We Keep Coding

Implicit conversions not considered in overload resolution?

I'm trying to wrap a type (outside of my control) so that it would seamlessly appear to implement an interface (also outside of my control).
Given these defintions
// External types. Not changable.
class Foo {
public int I { get; set; }
public int J { get; set; }
}
interface IGenerateSignature {
string Generate();
}
I would like to use a Foo instance to call a method with an IGenerateSignature parameter:
void Test() {
var foo = new Foo { I = 1, J = 2 };
GetSignature(foo);
}
void GetSignature(IGenerateSignature sig) {
Console.Write(sig.Generate());
}
I tried creating an intermediary struct like this:
struct FooSignaturizer : IGenerateSignature {
private readonly Foo _foo;
public FooSignaturizer(Foo f) {
_foo = f;
}
public static implicit operator FooSignaturizer(Foo f) {
return new FooSignaturizer(f);
}
public string Generate() {
return _foo.I + ":" + _foo.J;
}
}
But for some reason overload resolution fails to find the conversion from Foo to FooSignaturizer, and I get a "Cannot convert" compiler error. If I manually add a cast, GetSignature((FooSignaturizer) foo), it works. However, I need to also add support for the Bar and Qux types, with BarSignaturizer and QuxSignaturizer, so the cast won't work for those cases.
Is there a way to accomplish this?

As per 7.5.3.1 of the C# spec, only implicit conversions from argument expression to parameter type are considered.
7.5.3.1 Applicable function member
A function member is said to be an applicable function member with respect to an argument list A when all of the following are true:
Each argument in A corresponds to a parameter in the function member declaration as described in §7.5.1.1, and any parameter to which no argument corresponds is an optional parameter.
For each argument in A, the parameter passing mode of the argument (i.e., value, ref, or out) is identical to the parameter passing mode of the corresponding parameter, and
for a value parameter or a parameter array, an implicit conversion (§6.1) exists from the argument to the type of the corresponding parameter, or
for a ref or out parameter, the type of the argument is identical to the type of the corresponding parameter. After all, a ref or out parameter is an alias for the argument passed.
What you have here isn't an implicit conversion from Foo to IGenereateSignature, it's a wrapper.
As an explanation for this behaviour, you can't exect the compiler to go though every implementation of IGenerateSignature in scope to see whether it has an implicit conversion to/from Foo. What if there was more than one?
In terms of how you can achieve this for Foo, Bar and Qux...
What you're trying to achieve, one call to GetSignature(fooOrBarOrQux), isn't possible, because (based on your description of Foo) you can't have one variable that can be a Foo or a Bar or a Qux at compile time - they're unrelated. You'll always need three call sites, so there's no reason not to have three slightly-different conversions (wrapper class or overloaded method call or something) for the three cases.
... unless you use dynamic?

Rawling's answer gives a good explanation of why you have the problem. Since you can't fix this with implicit conversion, you could try extension methods to convert all types to IGenerateSignature like this:
void Test() {
var foo = new Foo { I = 1, J = 2 };
GetSignature(foo.AsIGenerateSignature());
}
void GetSignature(IGenerateSignature sig) {
Console.Write(sig.Generate());
}
public static class GenerateSignatureExtensions
{
public static IGenerateSignature AsIGenerateSignature(this IGenerateSignature me)
{
return me;
}
public static IGenerateSignature AsIGenerateSignature(this Foo me)
{
return new FooSignaturizer(me);
}
public static IGenerateSignature AsIGenerateSignature(this Bar me)
{
return new BarSignaturizer(me);
}
//....
}

Rawling's answer gives a great explanation of the why you are having a problem. As to how to achieve what you want. I might consider something like this:
public interface ISignaturizer
{
IGenerateSignature ToSignaturizer();
}
struct FooSignaturizer : IGenerateSignature, ISignaturizer{
private readonly Foo _foo;
public FooSignaturizer(Foo f) {
_foo = f;
}
public string Generate() {
return _foo.I + ":" + _foo.J;
}
public IGenerateSignature ToSignaturizer()
{
return (IGenerateSignature)this;
}
}
Now BarSignaturizer and QuxSignaturizer can implement the same interface. And then you can do:
GetSignature(((ISignaturizer)fooOrBarOrQux).ToSignaturizer());
Which isn't quite as elegant, but I think should accomplish what you need.

Return instance using reflection in C#

A sample code I tried to return an instance of class is given below.
public object getConstructorclass(int i)
{
if(i==1)
{
Type type = Type.GetType("test1");
}else
{
Type type = Type.GetType("test2");
}
return Activator.CreateInstance(type);
}
var objcls = getConstructorclass(1);
objcls.callclass();//error occured
How can I mention the class type here since the type is not known at compile time but it will decided at runtime.In the above example i just pass a value 1 (it can be anything and that class will be called accordingly), and the class test1 called.
here I will get an error on the line objcls.callclass(), because objcls is an object instance that doesn't have a callclass()method.
How can I restructure this piece of code? My aim is if I mention a class in the getConstructorclass() method, an object should be returned so as to use it in the further code to invoke the members of that class.

If you know that your classes will have this method, you should use a common interface for them and implement it accordingly. Then you will work with classes that you have made sure it will work.
It would look like this
IMyInterface objcls = getconstrorclass() as IMyInterface;
if (objcls != null)
objcls.callclass();
else
// we failed miserably and should do something about it
I don't think you should use some generic object returning constructor based on an int variable, if your classes don't have anything in common. It's really weird to handle it like this and it may lead to various problems (some of which you're currently already experiencing). Generic class constructors make sense if the classes are somewhat related and you can predict the outcome, but to create a do-it-all method.. Not so sure about correctness of such approach.
Anyway, if you insist (not recommended, but as you wish), you can create some checks for a type like this:
var createdObject = getConstructorclass(1);
if (createdObject is MyClass1)
{
var specificObject = (MyClass1)createdObject;
specificObject.callMethod1();
}
else if (createdObject is MyClass2)
{
var specificObject = (MyClass2)createdObject;
specificObject.callSomeOtherMethod();
}
...
But it gets very error prone soon, refactoring will probably be a nightmare etc., but it's your call..
Or you maybe can use solution from pwas, but to me it seems unnecessarily complicated for such a basic task. Looks nice and all, but it still returns only the type "object", so it doesn't really solve your specific problem.
Also, to address one issue I'm not sure you understand - you've already created the instance, you just return type object. That is why you can't call any specific methods on this object, because first you have to cast it to something, that actually has that method and make sure the cast can be done (inheritance etc).

If interface solution (see other answers) is enough, don't look at this answer. When you can't use common base class / interface and you still want call members, you can use solution with is keyword (and check types). Instead of writing many ifs for each case, you can use fluent API:
object obj = this.getConstructorclass();
obj.StronglyInvoke()
.When<int>(value => Console.WriteLine("Got {0} as int", value))
.When<string>(value => Console.WriteLine("Got {0} as string", value))
.OnFail(() => Debug.Write("No handle."))
.Invoke();
Solution:
public class GenericCaller
{
private IList<GenericInvoker> invokers = new List<GenericInvoker>();
private readonly object target;
private Action failAction;
public GenericCaller(object target)
{
if (target == null)
{
throw new ArgumentNullException("target");
}
this.target = target;
}
public GenericCaller OnFail(Action fail)
{
this.failAction = fail;
return this;
}
public GenericCaller When<T>(Action<T> then)
{
if (then == null)
{
throw new ArgumentNullException("then");
}
var invoker = new GenericInvoker<T>(this.target, then);
this.invokers.Add(invoker);
return this;
}
public void Invoke()
{
if (this.invokers.Any(invoker => invoker.Invoke()))
{
return;
}
if (this.failAction == null)
{
throw new InvalidOperationException("Handler not found");
}
this.failAction();
}
public abstract class GenericInvoker
{
protected readonly object target;
protected GenericInvoker(object target)
{
this.target = target;
}
public abstract bool Invoke();
}
public class GenericInvoker<T> : GenericInvoker
{
private readonly Action<T> then;
public GenericInvoker(object target, Action<T> then)
: base(target)
{
this.then = then;
}
public override bool Invoke()
{
if (this.target.GetType() == typeof(T))
{
this.then((T)this.target);
return true;
}
return false;
}
}
}
public static class Extensions
{
public static GenericCaller StronglyInvoke(this object o)
{
return new GenericCaller(o);
}
}
Remeber - it would be more elegant to use common interface (as other answers say) - my is only alternative way.

Declare your variable as dynamic
dynamic objcls = getconstrorclass();
Using this the will be determined at run-time, whatever the getconstrorclass method returns. You can access any member of the type and you won't get any error at compile-time. But if you try to access a member which doesn't exists you will get a RuntimeBinderException at runtime.

I would recommend using an interface and restricting the classes that you can instantiate this way to only those that implement the interface.
public interface IMyInterface
{
void callclass();
}
public <T> getConstructorClass()
{
T instance;
Type type = Type.GetType("test1");
// instance will be null if the object cannot be cast to type T.
instance = Activator.CreateInstance(type) as T;
return T;
}
IMyInterface objcls = getConstructorClass<IMyInterface>();
if(null != objcls)
{
objcls.callclass();
}

not sure what you want to achieve in the end, but this looks like a job for "Dependency Injection" - here is a nice sample using autofac

C# Explicit Operators and Inheritance

I'm sure this is a stupid question, but why does the following code not call the explicit operator for the cast on the child class MyBool?
public class DataType
{
public static explicit operator bool(DataType D)
{
return false;
}
public static explicit operator DataType(bool B)
{
return new DataType();
}
}
public class MyBool : DataType
{
public bool Value;
public MyBool()
{
Value = false;
}
public static explicit operator bool(MyBool B)
{
return B.Value;
}
public static explicit operator MyBool(bool B)
{
return new MyBool() { Value = B };
}
}
then:
List<DataType> Types = new List<DataType>();
Types.Add(new MyBool() { Value = true });
Types.Add(new MyBool() { Value = false });
foreach (DataType T in Types)
{
bool Value = (bool)T;
MessageBox.Show(Value.ToString());
}
Produces the output: false, false
Is my only option to write functions on each class to take the place of the explicit operator functions?

why does the following code not call the explicit operator for the cast on the child class MyBool?
Because the operator functions are static, hence also non-virtual and thus their target is resolved at compile time rather than runtime. This is the expected behaviour.
If you want to have polymorphic conversion operators you can call virtual functions inside the operators:
public abstract class DataType
{
public static explicit operator bool(DataType D)
{
return D.DoCastToBool();
}
public static explicit operator DataType(bool B)
{
// We haven’t got an instance of our class here.
// You can use a factory method pattern to emulate virtual constructors.
}
protected abstract bool DoCastToBool();
}

Operators are overloaded rather than overridden - in other words, the choice about which implementation to use is made at compile-time. The compiler only knows about T as DataType, so it calls the operator in DataType.
One option would be to remove the operator from MyBool, but add a virtual method in DataType, allowing for polymorphic behaviour:
public class DataType
{
public static explicit operator bool(DataType D)
{
// TODO: Decide how you want to handle null references
return D.ToBoolean();
}
protected virtual bool ToBoolean()
{
return false;
}
}
public class MyBool : DataType
{
// ...
protected override bool ToBoolean()
{
return Value;
}
}
Note that this won't work for the conversion from bool to a DataType, as in that case we don't have any information about which subtype of DataType you actually want to create.
(Side-note: your code would be easier to follow if you used the normal .NET naming conventions.)

Here's a garbage solution for you:
replace: bool Value = (bool)T;
with: bool Value = (bool)(T as MyBool);

C# Generic and method

How can I select the good method (I have in the example below show 2 differents way that doesn't work). I was using instead of a variable of type Object with a IF and IS to do the job but I am trying to avoid using Object and boxing/unboxing. So I thought that Generic could do the job but I am stuck here.
Here is a small snippet of code that illustrate my question:
class Program
{
static void Main(string[] args)
{
Parser p = new Parser();
ObjectType1 o1 = new ObjectType1();
p.execute(o1);
Console.Read();
}
}
class Parser
{
public T execute<T>(T obj)
{
/*
if (obj is ObjectType1)
this.action((ObjectType1)obj);
else if (obj is ObjectType2)
this.action((ObjectType2)obj);
*/
this.action(obj);
return obj;
}
private void action(ObjectType1 objectType1)
{
Console.WriteLine("1");
}
private void action(ObjectType2 objectType2)
{
Console.WriteLine("2");
}
}
class ObjectType1
{
}
class ObjectType2
{
}
Update
I do not want interface and class. Sorry. I knew that it's not the goal of the question.
Casting with (ObjectType)obj doesn't work but if you do :
if (obj is ObjectType1)
this.action(obj as ObjectType1);
else if (obj is ObjectType2)
this.action(obj as ObjectType1);
it works... why?
And... I cannot overload for all type the execute method because this method is from an Interface. This is why all need to be called from this method.

No, you can't do this. Generics don't work like C++ templates - the generic method is compiled just once. The only information that the compiler can use for overload resolution is the information it knows about within the generic method, regardless of what code uses it.
As an example to show this, here's a bit of code which may not work how you expect it to:
using System;
class Test
{
static void Main()
{
string x = "hello";
string y = string.Copy(x);
Console.WriteLine(x==y); // Overload used
Compare(x, y);
}
static void Compare<T>(T x, T y) where T : class
{
Console.WriteLine(x == y); // Reference comparison
}
}
It's hard to say the best way to proceed without knowing more about what you want to do.

Have you considered interfaces?
interface IAction
{
void action();
}
class ObjectType1 : IAction
{
void action() {
Console.WriteLine("1");
}
}
class ObjectType2 : IAction
{
void action() {
Console.WriteLine("2");
}
}
class Parser
{
public IAction execute(IAction obj)
{
obj.action();
return obj;
}
}
Edited by OP:
This solution would require to change all Business Logic Object to have this interface. This is really not a thing to do (in my situation). And, in other situation, I always prefer to have clean BusinessObject that doesn't have Interface not related with Business stuff. In my question, I want a solution that is more related with Generic/Object/Delegate method to achieve it. Thx you. This answer won't be accepted.

The class Parser has a lot of private method that are called by the execute method depending of the object type. It needs to redirect to the good method.
The compiler will do this work for you. Just use overloads.
class Parser
{
public ObjectType1 action(ObjectType1 objectType1)
{
Console.WriteLine("1");
return objectType1;
}
public ObjectType2 action(ObjectType2 objectType2)
{
Console.WriteLine("2");
return objectType2;
}
}
class ObjectType1 { }
struct ObjectType2 { }
Then, called with:
Parser p = new Parser();
p.action(new ObjectType1());
p.action(new ObjectType2());
There's no boxing/unboxing, and the appropriate method gets called.

I haven't tried it, but can you do this?
public T execute<T>(T obj)
{
this.action((T)obj);
return obj;
}
(according to comments, doesn't work)
or
public T execute<T>(T obj)
{
this.action(obj as T);
return obj;
}
(according to comments, works)

I know you're concerned about boxing/unboxing, so there could be ValueTypes involved here.
public T execute<T>(T obj)
{
this.action(obj);
return obj;
}
Supposing that action is modifying obj, and also supposing that modification is important to the caller (which is why you're returning the value back to the caller). This code has a nasty pass-by-value defect.
Consider this code:
public int execute(int obj)
{
this.action(obj);
return obj;
}
public void action(int obj)
{
obj = obj + 1;
}
Called in this way.
int x = p.execute(1);
x is 1, not 2.

Generics happens in compile time. It is best used when you want the same code to apply to different types. It is not dynamic, so it won't help you switch between methods depending on input types.
Overloading resolving as in David B's reply works, but also happens during compile time.
The code in your update does the same thing. It casts (after careful checking of types) and then uses overloading to resolve the method.
I feel that you want to switch methods based on runtime input.
You could get a more dynamic behaviour if you used Reflection.
public object execute(object obj)
{
MethodInfo m = typeof(Parser).GetMethod(
"action",
BindingFlags.Instance | BindingFlags.NonPublic,
null,
new Type[] { obj.GetType() },
null);
m.Invoke(this, new object[] { obj });
return obj;
}
It is perhaps a little fragile, but it works in the example.

IIRC you can use the "where" clause to allow this
public T execute<T>(T obj) where : /* somthing */
{
}
I always have to Google that one my self so I'll leave it at that.
edit: reading some comments. I would not advise calling type specific code. Rather put that code in a virtual function and call that. The call signature might get long, but that's what auto complete is for.
Koodos to joshua.ewer for finding the man page

Generics in c# & accessing the static members of T

My question concerns c# and how to access Static members ... Well I don't really know how to explain it (which kind of is bad for a question isn't it?) I will just give you some sample code:
Class test<T>{
int method1(Obj Parameter1){
//in here I want to do something which I would explain as
T.TryParse(Parameter1);
//my problem is that it does not work ... I get an error.
//just to explain: if I declare test<int> (with type Integer)
//I want my sample code to call int.TryParse(). If it were String
//it should have been String.TryParse()
}
}
So thank you guys for your answers (By the way the question is: how would I solve this problem without getting an error). This probably quite an easy question for you!
Edit: Thank you all for your answers!
Though I think the try - catch phrase is the most elegant, I know from my experience with vb that it can really be a bummer. I used it once and it took about 30 minutes to run a program, which later on only took 2 minutes to compute just because I avoided try - catch.
This is why I chose the switch statement as the best answer. It makes the code more complicated but on the other hand I imagine it to be relatively fast and relatively easy to read. (Though I still think there should be a more elegant way ... maybe in the next language I learn)
Though if you have some other suggestion I am still waiting (and willing to participate)

The problem is that TryParse isn't defined on an interface or base class anywhere, so you can't make an assumption that the type passed into your class will have that function. Unless you can contrain T in some way, you'll run into this a lot.
Constraints on Type Parameters

Short answer, you can't.
Long answer, you can cheat:
public class Example
{
internal static class Support
{
private delegate bool GenericParser<T>(string s, out T o);
private static Dictionary<Type, object> parsers =
MakeStandardParsers();
private static Dictionary<Type, object> MakeStandardParsers()
{
Dictionary<Type, object> d = new Dictionary<Type, object>();
// You need to add an entry for every type you want to cope with.
d[typeof(int)] = new GenericParser<int>(int.TryParse);
d[typeof(long)] = new GenericParser<long>(long.TryParse);
d[typeof(float)] = new GenericParser<float>(float.TryParse);
return d;
}
public static bool TryParse<T>(string s, out T result)
{
return ((GenericParser<T>)parsers[typeof(T)])(s, out result);
}
}
public class Test<T>
{
public static T method1(string s)
{
T value;
bool success = Support.TryParse(s, out value);
return value;
}
}
public static void Main()
{
Console.WriteLine(Test<int>.method1("23"));
Console.WriteLine(Test<float>.method1("23.4"));
Console.WriteLine(Test<long>.method1("99999999999999"));
Console.ReadLine();
}
}
I made a static dictionary holding a delegate for the TryParse method of every type I might want to use. I then wrote a generic method to look up the dictionary and pass on the call to the appropriate delegate. Since every delegate has a different type, I just store them as object references and cast them back to the appropriate generic type when I retrieve them. Note that for the sake of a simple example I have omitted error checking, such as to check whether we have an entry in the dictionary for the given type.

To access a member of a specific class or interface you need to use the Where keyword and specify the interface or base class that has the method.
In the above instance TryParse does not come from an interface or base class, so what you are trying to do above is not possible. Best just use Convert.ChangeType and a try/catch statement.
class test<T>
{
T Method(object P)
{
try {
return (T)Convert.ChangeType(P, typeof(T));
} catch(Exception e) {
return null;
}
}
}

One more way to do it, this time some reflection in the mix:
static class Parser
{
public static bool TryParse<TType>( string str, out TType x )
{
// Get the type on that TryParse shall be called
Type objType = typeof( TType );
// Enumerate the methods of TType
foreach( MethodInfo mi in objType.GetMethods() )
{
if( mi.Name == "TryParse" )
{
// We found a TryParse method, check for the 2-parameter-signature
ParameterInfo[] pi = mi.GetParameters();
if( pi.Length == 2 ) // Find TryParse( String, TType )
{
// Build a parameter list for the call
object[] paramList = new object[2] { str, default( TType ) };
// Invoke the static method
object ret = objType.InvokeMember( "TryParse", BindingFlags.InvokeMethod, null, null, paramList );
// Get the output value from the parameter list
x = (TType)paramList[1];
return (bool)ret;
}
}
}
// Maybe we should throw an exception here, because we were unable to find the TryParse
// method; this is not just a unable-to-parse error.
x = default( TType );
return false;
}
}
The next step would be trying to implement
public static TRet CallStaticMethod<TRet>( object obj, string methodName, params object[] args );
With full parameter type matching etc.

This isn't really a solution, but in certain scenarios it could be a good alternative: We can pass an additional delegate to the generic method.
To clarify what I mean, let's use an example. Let's say we have some generic factory method, that should create an instance of T, and we want it to then call another method, for notification or additional initialization.
Consider the following simple class:
public class Example
{
// ...
public static void PostInitCallback(Example example)
{
// Do something with the object...
}
}
And the following static method:
public static T CreateAndInit<T>() where T : new()
{
var t = new T();
// Some initialization code...
return t;
}
So right now we would have to do:
var example = CreateAndInit<Example>();
Example.PostInitCallback(example);
However, we could change our method to take an additional delegate:
public delegate void PostInitCallback<T>(T t);
public static T CreateAndInit<T>(PostInitCallback<T> callback) where T : new()
{
var t = new T();
// Some initialization code...
callback(t);
return t;
}
And now we can change the call to:
var example = CreateAndInit<Example>(Example.PostInitCallback);
Obviously this is only useful in very specific scenarios. But this is the cleanest solution in the sense that we get compile time safety, there is no "hacking" involved, and the code is dead simple.

Do you mean to do something like this:
Class test<T>
{
T method1(object Parameter1){
if( Parameter1 is T )
{
T value = (T) Parameter1;
//do something with value
return value;
}
else
{
//Parameter1 is not a T
return default(T); //or throw exception
}
}
}
Unfortunately you can't check for the TryParse pattern as it is static - which unfortunately means that it isn't particularly well suited to generics.

The only way to do exactly what you're looking for would be to use reflection to check if the method exists for T.
Another option is to ensure that the object you send in is a convertible object by restraining the type to IConvertible (all primitive types implement IConvertible). This would allow you to convert your parameter to the given type very flexibly.
Class test<T>
{
int method1(IConvertible Parameter1){
IFormatProvider provider = System.Globalization.CultureInfo.CurrentCulture.GetFormat(typeof(T));
T temp = Parameter1.ToType(typeof(T), provider);
}
}
You could also do a variation on this by using an 'object' type instead like you had originally.
Class test<T>
{
int method1(object Parameter1){
if(Parameter1 is IConvertible) {
IFormatProvider provider = System.Globalization.CultureInfo.CurrentCulture.GetFormat(typeof(T));
T temp = Parameter1.ToType(typeof(T), provider);
} else {
// Do something else
}
}
}

Ok guys: Thanks for all the fish. Now with your answers and my research (especially the article on limiting generic types to primitives) I will present you my solution.
Class a<T>{
private void checkWetherTypeIsOK()
{
if (T is int || T is float //|| ... any other types you want to be allowed){
return true;
}
else {
throw new exception();
}
}
public static a(){
ccheckWetherTypeIsOK();
}
}

You probably cant do it.
First of all if it should be possible you would need a tighter bound on T so the typechecker could be sure that all possible substitutions for T actually had a static method called TryParse.

You may want to read my previous post on limiting generic types to primitives. This may give you some pointers in limiting the type that can be passed to the generic (since TypeParse is obviously only available to a set number of primitives ( string.TryParse obviously being the exception, which doesn't make sense).
Once you have more of a handle on the type, you can then work on trying to parse it. You may need a bit of an ugly switch in there (to call the correct TryParse ) but I think you can achieve the desired functionality.
If you need me to explain any of the above further, then please ask :)

Best code: restrict T to ValueType this way:
class test1<T> where T: struct
A "struct" here means a value type.
String is a class, not a value type.
int, float, Enums are all value types.
btw the compiler does not accept to call static methods or access static members on 'type parameters' like in the following example which will not compile :(
class MyStatic { public static int MyValue=0; }
class Test<T> where T: MyStatic
{
public void TheTest() { T.MyValue++; }
}
=> Error 1 'T' is a 'type parameter', which is not valid in the given context
SL.

That is not how statics work. You have to think of statics as sort of in a Global class even if they are are spread across a whole bunch of types. My recommendation is to make it a property inside the T instance that can access the necessary static method.
Also T is an actual instance of something, and just like any other instance you are not able to access the statics for that type, through the instantiated value. Here is an example of what to do:
class a {
static StaticMethod1 ()
virtual Method1 ()
}
class b : a {
override Method1 () return StaticMethod1()
}
class c : a {
override Method1 () return "XYZ"
}
class generic<T>
where T : a {
void DoSomething () T.Method1()
}