Related
What is the real reason for that limitation? Is it just work that had to be done? Is it conceptually hard? Is it impossible?
Sure, one couldn't use the type parameters in fields, because they are allways read-write. But that can't be the answer, can it?
The reason for this question is that I'm writing an article on variance support in C# 4, and I feel that I should explain why it is restricted to delegates and interfaces. Just to inverse the onus of proof.
Update:
Eric asked about an example.
What about this (don't know if that makes sense, yet :-))
public class Lookup<out T> where T : Animal {
public T Find(string name) {
Animal a = _cache.FindAnimalByName(name);
return a as T;
}
}
var findReptiles = new Lookup<Reptile>();
Lookup<Animal> findAnimals = findReptiles;
The reason for having that in one class could be the cache that is held in the class itself. And please don't name your different type pets the same!
BTW, this brings me to optional type parameters in C# 5.0 :-)
Update 2: I'm not claiming the CLR and C# should allow this. Just trying to understand what led to that it doesnt.
First off, as Tomas says, it is not supported in the CLR.
Second, how would that work? Suppose you have
class C<out T>
{ ... how are you planning on using T in here? ... }
T can only be used in output positions. As you note, the class cannot have any field of type T because the field could be written to. The class cannot have any methods that take a T, because those are logically writes. Suppose you had this feature -- how would you take advantage of it?
This would be useful for immutable classes if we could, say, make it legal to have a readonly field of type T; that way we'd massively cut down on the likelihood that it be improperly written to. But it's quite difficult to come up with other scenarios that permit variance in a typesafe manner.
If you have such a scenario, I'd love to see it. That would be points towards someday getting this implemented in the CLR.
UPDATE: See
Why isn't there generic variance for classes in C# 4.0?
for more on this question.
As far as I know, this feature isn't supported by CLR, so adding this would require significant work on the CLR side as well. I believe that co- and contra-variance for interfaces and delegates was actually supported on CLR before the version 4.0, so this was a relatively straightforward extension to implement.
(Supporting this feature for classes would be definitely useful, though!)
If they were permitted, useful 100% type-safe (no internal typecasts) classes or structures could be defined which were covariant with regard to their type T, if their constructor accepted one or more T's or T supplier's. Useful, 100%-type-safe classes or structures could be defined which were contravariant with respect to T if their constructors accepted one or more T consumers. I'm not sure there's much advantage of a class over an interface, beyond the ability to use "new" rather than using a static factory method (most likely from a class whose name is similar to that of the interface), but I can certainly see usage cases for having immutable structures support covariance.
As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, visit the help center for guidance.
Closed 11 years ago.
Locked. This question and its answers are locked because the question is off-topic but has historical significance. It is not currently accepting new answers or interactions.
I was curious about how other people use the this keyword. I tend to use it in constructors, but I may also use it throughout the class in other methods. Some examples:
In a constructor:
public Light(Vector v)
{
this.dir = new Vector(v);
}
Elsewhere
public void SomeMethod()
{
Vector vec = new Vector();
double d = (vec * vec) - (this.radius * this.radius);
}
I don't mean this to sound snarky, but it doesn't matter.
Seriously.
Look at the things that are important: your project, your code, your job, your personal life. None of them are going to have their success rest on whether or not you use the "this" keyword to qualify access to fields. The this keyword will not help you ship on time. It's not going to reduce bugs, it's not going to have any appreciable effect on code quality or maintainability. It's not going to get you a raise, or allow you to spend less time at the office.
It's really just a style issue. If you like "this", then use it. If you don't, then don't. If you need it to get correct semantics then use it. The truth is, every programmer has his own unique programing style. That style reflects that particular programmer's notions of what the "most aesthetically pleasing code" should look like. By definition, any other programmer who reads your code is going to have a different programing style. That means there is always going to be something you did that the other guy doesn't like, or would have done differently. At some point some guy is going to read your code and grumble about something.
I wouldn't fret over it. I would just make sure the code is as aesthetically pleasing as possible according to your own tastes. If you ask 10 programmers how to format code, you are going to get about 15 different opinions. A better thing to focus on is how the code is factored. Are things abstracted right? Did I pick meaningful names for things? Is there a lot of code duplication? Are there ways I can simplify stuff? Getting those things right, I think, will have the greatest positive impact on your project, your code, your job, and your life. Coincidentally, it will probably also cause the other guy to grumble the least. If your code works, is easy to read, and is well factored, the other guy isn't going to be scrutinizing how you initialize fields. He's just going to use your code, marvel at it's greatness, and then move on to something else.
There are several usages of this keyword in C#.
To qualify members hidden by similar name
To have an object pass itself as a parameter to other methods
To have an object return itself from a method
To declare indexers
To declare extension methods
To pass parameters between constructors
To internally reassign value type (struct) value.
To invoke an extension method on the current instance
To cast itself to another type
To chain constructors defined in the same class
You can avoid the first usage by not having member and local variables with the same name in scope, for example by following common naming conventions and using properties (Pascal case) instead of fields (camel case) to avoid colliding with local variables (also camel case). In C# 3.0 fields can be converted to properties easily by using auto-implemented properties.
I only use it when absolutely necessary, ie, when another variable is shadowing another. Such as here:
class Vector3
{
float x;
float y;
float z;
public Vector3(float x, float y, float z)
{
this.x = x;
this.y = y;
this.z = z;
}
}
Or as Ryan Fox points out, when you need to pass this as a parameter. (Local variables have precedence over member variables)
Personally, I try to always use this when referring to member variables. It helps clarify the code and make it more readable. Even if there is no ambiguity, someone reading through my code for the first time doesn't know that, but if they see this used consistently, they will know if they are looking at a member variable or not.
I use it every time I refer to an instance variable, even if I don't need to. I think it makes the code more clear.
I can't believe all of the people that say using it always is a "best practice" and such.
Use "this" when there is ambiguity, as in Corey's example or when you need to pass the object as a parameter, as in Ryan's example. There is no reason to use it otherwise because being able to resolve a variable based on the scope chain should be clear enough that qualifying variables with it should be unnecessary.
EDIT: The C# documentation on "this" indicates one more use, besides the two I mentioned, for the "this" keyword - for declaring indexers
EDIT: #Juan: Huh, I don't see any inconsistency in my statements - there are 3 instances when I would use the "this" keyword (as documented in the C# documentation), and those are times when you actually need it. Sticking "this" in front of variables in a constructor when there is no shadowing going on is simply a waste of keystrokes and a waste of my time when reading it, it provides no benefit.
I use it whenever StyleCop tells me to. StyleCop must be obeyed. Oh yes.
Any time you need a reference to the current object.
One particularly handy scenario is when your object is calling a function and wants to pass itself into it.
Example:
void onChange()
{
screen.draw(this);
}
I tend to use it everywhere as well, just to make sure that it is clear that it is instance members that we are dealing with.
I use it anywhere there might be ambiguity (obviously). Not just compiler ambiguity (it would be required in that case), but also ambiguity for someone looking at the code.
Another somewhat rare use for the this keyword is when you need to invoke an explicit interface implementation from within the implementing class. Here's a contrived example:
class Example : ICloneable
{
private void CallClone()
{
object clone = ((ICloneable)this).Clone();
}
object ICloneable.Clone()
{
throw new NotImplementedException();
}
}
Here's when I use it:
Accessing Private Methods from within the class (to differentiate)
Passing the current object to another method (or as a sender object, in case of an event)
When creating extension methods :D
I don't use this for Private fields because I prefix private field variable names with an underscore (_).
[C++]
I agree with the "use it when you have to" brigade. Decorating code unnecessarily with this isn't a great idea because the compiler won't warn you when you forget to do it. This introduces potential confusion for people expecting this to always be there, i.e. they'll have to think about it.
So, when would you use it? I've just had a look around some random code and found these examples (I'm not passing judgement on whether these are good things to do or otherwise):
Passing "yourself" to a function.
Assigning "yourself" to a pointer or something like that.
Casting, i.e. up/down casting (safe or otherwise), casting away constness, etc.
Compiler enforced disambiguation.
You should always use it, I use it to diferantiate private fields and parameters (because our naming conventions state that we don't use prefixes for member and parameter names (and they are based on information found on the internet, so I consider that a best practice))
I use it when, in a function that accepts a reference to an object of the same type, I want to make it perfectly clear which object I'm referring to, where.
For example
class AABB
{
// ... members
bool intersects( AABB other )
{
return other.left() < this->right() &&
this->left() < other.right() &&
// +y increases going down
other.top() < this->bottom() &&
this->top() < other.bottom() ;
}
} ;
(vs)
class AABB
{
bool intersects( AABB other )
{
return other.left() < right() &&
left() < other.right() &&
// +y increases going down
other.top() < bottom() &&
top() < other.bottom() ;
}
} ;
At a glance which AABB does right() refer to? The this adds a bit of a clarifier.
In Jakub Šturc's answer his #5 about passing data between contructors probably could use a little explanation. This is in overloading constructors and is the one case where use of this is mandatory. In the following example we can call the parameterized constructor from the parameterless constructor with a default parameter.
class MyClass {
private int _x
public MyClass() : this(5) {}
public MyClass(int v) { _x = v;}
}
I've found this to be a particularly useful feature on occasion.
I got in the habit of using it liberally in Visual C++ since doing so would trigger IntelliSense ones I hit the '>' key, and I'm lazy. (and prone to typos)
But I've continued to use it, since I find it handy to see that I'm calling a member function rather than a global function.
I tend to underscore fields with _ so don't really ever need to use this. Also R# tends to refactor them away anyway...
I pretty much only use this when referencing a type property from inside the same type. As another user mentioned, I also underscore local fields so they are noticeable without needing this.
I use it only when required, except for symmetric operations which due to single argument polymorphism have to be put into methods of one side:
boolean sameValue (SomeNum other) {
return this.importantValue == other.importantValue;
}
[C++]
this is used in the assignment operator where most of the time you have to check and prevent strange (unintentional, dangerous, or just a waste of time for the program) things like:
A a;
a = a;
Your assignment operator will be written:
A& A::operator=(const A& a) {
if (this == &a) return *this;
// we know both sides of the = operator are different, do something...
return *this;
}
this on a C++ compiler
The C++ compiler will silently lookup for a symbol if it does not find it immediately. Sometimes, most of the time, it is good:
using the mother class' method if you did not overloaded it in the child class.
promoting a value of a type into another type
But sometimes, You just don't want the compiler to guess. You want the compiler to pick-up the right symbol and not another.
For me, those times are when, within a method, I want to access to a member method or member variable. I just don't want some random symbol picked up just because I wrote printf instead of print. this->printf would not have compiled.
The point is that, with C legacy libraries (§), legacy code written years ago (§§), or whatever could happen in a language where copy/pasting is an obsolete but still active feature, sometimes, telling the compiler to not play wits is a great idea.
These are the reasons I use this.
(§) it's still a kind of mystery to me, but I now wonder if the fact you include the <windows.h> header in your source, is the reason all the legacy C libraries symbols will pollute your global namespace
(§§) realizing that "you need to include a header, but that including this header will break your code because it uses some dumb macro with a generic name" is one of those russian roulette moments of a coder's life
'this.' helps find members on 'this' class with a lot of members (usually due to a deep inheritance chain).
Hitting CTRL+Space doesn't help with this, because it also includes types; where-as 'this.' includes members ONLY.
I usually delete it once I have what I was after: but this is just my style breaking through.
In terms of style, if you are a lone-ranger -- you decide; if you work for a company stick to the company policy (look at the stuff in source control and see what other people are doing). In terms of using it to qualify members, neither is right or wrong. The only wrong thing is inconsistency -- that is the golden rule of style. Leave the nit-picking others. Spend your time pondering real coding problems -- and obviously coding -- instead.
I use it every time I can. I believe it makes the code more readable, and more readable code equals less bugs and more maintainability.
When you are many developers working on the same code base, you need some code guidelines/rules. Where I work we've desided to use 'this' on fields, properties and events.
To me it makes good sense to do it like this, it makes the code easier to read when you differentiate between class-variables and method-variables.
It depends on the coding standard I'm working under. If we are using _ to denote an instance variable then "this" becomes redundant. If we are not using _ then I tend to use this to denote instance variable.
I use it to invoke Intellisense just like JohnMcG, but I'll go back and erase "this->" when I'm done. I follow the Microsoft convention of prefixing member variables with "m_", so leaving it as documentation would just be redundant.
1 - Common Java setter idiom:
public void setFoo(int foo) {
this.foo = foo;
}
2 - When calling a function with this object as a parameter
notifier.addListener(this);
There is one use that has not already been mentioned in C++, and that is not to refer to the own object or disambiguate a member from a received variable.
You can use this to convert a non-dependent name into an argument dependent name inside template classes that inherit from other templates.
template <typename T>
struct base {
void f() {}
};
template <typename T>
struct derived : public base<T>
{
void test() {
//f(); // [1] error
base<T>::f(); // quite verbose if there is more than one argument, but valid
this->f(); // f is now an argument dependent symbol
}
}
Templates are compiled with a two pass mechanism. During the first pass, only non-argument dependent names are resolved and checked, while dependent names are checked only for coherence, without actually substituting the template arguments.
At that step, without actually substituting the type, the compiler has almost no information of what base<T> could be (note that specialization of the base template can turn it into completely different types, even undefined types), so it just assumes that it is a type. At this stage the non-dependent call f that seems just natural to the programmer is a symbol that the compiler must find as a member of derived or in enclosing namespaces --which does not happen in the example-- and it will complain.
The solution is turning the non-dependent name f into a dependent name. This can be done in a couple of ways, by explicitly stating the type where it is implemented (base<T>::f --adding the base<T> makes the symbol dependent on T and the compiler will just assume that it will exist and postpones the actual check for the second pass, after argument substitution.
The second way, much sorter if you inherit from templates that have more than one argument, or long names, is just adding a this-> before the symbol. As the template class you are implementing does depend on an argument (it inherits from base<T>) this-> is argument dependent, and we get the same result: this->f is checked in the second round, after template parameter substitution.
You should not use "this" unless you absolutely must.
There IS a penalty associated with unnecessary verbosity. You should strive for code that is exactly as long as it needs to be, and no longer.
As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, visit the help center for guidance.
Closed 11 years ago.
Locked. This question and its answers are locked because the question is off-topic but has historical significance. It is not currently accepting new answers or interactions.
I was curious about how other people use the this keyword. I tend to use it in constructors, but I may also use it throughout the class in other methods. Some examples:
In a constructor:
public Light(Vector v)
{
this.dir = new Vector(v);
}
Elsewhere
public void SomeMethod()
{
Vector vec = new Vector();
double d = (vec * vec) - (this.radius * this.radius);
}
I don't mean this to sound snarky, but it doesn't matter.
Seriously.
Look at the things that are important: your project, your code, your job, your personal life. None of them are going to have their success rest on whether or not you use the "this" keyword to qualify access to fields. The this keyword will not help you ship on time. It's not going to reduce bugs, it's not going to have any appreciable effect on code quality or maintainability. It's not going to get you a raise, or allow you to spend less time at the office.
It's really just a style issue. If you like "this", then use it. If you don't, then don't. If you need it to get correct semantics then use it. The truth is, every programmer has his own unique programing style. That style reflects that particular programmer's notions of what the "most aesthetically pleasing code" should look like. By definition, any other programmer who reads your code is going to have a different programing style. That means there is always going to be something you did that the other guy doesn't like, or would have done differently. At some point some guy is going to read your code and grumble about something.
I wouldn't fret over it. I would just make sure the code is as aesthetically pleasing as possible according to your own tastes. If you ask 10 programmers how to format code, you are going to get about 15 different opinions. A better thing to focus on is how the code is factored. Are things abstracted right? Did I pick meaningful names for things? Is there a lot of code duplication? Are there ways I can simplify stuff? Getting those things right, I think, will have the greatest positive impact on your project, your code, your job, and your life. Coincidentally, it will probably also cause the other guy to grumble the least. If your code works, is easy to read, and is well factored, the other guy isn't going to be scrutinizing how you initialize fields. He's just going to use your code, marvel at it's greatness, and then move on to something else.
There are several usages of this keyword in C#.
To qualify members hidden by similar name
To have an object pass itself as a parameter to other methods
To have an object return itself from a method
To declare indexers
To declare extension methods
To pass parameters between constructors
To internally reassign value type (struct) value.
To invoke an extension method on the current instance
To cast itself to another type
To chain constructors defined in the same class
You can avoid the first usage by not having member and local variables with the same name in scope, for example by following common naming conventions and using properties (Pascal case) instead of fields (camel case) to avoid colliding with local variables (also camel case). In C# 3.0 fields can be converted to properties easily by using auto-implemented properties.
I only use it when absolutely necessary, ie, when another variable is shadowing another. Such as here:
class Vector3
{
float x;
float y;
float z;
public Vector3(float x, float y, float z)
{
this.x = x;
this.y = y;
this.z = z;
}
}
Or as Ryan Fox points out, when you need to pass this as a parameter. (Local variables have precedence over member variables)
Personally, I try to always use this when referring to member variables. It helps clarify the code and make it more readable. Even if there is no ambiguity, someone reading through my code for the first time doesn't know that, but if they see this used consistently, they will know if they are looking at a member variable or not.
I use it every time I refer to an instance variable, even if I don't need to. I think it makes the code more clear.
I can't believe all of the people that say using it always is a "best practice" and such.
Use "this" when there is ambiguity, as in Corey's example or when you need to pass the object as a parameter, as in Ryan's example. There is no reason to use it otherwise because being able to resolve a variable based on the scope chain should be clear enough that qualifying variables with it should be unnecessary.
EDIT: The C# documentation on "this" indicates one more use, besides the two I mentioned, for the "this" keyword - for declaring indexers
EDIT: #Juan: Huh, I don't see any inconsistency in my statements - there are 3 instances when I would use the "this" keyword (as documented in the C# documentation), and those are times when you actually need it. Sticking "this" in front of variables in a constructor when there is no shadowing going on is simply a waste of keystrokes and a waste of my time when reading it, it provides no benefit.
I use it whenever StyleCop tells me to. StyleCop must be obeyed. Oh yes.
Any time you need a reference to the current object.
One particularly handy scenario is when your object is calling a function and wants to pass itself into it.
Example:
void onChange()
{
screen.draw(this);
}
I tend to use it everywhere as well, just to make sure that it is clear that it is instance members that we are dealing with.
I use it anywhere there might be ambiguity (obviously). Not just compiler ambiguity (it would be required in that case), but also ambiguity for someone looking at the code.
Another somewhat rare use for the this keyword is when you need to invoke an explicit interface implementation from within the implementing class. Here's a contrived example:
class Example : ICloneable
{
private void CallClone()
{
object clone = ((ICloneable)this).Clone();
}
object ICloneable.Clone()
{
throw new NotImplementedException();
}
}
Here's when I use it:
Accessing Private Methods from within the class (to differentiate)
Passing the current object to another method (or as a sender object, in case of an event)
When creating extension methods :D
I don't use this for Private fields because I prefix private field variable names with an underscore (_).
[C++]
I agree with the "use it when you have to" brigade. Decorating code unnecessarily with this isn't a great idea because the compiler won't warn you when you forget to do it. This introduces potential confusion for people expecting this to always be there, i.e. they'll have to think about it.
So, when would you use it? I've just had a look around some random code and found these examples (I'm not passing judgement on whether these are good things to do or otherwise):
Passing "yourself" to a function.
Assigning "yourself" to a pointer or something like that.
Casting, i.e. up/down casting (safe or otherwise), casting away constness, etc.
Compiler enforced disambiguation.
You should always use it, I use it to diferantiate private fields and parameters (because our naming conventions state that we don't use prefixes for member and parameter names (and they are based on information found on the internet, so I consider that a best practice))
I use it when, in a function that accepts a reference to an object of the same type, I want to make it perfectly clear which object I'm referring to, where.
For example
class AABB
{
// ... members
bool intersects( AABB other )
{
return other.left() < this->right() &&
this->left() < other.right() &&
// +y increases going down
other.top() < this->bottom() &&
this->top() < other.bottom() ;
}
} ;
(vs)
class AABB
{
bool intersects( AABB other )
{
return other.left() < right() &&
left() < other.right() &&
// +y increases going down
other.top() < bottom() &&
top() < other.bottom() ;
}
} ;
At a glance which AABB does right() refer to? The this adds a bit of a clarifier.
In Jakub Šturc's answer his #5 about passing data between contructors probably could use a little explanation. This is in overloading constructors and is the one case where use of this is mandatory. In the following example we can call the parameterized constructor from the parameterless constructor with a default parameter.
class MyClass {
private int _x
public MyClass() : this(5) {}
public MyClass(int v) { _x = v;}
}
I've found this to be a particularly useful feature on occasion.
I got in the habit of using it liberally in Visual C++ since doing so would trigger IntelliSense ones I hit the '>' key, and I'm lazy. (and prone to typos)
But I've continued to use it, since I find it handy to see that I'm calling a member function rather than a global function.
I tend to underscore fields with _ so don't really ever need to use this. Also R# tends to refactor them away anyway...
I pretty much only use this when referencing a type property from inside the same type. As another user mentioned, I also underscore local fields so they are noticeable without needing this.
I use it only when required, except for symmetric operations which due to single argument polymorphism have to be put into methods of one side:
boolean sameValue (SomeNum other) {
return this.importantValue == other.importantValue;
}
[C++]
this is used in the assignment operator where most of the time you have to check and prevent strange (unintentional, dangerous, or just a waste of time for the program) things like:
A a;
a = a;
Your assignment operator will be written:
A& A::operator=(const A& a) {
if (this == &a) return *this;
// we know both sides of the = operator are different, do something...
return *this;
}
this on a C++ compiler
The C++ compiler will silently lookup for a symbol if it does not find it immediately. Sometimes, most of the time, it is good:
using the mother class' method if you did not overloaded it in the child class.
promoting a value of a type into another type
But sometimes, You just don't want the compiler to guess. You want the compiler to pick-up the right symbol and not another.
For me, those times are when, within a method, I want to access to a member method or member variable. I just don't want some random symbol picked up just because I wrote printf instead of print. this->printf would not have compiled.
The point is that, with C legacy libraries (§), legacy code written years ago (§§), or whatever could happen in a language where copy/pasting is an obsolete but still active feature, sometimes, telling the compiler to not play wits is a great idea.
These are the reasons I use this.
(§) it's still a kind of mystery to me, but I now wonder if the fact you include the <windows.h> header in your source, is the reason all the legacy C libraries symbols will pollute your global namespace
(§§) realizing that "you need to include a header, but that including this header will break your code because it uses some dumb macro with a generic name" is one of those russian roulette moments of a coder's life
'this.' helps find members on 'this' class with a lot of members (usually due to a deep inheritance chain).
Hitting CTRL+Space doesn't help with this, because it also includes types; where-as 'this.' includes members ONLY.
I usually delete it once I have what I was after: but this is just my style breaking through.
In terms of style, if you are a lone-ranger -- you decide; if you work for a company stick to the company policy (look at the stuff in source control and see what other people are doing). In terms of using it to qualify members, neither is right or wrong. The only wrong thing is inconsistency -- that is the golden rule of style. Leave the nit-picking others. Spend your time pondering real coding problems -- and obviously coding -- instead.
I use it every time I can. I believe it makes the code more readable, and more readable code equals less bugs and more maintainability.
When you are many developers working on the same code base, you need some code guidelines/rules. Where I work we've desided to use 'this' on fields, properties and events.
To me it makes good sense to do it like this, it makes the code easier to read when you differentiate between class-variables and method-variables.
It depends on the coding standard I'm working under. If we are using _ to denote an instance variable then "this" becomes redundant. If we are not using _ then I tend to use this to denote instance variable.
I use it to invoke Intellisense just like JohnMcG, but I'll go back and erase "this->" when I'm done. I follow the Microsoft convention of prefixing member variables with "m_", so leaving it as documentation would just be redundant.
1 - Common Java setter idiom:
public void setFoo(int foo) {
this.foo = foo;
}
2 - When calling a function with this object as a parameter
notifier.addListener(this);
There is one use that has not already been mentioned in C++, and that is not to refer to the own object or disambiguate a member from a received variable.
You can use this to convert a non-dependent name into an argument dependent name inside template classes that inherit from other templates.
template <typename T>
struct base {
void f() {}
};
template <typename T>
struct derived : public base<T>
{
void test() {
//f(); // [1] error
base<T>::f(); // quite verbose if there is more than one argument, but valid
this->f(); // f is now an argument dependent symbol
}
}
Templates are compiled with a two pass mechanism. During the first pass, only non-argument dependent names are resolved and checked, while dependent names are checked only for coherence, without actually substituting the template arguments.
At that step, without actually substituting the type, the compiler has almost no information of what base<T> could be (note that specialization of the base template can turn it into completely different types, even undefined types), so it just assumes that it is a type. At this stage the non-dependent call f that seems just natural to the programmer is a symbol that the compiler must find as a member of derived or in enclosing namespaces --which does not happen in the example-- and it will complain.
The solution is turning the non-dependent name f into a dependent name. This can be done in a couple of ways, by explicitly stating the type where it is implemented (base<T>::f --adding the base<T> makes the symbol dependent on T and the compiler will just assume that it will exist and postpones the actual check for the second pass, after argument substitution.
The second way, much sorter if you inherit from templates that have more than one argument, or long names, is just adding a this-> before the symbol. As the template class you are implementing does depend on an argument (it inherits from base<T>) this-> is argument dependent, and we get the same result: this->f is checked in the second round, after template parameter substitution.
You should not use "this" unless you absolutely must.
There IS a penalty associated with unnecessary verbosity. You should strive for code that is exactly as long as it needs to be, and no longer.
In earlier versions of C# IEnumerable was defined like this:
public interface IEnumerable<T> : IEnumerable
Since C# 4 the definition is:
public interface IEnumerable<out T> : IEnumerable
Is it just to make the annoying casts in LINQ expressions go away?
Won't this introduce the same problems like with string[] <: object[] (broken array variance) in C#?
How was the addition of the covariance done from a compatibility point of view? Will earlier code still work on later versions of .NET or is recompilation necessary here? What about the other way around?
Was previous code using this interface strictly invariant in all cases or is it possible that certain use cases will behave different now?
Marc's and CodeInChaos's answers are pretty good, but just to add a few more details:
First off, it sounds like you are interested in learning about the design process we went through to make this feature. If so, then I encourage you to read my lengthy series of articles that I wrote while designing and implementing the feature. Start from the bottom of the page:
Covariance and contravariance blog posts
Is it just to make the annoying casts in LINQ expressions go away?
No, it is not just to avoid Cast<T> expressions, but doing so was one of the motivators that encouraged us to do this feature. We realized that there would be an uptick in the number of "why can't I use a sequence of Giraffes in this method that takes a sequence of Animals?" questions, because LINQ encourages the use of sequence types. We knew that we wanted to add covariance to IEnumerable<T> first.
We actually considered making IEnumerable<T> covariant even in C# 3 but decided that it would be strange to do so without introducing the whole feature for anyone to use.
Won't this introduce the same problems like with string[] <: object[] (broken array variance) in C#?
It does not directly introduce that problem because the compiler only allows variance when it is known to be typesafe. However, it does preserve the broken array variance problem. With covariance, IEnumerable<string[]> is implicitly convertible to IEnumerable<object[]>, so if you have a sequence of string arrays, you can treat that as a sequence of object arrays, and then you have the same problem as before: you can try to put a Giraffe into that string array and get an exception at runtime.
How was the addition of the covariance done from a compatibility point of view?
Carefully.
Will earlier code still work on later versions of .NET or is recompilation necessary here?
Only one way to find out. Try it and see what fails!
It's often a bad idea to try to force code compiled against .NET X to run against .NET Y if X != Y, regardless of changes to the type system.
What about the other way around?
Same answer.
Is it possible that certain use cases will behave different now?
Absolutely. Making an interface covariant where it was invariant before is technically a "breaking change" because it can cause working code to break. For example:
if (x is IEnumerable<Animal>)
ABC();
else if (x is IEnumerable<Turtle>)
DEF();
When IE<T> is not covariant, this code chooses either ABC or DEF or neither. When it is covariant, it never chooses DEF anymore.
Or:
class B { public void M(IEnumerable<Turtle> turtles){} }
class D : B { public void M(IEnumerable<Animal> animals){} }
Before, if you called M on an instance of D with a sequence of turtles as the argument, overload resolution chooses B.M because that is the only applicable method. If IE is covariant, then overload resolution now chooses D.M because both methods are applicable, and an applicable method on a more-derived class always beats an applicable method on a less-derived class, regardless of whether the argument type match is exact or not.
Or:
class Weird : IEnumerable<Turtle>, IEnumerable<Banana> { ... }
class B
{
public void M(IEnumerable<Banana> bananas) {}
}
class D : B
{
public void M(IEnumerable<Animal> animals) {}
public void M(IEnumerable<Fruit> fruits) {}
}
If IE is invariant then a call to d.M(weird) resolves to B.M. If IE suddenly becomes covariant then both methods D.M are applicable, both are better than the method on the base class, and neither is better than the other, so, overload resolution becomes ambiguous and we report an error.
When we decided to make these breaking changes, we were hoping that (1) the situations would be rare, and (2) when situations like this arise, almost always it is because the author of the class is attempting to simulate covariance in a language that doesn't have it. By adding covariance directly, hopefully when the code "breaks" on recompilation, the author can simply remove the crazy gear trying to simulate a feature that now exists.
In order:
Is it just to make the annoying casts in LINQ expressions go away?
It makes things behave like people generally expect ;p
Won't this introduce the same problems like with string[] <: object[] (broken array variance) in C#?
No; since it doesn't expose any Add mechanism or similar (and can't; out and in are enforced at the compiler)
How was the addition of the covariance done from a compatibility point of view?
The CLI already supported it, this merely makes C# (and some of the existing BCL methods) aware of it
Will earlier code still work on later versions of .NET or is recompilation necessary here?
It is entirely backwards compatible, however: C# that relies on C# 4.0 variance won't compile in a C# 2.0 etc compiler
What about the other way around?
That is not unreasonable
Was previous code using this interface strictly invariant in all cases or is it possible that certain use cases will behave different now?
Some BCL calls (IsAssignableFrom) may return differently now
Is it just to make the annoying casts in LINQ expressions go away?
Not only when using LINQ. It's useful everywhere you have an IEnumerable<Derived> and the code expects a IEnumerable<Base>.
Won't this introduce the same problems like with string[] <: object[] (broken array variance) in C#?
No, because covariance is only allowed on interfaces that return values of that type, but don't accept them. So it's safe.
How was the addition of the covariance done from a compatibility point of view? Will earlier code still work on later versions of .NET or is recompilation necessary here? What about the other way around?
I think already compiled code will mostly work as is. Some runtime type-checks (is, IsAssignableFrom, ...) will return true where they returned false earlier.
Was previous code using this interface strictly invariant in all cases or is it possible that certain use cases will behave different now?
Not sure what you mean by that
The biggest problems are related to overload resolution. Since now additional implicit conversions are possible a different overload might be chosen.
void DoSomething(IEnumerabe<Base> bla);
void DoSomething(object blub);
IEnumerable<Derived> values = ...;
DoSomething(values);
But of course, if these overload behave differently, the API is already badly designed.
What is the real reason for that limitation? Is it just work that had to be done? Is it conceptually hard? Is it impossible?
Sure, one couldn't use the type parameters in fields, because they are allways read-write. But that can't be the answer, can it?
The reason for this question is that I'm writing an article on variance support in C# 4, and I feel that I should explain why it is restricted to delegates and interfaces. Just to inverse the onus of proof.
Update:
Eric asked about an example.
What about this (don't know if that makes sense, yet :-))
public class Lookup<out T> where T : Animal {
public T Find(string name) {
Animal a = _cache.FindAnimalByName(name);
return a as T;
}
}
var findReptiles = new Lookup<Reptile>();
Lookup<Animal> findAnimals = findReptiles;
The reason for having that in one class could be the cache that is held in the class itself. And please don't name your different type pets the same!
BTW, this brings me to optional type parameters in C# 5.0 :-)
Update 2: I'm not claiming the CLR and C# should allow this. Just trying to understand what led to that it doesnt.
First off, as Tomas says, it is not supported in the CLR.
Second, how would that work? Suppose you have
class C<out T>
{ ... how are you planning on using T in here? ... }
T can only be used in output positions. As you note, the class cannot have any field of type T because the field could be written to. The class cannot have any methods that take a T, because those are logically writes. Suppose you had this feature -- how would you take advantage of it?
This would be useful for immutable classes if we could, say, make it legal to have a readonly field of type T; that way we'd massively cut down on the likelihood that it be improperly written to. But it's quite difficult to come up with other scenarios that permit variance in a typesafe manner.
If you have such a scenario, I'd love to see it. That would be points towards someday getting this implemented in the CLR.
UPDATE: See
Why isn't there generic variance for classes in C# 4.0?
for more on this question.
As far as I know, this feature isn't supported by CLR, so adding this would require significant work on the CLR side as well. I believe that co- and contra-variance for interfaces and delegates was actually supported on CLR before the version 4.0, so this was a relatively straightforward extension to implement.
(Supporting this feature for classes would be definitely useful, though!)
If they were permitted, useful 100% type-safe (no internal typecasts) classes or structures could be defined which were covariant with regard to their type T, if their constructor accepted one or more T's or T supplier's. Useful, 100%-type-safe classes or structures could be defined which were contravariant with respect to T if their constructors accepted one or more T consumers. I'm not sure there's much advantage of a class over an interface, beyond the ability to use "new" rather than using a static factory method (most likely from a class whose name is similar to that of the interface), but I can certainly see usage cases for having immutable structures support covariance.