Mozilla PDF.js is a comprehensive, web-standards-based PDF rendering library and viewer built entirely in JavaScript. It parses and renders PDF documents in the browser using HTML5 Canvas and SVG, with a worker thread for heavy parsing to keep the UI responsive. The library supports PDF versions 1.0 through 2.0, including encryption, digital signatures, XFA forms, Acrobat JavaScript scripting, and a full annotation editing suite (free text, ink, highlight, stamp, signature). It also provides a complete web viewer application with sidebar, find bar, presentation mode, and printing, plus a Chrome extension and Firefox integration. The architecture is modular: a core worker thread handles parsing, font rendering, and image decoding (using WebAssembly for JPEG2000, JBIG2, Brotli, and ICC color management), while the display layer manages canvas rendering, text extraction, and annotation overlays. The viewer layer provides UI components with an event bus for decoupled communication. The codebase demonstrates exceptional craftsmanship with clean abstractions (factory patterns, base classes, lazy initialization), extensive unit and integration tests, and a sophisticated build system that targets generic web, legacy browsers, Chrome, and Firefox. Key reusable patterns include a PostScript-to-typed-array compiler with fallback interpreter, a TrueType glyf table parser/writer, an operator list optimizer using state machines, a generic message handler for worker communication, and image processing utilities for signature extraction. The project is actively maintained with frequent commits, low technical drift, and a large community.

Nunjucks is a full-featured JavaScript templating engine heavily inspired by Jinja2 (Python). It compiles templates into JavaScript functions for high-performance rendering, supporting template inheritance, macros, filters, tests, async control, and custom extensions. The library runs in both Node.js and browser environments, with a slim build for precompiled templates. It integrates with Express as a view engine and provides a CLI tool for precompiling templates offline. The architecture follows a classic compiler pipeline: lexer → parser → AST → compiler → JavaScript function. The public API exposes Environment, Template, Loader, and various configuration methods. Templates are loaded via loaders (FileSystemLoader, WebLoader, PrecompiledLoader) and rendered synchronously or asynchronously. The library includes a rich set of built-in filters (escape, default, sort, groupby, etc.) and tests (defined, odd, mapping, etc.), all extensible via addFilter/addTest. Key entry points include nunjucks/index.js (main module), bin/precompile (CLI), and browser/nunjucks.js (UMD build). External dependencies are minimal: a-sync-waterfall for async control flow, asap for microtask scheduling, and chokidar for file watching. The codebase is well-tested with Mocha and has CI via GitHub Actions. While the core engine is well-crafted and actively maintained, the repository carries technical debt from vendored documentation dependencies (Bootstrap 3, jQuery 2, Bower) and an outdated build toolchain (Babel beta, PhantomJS). The library itself is production-ready and widely used (8.9k stars).

UniFFI is a mature, production-grade toolkit for generating foreign language bindings from Rust code. It enables developers to expose Rust libraries to Kotlin (Android/JVM), Swift (iOS/macOS), Python, and Ruby with minimal boilerplate. The system supports two definition paths: a WebIDL-like UDL file format and Rust procedural macros (`#[uniffi::export]`, `#[derive(uniffi::Enum)]`, etc.), both of which produce metadata consumed by a multi-stage pipeline. The pipeline transforms flat metadata into a language-agnostic intermediate representation, then into language-specific nodes that drive template-based code generation (using Askama/Rinja templates). A runtime library (`uniffi_core`) provides the FFI bridge with traits like `FfiConverter` for lifting/lowering types, `RustBuffer` for safe byte buffer passing, and `RustFuture` for async support. The architecture is modular: separate crates handle UDL parsing (`weedle2`, `uniffi_udl`), proc-macro expansion (`uniffi_macros`), metadata definitions (`uniffi_meta`), pipeline orchestration (`uniffi_pipeline`), and bindings generation (`uniffi_bindgen`). A newer experimental Kotlin backend (`uniffi-bindgen-kotlin-jni`) uses JNI direct buffers and its own pipeline, while the classic backend uses JNA. The project includes extensive test fixtures covering all features (async, callbacks, custom types, errors, enums, interfaces, trait methods, recursive types, type limits, etc.) and benchmarks for performance across languages. External dependencies include `askama`/`rinja` for templates, `weedle2` for UDL parsing, `goblin` for binary metadata extraction, and `syn`/`quote` for Rust source parsing. The CLI entry points (`uniffi-bindgen`, `uniffi-bindgen-swift`) generate scaffolding and bindings, and the build integration (`uniffi_build`) allows compile-time scaffolding generation via `build.rs`.

Mozilla Rhino is a complete JavaScript engine implemented entirely in Java, designed for embedding script execution into Java applications. It provides a full ECMAScript runtime, including ES6+ features (BigInt, Promises, typed arrays, classes, modules), E4X (XML) support, LiveConnect (Java-JS interop), and a JSR-223 ScriptEngine implementation. The architecture is modular, with separate modules for the core engine, tools (shell, debugger, compiler), XML support, and Kotlin integration. The engine supports multiple execution modes: an interpreted bytecode (interpreter v1 and v2), a just-in-time compiler via the optimizer (using jdk.dynalink), and a classfile-based compiler. Key entry points include the `Context` class for script execution, `RhinoScriptEngine` for Java scripting API, and the `Main` shell tool. The project is well-tested with thousands of regression tests, test262 conformance, and JMH benchmarks. It is actively maintained, with recent releases and CI pipelines for Gradle and Bazel, targeting Java 17+ and Android.

cubeb is a cross-platform C library for low-latency audio playback and recording, originally developed by Mozilla for Firefox. It provides a simple callback-based API and supports a wide range of backends including ALSA, PulseAudio, WASAPI, CoreAudio, OpenSL ES, AAudio, JACK, and others. The library handles device enumeration, resampling (via vendored Speex), channel mixing, and real-time streaming. Its architecture uses a vtable pattern ('cubeb_ops') to abstract backend implementations, and includes several high-quality lock-free data structures. While primarily a production library, it includes extensive test suites and debugging tools like audio dumping. The codebase is actively maintained, with ongoing migrations of some backends to Rust, but retains legacy backends (WinMM, OSS) with limited functionality.