LEGACY C++ REFACTORING AND TESTING IN PRACTICE

Every engineering team has a codebase like this one: old enough to carry real institutional knowledge, tangled enough that nobody fully trusts it, and critical enough that nobody is willing to rewrite it from scratch. Legacy C++ Refactoring and Testing in Practice is a hands-on, technically rigorous guide to safely transforming exactly that kind of system - without breaking what already works.

Built around a single recurring, realistic C++ codebase, this book walks through the complete discipline of legacy modernization: exploring unfamiliar code with confidence, building characterization tests that pin down real behavior before you touch it, introducing seams to make untestable code testable, and refactoring safely with modern C++ - move semantics, smart pointers, enum class, RAII, and more. From there it goes deeper: untangling raw memory management with unique_ptr, shared_ptr, and weak_ptr; breaking apart oversized "God classes"; modernizing a legacy CMake build into a clean, target-based one; fixing real concurrency bugs with ThreadSanitizer; redesigning error handling with explicit result types; and wiring the whole safety net into continuous integration.

The book closes with a full capstone project that applies every technique, in sequence, to a complete production-style system - and a chapter on planning and sequencing a modernization effort across an entire team, not just a single file.

This is not a beginner's C++ tutorial, a design-pattern catalog, or a call to rewrite everything from scratch. It's a practical playbook for engineers who inherit large, aging C++ systems and need to make them safer to change - one verified step at a time.

What you'll learn:

• How to investigate and map an unfamiliar, undocumented C++ codebase
• How to write characterization tests that capture real behavior before refactoring
• How to introduce testing seams without changing production behavior
• How to modernize memory management, build systems, and error handling safely
• How to find and fix real concurrency bugs with sanitizers
• How to build CI pipelines that enforce the safety net automatically
• How to plan and sequence a legacy modernization effort across a team

Written for working C++ engineers - not a textbook author's blank-page example, but a guide grounded in the fact that your production system already has users, assumptions, and behavior that can't simply be thrown away.