We Keep Coding

How to Build Expression Tree for Invoke Method from Generic Interface

I'm sorry to ask, how can i invoke method using other that method.Invoke, because some article said, method.Invoke has slower performance.
Actually i'm using .NET Core 3.1.
Example, i have a structure code something like this.
public class Message { }
public class MyMessageType : Message { }
public class MyMessageResult
{
public bool Status { get; set; }
public DateTime Date => DateTime.Now;
}
public interface IEncapsulatedMessageHandlerV2<T, TResult> where T : class
{
Task<TResult> HandleMessageResultAsync(T message);
}
public abstract class EncasulatedMessageHandlerV2<T, TResult> : IEncapsulatedMessageHandlerV2<T, TResult> where T : Message
{
public abstract Task<TResult> HandleMessageResultExecAsync(T message);
async Task<TResult> IEncapsulatedMessageHandlerV2<T, TResult>.HandleMessageResultAsync(T message)
{
var msg = message as T;
if (msg != null)
return await HandleMessageResultExecAsync(msg);
return default;
}
}
public class HandlerV2 : EncasulatedMessageHandlerV2<MyMessageType, MyMessageResult>
{
public override Task<MyMessageResult> HandleMessageResultExecAsync(MyMessageType message)
{
Console.WriteLine("Yo Async!");
return Task.FromResult(new MyMessageResult
{
Status = true
});
}
}
And i can successfully call using method.Invoke
static TResponse UsingMethodInvoke<TResponse>()
{
// Assume, i was build this using MakeGenericMethod
var type = typeof(IEncapsulatedMessageHandlerV2<MyMessageType, MyMessageResult>);
var typeActivator = typeof(HandlerV2);
var instance = Activator.CreateInstance(typeActivator);
var method = type.GetMethod("HandleMessageResultAsync");
var tsk = (Task<TResponse>)method.Invoke(instance, new[] { new MyMessageType() });
var result = tsk.GetAwaiter().GetResult();
return result;
}
And i try to using Dynamic too, unfortunately they can't calling through abstract HandleMessageResultAsync instead only through implemented class HandleMessageResultExecAsync
static TResponse UsingDynamicInvoke<TResponse>()
{
// Assume, i was build this using MakeGenericMethod
var typeActivator = typeof(HandlerV2);
var instance = Activator.CreateInstance(typeActivator);
var tsk = (Task<TResponse>)((dynamic)instance).HandleMessageResultExecAsync(new MyMessageType());
var result = tsk.GetAwaiter().GetResult();
return result;
}
And i was follow to with stackoverflow Speeding up Reflection Invoke C#/.NET, and i get stuck
static void ActivatorMyMessageResultAsnc()
{
var type = typeof(HandlerV2);
var instance = Activator.CreateInstance(type);
var method = type.GetMethod("HandleMessageResultAsync", BindingFlags.Instance | BindingFlags.Public);
var originalType = type;
// Loop until we hit the type we want.
while (!(type.IsGenericType) || type.GetGenericTypeDefinition() != typeof(EncasulatedMessageHandlerV2<,>))
{
type = type.BaseType;
if (type == null)
throw new ArgumentOutOfRangeException("type");
}
var messageType = type.GetGenericArguments()[0]; // MyMessageType
// Use expression to create a method we can.
var instExpr = Expression.Parameter(typeof(object), "instance");
var paramExpr = Expression.Parameter(typeof(Message), "message");
// (Handler)instance;
var instCastExpr = Expression.Convert(instExpr, originalType);
// (MyMessageType)message
var castExpr = Expression.Convert(paramExpr, messageType);
// ((Handler)inst).HandleMessage((MyMessageType)message)
var invokeExpr = Expression.Call(instCastExpr, method, castExpr); // <--- this give me error
// i'm stuck, i don't know what should i do next
////// Assume this is build from MakeGeneric too
////var delType = typeof(Func<object, Message, Task<MessageResult>>);
//var lambda = Expression.Lambda<Func<object, Message, Task<object>>>(invokeExpr, instExpr, paramExpr);
//var compiled = lambda.Compile();
//Func<Message, Task<object>> hook = x => compiled(instance, x);
Or, is there any other ways to Invoke method by dynamicaly, which is faster that method.Invoke
Thanks in advance,
PS: Sorry for my Bad English

I had a similar problem I was trying to solve a while back. I dug into the Net Core codebase and found a class which is really helpful for invoking methods at runtime without being strongly typed.
https://github.com/dotnet/extensions/blob/ff87989d893b000aac1bfef0157c92be1f04f714/shared/Microsoft.Extensions.ObjectMethodExecutor.Sources/ObjectMethodExecutor.cs
ObjectMethodExecutor lets you call a method on an object using its name. An example of usage is:
var methodInfo = [myInterface].GetMethod("[method-you-want-call]");
var classTypeInfo = [myClass].GetTypeInfo();
var executor = ObjectMethodExecutor.Create(methodInfo, classTypeInfo);
await executor.ExecuteAsync(handler, new[] { [data-you-want-to-pass-to-your-object] });
Its used extensively in the SignalR codebase for matching SignalR messages to the internal methods on a hub. Its heavily optimised and has the added benefit of allowing async methods to be called too

How to get MethodInfo for a generic method by name and its generic parameters?

I have this interface:
interface IRepository
{
string GetId<T>(T obj) where T : IDomainObject;
string GetId<T>(Reference<T> reference) where T : IDomainObject;
}
What exactly IDomainObject and Reference are is not relevant to this question. So assume they are completely empty:
interface IDomainObject
{
// something
}
class Reference<T> where T : IDomainObject
{
// something
}
My question is: How do I get the MethodInfo for the GetId<T> method in IRepository that accepts a Reference<T>?
Here's what I've tried:
public MethodInfo GetReferenceAcceptingGetIdMethod()
{
// We want to return the MethodInfo for the GetId<T> method of IRepository
// that accepts a Reference<T> argument.
var repositoryInterfaceType = typeof(IRepository);
// Look through all methods of IRepository...
foreach (var m in repositoryInterfaceType.GetMethods())
{
// ... to find a candidate method, going by Genericness and Name ...
if (m.IsGenericMethodDefinition && m.Name == nameof(IRepository.GetId))
{
// ... and to narrow it further down by looking at the parameters ...
var parameters = m.GetParameters();
if (parameters.Length == 1)
{
// ... to check if the one and only parameter is a generic Reference<>.
var firstParamType = parameters[0].ParameterType;
var genericReferenceType = typeof(Reference<>);
if (firstParamType == genericReferenceType)
{
// !!! This if will never be true.
// Why?
// And what do I have to change the condition into to make it work?
return m;
}
}
}
}
throw new Exception();
}
It appears that method's parameter type is somehow different from a completely open generic type. I guess and it seems that the type of the method's parameter is somehow linked to the method's generic type parameter.
So how can I get the MethodInfo in such a case?

The parameter type can't be the generic type definition. It has a type argument - which is the type parameter for the method. You can get that type parameter from MethodInfo.GetGenericArguments(), and then use it with typeof(Reference<>).MakeGenericType(...) to get the expected parameter type.
Here's a complete example, basically adapting your original code:
using System;
using System.Reflection;
class Reference<T> {}
interface IDomainObject {}
interface IRepository
{
string GetId<T>(T obj) where T : IDomainObject;
string GetId<T>(Reference<T> reference) where T : IDomainObject;
}
class Test
{
static void Main()
{
var method = GetReferenceAcceptingGetIdMethod();
Console.WriteLine(method);
}
public static MethodInfo GetReferenceAcceptingGetIdMethod()
{
var repositoryInterfaceType = typeof(IRepository);
foreach (var m in repositoryInterfaceType.GetMethods())
{
if (m.IsGenericMethodDefinition && m.Name == nameof(IRepository.GetId))
{
var typeParameter = m.GetGenericArguments()[0];
var expectedParameterType = typeof(Reference<>).MakeGenericType(typeParameter);
var parameters = m.GetParameters();
if (parameters.Length == 1)
{
var firstParamType = parameters[0].ParameterType;
if (firstParamType == expectedParameterType)
{
return m;
}
}
}
}
throw new Exception();
}
}

How can I convert C# methods to compiled expressions?

I have the following class hierarchy:
public class Parent
{
[DebuggerStepThrough]
public void SayParent()
{
Console.WriteLine("Parent");
}
}
public sealed class Child : Parent
{
private static int _number = 0;
public Child() // May contain parameter i.e. not always parameterless consctructor
{
_number++;
}
[DebuggerStepThrough]
public void SayInstance()
{
Console.WriteLine("{0}-Say", _number);
}
[DebuggerStepThrough]
public void SayInstanceWithArg(string input)
{
Console.WriteLine("{0}-Say: {1}", _number, input);
}
[DebuggerStepThrough]
public static void SayStatic()
{
Console.WriteLine("{0}-Say", _number);
}
[DebuggerStepThrough]
public static void SayStaticWithArg(string input)
{
Console.WriteLine("{0}-Say: {1}", _number, input);
}
[DebuggerStepThrough]
public static Task SayStaticWithArgAndReturn(string input)
{
Console.WriteLine("{0}-Say: {1}", _number, input);
return null;
}
}
I need to be able to invoke any of these methods for a new instance of Child at any given time using reflection however to improve performance I need to resort to Delegate and/or Compiled Expressions.
So for example I can have:
var instanceOne = new Child();
var instanceTwo = new Child();
for which I would need to at runtime invoke these methods passing the arguments for those that need it. Note they include both static and instance methods with some accepting a parameter.
I have so far tried the following for the "SayInstance" method:
var sayInstanceMethod = typeof(Child)
.GetMethods(BindingFlags.Public | BindingFlags.Instance | BindingFlags.Static)
.Where(m => m.GetCustomAttributes(typeof(DebuggerStepThroughAttribute), true).Length > 0)
.Where(t => t.Name == "SayInstance")
.First()
And then:
var instance = Expression.Constant(new Child()); // This should NOT be Constant, but then what should it be?!
var mCallInstance = Expression.Call(instance, sayInstanceMethod);
Action action = Expression.Lambda<Action>(mCallInstance).Compile();
action();
action(); // I need to pass in a new instance of Child to this method somehow
However I am getting:
1-Say
1-Say
instead of:
1-Say
2-Say
I suspect this is due to Expression.Constant but I cannot figure out how I could let it accept an instance of Child as its target at runtime.
I am hopeless when it comes to Expressions :-(
I am basically trying to implement what Jon Skeet mentions HERE either using Delegates or Compiled Expressions.
Any help is very much appreciated.

If I understood correctly, you need to use parameters, like this:
var instanceOne = new Child();
var instanceTwo = new Child();
var instance = Expression.Parameter(typeof(Child), "c"); // This should NOT be Constant, but then what should it be?!
var mCallInstance = Expression.Call(instance, sayInstanceMethod);
Action<Child> action = Expression.Lambda<Action<Child>>(mCallInstance, instance).Compile();
action(instanceOne);
action(instanceTwo); // I need to pass in a new instance of Child to this method somehow
Of course this will not output 1, 2 because your _number field is static and after creation of two instances has value 2 for both.
EDIT. If you need to call method with arguments - declare more parameters. For example if SayInstance has one argument of type string, then:
var instanceOne = new Child();
var instanceTwo = new Child();
var instance = Expression.Parameter(typeof(Child), "instance");
var arg = Expression.Parameter(typeof(string), "arg");
var mCallInstance = Expression.Call(instance, sayInstanceMethod, arg);
Action<Child,string> action = Expression.Lambda<Action<Child,string>>(mCallInstance, instance, arg).Compile();
action(instanceOne, "one");
action(instanceTwo, "two");

Try this out, this workf for me for a parameterless constructor, but this is that you need:
var instance = Expression.New(typeof(Child).GetConstructor(new Type[0]));
var mCallInstance = Expression.Call(instance, sayInstanceMethod);
Action action = Expression.Lambda<Action>(mCallInstance).Compile();
action();
action(); // I need to pass in a new instance of Child to this method someh

Add intellisense when writing lambda's for an Action<dynamic>

This question is related to this other question.
I have the following method:
public static T GetNewData<T>(params Action<dynamic>[] actions) where T : class, new()
{
dynamic dynamicData = new DeepObject();
foreach (var action in actions)
{
action(dynamicData);
}
return Converter.Convert<T>(dynamicData);
}
The users of this method will include less technical people, even non-developers and as such the easier writing calls to this method is the better. My sticking point right now is that by using Action<dynamic> as the parameter type there is no intellisense provided to the user. In the context I know that the intellisense should be acting as if the dynamic was in fact T.
So is their a way I could either: Tell Visual Studio to use type T for the intellisense or change the parameter to be Action<T> and somehow programmatically change it to be Action<dynamic> or Action<DeepObject> so that the call to it will succeed?
EDIT: To clarify, the types that I am using for T are not of type DeepObject and they do not inherit any standard interface, the use of DeepObject is to allow setting up nested types without the user needing to explicitly instantiate at each level. This was the original usage before adding the dynamic and DeepObject code:
ExampleDataFactory.GetNewData<ServicesAndFeaturesInfo>(
x => x.Property1 = ExampleDataFactory.GetNewData<Property1Type>(),
x => x.Property1.Property2 = ExampleDataFactory.GetNewData<Property2Type>(),
x => x.Property1.Property2.Property3 = ExampleDataFactory.GetNewData<Property3Type>(),
x => x.Property1.Property2.Property3.Property4 = true);
Here is what it looks like now:
ExampleDataFactory.GetNewData<ServicesAndFeaturesInfo>(
x => x.Property1.Property2.Property3.Property4 = true);
EDIT: Here is the fully implemented solution based on nmclean's answer
public static DataBuilder<T> GetNewData<T>() where T : class, new()
{
return new DataBuilder<T>();
}
The DataBuilder Class:
public class DataBuilder<T>
{
public readonly T data;
public DataBuilder()
{
data = Activator.CreateInstance<T>();
}
public DataBuilder(T data)
{
this.data = data;
}
public DataBuilder<T> SetValue<T2>(Expression<Func<T, T2>> expression, T2 value)
{
var mExpr = GetMemberExpression(expression);
var obj = Recurse(mExpr);
var p = (PropertyInfo)mExpr.Member;
p.SetValue(obj, value);
return this;
}
public T Build()
{
return data;
}
public object Recurse(MemberExpression expr)
{
if (expr.Expression.Type != typeof(T))
{
var pExpr = GetMemberExpression(expr.Expression);
var parent = Recurse(pExpr);
var pInfo = (PropertyInfo) pExpr.Member;
var obj = pInfo.GetValue(parent);
if (obj == null)
{
obj = Activator.CreateInstance(pInfo.PropertyType);
pInfo.SetValue(parent, obj);
}
return obj;
}
return data;
}
private static MemberExpression GetMemberExpression(Expression expr)
{
var member = expr as MemberExpression;
var unary = expr as UnaryExpression;
return member ?? (unary != null ? unary.Operand as MemberExpression : null);
}
private static MemberExpression GetMemberExpression<T2>(Expression<Func<T, T2>> expr)
{
return GetMemberExpression(expr.Body);
}
}
The Usage:
ExampleDataFactory.GetNewData<ServicesAndFeaturesInfo>()
.SetValue(x=> x.Property1.EnumProperty, EnumType.Own)
.SetValue(x=> x.Property2.Property3.Property4.BoolProperty, true)
.Build();

Do not use Action<dynamic>, use Action<T>with method's constraint where T:DeepObject. Users will get intellisence and ability to use strongly typed objects:
public static DeepObject GetNewData<T>(params Action<T>[] actions)
where T : DeepObject, //restrict user only inheritors of DeepObject
new() //and require constructor
{
var data = new T();
foreach (var action in actions)
{
action(data);
}
return data;
}

Does the user need to access unknown properties or add new ones? If not, using dynamic objects seems like a step backwards. If your desired syntax does compile as an Action<T>, I think you should just declare it that way and then go with your first instinct of using the LINQ Expression API to decide how to interpret the code.
Unfortunately, although statements, such as an assignment, are part of the API, C# doesn't support converting them to expression trees. This is not allowed:
public static T GetNewData<T>(params Expression<Action<T>>[] actions)
where T : class, new() {
...
}
...
ExampleDataFactory.GetNewData<ServicesAndFeaturesInfo>(
x => x.Property1.Property2.Property3.Property4 = true);
Only single-line expressions that would have a return a value are supported. So I think the best you could do is something like this:
public class Assignment<T> {
public readonly Expression Expression;
public readonly object Value;
public Assignment(Expression<Func<T, object>> expression, object value) {
Expression = expression;
Value = value;
}
}
...
public static T GetNewData<T>(params Assignment<T>[] assignments)
where T : class, new() {
var data = Activator.CreateInstance<T>();
foreach (var assignment in assignments) {
// todo:
// - pull property names from assignment.Expression
// - initialize nested properties / assign to assignment.Value
}
return data;
}
...
ExampleDataFactory.GetNewData<ServicesAndFeaturesInfo>(
new Assignment<ServicesAndFeaturesInfo>(
x => x.Property1.Property2.Property3.Property4, true));
Getting the property names from an expression tree of chained property access is not too complicated. Here is one implementation.
Of course, the new Assignment<ServicesAndFeaturesInfo>(...) is ugly and repetitive, so maybe it could be restructured to something like this:
var newData = ExampleDataFactory.NewData<ServicesAndFeaturesInfo>();
newData.Add(x => x.Property1.Property2.Property3.Property4, true);
newData.Add(...);
...
newData.Get();

Create instance of generic type whose constructor requires a parameter?

If BaseFruit has a constructor that accepts an int weight, can I instantiate a piece of fruit in a generic method like this?
public void AddFruit<T>()where T: BaseFruit{
BaseFruit fruit = new T(weight); /*new Apple(150);*/
fruit.Enlist(fruitManager);
}
An example is added behind comments. It seems I can only do this if I give BaseFruit a parameterless constructor and then fill in everything through member variables. In my real code (not about fruit) this is rather impractical.
-Update-
So it seems it can't be solved by constraints in any way then. From the answers there are three candidate solutions:
Factory Pattern
Reflection
Activator
I tend to think reflection is the least clean one, but I can't decide between the other two.

Additionally a simpler example:
return (T)Activator.CreateInstance(typeof(T), new object[] { weight });
Note that using the new() constraint on T is only to make the compiler check for a public parameterless constructor at compile time, the actual code used to create the type is the Activator class.
You will need to ensure yourself regarding the specific constructor existing, and this kind of requirement may be a code smell (or rather something you should just try to avoid in the current version on c#).

You can't use any parameterised constructor. You can use a parameterless constructor if you have a "where T : new()" constraint.
It's a pain, but such is life :(
This is one of the things I'd like to address with "static interfaces". You'd then be able to constrain T to include static methods, operators and constructors, and then call them.

Yes; change your where to be:
where T:BaseFruit, new()
However, this only works with parameterless constructors. You'll have to have some other means of setting your property (setting the property itself or something similar).

Most simple solution
Activator.CreateInstance<T>()

As Jon pointed out this is life for constraining a non-parameterless constructor. However a different solution is to use a factory pattern. This is easily constrainable
interface IFruitFactory<T> where T : BaseFruit {
T Create(int weight);
}
public void AddFruit<T>( IFruitFactory<T> factory ) where T: BaseFruit {
BaseFruit fruit = factory.Create(weight); /*new Apple(150);*/
fruit.Enlist(fruitManager);
}
Yet another option is to use a functional approach. Pass in a factory method.
public void AddFruit<T>(Func<int,T> factoryDel) where T : BaseFruit {
BaseFruit fruit = factoryDel(weight); /* new Apple(150); */
fruit.Enlist(fruitManager);
}

You can do by using reflection:
public void AddFruit<T>()where T: BaseFruit
{
ConstructorInfo constructor = typeof(T).GetConstructor(new Type[] { typeof(int) });
if (constructor == null)
{
throw new InvalidOperationException("Type " + typeof(T).Name + " does not contain an appropriate constructor");
}
BaseFruit fruit = constructor.Invoke(new object[] { (int)150 }) as BaseFruit;
fruit.Enlist(fruitManager);
}
EDIT: Added constructor == null check.
EDIT: A faster variant using a cache:
public void AddFruit<T>()where T: BaseFruit
{
var constructor = FruitCompany<T>.constructor;
if (constructor == null)
{
throw new InvalidOperationException("Type " + typeof(T).Name + " does not contain an appropriate constructor");
}
var fruit = constructor.Invoke(new object[] { (int)150 }) as BaseFruit;
fruit.Enlist(fruitManager);
}
private static class FruitCompany<T>
{
public static readonly ConstructorInfo constructor = typeof(T).GetConstructor(new Type[] { typeof(int) });
}

As an addition to user1471935's suggestion:
To instantiate a generic class by using a constructor with one or more parameters, you can now use the Activator class.
T instance = Activator.CreateInstance(typeof(T), new object[] {...})
The list of objects are the parameters you want to supply. According to Microsoft:
CreateInstance [...] creates an instance of the specified type using the constructor that best matches the specified parameters.
There's also a generic version of CreateInstance (CreateInstance<T>()) but that one also does not allow you to supply constructor parameters.

I created this method:
public static V ConvertParentObjToChildObj<T,V> (T obj) where V : new()
{
Type typeT = typeof(T);
PropertyInfo[] propertiesT = typeT.GetProperties();
V newV = new V();
foreach (var propT in propertiesT)
{
var nomePropT = propT.Name;
var valuePropT = propT.GetValue(obj, null);
Type typeV = typeof(V);
PropertyInfo[] propertiesV = typeV.GetProperties();
foreach (var propV in propertiesV)
{
var nomePropV = propV.Name;
if(nomePropT == nomePropV)
{
propV.SetValue(newV, valuePropT);
break;
}
}
}
return newV;
}
I use that in this way:
public class A
{
public int PROP1 {get; set;}
}
public class B : A
{
public int PROP2 {get; set;}
}
Code:
A instanceA = new A();
instanceA.PROP1 = 1;
B instanceB = new B();
instanceB = ConvertParentObjToChildObj<A,B>(instanceA);

You can use the following command:
T instance = (T)typeof(T).GetConstructor(new Type[0]).Invoke(new object[0]);
Be sure to see the following
reference.

Recently I came across a very similar problem. Just wanted to share our solution with you all. I wanted to I created an instance of a Car<CarA> from a json object using which had an enum:
Dictionary<MyEnum, Type> mapper = new Dictionary<MyEnum, Type>();
mapper.Add(1, typeof(CarA));
mapper.Add(2, typeof(BarB));
public class Car<T> where T : class
{
public T Detail { get; set; }
public Car(T data)
{
Detail = data;
}
}
public class CarA
{
public int PropA { get; set; }
public CarA(){}
}
public class CarB
{
public int PropB { get; set; }
public CarB(){}
}
var jsonObj = {"Type":"1","PropA":"10"}
MyEnum t = GetTypeOfCar(jsonObj);
Type objectT = mapper[t]
Type genericType = typeof(Car<>);
Type carTypeWithGenerics = genericType.MakeGenericType(objectT);
Activator.CreateInstance(carTypeWithGenerics , new Object[] { JsonConvert.DeserializeObject(jsonObj, objectT) });

If you are willing to use a c# precompiler, you could resolve this so that it does have compile time constraints:
// Used attribute
[AttributeUsage(AttributeTargets.Parameter)]
class ResolvedAsAttribute : Attribute
{
public string Expression;
public ResolvedAsAttribute(string expression)
{
this.Expression = expression;
}
}
// Fruit manager source:
class FruitManager {
...
public void AddFruit<TFruit>([ResolvedAs("(int p) => new TFruit(p)")] Func<int,TFruit> ctor = null)where TFruit: BaseFruit{
BaseFruit fruit = ctor(weight); /*new Apple(150);*/
fruit.Enlist(fruitManager);
}
}
// Fruit user source:
#ResolveInclude ../Managers/FruitManager.cs
...
fruitManager.AddFruit<Apple>();
...
Your precompiler would then turn the Fruit user source into:
...
fruitManager.AddFruit<Apple>((int p) => new Apple(p));
...
Using Roslyn, your precompiler could look something like this (here is room for improvement):
using System;
using System.Collections.Generic;
using System.IO;
using System.Linq;
using System.Text;
using Microsoft.CodeAnalysis;
using Microsoft.CodeAnalysis.CSharp;
using Microsoft.CodeAnalysis.CSharp.Syntax;
using Microsoft.CodeAnalysis.CSharp.Symbols;
using System.Threading;
using System.Text.RegularExpressions;
public class CsResolveIncludeAnalyser : CSharpSyntaxWalker
{
private List<(string key, MethodDeclarationSyntax node)> methodsToResolve = new List<(string key, MethodDeclarationSyntax node)>();
public List<(string key, MethodDeclarationSyntax node)> Analyse(string source)
{
var tree = CSharpSyntaxTree.ParseText(source);
var syntaxRoot = tree.GetRoot();
Visit(tree.GetRoot());
return methodsToResolve;
}
public override void VisitMethodDeclaration(MethodDeclarationSyntax methodDeclaration)
{
base.VisitMethodDeclaration(methodDeclaration);
if (methodDeclaration.ParameterList.Parameters.Count > 0)
{
foreach (var parm in methodDeclaration.ParameterList.Parameters)
{
var parmHasResolvedAs = parm.AttributeLists.Where((el) => el.Attributes.Where((attr) => attr.Name is IdentifierNameSyntax && ((IdentifierNameSyntax)attr.Name).Identifier.Text.Contains("ResolvedAs")).Any()).Any();
if (parmHasResolvedAs)
{
var name = methodDeclaration.Identifier.ValueText;
methodsToResolve.Add((name, methodDeclaration));
return;
}
}
}
}
}
public class CsSwiftRewriter : CSharpSyntaxRewriter
{
private string currentFileName;
private bool withWin32ErrorHandling;
private Dictionary<string,MethodDeclarationSyntax> methodsToResolve = new Dictionary<string, MethodDeclarationSyntax>();
private Dictionary<string, MethodDeclarationSyntax> getMethodsToResolve(string source, string fileName)
{
Dictionary<string, MethodDeclarationSyntax> methodsToResolve = new Dictionary<string, MethodDeclarationSyntax>();
var path = Path.GetDirectoryName(fileName);
var lines = source.Split(new[] { '\r', '\n' });
var resolveIncludes = (from el in lines where el.StartsWith("#ResolveInclude") select el.Substring("#ResolveInclude".Length).Trim()).ToList();
var analyser = new CsResolveIncludeAnalyser();
foreach (var resolveInclude in resolveIncludes)
{
var src = File.ReadAllText(path + "/" + resolveInclude);
var list = analyser.Analyse(src);
foreach (var el in list)
{
methodsToResolve.Add(el.key, el.node);
}
}
return methodsToResolve;
}
public static string Convert(string source, string fileName)
{
return Convert(source, fileName, false);
}
public static string Convert(string source, string fileName, bool isWithWin32ErrorHandling)
{
var rewriter = new CsSwiftRewriter() { currentFileName = fileName, withWin32ErrorHandling = isWithWin32ErrorHandling };
rewriter.methodsToResolve = rewriter.getMethodsToResolve(source, fileName);
var resolveIncludeRegex = new Regex(#"(\#ResolveInclude)\b");
source = resolveIncludeRegex.Replace(source, "//$1");
var tree = CSharpSyntaxTree.ParseText(source);
var syntaxRoot = tree.GetRoot();
var result = rewriter.Visit(tree.GetRoot());
return "#line 1 \"" + Path.GetFileName(fileName) + "\"\r\n" + result.ToFullString();
}
internal List<string> transformGenericArguments(List<string> arguments, GenericNameSyntax gName, TypeParameterListSyntax typeParameterList)
{
var res = new List<string>();
var typeParameters = typeParameterList.ChildNodes().ToList();
foreach (var argument in arguments)
{
var arg = argument;
for (int i = 0; i < gName.TypeArgumentList.Arguments.Count; i++)
{
var key = typeParameters[i];
var replacement = gName.TypeArgumentList.Arguments[i].ToString();
var regex = new System.Text.RegularExpressions.Regex($#"\b{key}\b");
arg = regex.Replace(arg, replacement);
}
res.Add(arg);
}
return res;
}
const string prefix = "";
internal List<string> extractExtraArguments(MethodDeclarationSyntax methodDeclaration)
{
var res = new List<String>();
foreach (var parm in methodDeclaration.ParameterList.Parameters)
{
foreach (var attrList in parm.AttributeLists)
{
foreach (var attr in attrList.Attributes)
{
if (attr.Name is IdentifierNameSyntax && string.Compare(((IdentifierNameSyntax)attr.Name).Identifier.Text, "ResolvedAs") == 0)
{
var programmCode = attr.ArgumentList.Arguments.First().ToString().Trim();
var trimmedProgrammCode = (programmCode.Length >= 2 && programmCode[0] == '"' && programmCode[programmCode.Length - 1] == '"') ? programmCode.Substring(1, programmCode.Length - 2) : programmCode;
res.Add(prefix + parm.Identifier.Text + ":" + trimmedProgrammCode);
}
}
}
}
return res;
}
internal List<string> extractExtraArguments(MethodDeclarationSyntax methodDeclaration, SimpleNameSyntax name)
{
var arguments = extractExtraArguments(methodDeclaration);
if (name != null && name is GenericNameSyntax)
{
var gName = name as GenericNameSyntax;
return transformGenericArguments(arguments, gName, methodDeclaration.TypeParameterList);
}
return arguments;
}
public override SyntaxNode VisitInvocationExpression(InvocationExpressionSyntax c_expressionStatement)
{
InvocationExpressionSyntax expressionStatement = (InvocationExpressionSyntax) base.VisitInvocationExpression(c_expressionStatement);
List<string> addedArguments = null;
switch (expressionStatement.Expression)
{
case MemberAccessExpressionSyntax exp:
if (methodsToResolve.ContainsKey(exp.Name?.Identifier.ValueText))
{
addedArguments = extractExtraArguments(methodsToResolve[exp.Name.Identifier.ValueText], exp.Name);
}
break;
case GenericNameSyntax gName:
if (methodsToResolve.ContainsKey(gName.Identifier.ValueText))
{
addedArguments = extractExtraArguments(methodsToResolve[gName.Identifier.ValueText], gName);
}
break;
default:
var name = (from el in expressionStatement.ChildNodes()
where el is GenericNameSyntax
select (el as GenericNameSyntax)).FirstOrDefault();
if (name != default(GenericNameSyntax))
{
if (methodsToResolve.ContainsKey(name.Identifier.ValueText))
{
addedArguments = extractExtraArguments(methodsToResolve[name.Identifier.ValueText], name);
}
}
break;
}
if (addedArguments?.Count > 0)
{
var addedArgumentsString = string.Join(",", addedArguments);
var args = expressionStatement.ArgumentList.ToFullString();
var paras = $"({(expressionStatement.ArgumentList.Arguments.Count > 0 ? string.Join(",", args.Substring(1,args.Length - 2), addedArgumentsString) : addedArgumentsString)})" ;
var argList = SyntaxFactory.ParseArgumentList(paras);
return expressionStatement.WithArgumentList(argList);
}
return expressionStatement;
}
}
The Precompiler could be called using a T4 script, optionally regenerating the source at compile time.

It is still possible, with high performance, by doing the following:
//
public List<R> GetAllItems<R>() where R : IBaseRO, new() {
var list = new List<R>();
using ( var wl = new ReaderLock<T>( this ) ) {
foreach ( var bo in this.items ) {
T t = bo.Value.Data as T;
R r = new R();
r.Initialize( t );
list.Add( r );
}
}
return list;
}
and
//
///<summary>Base class for read-only objects</summary>
public partial interface IBaseRO {
void Initialize( IDTO dto );
void Initialize( object value );
}
The relevant classes then have to derive from this interface and initialize accordingly.
Please note, that in my case, this code is part of a surrounding class, which already has <T> as generic parameter.
R, in my case, also is a read-only class. IMO, the public availability of Initialize() functions has no negative effect on the immutability. The user of this class could put another object in, but this would not modify the underlying collection.