Implementing abstract classes while adding the necessary using directives - c#

Lets assume we have the following abstract base class that declares a parameterized method, whereas the parameter lives in another namespace:
using Example.SubNamespace;
namespace Example
{
public abstract class BaseClass
{
public abstract void Method(Parameter param);
}
}
Following a simple class that acts as our parameter (note the different namespace):
namespace Example.SubNamespace
{
public class Parameter
{
}
}
When implementing a sub class, visual studio offers the option "Implement Abstract Class" when right-clicking the derived class. I've done that in the following example:
namespace Example
{
public class SubClass : BaseClass
{
public override void Method(SubNamespace.Parameter param)
{
throw new System.NotImplementedException();
}
}
}
Call me petty but sometimes it seems very annoying to me having the fully qualified namespaces within the methods signature.
Is there any possibility to implement an abstract class while automatically inserting the correct using directives? Changing these things manually seems like an unnecessary effort to me.

No way to do this. Consider placing your SubClass and BaseClass in the same namespace. I don't think this is a big deal.
You can also use intellisense to add namespaces:
Visual Studio keyboard shortcut to automatically add the needed 'using' statement

Related

How to extend System.Windows.Input.Keyboard class

I would like to add some additional methods to System.Windows.Input.Keyboard.I have tried to create a partial class like this;
namespace System.Windows.Input
{
public static partial class Keyboard
{
//some code...
}
}
However when I try this Resharper informs me that this is not a partial file and when I try to use it I get an ambiguous reference error although both Keyboards are shown as being in the same namespace. Is what I'm trying to do even possible and if not, why not?
No, it's not possible. Keyboard is a static class.
Static classes cannot be instantiated and they cannot be extended.
You can always write your own (static) class and put your methods in there.
You shouldn't use the same namespace as Keyboard, the class should be only static, you can name the class whatever you want "KeyboardExtensions" for exemple
See this for extension methods https://msdn.microsoft.com/en-us/library/bb383977.aspx
What you are trying to do is not possible because the Keyboard class is not partial to begin with.
The ambiguous reference is caused by the fact that you're creating a class that already exists in the same namespace so the compiler doesn't know which one to use.
You cannot create a static class that inherits from another static class as suggested by #harmoniemand. Static classes can only inherit from Object.
And you cannot create extension methods on a static class as #Hamza_L seems to be suggesting because you can't use static types as parameters in extension methods.
with a look at the manual, you could see that the System.Windows.Input.Keyboard is not defined as partial (look here) so you are not able to extend it this way.
The better way to do it, is to write an inherited class.
namespace MyApplication.Wrapper
{
public static class MyKeyboard : System.Windows.Input.Keyboard
{
//some code...
}
}

Foward Declarations in C#

I really like declaring all of my methods at the start of a class and would like to do so with forward declarations and then implement them further down. Is this possible in C#?
Ex:
private void Test();
private void Test()
{
}
Yes you can do this. Or you can kinda sorta do this. But for each of these possible "Solutions", if it is for cosmetic reasons only don't use these constructs.
One trick is to use a partial class with partial methods.
partial class A
{
partial void OnSomethingHappened(string s);
}
// This part can be in a separate file.
partial class A
{
/* Comment out this method and the program
will still compile.*/
partial void OnSomethingHappened(string s)
{
Console.WriteLine("Something happened: {0}", s);
}
}
As highlighted in the referenced documentation, its uses are limited:
A partial method has its signature defined in one part of a partial type, and its implementation defined in another part of the type. Partial methods enable class designers to provide method hooks, similar to event handlers, that developers may decide to implement or not. If the developer does not supply an implementation, the compiler removes the signature at compile time.
The following conditions apply to partial methods:
Signatures in both parts of the partial type must match.
The method must return void.
No access modifiers are allowed. Partial methods are implicitly private.
As pointed out in #Serv and #StanimirYakimov answers, another "kinda sorta" construct that can be used is the declaration of an interface or pure abstract class, very similar to how you would declare one in c++:
public interface IA
{
int GetTheOneAndOnlyNumber();
}
public abstract class AA
{
protected abstract void OnSomethingHappened(string s);
}
public class A : AA, IA
{
public int GetTheOneAndOnlyNumber()
{
return 42;
}
protected override void OnSomethingHappened(string s)
{
Console.WriteLine(s);
}
}
When switching from a language we know well to a new language, there are always specific idioms and constructs that we will miss. This does not mean that it is wise to try to emulate them with non-idiomatic constructs in the new language.
If you would like to have a condensed view of the structure of your code in Visual Studio, there are several standard ways of doing that, such as the Object Browser and Class View.
And if you search and look around, you will find other available tools that can help you with getting a quick overview of the structure of your code.
You can do it using interfaces
Just define which methods your class will use by inheriting from an interface.
See this interface as a mixture of forward declarations and a contract. Classes inherting from an interface must implement all of its members.
public interface iSomeClass
{
void MyMethod1();
bool MyBoolMethod();
}
public class MyClass : iSomeClass
{
public void MyMethod1()
{
//...
}
public bool MyBoolMethod()
{
//...
return true;
}
}
Simple answer no it is not possible.
The reason is that there are no standalone functions in C# but classes with methods.
Also in other languages like C++ once you start using classes forward declaration of the methods themselves are not needed. It can be that the classes themselves need forward declarations in C++ but since you are talking about methods the comparison still stands.
Bottom line a class is completely defined by its methods in any order they are defined.
You can't really do things like this in C#. I am not sure why you would really want to anyway. Having the declaration and definition in one place is simple and easier to understand
Doing something like this is required in some other languages, but forcing that pattern in C# probably isn't a good idea.
No you can not do it. You can only declare abstract methods like this. Otherwise it will give you an error.
I tried and it gave me
Error 1 'Project.SomeClass.ABC()' must declare a body because it is not marked abstract, extern, or partial
Only abstract methods can be defined like this, abstract methods need only a prototype to be overridden.
This is valid:
public abstract void ABC();
Forward declaration is not possible in C# classes. It's possible only in Interfaces and if your method is abstract. Something like this is allowed
abstract void Draw();

Why does my C# TestClass class complain (wrongly) that its parent doesn't have a parameterless constructor?

I am doing some TDD in C# (Visual Studio 2012) using MSTest.
I have a class declared with the [TestClass] attribute.
My test class inherits from the class I am trying to test.
The parent class has a parameterless constructor, and yet I am getting a build error
'RatingsClass.OutputLine' does not contain a constructor that takes 0 arguments
Here is my base class with its parameterless constructor:
namespace RatingsClasses
{
public class OutputLine
{
public OutputLine()
{
Initialise("Parameterless constructor called");
}
(and so on)
Here is the test class which inherits from the base class. This causes the following build error:
'RatingsClass.OutputLine' does not contain a constructor that takes 0 arguments
Code:
using RatingsClasses;
namespace RatingsKataV2
{
[TestClass]
public class RatingsTests: OutputLine
{
[TestMethod]
public void TestSingleTripRatingIs20()
{
(and so on)
Also when I try to call the base class's parameterless constructor directly from the derived class's constructor, I get lots of red squiggly lines and syntax errors such as Constructor must have body and Unexpected token.
Here is my attempt to call the base constructor explicitly:
using RatingsClasses;
namespace RatingsKataV2
{
[TestClass]
public class RatingsTests: OutputLine
{
public RatingsTests(): OutputLine()
{
}
(the rest of the class goes here)
What am I doing wrong?
The reason I am doing it this way is that I would like to test the various private members of my base class are being populated correctly. Rather than provide getters and setters or make those members public, it seemed to make sense that my test class simply inherit from the class it is testing, so that it can directly access those private members. Is there a better way of doing this?
Your syntax is wrong for calling the base constructor. You should use the base keyword:
public RatingsTest()
: base()
{
}
The TestClass attribute is for classes that contain unit tests, not for classes that are the targets of those tests. By deriving from the target under test, you are also extending that target to RatingsTests.
RatingsTests should not need to inherit from OutputLine in order to test, but the constructor shoud read.
public RatingsTests(): base() {}
You use base to explicitly choose the constructor of the the base class that should be used:
public RatingsTests() : base()
{
}
However, this is unnecessary, because the parameterless constructor is the default one being used when this is omitted.
The first code you showed should work, so I guess that your actual code is different or you have a second class called OutputLine somewhere in your code base.
Make sure that you are using the correct OutputLine class and that this class actually has a public parameterless constructor. If public is missing, it is automatically private and can't be used by a derived class.

Change the default base class instead of "object"

Not sure if it is not logical but want to learn. If we don't declare our base class when creating a new class, both visual studio and compiler knows it inherits "object". is there any way to tell the compiler and visual studio(for intellisense) that my base class is not "object", it is "myobject" for example? maybe any configuration on visual studio?
you can say "inherit all your classes from 'myobject'", i know it but just want to know is it possible.
No, absolutely not. This is part of the C# spec, and is in no way optional. From section 10.1.4.1:
If a class declaration has no class-base, or if the class-base lists only interface types, the direct base class is assumed to be object.
No implementation-specific wiggle-room - that's just the way it is. Personally I'm glad - I want to be able to tell the direct base class just from looking at the source code, without knowing any project configuration.
EDIT: Just to be clear, C# could have been designed such that a different type could be specified as the base type. For example, I could imagine (but dislike) a language where:
class Foo
{
}
compiled with:
csc /evil:DefaultBaseClass=System.IO.Stream Foo.cs
was equivalent to:
class Foo : Stream
{
}
That doesn't break .NET at all - it's a purely language decision. What is enforced by .NET is that most types end up with at least an indirect base class of System.Object. I don't think a language could be designed to allow you to set up a "parallel" type hierarchy.
That's not possible, the way C# is designed, (almost) everything derives from Object. If this were not the case you couldn't rely on the methods Object provides being there, which are pretty fundamental/useful.
Note that it is .Net that requires all objects to be derived from Object (directly or indirectly). C# specifies that objects declared with no base class inherit from Object, other languages are free to specify another type, however that type must inherit (directly or indirectly) from Object as per .net requirements.
Obviously you're free to have your own base class for all types you create, however that base class has to inherit from object (either directly or indirectly).
Do you want to add methods to all/existing types? In which case use extension methods.
You could modify the existing item templates (or create a new one) so that new classes automatically inherit from some base class:
using System;
using System.Collections.Generic;
$if$ ($targetframeworkversion$ >= 3.5)using System.Linq;
$endif$using System.Text;
namespace $rootnamespace$
{
class $safeitemrootname$ : SomeBaseClass
{
}
}
See Customizing Project and Item Templates
You can create your own custom abstract class and override the 4 methods defined by the Object class and further add new virtual and abstract methods to your custom base class. Then, you can use this class as the base class explicitly for all your other custom classes. But, that will still mean the super base class of all the classes will be Object only (you can't change that). e.g.
//your custom base class
public abstract class MyObject
{
public virtual void MyCustomMethod()
{
//Your custom method implementation
}
public abstract void MyCustomAbstractMethod();
public override string ToString( )
{
//your custom implementation for override
}
public override string Equals( )
{
//your custom implementation for override
}
public override string GetHashCode( )
{
//your custom implementation for override
}
public override string GetType( )
{
//your custom implementation for override
}
}
//your custom child class
public class CustomClass1 : MyObject //still derived from object
{
//implement and override the MyObject and object methods
}
//your custom child class
public class CustomClass2 : MyObject //still derived from object
{
//implement and override the MyObject and object methods
}
But all the built-in types provided by the .Net BCL will not be able to use your MyObject class, they still will be known to be derived from object.
Ok one thing: In C# and Java all classes declared are derived automatically from Object. There is no changing this. Object is the root of the class hierarchy. Now, you can have your classes inherit another class like:
public class A : B // A inherits B
but up the chain, you will still find Object at the top because anyway B inherited Object.
i think every thing in .NET inherites from object class. 4 example, becouse it, every object have ToString() method.
All the classes and struct inheret from Object, it is the parent of all, even you inherited from a class
public class A : B // A inherits B

Partial classes/partial methods vs base/inherited classes

a question about class design. Currently I have the following structure:
abstract Base Repository Class
Default Repository implementation class (implements some abstract methods, where logic is common thru all of the Specific classes but leaves other empty)
Specific Repository implementation Class (implements what is left empty in the above Default class)
I've now came to the problem where I have a specific Update() method in Specific class but when all the code in this method executes some code from the base Default class should be executed too.
I could do it like this
public override Update()
{
// do Specific class actions and updates
// ....
// follow with base.Update()
base.Update();
}
but this requires those base.XYZ() calls in all the inherited methods. Could I go around that somehow with partials?
So the requirement is to have code in both parent and inherited class (or to make those two one class using partials) and code from method implementation in both places should be executed. Also what about if I wanted to turn it around and execute base class code first followed by the inherited class code?
thanks
Have you considered something like:
public abstract class YourBaseClass
{
public void Update()
{
// Do some stuff
//
// Invoke inherited class's method
UpdateCore();
}
protected abstract void UpdateCore();
}
public class YourChildClass : YourBaseClass
{
protected override void UpdateCore()
{
//Do the important stuff
}
}
//Somewhere else in code:
var ycc = new YourChildClass();
ycc.Update();
All the partial keyword means is that the definition of the class is split between source files:
It is possible to split the definition of a class or a struct, an interface or a method over two or more source files. Each source file contains a section of the type or method definition, and all parts are combined when the application is compiled.
There still has to be a complete definition of the class in the project.
You'd be better off creating a subclass, that way you can override specific methods.
As far as partial methods go (from the same link as above):
A partial method declaration consists of two parts: the definition, and the implementation. These may be in separate parts of a partial class, or in the same part. If there is no implementation declaration, then the compiler optimizes away both the defining declaration and all calls to the method.
// Definition in file1.cs
partial void onNameChanged();
// Implementation in file2.cs
partial void onNameChanged()
{
// method body
}
You can't have half the method in one file and the other half in another.
Here's how you can do this:
public sealed override void Update()
{
UpdateCore();
base.Update();
}
public abstract /* or virtual */ void UpdateCore()
{
// Class-specific stuff
}
Forget about partial, that has entirely different semantics. Whether the overrider of your virtual base class method should call the base class method is not automatic. It needs to be part of your documentation. A good example are the OnXxxx() methods in the Control class, the MSDN library docs have a "Note to implementer" comment that warns that calling the base class method is usually necessary.
If you make the base class method abstract then it is crystal-clear to the overrider. If it is not, you are dropping a strong hint that it ought to be done. If you expect the override to completely replace your base implementation then you should really consider making it abstract. This ambiguity, combined with the odds that the overrider breaks your base class by overriding incorrectly, is definitely one of the weak points of polymorphism.
Add a new virtual (not abstract, since not all specific implementation need to override it?) method to your Default implementation to accommodate inheritors having their own additional steps on top of it's own implementation.
Call the virtual method at the appropriate point in the Default Implementation's abstract method implementation.
You've left your abstract class as it is and transparently grown the flexibility of the default implementation.

Categories