclusions
- Format all code with cargo fmt (Rust) and Biome (TypeScript)
All changes maintain backward compatibility and can be tested
incrementally.
Add batch upsert functionality to reduce database round-trips from N to 1 when inserting courses. Create comprehensive database indexes for common query patterns including term/subject lookups, time-series metrics, and job scheduling. Remove redundant indexes and add monitoring guidance for BRIN index effectiveness.
Add retry tracking to scrape jobs with configurable max retries (default 5), implement
automatic database migrations on startup, and significantly reduce logging noise from
infrastructure layers. Enhanced tracing with structured spans for better debugging while
keeping output readable by suppressing verbose trace logs from rate limiters and session
management. Improved error handling with detailed retry context and proper session cookie
validation.
Implements graceful shutdown with broadcast channels and proper timeout handling
for scraper workers, scheduler, bot service, and status update tasks. Introduces
centralized shutdown utilities and improves service manager to handle parallel
shutdown with per-service timeouts instead of shared timeout budgets.
Key changes:
- Add utils module with shutdown helper functions
- Update ScraperService to return errors on shutdown failures
- Refactor scheduler with cancellable work tasks and 5s grace period
- Extract worker shutdown logic into helper methods for clarity
- Add broadcast channel shutdown support to BotService and status task
- Improve ServiceManager to shutdown services in parallel with individual timeouts
Replace task abortion with broadcast-based graceful shutdown for scheduler and workers. Implement cancellation tokens for in-progress work with 5s timeout. Add tokio-util dependency for CancellationToken support. Update ServiceManager to use completion channels and abort handles for better service lifecycle control.
