Каким образом реализовать механизм Copy-on-write в практике программирования?

The Copy-on-write (COW) mechanism in programming is implemented so that when copying an object, a new copy of the data is not actually created; instead, both variables refer to the same memory location. Actual copying occurs only when an attempt is made to modify the data (record), at which point a separate copy is created for modification.

In iOS and Swift, this is often implemented using classes and structures, where structures — value types — use COW for optimization. For example, the standard Array type in Swift uses COW: when copying an array, the data is not duplicated until a change occurs.

An example of a simplified COW implementation in Swift:

class Storage {
    var data: [Int]
    init(data: [Int]) { self.data = data }
}

struct CowArray {
    private var storage: Storage

    init(_ data: [Int]) {
        storage = Storage(data: data)
    }

    var count: Int { storage.data.count }

    mutating func append(_ element: Int) {
        if !isKnownUniquelyReferenced(&storage) {
            storage = Storage(data: storage.data) // copy on write
        }
        storage.data.append(element)
    }

    subscript(index: Int) -> Int {
        get { storage.data[index] }
        set {
            if !isKnownUniquelyReferenced(&storage) {
                storage = Storage(data: storage.data) // copy on write
            }
            storage.data[index] = newValue
        }
    }
}

Here, isKnownUniquelyReferenced checks if there are still references to the Storage object. If yes — a copy is made before modification.

Thus, COW helps save memory and improve performance by avoiding unnecessary copies until the data is modified.