LogTape 1.2.0: Nested property access and context isolation

[dahlia]

LogTape is a logging library for JavaScript and TypeScript that works across Deno, Node.js, Bun, and browsers. It's built around structured logging with zero dependencies and flexible configuration.

We just released version 1.2.0 with several features that make working with structured data even more powerful. Here's what's new.

Nested property access in templates

You can now access nested properties directly in your message templates without having to flatten your data first. This was one of the most requested features, and it makes logging structured data much cleaner.

logger.info("User {user.name} ordered {order.items[0].product}", {
  user: { name: "Alice" },
  order: { items: [{ product: "Coffee" }] }
});

The syntax supports everything you'd expect: dot notation for nested objects, array indexing, bracket notation with quotes for property names that contain special characters, and optional chaining when you're not sure if a property exists. We also built in protection against prototype pollution by blocking access to __proto__, prototype, and constructor at any depth.

For more details, check out the structured logging documentation.

Context-based isolation for fingers crossed sink

If you're using the fingers crossed sink (which buffers logs and only outputs them when something goes wrong), you can now isolate buffers by context. This is particularly useful for HTTP request tracing where you want each request's logs to be independent.

import { fingersCrossed } from "@logtape/logtape";

const sink = fingersCrossed(consoleSink, {
  isolateByContext: ["requestId"],
  bufferTtlMs: 60000,  // Clean up old buffers after 1 minute
  maxContexts: 1000    // Limit memory usage
});

Each isolated context maintains its own trigger state and buffer, so logs from one request won't accidentally flush logs from another. The new memory management options (TTL, cleanup intervals, and LRU limits) ensure this won't leak memory in high-traffic applications.

Learn more about context isolation in the documentation.

Better custom formatting support

The text formatter's value callback now receives the default cross-runtime inspect() function as a second parameter. This makes it much easier to customize formatting for specific types while falling back to sensible defaults for everything else.

import { getTextFormatter } from "@logtape/logtape";

const formatter = getTextFormatter({
  value: (val, inspect) => {
    if (val instanceof Date) return val.toISOString();
    return inspect(val);  // Use default for everything else
  }
});

You no longer need to reimplement the complex runtime detection logic yourself. See the TextFormatterOptions.value documentation for more details.

Additional improvements

The @logtape/pretty package now supports getters and showProxy options in inspectOptions, giving you fine-grained control over how objects and proxies appear in your logs. Check out the PrettyFormatterOptions.inspectOptions documentation to learn more.

The @logtape/redaction package now recursively redacts sensitive fields in objects nested within arrays, so your data stays protected even in complex structures.

Upgrading

If you're already using LogTape, upgrading to 1.2.0 is straightforward. The changes are backward compatible, so your existing code will continue to work. The only breaking change is for the TextFormatterOptions.value callback, but the second parameter is optional, so even that won't break existing code.

Just update @logtape/logtape and any other @logtape/* packages you're using to version 1.2.0. All the satellite packages (@logtape/file, @logtape/otel, @logtape/pretty, @logtape/redaction, @logtape/sentry, etc.) share the same version number with the core package, so keeping them in sync is straightforward.