anandabits

/*
 This gist contains an analysis of the design space for abstractions in a type system like Swift's.
 It focuses on Monoid as an example but the same set of tradeoffs apply to any abstraction.
 As such, it focuses on a single abstraction and does not explore issues that arise when designing
 an abstraction hierarchy in depth.

 The matrix below describes some of the design important design tradeoffs for various approaches including:
 - OO style protocols such as `Monoid { var identity: Self ... }
    (Haskell also adopts this approach for many abstractions, including for Monoid)
 - ML signature style static protocols with empty enum conformances analagous with ML structures

IanKeen

public protocol KeypathUpdatable {
    func update<T>(_ keyPath: WritableKeyPath<Self, T>, to value: T) -> Self
}

public extension KeypathUpdatable {
    public func update<T>(_ keyPath: WritableKeyPath<Self, T>, to value: T) -> Self {
        var copy = self
        copy[keyPath: keyPath] = value
        return copy
    }

anandabits

// This example shows how closed protocols can be emulated in Swift today.
// The technique leverages Swift's support for public (but not open) classes.

// First, it's worth observing that there is an almost trivial technique that can be used when
// it is possible to specify a (possibly abstract) superclass for all conforiming types.
// Simply declare a public (not open) base class and a protocol with a Self inheritance constraint.
// Swift does not support open subclasses of a public superclass so no classes outside the module
// will be able to meet the self inheritance constraint.
public class FooBase {}
public protocol Foo where Self: FooBase {}

anandabits

// This example shows how higher-kinded types can be emulated in Swift today.
// It acheives correct typing at the cost of some boilerplate, manual lifting and an existential representation.
// The technique below was directly inspired by the paper Lightweight Higher-Kinded Polymorphism
// by Jeremy Yallop and Leo White found at http://ocamllabs.io/higher/lightweight-higher-kinded-polymorphism.pdf

/// `ConstructorTag` represents a type constructor.
/// `Argument` represents an argument to the type constructor.
struct Apply<ConstructorTag, Argument> {
    /// An existential containing a value of `Constructor<Argument>`
    /// Where `Constructor` is the type constructor represented by `ConstructorTag`

IanKeen

1. Create a new xcode project
  - You can skip this if you're adding the playground to an existing project
2. File > Save as workspace - save workspace in project folder
  - You can skip this if you're adding the playground to an existing workspace
3. File > Playground - create new playground in project folder
4. Drag playground in at the _workspace_ level (root item, i.e. _not_ under the project)
5. Add a new Framework to your project (I usually name mine something like 'PlaygroundKit')
  - You don't need to include tests
  - Choose 'None' for Enbed in Application
6. Make all project files you want to be accessible to the playground are members of the new target 

ZevEisenberg

import QuartzCore

extension CGFloat {
    func map(from from: ClosedInterval<CGFloat>, to: ClosedInterval<CGFloat>) -> CGFloat {
        let result = ((self - from.start) / (from.end - from.start)) * (to.end - to.start) + to.start
        return result
    }
}

extension Double {

natecook1000

// NSScanner+Swift.swift
// A set of Swift-idiomatic methods for NSScanner
//
// (c) 2015 Nate Cook, licensed under the MIT license

import Foundation

extension NSScanner {
    
    // MARK: Strings

jboner

Latency Comparison Numbers (~2012)
----------------------------------
L1 cache reference                           0.5 ns
Branch mispredict                            5   ns
L2 cache reference                           7   ns                      14x L1 cache
Mutex lock/unlock                           25   ns
Main memory reference                      100   ns                      20x L2 cache, 200x L1 cache
Compress 1K bytes with Zippy             3,000   ns        3 us
Send 1K bytes over 1 Gbps network       10,000   ns       10 us
Read 4K randomly from SSD*             150,000   ns      150 us          ~1GB/sec SSD