Let's start with what goes wrong when you manage locks manually. Consider this seemingly simple code.

The bug is obvious once you see it, but in real code with multiple conditions, loops, and exception-throwing functions, it's easy to miss a path. Every branch needs an unlock. Every exception handler needs an unlock. And when someone adds a new early return six months from now, they might not even realize there's a mutex involved.

The fix is RAII: let the compiler handle unlocking.

The following diagram illustrates the difference between manual locking and RAII-based locking. Notice how the manual approach leaves a dangerous path where the mutex stays locked, while RAII guarantees cleanup regardless of how the scope exits.

The RAII Principle