standard-typescript-best-practices

Standard: TypeScript Best Practices

This standard covers high-impact TypeScript patterns for runtime boundaries, error and async handling, resource lifecycle, configuration access, logging, retries/timeouts, concurrency safety, and test... :

• Implement retries only for idempotent operations with bounded attempts and backoff; avoid unbounded loops or retrying non-idempotent side effects.
• Keep tests isolated by controlling time, randomness, and global state; avoid tests that depend on real timeouts, Math.random, or shared module singletons.
• Manage resources with deterministic cleanup using try/finally or using; avoid leaving timers, event listeners, streams, or file handles open.
• Model config access with a typed loader that validates required keys and parses types; avoid reading process.env directly across the codebase.
• Prefer immutable data with readonly and avoid exporting mutable singletons; isolate shared state behind functions or classes with explicit update methods.
• Set explicit timeouts and cancellation for all network or long-running async operations using AbortSignal; avoid Promises that can hang indefinitely.
• Type async callbacks as returning Promise and handle rejections at the boundary; avoid floating Promises inside event handlers and timers.
• Use structured logging with stable keys and include error stack and correlation context; avoid concatenated strings and logging only error.message.
• Use typed errors or discriminated Result types for expected failures; avoid throwing raw strings or relying on message matching.
• Validate all untrusted inputs at module boundaries using runtime schemas and return typed values; avoid casting unknown data to domain types.

Full standard is available here for further request: TypeScript Best Practices