Swift class introspection & generics - introspection

I am trying to dynamically create a class instance based type using generics, however I am encountering difficulty with class introspection.
Here are the questions:
Is there a Swift-equivalent to Obj-C's self.class?
Is there a way to instantiate a class using the AnyClass result from NSClassFromString?
Is there a way to get AnyClass or otherwise type information strictly from a generic parameter T? (Similar to C#'s typeof(T) syntax)

Well, for one, the Swift equivalent of [NSString class] is .self (see Metatype docs, though they're pretty thin).
In fact, NSString.class doesn't even work! You have to use NSString.self.
let s = NSString.self
var str = s()
str = "asdf"
Similarly, with a swift class I tried...
class MyClass {
}
let MyClassRef = MyClass.self
// ERROR :(
let my_obj = MyClassRef()
Hmm… the error says:
Playground execution failed: error: :16:1: error: constructing an object of class type 'X' with a metatype value requires an '#required' initializer
Y().me()
^
<REPL>:3:7: note: selected implicit initializer with type '()'
class X {
^
It took me a while to figure out what this means… turns out it wants the class to have a #required init()
class X {
func me() {
println("asdf")
}
required init () {
}
}
let Y = X.self
// prints "asdf"
Y().me()
Some of the docs refer to this as .Type, but MyClass.Type gives me an error in the playground.

Here's how to use NSClassFromString. You have to know the superclass of what you're going to end up with. Here are a superclass-subclass pair that know how to describe themselves for println:
#objc(Zilk) class Zilk : NSObject {
override var description : String {return "I am a Zilk"}
}
#objc(Zork) class Zork : Zilk {
override var description : String {return "I am a Zork"}
}
Notice the use of the special #obj syntax to dictate the Objective-C munged name of these classes; that's crucial, because otherwise we don't know the munged string that designates each class.
Now we can use NSClassFromString to make the Zork class or the Zilk class, because we know we can type it as an NSObject and not crash later:
let aClass = NSClassFromString("Zork") as NSObject.Type
let anObject = aClass()
println(anObject) // "I am a Zork"
And it's reversible; println(NSStringFromClass(anObject.dynamicType)) also works.
Modern version:
if let aClass = NSClassFromString("Zork") as? NSObject.Type {
let anObject = aClass.init()
print(anObject) // "I am a Zork"
print(NSStringFromClass(type(of:anObject))) // Zork
}

If I'm reading the documentation right, if you deal with instances and e.g. want to return a new instance of the same Type than the object you have been given and the Type can be constructed with an init() you can do:
let typeOfObject = aGivenObject.dynamicType
var freshInstance = typeOfObject()
I quickly tested it with String:
let someType = "Fooo".dynamicType
let emptyString = someType()
let threeString = someType("Three")
which worked fine.

In swift 3
object.dynamicType
is deprecated.
Instead use:
type(of:object)

Swift implementation of comparing types
protocol Decoratable{}
class A:Decoratable{}
class B:Decoratable{}
let object:AnyObject = A()
object.dynamicType is A.Type//true
object.dynamicType is B.Type//false
object.dynamicType is Decoratable.Type//true
NOTE: Notice that it also works with protocols the object may or may not extend

Finally got something to work. Its a bit lazy but even the NSClassFromString() route did not work for me...
import Foundation
var classMap = Dictionary<String, AnyObject>()
func mapClass(name: String, constructor: AnyObject) -> ()
{
classMap[name] = constructor;
}
class Factory
{
class func create(className: String) -> AnyObject?
{
var something : AnyObject?
var template : FactoryObject? = classMap[className] as? FactoryObject
if (template)
{
let somethingElse : FactoryObject = template!.dynamicType()
return somethingElse
}
return nil
}
}
import ObjectiveC
class FactoryObject : NSObject
{
#required init() {}
//...
}
class Foo : FactoryObject
{
class override func initialize()
{
mapClass("LocalData", LocalData())
}
init () { super.init() }
}
var makeFoo : AnyObject? = Factory.create("Foo")
and bingo, "makeFoo" contains a Foo instance.
The downside is your classes must derrive from FactoryObject and they MUST have the Obj-C +initialize method so your class gets automagically inserted in the class map by global function "mapClass".

Here is another example showing class hierarchy implementation, similar to accepted answer, updated for the first release of Swift.
class NamedItem : NSObject {
func display() {
println("display")
}
required override init() {
super.init()
println("base")
}
}
class File : NamedItem {
required init() {
super.init()
println("folder")
}
}
class Folder : NamedItem {
required init() {
super.init()
println("file")
}
}
let y = Folder.self
y().display()
let z = File.self
z().display()
Prints this result:
base
file
display
base
folder
display

Related

Get filename of derived class from base class in typescript running on node.js?

I'm looking for a way to get the filename of a derived class from a base class in typescript running on node.js. An example of this would be:
Foo.ts
export abstract class Foo {
constructor() { }
name() { return (__filename); }
print() { console.log(this.name()); }
}
Bar.ts
import { Foo } from './Foo';
export class Bar extends Foo {
constructor() { super(); }
}
main.ts
import { Bar } from './Bar';
let bar = new Bar();
bar.print(); // should yield the location of Bar.ts
Due to the number of files involved and just cleanliness I'd like this to be confined to the Foo class rather than having an override of the name() function in each derived class.
I was able to sort-of solve this with the code:
private getDerivedFilePath(): string {
let errorStack: string[] = new Error().stack.split('\n');
let ret: string = __filename;
let baseClass: any = ThreadPoolThreadBase;
for (let i: number = 3; i < errorStack.length; i++) {
let filename: string = errorStack[i].slice(
errorStack[i].lastIndexOf('(') + 1,
Math.max(errorStack[i].lastIndexOf('.js'), errorStack[i].lastIndexOf('.ts')) + 3
);
let other: any = require(filename);
if (other.__proto__ === baseClass) {
ret = filename;
baseClass = other;
} else {
break;
}
}
return (ret || '');
}
Added to Foo, which will work when called from the constructor to set a private _filename property, for inheritance chains beyond the example above so long as the files are structured with a default export of the class being used. There may also be a caveat that if a base class from which a derived object is inheriting directly is initialized as a separate instance within the constructor of any member of the inheritance chain it could get confused and jump to another independent derived class - so it's a bit of a hacky work-around and I'd be interested if someone comes up with something better, but wanted to post this in case someone stumbles across this question and it works for them.
You can use require.cache to get all cached NodeModule objects and filter it to find your module.
https://nodejs.org/api/modules.html#requirecache
class ClassA {
public static getFilePath():string{
const nodeModule = this.getNodeModule();
return (nodeModule) ? nodeModule.filename : "";
}
public static getNodeModule(): NodeModule | undefined{
const nodeModule = Object.values(require.cache)
.filter((chl) => chl?.children.includes(module))
.filter((mn)=> mn?.filename.includes(this.name))
.shift();
return nodeModule;
}
}
class ClassB extends ClassA {
constructor(){}
}
const pathA = ClassA.getFilePath(); //Must return the absolute path of ClassA
const pathB = ClassB.getFilePath(); //Must return the absolute path of ClassB

haxe get type info for fields of a function type parameter

Reflect can list an object's fields at runtime, but doesn't have type info. rtti has type info, but doesn't seem to work for type parameters.
In this example, I want MyLibrary to be able to figure out that T's val field is an Int.
import haxe.Constraints;
class Main {
static function main() {
var obj :IntStruct = MyLibrary.foo();
}
}
#:rtti
class IntStruct {
public var val :Int;
public function new() { }
}
class MyLibrary {
#:generic
public static function foo<T:Constructible<()->Void>>() :T {
var something = new T();
trace(Reflect.fields(something)); // [val]
trace(haxe.rtti.Rtti.getRtti(T)); // fails here
return something;
}
}
this gives this compile error:
Main.hx:20: characters 38-39 : Only #:const type parameters on #:generic classes can be used as value
if I add #:const to the type parameter declaration, I get this instead
Main.hx:20: characters 38-39 : foo.T should be Class<Unknown<0>>
Main.hx:20: characters 38-39 : ... For function argument 'c'
Since foo is generic, the compiler will write a separate function for each T, so it should be no problem for it to tell me about T inside foo. How can it be done?

Kotlin thread safe native lazy singleton with parameter

In java we can write thead-safe singletons using double Checked Locking & volatile:
public class Singleton {
private static volatile Singleton instance;
public static Singleton getInstance(String arg) {
Singleton localInstance = instance;
if (localInstance == null) {
synchronized (Singleton.class) {
localInstance = instance;
if (localInstance == null) {
instance = localInstance = new Singleton(arg);
}
}
}
return localInstance;
}
}
How we can write it in kotlin?
About object
object A {
object B {}
object C {}
init {
C.hashCode()
}
}
I used kotlin decompiler to get that
public final class A {
public static final A INSTANCE;
private A() {
INSTANCE = (A)this;
A.C.INSTANCE.hashCode();
}
static {
new A();
}
public static final class B {
public static final A.B INSTANCE;
private B() {
INSTANCE = (A.B)this;
}
static {
new A.B();
}
}
public static final class C {
public static final A.C INSTANCE;
private C() {
INSTANCE = (A.C)this;
}
static {
new A.C();
}
}
}
All of object have constructor invoke in static block. Based on it, we can think that it's not lazy.
Сlose to the right answer.
class Singleton {
companion object {
val instance: Singleton by lazy(LazyThreadSafetyMode.PUBLICATION) { Singleton() }
}
}
Decompiled:
public static final class Companion {
// $FF: synthetic field
private static final KProperty[] $$delegatedProperties = new KProperty[]{(KProperty)Reflection.property1(new PropertyReference1Impl(Reflection.getOrCreateKotlinClass(Singleton.Companion.class), "instance", "getInstance()Lru/example/project/tech/Singleton;"))};
#NotNull
public final Singleton getInstance() {
Lazy var1 = Singleton.instance$delegate;
KProperty var3 = $$delegatedProperties[0];
return (Singleton)var1.getValue();
}
private Companion() {
}
// $FF: synthetic method
public Companion(DefaultConstructorMarker $constructor_marker) {
this();
}
}
I hope Kotlin developers will make non reflection implementation in future...
Kotlin has an equivalent of your Java code, but more safe. Your double lock check is not recommended even for Java. In Java you should use an inner class on the static which is also explained in Initialization-on-demand holder idiom.
But that's Java. In Kotlin, simply use an object (and optionally a lazy delegate):
object Singletons {
val something: OfMyType by lazy() { ... }
val somethingLazyButLessSo: OtherType = OtherType()
val moreLazies: FancyType by lazy() { ... }
}
You can then access any member variable:
// Singletons is lazy instantiated now, then something is lazy instantiated after.
val thing = Singletons.something // This is Doubly Lazy!
// this one is already loaded due to previous line
val eager = Singletons.somethingLazyButLessSo
// and Singletons.moreLazies isn't loaded yet until first access...
Kotlin intentionally avoids the confusion people have with singletons in Java. And avoids the "wrong versions" of this pattern -- of which there are many. It instead provides the simpler and the safest form of singletons.
Given the use of lazy(), if you have other members each would individually be lazy. And since they are initialized in the lambda passed to lazy() you can do things that you were asking about for about customizing the constructor, and for each member property.
As a result you have lazy loading of Singletons object (on first access of instance), and then lazier loading of something (on first access of member), and complete flexibility in object construction.
See also:
lazy() function
Lazy thread safe mode options
Object declarations
As a side note, look at object registry type libraries for Kotlin that are similar to dependency injection, giving you singletons with injection options:
Injekt - I'm the author
Kodein - Very similar and good
Object declaration is exactly for this purpose:
object Singleton {
//singleton members
}
It is lazy and thread-safe, it initializes upon first call, much as Java's static initializers.
You can declare an object at top level or inside a class or another object.
For more info about working with objects from Java, please refer to this answer.
As to the parameter, if you want to achieve exactly the same semantics (first call to getInstance takes its argument to initialize the singleton, following calls just return the instance, dropping the arguments), I would suggest this construct:
private object SingletonInit { //invisible outside the file
lateinit var arg0: String
}
object Singleton {
val arg0: String = SingletonInit.arg0
}
fun Singleton(arg0: String): Singleton { //mimic a constructor, if you want
synchronized(SingletonInit) {
SingletonInit.arg0 = arg0
return Singleton
}
}
The main flaw of this solution is that it requires the singleton to be defined in a separate file to hide the object SingletonInit, and you cannot reference Singleton directly until it's initialized.
Also, see a similar question about providing arguments to a singleton.
I recently wrote an article on that topic.
TL;DR Here's the solution I came up to:
1) Create a SingletonHolder class. You only have to write it once:
open class SingletonHolder<out T, in A>(creator: (A) -> T) {
private var creator: ((A) -> T)? = creator
#Volatile private var instance: T? = null
fun getInstance(arg: A): T {
val i = instance
if (i != null) {
return i
}
return synchronized(this) {
val i2 = instance
if (i2 != null) {
i2
} else {
val created = creator!!(arg)
instance = created
creator = null
created
}
}
}
}
2) Use it like this in your singletons:
class MySingleton private constructor(arg: ArgumentType) {
init {
// Init using argument
}
companion object : SingletonHolder<MySingleton, ArgumentType>(::MySingleton)
}
The singleton initialization will be lazy and thread-safe.

Casting Dynamic to an other class

I would like to know if that's possible to cast a Dynamic to an other class (partially or totally)
For example, this code breaks :
class Test {
public function new() {}
public var id: String;
}
class Main {
public static function main() {
var x:Dynamic = JsonParser.parse("{\"id\":\"sdfkjsdflk\"}");
var t:Test = cast(x, Test);
}
}
with the following message
Class cast error
However, my "Test" class has an "id" field like the dynamic object. (That's an example, my use case is more complexe than that ^^)
So, I don't understand how to get an object from my Dynamic one.
This isn't exactly casting a dynamic to a class instance but may accomplish the same thing:
create an empty instance of the class with Type.createEmptyInstance
set all of the fields from the Dynamic object on the new class instance using Reflect
Example:
import haxe.Json;
class Test {
public function new() {}
public var id: String;
}
class Main {
public static function main() {
var x:Dynamic = Json.parse("{\"id\":\"sdfkjsdflk\"}");
var t:Test = Type.createEmptyInstance(Test);
for (field in Type.getInstanceFields(Test))
if (Reflect.hasField(x, field))
Reflect.setProperty(t, field, Reflect.getProperty(x, field));
trace(t.id);
}
}
You could use typedef
typedef Test = {
public var id: String;
}
class Main {
public static function main() {
var t:Test = JsonParser.parse("{\"id\":\"sdfkjsdflk\"}");
}
}
Json.parse returns anonymous structure(implementation platform dependent), typed as Dynamic. There isn't a single chance to cast it to anything but Dynamic, unless Json.parse returns Int, Float or String, which some parsers permit, but which isn't actually permitted by JSON specification.
That is this way because, the operation of casting doesn't check what fields some object have. Operation of casting only checks if the object is an instance of class you are casting to. Obviously, anonymous structure can't be an instance of any class(inside haxe abstractions at least).
However, the right way to perform the thing you seem to be trying to perform is the way stated by #Ben Morris, in his answer.

Using Swift protocols with generics

I have a simple example that seems like it should work:
import CoreData
#objc protocol CoreDataModel {
#optional class func entityName() -> String
}
class AbstractModel: NSManagedObject, CoreDataModel {
class func create<T : CoreDataModel>(context:NSManagedObjectContext) -> T {
var name = T.entityName?()
var object = NSEntityDescription.insertNewObjectForEntityForName(name, inManagedObjectContext: context) as T
return object
}
}
So we have a class called AbstractModel which conforms to the protocol CoreDataModel, and CoreDataModel defines an optional class method called entityName.
However, this line:
var name = T.entityName?()
causes the error:
Expected member name or constructor call after type name
Any idea what I'm doing wrong?
Edit
Removing the word #optional from the declaration and changing the function a bit allows the code to compile, but now I get a runtime error saying that the
'Swift dynamic cast failed'
#objc protocol CoreDataModel {
class func entityName() -> String
}
class AbstractModel: NSManagedObject, CoreDataModel {
class func entityName() -> String {
return "AbstractModel"
}
class func create<T : CoreDataModel>(aClass:T.Type, context:NSManagedObjectContext) -> T {
var name = aClass.entityName()
var object = NSEntityDescription.insertNewObjectForEntityForName(name, inManagedObjectContext: context) as T
return object
}
}
I cannot explain why your code causes a runtime exception. But it works if you change
the function prototype
class func create<T : CoreDataModel>(aClass:T.Type, context:NSManagedObjectContext) -> T
to
class func create<T : NSManagedObject where T: CoreDataModel>(aClass:T.Type, context:NSManagedObjectContext) -> T
Assuming that your managed object subclass conforms to the protocol, for example
extension Event : CoreDataModel {
class func entityName() -> String {
return "Event"
}
}
then this works and creates a new object:
let newManagedObject = AbstractModel.create(Event.self, context: context)
Alternatively, you could use the approach from the answer to
"Swift: return Array of type self" and
define an extension to the NSManagedObjectContext class:
extension NSManagedObjectContext {
func create<T : NSManagedObject where T : CoreDataModel >(entity: T.Type) -> T {
var classname = entity.entityName()
var object = NSEntityDescription.insertNewObjectForEntityForName(classname, inManagedObjectContext: self) as T
return object
}
}
Then a new object would be created as
let newManagedObject = context.create(Event.self)
From "The Swift Programming Language"
Because T is a placeholder, Swift does not look for an actual type called T.
As T is not a real type, it is maybe not useful to cast to T.

Resources