We Keep Coding

C# Reflection dynamically create concrete class instance of T

I want to do something at runtime with objects based on there type.
Depending on the type I need to do different operations so I thought I would create
an abstract class like this:
internal abstract class DataOperator<T>
where T : class
{
public abstract void DoSomething(IList<T> data);
}
with multiple concrete implementations like:
class MyClassOperator : DataOperator<MyClass>
{
public override void DoSomething(IList<MyClass> data)
{
throw new NotImplementedException();
}
}
But how would I actually create an instance of the concrete class at runtime?

The problem is that DataOperator<A> and DataOperator<B> are not assignment compatible. Not even if A and B are. So, you cannot assign your classes to, say DataOperator<object>. Therefore, you will have to treat them individually as non related types.
A way out is to have a non generic base type or interface.
interface IDataOperator
{
public void DoSomething(IList data);
}
We create an abstract class that implements it explicitly in order to hide this weakly typed DoSomething when not called through the interface. We also introduce the generic type parameter.
abstract class DataOperator<T> : IDataOperator
where T : class
{
void IDataOperator.DoSomething(IList data)
{
if (data is IList<T> listT) {
DoSomething(listT);
} else {
throw new ArgumentException("The List is not compatible.", nameof(data));
}
}
abstract public void DoSomething(IList<T> data);
}
We implement concrete types like this:
class DataOperatorA : DataOperator<MyClassA>
{
public override void DoSomething(IList<MyClassA> data)
{
throw new NotImplementedException();
}
}
class DataOperatorB : DataOperator<MyClassB>
{
public override void DoSomething(IList<MyClassB> data)
{
throw new NotImplementedException();
}
}
These implementations are assignment compatible to our interface. Example:
IDataOperator[] operators = { new DataOperatorA(), new DataOperatorB() };
If you want to create different operators depending on other data, you can create a factory class. In this example, we use a string as discriminator, but it could be anything else, like a System.Type or an enum, etc.
static class DataOperator
{
public static IDataOperator Create(string op)
{
return op switch {
"A" => new DataOperatorA(),
"B" => new DataOperatorB(),
_ => throw new ArgumentException("Unknown operator.", "op")
};
}
}
Same example as above but with the factory:
IDataOperator[] operators = { DataOperator.Create("A"), DataOperator.Create("B") };

I resolved this issue using a factory as suggested by Matthew Watson.
DataOperator is actually AggregateRootMigrator
public static class AggregateRootMigratorFactoryProvider
{
public static AggregateRootMigrator<T> GetAggregateRootMigrator<T>()
where T : AggregateRoot
{
var type = typeof(AggregateRootMigrator<T>).Assembly.GetTypes().Where(
t =>
{
if (t.IsAbstract == false && t.IsSubclassOf(typeof(AggregateRootMigrator<T>)))
{
return true;
}
return false;
}).SingleOrDefault();
if (type == null)
{
throw new InvalidOperationException($"No Factory found for eventType: {typeof(T).Name}");
}
return (AggregateRootMigrator<T>)Activator.CreateInstance(type);
}
}
now I can use it like this:
var dataMigrator = AggregateRootMigratorFactoryProvider.GetAggregateRootMigrator();
var result = dataMigrator.Migrate(data);

How to avoid casting to a specific interface

Background info
I have a set of interfaces/classes as follows. For the sake of simplicity imagine more properties, collections etc.
interface IMaster
{
//Some properties
}
interface IB : IMaster
{
string PropOnA { get; set }
}
interface IC : IMaster
{
string PropOnB { get; set }
}
class B : IB
class C : IC
...
These contracts were designed to store data(which is held in a slightly different format in each case). There is a lot of code that uses these contracts to get the data, format it, process it, write etc.
We have developed an entire library that does not see the concrete implementations(B,C) of any of these contracts by inverting control and allow the user to use our 'default implementations' for each contract or just loading in their own. We have registry where the user can register a different implementation.
To this end I have implemented a kind of strategy pattern where there exists a strategy for each contract type based on the task at hand. For the sake of simplicity lets say the task is writing, in reality it is much more complicated.
interface IWriteStrategy
{
public Write(IMaster thing);
}
class WriterA : IWriteStrategy
class WriterB : IWriteStrategy
etc
The above concrete strategies are also never 'seen' in our library, the client must register their own implementation or our default version.
Design flaw??
I am not liking the cast in every strategy that is now necessary.
public classWriterA : IWriteStrategy
{
public void Write(IMaster thing)
{
if(thing is IA thingA)
//do some work
}
}
public classWriterB : IWriteStrategy
{
public void Write(IMaster thing)
{
if(thing is IB thingB)
//do some work
}
}
What we want to do is be able to loop through a list of IMaster objects and run some operations.
foreach(var thing in Things)
{
var strategy = GetStrategy(thing.GetType()); //this gets the strategy object from our `registry` if one exists
strategy.Execute(thing);
}
The above design allows this but there seems to be a flaw which I cant for the life of me spot a solution to. We have to cast to the specific interface within each strategy implementation.
I have tried with generics, but just cant seem to nail it.
Question
What would be a better way of designing this to avoid the cast but still be able to loop through a list of IMaster things and treat them the same? Or is the cast absolutely necessary here?
I am trying to follow a SOLID design but feel the cast is messing with this as the client implementing the strategies will have to do the cast in order to get anything to work within the Write method.
[Edit]
I have updated the classes implementing the IWriteStrategy.

If you rarely add new IMaster specializations, but often add new operations OR need to make sure operation providers (e.g writer) needs to support ALL specializations then the Visitor Pattern is a perfect fit.
Otherwise you basically need some kind of service locator & registration protocol to map operation providers/strategies to IMaster specializations.
One way you could do it is define generic interfaces such as IMasterWriter<T> where T:IMaster which can then be implemented like IBWriter : IMasterWriter<IB> which defines the mapping.
From that point you only need a mechanism that uses reflection to find a specific IMasterWriter implementor for a given type of IMaster and decide what to do if it's missing. You could scan assemblies early to detect missing implementations at boot rather than failing later too.

Maybe it is appropriate to use Strategy pattern and just give an implementation and execute it. Let me show an example.
interface IMaster
{
void ExecuteMaster();
}
class MasterOne : IMaster
{
public void ExecuteMaster()
{
Console.WriteLine("Master One");
}
}
class MasterTwo : IMaster
{
public void ExecuteMaster()
{
Console.WriteLine("Master Two");
}
}
and
interface IWriteStrategy
{
void Write(IMaster thing);
}
class WriterA : IWriteStrategy
{
public void Write(IMaster thing)
{
Console.WriteLine("Writer A");
thing.ExecuteMaster();
}
}
class WriterB : IWriteStrategy
{
public void Write(IMaster thing)
{
Console.WriteLine("Writer B");
thing.ExecuteMaster();
}
}
and code to execute:
static void Main(string[] args)
{
List<IWriteStrategy> writeStrategies = new List<IWriteStrategy>()
{
new WriterA(),
new WriterB()
};
List<IMaster> executes = new List<IMaster>()
{
new MasterOne(),
new MasterTwo()
};
for (int i = 0; i < writeStrategies.Count(); i++)
{
writeStrategies[i].Write(executes[i]);
}
}

what about this, you will have all your casts in one strategy factory method:
public interface IWriterStrategy
{
void Execute();
}
public class WriterA : IWriterStrategy
{
private readonly IA _thing;
public WriterA(IA thing)
{
_thing = thing;
}
public void Execute()
{
Console.WriteLine(_thing.PropOnA);
}
}
public class WriterB : IWriterStrategy
{
private readonly IB _thing;
public WriterB(IB thing)
{
_thing = thing;
}
public void Execute()
{
Console.WriteLine(_thing.PropOnB);
}
}
public static class WriterFactory
{
public static List<(Type Master, Type Writer)> RegisteredWriters = new List<(Type Master, Type Writer)>
{
(typeof(IA), typeof(WriterA)),
(typeof(IB), typeof(WriterB))
};
public static IWriterStrategy GetStrategy(IMaster thing)
{
(Type Master, Type Writer) writerTypeItem = RegisteredWriters.Find(x => x.Master.IsAssignableFrom(thing.GetType()));
if (writerTypeItem.Master != null)
{
return (IWriterStrategy)Activator.CreateInstance(writerTypeItem.Writer, thing);
}
throw new Exception("Strategy not found!");
}
}
public interface IMaster
{
//Some properties
}
public interface IA : IMaster
{
string PropOnA { get; set; }
}
public interface IB : IMaster
{
string PropOnB { get; set; }
}
public interface IC : IMaster
{
string PropOnC { get; set; }
}
public class ThingB : IB
{
public string PropOnB { get => "IB"; set => throw new NotImplementedException(); }
}
public class ThingA : IA
{
public string PropOnA { get => "IA"; set => throw new NotImplementedException(); }
}
public class ThingC : IC
{
public string PropOnC { get => "IC"; set => throw new NotImplementedException(); }
}
internal static class Program
{
private static void Main(string[] args)
{
var things = new List<IMaster> {
new ThingA(),
new ThingB()//,
//new ThingC()
};
foreach (var thing in things)
{
var strategy = WriterFactory.GetStrategy(thing); //this gets the strategy object from our `registry` if one exists
strategy.Execute();
}
}
}

How do you return a class that inherits a base class with a generic response? [duplicate]

I have the following classes
public abstract class BaseViewPresenter { }
public abstract class BaseView<T> : UserControl
where T : BaseViewPresenter { }
public class LoginPresenter : BaseViewPresenter { }
public partial class LoginView : BaseView<LoginPresenter> { }
I have a method that looks like this (simplified)
public BaseView<BaseViewPresenter> Resolve(BaseViewPresenter model)
{
var type = model.GetType();
var viewType = _dataTemplates[type];
// Correctly creates BaseView object
var control = Activator.CreateInstance(viewType);
// Fails to cast as BaseView<BaseViewPresenter> so returns null
return control as BaseView<BaseViewPresenter>;
}
When I call this using an instances of LoginPresenter
var login = new LoginPresenter();
var ctl = Resolve(login);
The line Activator.CreateInstance(viewType) correctly resolves into a new instances of my LoginView, however control as BaseView<BaseViewPresenter> can't do the cast correctly so returns null.
Is there a way to correctly cast the control into BaseView<BaseViewPresenter> without using specific type generics?
Since LoginView inherits from BaseView<LoginPresenter>, and LoginPresenter inherits from BaseViewPresenter, I would assume there's a way to convert LoginView to BaseView<BaseViewPresenter>.
I am stuck with using .Net 3.5

This is a very frequently asked question. Let's rename your types:
abstract class Fruit { } // was BaseViewPresenter
abstract class FruitBowl<T> where T : Fruit // was BaseView
class Apple : Fruit { } // was LoginPresenter
class BowlOfApples : FruitBowl<Apple> { } // was LoginView
Your question now is:
I have a BowlOfApples, which inherits from FruitBowl<Apple>. Why can I not use it as a FruitBowl<Fruit>? An apple is a fruit, so a bowl of apples is a bowl of fruit.
No, it isn't. You can put a banana in a bowl of fruit, but you can't put a banana in a bowl of apples, and therefore a bowl of apples is not a bowl of fruit. (And by similar argument, a bowl of fruit is not a bowl of apples either.) Since the operations you can legally perform on the two types are different, they cannot be compatible.
Here is a photo of StackOverflow legend Jon Skeet demonstrating this fact:
The feature you want is called generic contravariance, and it is supported only on interfaces and delegate types when the compiler can prove that the variance is safe, and when the varying type is a reference type. For example, you can use an IEnumerable<Apple> in a context where IEnumerable<Fruit> is needed because the compiler can verify that there is no way that you can put a Banana into a sequence of fruit.
Do a search on "C# covariance and contravariance" on this site or on the web and you'll find many more details about how this feature works. In particular, my series of articles on how we designed and implemented this feature in C# 4 starts here: http://blogs.msdn.com/b/ericlippert/archive/2007/10/16/covariance-and-contravariance-in-c-part-one.aspx

I accepted Eric's answer since it provides a great explanation of why what I wanted wasn't possible, but I also thought I'd share my solution in case anyone else runs into this same problem.
I removed the generic type parameter from my original BaseView class, and created a 2nd version of the BaseView class that included the generic type parameter and specifics for it.
The first version is used by my .Resolve() method or other code that doesn't care about the specific types, and the second version is used by any code that does care, such as the implentation of a BaseView
Here's an example of how my code ended up looking
// base classes
public abstract class BaseViewPresenter { }
public abstract class BaseView : UserControl
{
public BaseViewPresenter Presenter { get; set; }
}
public abstract class BaseView<T> : BaseView
where T : BaseViewPresenter
{
public new T Presenter
{
get { return base.Presenter as T; }
set { base.Presenter = value; }
}
}
// specific classes
public class LoginPresenter : BaseViewPresenter { }
public partial class LoginView : BaseView<LoginPresenter>
{
// Can now call things like Presenter.LoginPresenterMethod()
}
// updated .Resolve method used for obtaining UI object
public BaseView Resolve(BaseViewPresenter presenter)
{
var type = model.GetType();
var viewType = _dataTemplates[type];
BaseView view = Activator.CreateInstance(viewType) as BaseView;
view.Presenter = presenter;
return view;
}

You're expecting to treat the type as being covariant with respect to the generic argument. Classes can never be covariant; you'd need to use an interface rather than (or in addition to) an abstract class to make it covariant with respect to T. You'd also need to be using C# 4.0.

My usual solution to this problem is to create an intermediary class that has access to the type-parametric class's methods through delegates. Fields can also be accessed through getters/setters.
The general pattern goes:
public abstract class Super {}
public abstract class MyAbstractType<T> where T : Super {
public MyGeneralType AsGeneralType() {
return MyGeneralType.Create(this);
}
// Depending on the context, an implicit cast operator might make things
// look nicer, though it might be too subtle to some tastes.
public static implicit operator MyGeneralType(MyAbstractType<T> t) {
return MyGeneralType.Create(t);
}
public int field;
public void MyMethod1() {}
public void MyMethod2(int argument) {}
public abstract bool MyMethod3(string argument);
}
public delegate T Getter<T>();
public delegate void Setter<T>(T value);
public delegate void MyMethod1Del();
public delegate void MyMethod2Del(int argument);
public delegate bool MyMethod3Del(string argument);
public class MyGeneralType {
public Getter<int> FieldGetter;
public Setter<int> FieldSetter;
public MyMethod1Del MyMethod1;
public MyMethod2Del MyMethod2;
public MyMethod3Del MyMethod3;
public static MyGeneralType Create<T>(MyAbstractType<T> t) where T : Super {
var g = new MyGeneralType();
g.FieldGetter = delegate { return t.field; };
g.FieldSetter = value => { t.field = value; };
g.MyMethod1 = t.MyMethod1;
g.MyMethod2 = t.MyMethod2;
g.MyMethod3 = t.MyMethod3;
return g;
}
public int field {
get { return FieldGetter(); }
set { FieldSetter(value); }
}
}
The above exemplifies getting all the methods and fields but normally I only need a few of them. This is a general solution to the problem and one could feasibly write a tool to generate these intermediary classes automatically, which I might at some point.
Try it here: https://dotnetfiddle.net/tLkmgR
Note that this is enough for all my cases, but you can be extra hacky with this:
public abstract class MyAbstractType<T> where T : Super {
// ... Same everything else ...
// data fields must become abstract getters/setters, unfortunate
public abstract int field {
get;
set;
}
public static implicit operator MyAbstractType<Super>(MyAbstractType<T> t) {
return MyGeneralType.Create(t);
}
}
public class MyGeneralType : MyAbstractType<Super> {
// ... same constructors and setter/getter
// fields but only keep method fields
// that contain the method references for
// implementations of abstract classes,
// and rename them not to clash with the
// actual method names ...
public MyMethod3Del myMethod3Ref;
// Implement abstract methods by calling the corresponding
// method references.
public override bool MyMethod3(string argument) {
return myMethod3Ref(argument);
}
// Same getters/setters but with override keyword
public override int field {
get { return FieldGetter(); }
set { FieldSetter(value); }
}
}
And there you go, now you can literally cast a MyAbstractType<Sub> where Sub : Super to a MyAbstractType<Super>, although it's no longer the same object anymore, but it does retain the same methods and data, it's sort of a complex pointer.
public class Sub : Super {}
public class MySubType : MyAbstractType<Sub> {
public int _field;
public override int field {
get { return _field; }
set { _field = value; }
}
public override bool MyMethod3(string argument) {
Console.WriteLine("hello " + argument);
return argument == "world";
}
}
public class MainClass {
public static void Main() {
MyAbstractType<Sub> sub = new MyAbstractType<Sub>();
MyAbstractType<Super> super = sub;
super.MyMethod3("hello"); // calls sub.MyMethod3();
super.field = 10; // sets sub.field
}
}
This isn't as good in my opinion, the other version of MyGeneralType is a more straighforward layer over the concrete types, plus it doesn't require rewriting the data fields, but it does actually answer the question, technically. Try it here: https://dotnetfiddle.net/S3r3ke
Example
Using these abstract classes:
public abstract class Animal {
public string name;
public Animal(string name) {
this.name = name;
}
public abstract string Sound();
}
public abstract class AnimalHouse<T> where T : Animal {
List<T> animals;
public AnimalHouse(T[] animals) {
this.animals = animals.ToList();
}
public static implicit operator GeneralAnimalHouse(AnimalHouse<T> house) {
return GeneralAnimalHouse.Create(house);
}
public List<string> HouseSounds() {
return animals.Select(animal => animal.Sound()).ToList();
}
}
We make this "general" variant:
public delegate List<string> HouseSoundsDel();
public class GeneralAnimalHouse {
public HouseSoundsDel HouseSounds;
public static GeneralAnimalHouse Create<T>(AnimalHouse<T> house) where T : Animal {
var general = new GeneralAnimalHouse();
general.HouseSounds = house.HouseSounds;
return general;
}
}
And finally with these inheritors:
public class Dog : Animal {
public Dog(string name) : base(name) {}
public override string Sound() {
return name + ": woof";
}
}
public class Cat : Animal {
public Cat(string name) : base(name) {}
public override string Sound() {
return name + ": meow";
}
}
public class DogHouse : AnimalHouse<Dog> {
public DogHouse(params Dog[] dogs) : base(dogs) {}
}
public class CatHouse : AnimalHouse<Cat> {
public CatHouse(params Cat[] cats) : base(cats) {}
}
We use it like this:
public class AnimalCity {
List<GeneralAnimalHouse> houses;
public AnimalCity(params GeneralAnimalHouse[] houses) {
this.houses = houses.ToList();
}
public List<string> CitySounds() {
var random = new Random();
return houses.SelectMany(house => house.HouseSounds())
.OrderBy(x => random.Next())
.ToList();
}
}
public class MainClass {
public static void Main() {
var fluffy = new Cat("Fluffy");
var miu = new Cat("Miu");
var snuffles = new Cat("Snuffles");
var snoopy = new Dog("Snoopy");
var marley = new Dog("Marley");
var megan = new Dog("Megan");
var catHouse = new CatHouse(fluffy, miu, snuffles);
var dogHouse = new DogHouse(snoopy, marley, megan);
var animalCity = new AnimalCity(catHouse, dogHouse);
foreach (var sound in animalCity.CitySounds()) {
Console.WriteLine(sound);
}
}
}
Output:
Miu: meow
Snoopy: woof
Snuffles: meow
Fluffy: meow
Marley: woof
Megan: woof
Notes:
I added names so it's clear that the method references carry their owner's data with them, for those unfamiliar with delegates.
The required using statements for this code are System, System.Collections.Generic, and System.Linq.
You can try it here: https://dotnetfiddle.net/6qkHL3#
A version that makes GeneralAnimalHouse a subclass of AnimalHouse<Animal> can be found here: https://dotnetfiddle.net/XS0ljg

C# inheritance - save reference to last instance

I just came up with a really odd problem and I wasn't able to figure out how to solve it.
I have 3 classes, the class A is the base for B and C, that is:
class A { ... }
class B : A { ... }
class C : B { ... }
Now I would like to have a static property in these classes that stores the last object of each classes created, for example:
class A
{
static public A lastInstance;
}
class B : A
{
public B()
{
lastInstance = this;
}
}
class C : A
{
public C()
{
lastInstance = this;
}
}
What I would like to achieve is to be able to retrieve an instance for each subclass, for example:
var v1 = new B();
var v2 = new C();
var v3 = B.lastInstance; // v3 == v1 and v3 != v2
var v4 = C.lastInstance; // v4 == v2 and v4 != v3
Is it possible anyhow?
The only approach that seems promising to me shown in C# Static instance members for each inherited class: is it really the only chance I have to avoid defining a static member manually for each class?

I think this could be done with Dictionary and that's the only way i can think of right now:
class A {
static Dictionary<Type, A> _LastInstances = new Dictionary<Type, A>(); // because every subclass will inherit from A
public static A LastInstance {
get {
if ( _LastInstances.ContainsKey(GetType()) ) {
return _LastInstances[GetType()];
}
return null;
}
protected set {
if ( _LastInstances.ContainsKey(GetType()) ) {
_LastInstances[GetType()] = value;
} else {
_LastInstances.Add(GetType(), value);
}
}
}
class B : A {
public B(){
LastInstance = this;
}
}

At first: yes you can. But you missed two points with your implementation.
As you declared the lastInstance as public in class A every derived class can use it. As you declared it as static every instance of A will copy itself into it. But so will every instance of B, C and every other class with id derived from A: they all use the same instance. Thus the last instatiated class is saved and everything instantiated before is overwritten.
To overcome this, you must have a static property LastInstance (I switched to my naming convention) on every class, which you can accomplish by using the new modifier on derived classes
public class A
{
public static A LastInstance { get; private set; }
...
}
public class B : A
{
public static new B LastInstance { get; private set; }
...
}
But you're not done with that alone, because
When you create a new instance of B the (default-)construstor first makes a call into the construtor of A. Thus a reference to any already created instance of a base class is overwritten by the currently created instance of the derived class. So your constructors should look like this:
public class A
{
public static A LastInstance { get; private set; }
public A()
{
if (this.GetType() == typeof(A))
{
LastInstance = this;
}
}
}
public class B : A
{
public static new B LastInstance { get; private set; }
public B()
{
if (this.GetType() == typeof(B))
{
LastInstance = this;
}
}
}
This way you will get the correct lastly created instance (if any) in each classes static LastInstance.
Hope this helps

Because static members aren't inherited, you won't be able to access B.lastInstance if class A defines lastInstance. The suggestion you linked to seems reasonable. Although I don't have enough information on why you're attempting this, you could consider using a factory class that holds onto the latest created object.
Here's an example. This is not a good long term solution if you plan to have many classes deriving from A.
class HoldLastKnownFactory
{
B CreateB() { ... }
C CreateC() { ... }
B LastB { get {...} }
C LastC { get {...} }
}

How to correctly cast a class to an abstract class when using type generics?

I have the following classes
public abstract class BaseViewPresenter { }
public abstract class BaseView<T> : UserControl
where T : BaseViewPresenter { }
public class LoginPresenter : BaseViewPresenter { }
public partial class LoginView : BaseView<LoginPresenter> { }
I have a method that looks like this (simplified)
public BaseView<BaseViewPresenter> Resolve(BaseViewPresenter model)
{
var type = model.GetType();
var viewType = _dataTemplates[type];
// Correctly creates BaseView object
var control = Activator.CreateInstance(viewType);
// Fails to cast as BaseView<BaseViewPresenter> so returns null
return control as BaseView<BaseViewPresenter>;
}
When I call this using an instances of LoginPresenter
var login = new LoginPresenter();
var ctl = Resolve(login);
The line Activator.CreateInstance(viewType) correctly resolves into a new instances of my LoginView, however control as BaseView<BaseViewPresenter> can't do the cast correctly so returns null.
Is there a way to correctly cast the control into BaseView<BaseViewPresenter> without using specific type generics?
Since LoginView inherits from BaseView<LoginPresenter>, and LoginPresenter inherits from BaseViewPresenter, I would assume there's a way to convert LoginView to BaseView<BaseViewPresenter>.
I am stuck with using .Net 3.5

This is a very frequently asked question. Let's rename your types:
abstract class Fruit { } // was BaseViewPresenter
abstract class FruitBowl<T> where T : Fruit // was BaseView
class Apple : Fruit { } // was LoginPresenter
class BowlOfApples : FruitBowl<Apple> { } // was LoginView
Your question now is:
I have a BowlOfApples, which inherits from FruitBowl<Apple>. Why can I not use it as a FruitBowl<Fruit>? An apple is a fruit, so a bowl of apples is a bowl of fruit.
No, it isn't. You can put a banana in a bowl of fruit, but you can't put a banana in a bowl of apples, and therefore a bowl of apples is not a bowl of fruit. (And by similar argument, a bowl of fruit is not a bowl of apples either.) Since the operations you can legally perform on the two types are different, they cannot be compatible.
Here is a photo of StackOverflow legend Jon Skeet demonstrating this fact:
The feature you want is called generic contravariance, and it is supported only on interfaces and delegate types when the compiler can prove that the variance is safe, and when the varying type is a reference type. For example, you can use an IEnumerable<Apple> in a context where IEnumerable<Fruit> is needed because the compiler can verify that there is no way that you can put a Banana into a sequence of fruit.
Do a search on "C# covariance and contravariance" on this site or on the web and you'll find many more details about how this feature works. In particular, my series of articles on how we designed and implemented this feature in C# 4 starts here: http://blogs.msdn.com/b/ericlippert/archive/2007/10/16/covariance-and-contravariance-in-c-part-one.aspx

I accepted Eric's answer since it provides a great explanation of why what I wanted wasn't possible, but I also thought I'd share my solution in case anyone else runs into this same problem.
I removed the generic type parameter from my original BaseView class, and created a 2nd version of the BaseView class that included the generic type parameter and specifics for it.
The first version is used by my .Resolve() method or other code that doesn't care about the specific types, and the second version is used by any code that does care, such as the implentation of a BaseView
Here's an example of how my code ended up looking
// base classes
public abstract class BaseViewPresenter { }
public abstract class BaseView : UserControl
{
public BaseViewPresenter Presenter { get; set; }
}
public abstract class BaseView<T> : BaseView
where T : BaseViewPresenter
{
public new T Presenter
{
get { return base.Presenter as T; }
set { base.Presenter = value; }
}
}
// specific classes
public class LoginPresenter : BaseViewPresenter { }
public partial class LoginView : BaseView<LoginPresenter>
{
// Can now call things like Presenter.LoginPresenterMethod()
}
// updated .Resolve method used for obtaining UI object
public BaseView Resolve(BaseViewPresenter presenter)
{
var type = model.GetType();
var viewType = _dataTemplates[type];
BaseView view = Activator.CreateInstance(viewType) as BaseView;
view.Presenter = presenter;
return view;
}

You're expecting to treat the type as being covariant with respect to the generic argument. Classes can never be covariant; you'd need to use an interface rather than (or in addition to) an abstract class to make it covariant with respect to T. You'd also need to be using C# 4.0.

My usual solution to this problem is to create an intermediary class that has access to the type-parametric class's methods through delegates. Fields can also be accessed through getters/setters.
The general pattern goes:
public abstract class Super {}
public abstract class MyAbstractType<T> where T : Super {
public MyGeneralType AsGeneralType() {
return MyGeneralType.Create(this);
}
// Depending on the context, an implicit cast operator might make things
// look nicer, though it might be too subtle to some tastes.
public static implicit operator MyGeneralType(MyAbstractType<T> t) {
return MyGeneralType.Create(t);
}
public int field;
public void MyMethod1() {}
public void MyMethod2(int argument) {}
public abstract bool MyMethod3(string argument);
}
public delegate T Getter<T>();
public delegate void Setter<T>(T value);
public delegate void MyMethod1Del();
public delegate void MyMethod2Del(int argument);
public delegate bool MyMethod3Del(string argument);
public class MyGeneralType {
public Getter<int> FieldGetter;
public Setter<int> FieldSetter;
public MyMethod1Del MyMethod1;
public MyMethod2Del MyMethod2;
public MyMethod3Del MyMethod3;
public static MyGeneralType Create<T>(MyAbstractType<T> t) where T : Super {
var g = new MyGeneralType();
g.FieldGetter = delegate { return t.field; };
g.FieldSetter = value => { t.field = value; };
g.MyMethod1 = t.MyMethod1;
g.MyMethod2 = t.MyMethod2;
g.MyMethod3 = t.MyMethod3;
return g;
}
public int field {
get { return FieldGetter(); }
set { FieldSetter(value); }
}
}
The above exemplifies getting all the methods and fields but normally I only need a few of them. This is a general solution to the problem and one could feasibly write a tool to generate these intermediary classes automatically, which I might at some point.
Try it here: https://dotnetfiddle.net/tLkmgR
Note that this is enough for all my cases, but you can be extra hacky with this:
public abstract class MyAbstractType<T> where T : Super {
// ... Same everything else ...
// data fields must become abstract getters/setters, unfortunate
public abstract int field {
get;
set;
}
public static implicit operator MyAbstractType<Super>(MyAbstractType<T> t) {
return MyGeneralType.Create(t);
}
}
public class MyGeneralType : MyAbstractType<Super> {
// ... same constructors and setter/getter
// fields but only keep method fields
// that contain the method references for
// implementations of abstract classes,
// and rename them not to clash with the
// actual method names ...
public MyMethod3Del myMethod3Ref;
// Implement abstract methods by calling the corresponding
// method references.
public override bool MyMethod3(string argument) {
return myMethod3Ref(argument);
}
// Same getters/setters but with override keyword
public override int field {
get { return FieldGetter(); }
set { FieldSetter(value); }
}
}
And there you go, now you can literally cast a MyAbstractType<Sub> where Sub : Super to a MyAbstractType<Super>, although it's no longer the same object anymore, but it does retain the same methods and data, it's sort of a complex pointer.
public class Sub : Super {}
public class MySubType : MyAbstractType<Sub> {
public int _field;
public override int field {
get { return _field; }
set { _field = value; }
}
public override bool MyMethod3(string argument) {
Console.WriteLine("hello " + argument);
return argument == "world";
}
}
public class MainClass {
public static void Main() {
MyAbstractType<Sub> sub = new MyAbstractType<Sub>();
MyAbstractType<Super> super = sub;
super.MyMethod3("hello"); // calls sub.MyMethod3();
super.field = 10; // sets sub.field
}
}
This isn't as good in my opinion, the other version of MyGeneralType is a more straighforward layer over the concrete types, plus it doesn't require rewriting the data fields, but it does actually answer the question, technically. Try it here: https://dotnetfiddle.net/S3r3ke
Example
Using these abstract classes:
public abstract class Animal {
public string name;
public Animal(string name) {
this.name = name;
}
public abstract string Sound();
}
public abstract class AnimalHouse<T> where T : Animal {
List<T> animals;
public AnimalHouse(T[] animals) {
this.animals = animals.ToList();
}
public static implicit operator GeneralAnimalHouse(AnimalHouse<T> house) {
return GeneralAnimalHouse.Create(house);
}
public List<string> HouseSounds() {
return animals.Select(animal => animal.Sound()).ToList();
}
}
We make this "general" variant:
public delegate List<string> HouseSoundsDel();
public class GeneralAnimalHouse {
public HouseSoundsDel HouseSounds;
public static GeneralAnimalHouse Create<T>(AnimalHouse<T> house) where T : Animal {
var general = new GeneralAnimalHouse();
general.HouseSounds = house.HouseSounds;
return general;
}
}
And finally with these inheritors:
public class Dog : Animal {
public Dog(string name) : base(name) {}
public override string Sound() {
return name + ": woof";
}
}
public class Cat : Animal {
public Cat(string name) : base(name) {}
public override string Sound() {
return name + ": meow";
}
}
public class DogHouse : AnimalHouse<Dog> {
public DogHouse(params Dog[] dogs) : base(dogs) {}
}
public class CatHouse : AnimalHouse<Cat> {
public CatHouse(params Cat[] cats) : base(cats) {}
}
We use it like this:
public class AnimalCity {
List<GeneralAnimalHouse> houses;
public AnimalCity(params GeneralAnimalHouse[] houses) {
this.houses = houses.ToList();
}
public List<string> CitySounds() {
var random = new Random();
return houses.SelectMany(house => house.HouseSounds())
.OrderBy(x => random.Next())
.ToList();
}
}
public class MainClass {
public static void Main() {
var fluffy = new Cat("Fluffy");
var miu = new Cat("Miu");
var snuffles = new Cat("Snuffles");
var snoopy = new Dog("Snoopy");
var marley = new Dog("Marley");
var megan = new Dog("Megan");
var catHouse = new CatHouse(fluffy, miu, snuffles);
var dogHouse = new DogHouse(snoopy, marley, megan);
var animalCity = new AnimalCity(catHouse, dogHouse);
foreach (var sound in animalCity.CitySounds()) {
Console.WriteLine(sound);
}
}
}
Output:
Miu: meow
Snoopy: woof
Snuffles: meow
Fluffy: meow
Marley: woof
Megan: woof
Notes:
I added names so it's clear that the method references carry their owner's data with them, for those unfamiliar with delegates.
The required using statements for this code are System, System.Collections.Generic, and System.Linq.
You can try it here: https://dotnetfiddle.net/6qkHL3#
A version that makes GeneralAnimalHouse a subclass of AnimalHouse<Animal> can be found here: https://dotnetfiddle.net/XS0ljg