Constructors, Destructors, and const Methods

Constructors: born valid

A constructor runs automatically when an object is created. Its job is to leave the object valid — no "uninitialized then fixed later" window:

class Timer {
public:
    explicit Timer(std::string name) : name_(std::move(name)), start_(std::chrono::steady_clock::now()) {}

private:
    std::string name_;
    std::chrono::steady_clock::time_point start_;
};

• The member initializer list (after :) initializes members directly — prefer it over assigning in the body.
• explicit stops surprise conversions from single arguments (Timer t = "x"; will not compile — good).
• std::move(name) passes the string's guts into the member instead of copying (the standard library does this everywhere; you use std::move in constructors, never on const objects, and never manually on locals right before last use without a reason — module 12 revisits).

Destructors: born valid, die clean

The destructor ~Timer() runs automatically at scope exit, in reverse construction order. You rarely write one in Beginner (members clean themselves up — the RAII promise of module 12), but you must know it exists and runs deterministically: this is why C++ has no finally and does not need one.

const member functions

long long balance() const { return balance_cents_; }   // promises: no mutation

const after the parameter list means "calling this does not modify the object". Non-const objects and const references (i.e., nearly every parameter in well-written code) can only call const methods — so mark every read-only method const, or your API becomes unusable from const BankAccount& parameters.

The discipline, summarized

• Constructor: make every member valid.
• Methods: enforce the invariant on every mutation.
• const on every read-only method.
• Everything else stays private.