stderr-lib

Type-safe, standardized error handling for TypeScript/JavaScript

Clean, opinionated error handling with explicit Result pattern and structured error normalization.

---

Overview

stderr-lib gives you two focused tools:

1. stderr() – normalize any error-like value into a standard StdError
2. tryCatch() – wrap sync/async code in a type-safe Result union

Use it when you want:

• Consistent, serializable errors (including cause chains and nested errors)
• Explicit, type-checked error handling instead of ad-hoc try/catch
• Built‑in defenses against runaway structures (max depth, properties, array length)

Advanced patterns (timeouts, limit tuning, security, performance, testing) live in docs/BestPractices.md.

---

Installation

npm install stderr-lib

pnpm add stderr-lib

yarn add stderr-lib

---

Quick Start

Normalize Any Error for Logging

import { stderr } from 'stderr-lib';

try {
    await riskyOperation();
} catch (error: unknown) {
    const err = stderr(error);

    console.log(err.toString());
    // Includes message, stack (if present), cause chain, custom properties, everything!

    logger.error('Operation failed', err); // Works with typical loggers
}

Type-Safe Error Handling with Result Pattern

import { tryCatch, type Result } from 'stderr-lib';

interface UserDto {
    id: string;
    name: string;
}

// You can pass an async function - type is inferred as Promise<Result<UserDto>>
const result = await tryCatch<UserDto>(async () => {
    const response = await fetch('/api/user/123');
    if (!response.ok) {
        throw new Error(`Request failed - ${response.status}`); // will be converted to StdError
    }
    return response.json() as Promise<UserDto>;
});

if (!result.ok) {
    // You are forced to handle the error explicitly
    console.error('Request failed:', result.error.toString());
    return null;
}

// In the success branch, value is non-null and correctly typed as UserDto
console.log('User name:', result.value.name);

---

Core API (Surface)

stderr(input, options?)

Normalize any value into a StdError instance.

stderr(input: unknown, options?: NormalizeOptions): StdError;

Key points:

• Accepts anything: native Error, strings, objects, null/undefined, arrays, third‑party errors
• Preserves:
  • name, message, stack (if present)
  • cause, errors (AggregateError / validation errors)
  • Custom properties (e.g., code, statusCode, metadata)
• Adds safe defaults to avoid DoS:
  • maxDepth (default 8, range 1–1000)
  • maxProperties (default 1000, range 1–100000)
  • maxArrayLength (default 10000, range 1–1000000)

See Best Practices – Normalization & Limits for detailed guidance and examples.

StdError

All stderr() results are instances of StdError.

Important properties (non‑exhaustive):

• name?: string
• message?: string
• stack?: string
• cause?: unknown
• errors?: unknown (arrays or maps of nested errors)
• [key: string]: unknown – arbitrary extra metadata

Important methods:

• toString(): string – human‑readable, multi‑line representation with cause chain and nested errors
• toJSON(): object – JSON‑safe representation suitable for logging systems and transports

Deep dives:

• Best Practices – Error Immutability
• Custom Error Classes vs StdError

tryCatch(fn, mapError?)

Wrap a function, async function, or Promise and always get a Result instead of thrown exceptions.

// Sync function -> Result
tryCatch<T>(fn: () => T): Result<T, StdError>;

// Async function or Promise -> Promise<Result>
tryCatch<T>(fn: Promise<T> | (() => Promise<T>)): Promise<Result<T, StdError>>;

// Sync function with custom error -> Result
tryCatch<T, E>(fn: () => T, mapError: (err: StdError) => E): Result<T, E>;

// Async function or Promise with custom error -> Promise<Result>
tryCatch<T, E>(fn: Promise<T> | (() => Promise<T>), mapError: (err: StdError) => E): Promise<Result<T, E>>;

Key features:

• Accepts Promises directly: Pass fetch(url) or () => fetch(url) – both work
• Automatic sync/async detection: Returns Result<T> for sync, Promise<Result<T>> for async
• Optional error transformation: Use mapError to convert StdError to your custom error type

Result<T, E> is:

type Result<T, E = StdError> = { ok: true; value: T; error: null } | { ok: false; value: null; error: E };

This forces you to handle both branches explicitly in TypeScript.

More patterns:

• Promise Handling
• Common Patterns – API Error Handling
• Exception Usage – When to Use stderr vs tryCatch

---

Global Configuration (Limits)

You can tune normalization limits globally via properties on stderr:

import { stderr } from 'stderr-lib';

stderr.maxDepth = 10; // Default: 8, range: 1–1000
stderr.maxProperties = 500; // Default: 1000, range: 1–100000
stderr.maxArrayLength = 5000; // Default: 10000, range: 1–1000000

Invalid values throw TypeError/RangeError instead of silently misconfiguring:

stderr.maxDepth = 0; // RangeError
stderr.maxDepth = 3.5; // TypeError
stderr.maxProperties = 200000; // RangeError
stderr.maxArrayLength = 2000000; // RangeError

For environment‑specific tuning, see Best Practices – Security & DoS Protection and Performance Tips.

---

When to Use stderr-lib

A Good Fit

• Web apps (frontend and backend)
• Node.js services and APIs
• Business, financial, or e‑commerce systems
• Logging / monitoring / error‑reporting services
• Codebases that prefer explicit Result‑style error handling

Not a Good Fit

• Hard real‑time or safety‑critical systems (avionics, life‑critical medical devices, automotive safety, etc.)
• Environments that require certified runtimes or deterministic timing

(These environments generally should not use JavaScript/TypeScript at all.)

---

Further Reading

• BestPractices.md – deep dive:
  • Promise handling patterns and timeouts
  • Exception usage, mutation, and error immutability
  • Security/sanitization and DoS protections
  • Performance tuning and environment‑specific limits
  • Testing strategies and property‑based tests
  • Custom error classes vs StdError, common patterns, and anti‑patterns
• ADR.md – architectural decisions
• TypeScript Coding Standard – broader Result‑pattern guidance

---

Contributing

Contributions are welcome. Please:

1. Follow existing code style and patterns
2. Add tests for new behavior
3. Update documentation where relevant
4. Run linting and tests before submitting (see package.json scripts)

---

License

MIT © Michael L. Hobbs