I recently came to know that in C++ pure virtual functions can optionally have a body.
What are the real-world use cases for such functions?
The classic is a pure virtual destructor:
class abstract {
public:
virtual ~abstract() = 0;
};
abstract::~abstract() {}
You make it pure because there's nothing else to make so, and you want the class to be abstract, but you have to provide an implementation nevertheless, because the derived classes' destructors call yours explicitly. Yeah, I know, a pretty silly textbook example, but as such it's a classic. It must have been in the first edition of The C++ Programming Language.
Anyway, I can't remember ever really needing the ability to implement a pure virtual function. To me it seems the only reason this feature is there is because it would have had to be explicitly disallowed and Stroustrup didn't see a reason for that.
If you ever feel you need this feature, you're probably on the wrong track with your design.
Pure virtual functions with or without a body simply mean that the derived types must provide their own implementation.
Pure virtual function bodies in the base class are useful if your derived classes wants to call your base class implementation.
One reason that an abstract base class (with a pure virtual function) might provide an implementation for a pure virtual function it declares is to let derived classes have an easy 'default' they can choose to use. There isn't a whole lot of advantage to this over a normal virtual function that can be optionally overridden - in fact, the only real difference is that you're forcing the derived class to be explicit about using the 'default' base class implementation:
class foo {
public:
virtual int interface();
};
int foo::interface()
{
printf( "default foo::interface() called\n");
return 0;
};
class pure_foo {
public:
virtual int interface() = 0;
};
int pure_foo::interface()
{
printf( "default pure_foo::interface() called\n");
return 42;
}
//------------------------------------
class foobar : public foo {
// no need to override to get default behavior
};
class foobar2 : public pure_foo {
public:
// need to be explicit about the override, even to get default behavior
virtual int interface();
};
int foobar2::interface()
{
// foobar is lazy; it'll just use pure_foo's default
return pure_foo::interface();
}
I'm not sure there's a whole lot of benefit - maybe in cases where a design started out with an abstract class, then over time found that a lot of the derived concrete classes were implementing the same behavior, so they decided to move that behavior into a base class implementation for the pure virtual function.
I suppose it might also be reasonable to put common behavior into the pure virtual function's base class implementation that the derived classes might be expected to modify/enhance/augment.
One use case is calling the pure virtual function from the constructor or the destructor of the class.
The almighty Herb Sutter, former chair of the C++ standard committee, did give 3 scenarios where you might consider providing implementations for pure virtual methods.
Gotta say that personally – I find none of them convincing, and generally consider this to be one of C++'s semantic warts. It seems C++ goes out of its way to build and tear apart abstract-parent vtables, then briefly exposes them only during child construction/destruction, and then the community experts unanimously recommend never to use them.
The only difference of virtual function with body and pure virtual function with body is that existence of second prevent instantiation. You can't mark class abstract in c++.
This question can really be confusing when learning OOD and C++. Personally, one thing constantly coming in my head was something like:
If I needed a Pure Virtual function to also have an implementation, so why make it "Pure" in first place ? Why not just leaving it only "Virtual" and have derivatives, both benefit and override the base implementation ?
The confusion comes to the fact that many developers consider the no body/implementation as the primary goal/benefit of defining a pure virtual function. This is not true!
The absence of body is in most cases a logical consequence of having a pure virtual function. The main benefit of having a pure virtual function is defining a contract: By defining a pure virtual function, you want to force every derivative to always provide their own implementation of the function. This "contract aspect" is very important especially if you are developing something like a public API. Making the function only virtual is not so sufficient because derivatives are no longer forced to provide their own implementation, therefore you may loose the contract aspect (which can be limiting in the case of a public API).
As commonly said :
"Virtual functions can be overrided, Pure Virtual functions must be overrided."
And in most cases, contracts are abstract concepts so it doesn't make sense for the corresponding pure virtual functions to have any implementation.
But sometimes, and because life is weird, you may want to establish a strong contract among derivatives and also want them to somehow benefit from some default implementation while specifying their own behavior for the contract. Even if most book authors recommend to avoid getting yourself into these situations, the language needed to provide a safety net to prevent the worst! A simple virtual function wouldn't be enough since there might be risk of escaping the contract. So the solution C++ provided was to allow pure virtual functions to also be able to provide a default implementation.
The Sutter article cited above gives interesting use cases of having Pure Virtual functions with body.
Related
Is there any reason to declare a method virtual if a class has no subclasses, and is always used directly?
For example:
class Foo {
public:
virtual void DoBar {
// Do something here.
}
}
I came across this in some code I was reading, and couldn't find any justification.
Thanks!
Well the essence of virtual keyword is directly related to inheritance. This is an extract from CPP Ref:-
Virtual members A virtual member is a member function that can be
redefined in a derived class, while preserving its calling properties
through references. The syntax for a function to become virtual is to
precede its declaration with the virtual keyword
So IMHO - the ans to your question is no - it makes no sense - unless the code has changed from initial implementation - and trust me that happens a lot!
It is useful when writing library code to keep the future programmer in mind who may want to extend the class and provide their own behaviour. For example it is common to have a virtual Paint() function or virtual mouse handling functions in GUI libraries. They provide default implementations, but they allow the possibility of extension.
If that class is meant to be derive from then yes it makes sense. These decisions should be made when deciding the architecture of a program, and defining what can be done with the interfaces. If they do not want this to be derived from then it should not be virtual. If they do want it to be derived from then it should be virtual (and it should also make the destructor virtual).
Normally calling virtual functions from constructors is considered bad practice, because overridden functions in sub-objects will not be called as the objects have not been constructed yet.
But, Consider the following classes:
class base
{
public:
base() {}
~base() {}
private:
virtual void startFSM() = 0;
};
class derived final : public base
, public fsm_action_interface
{
public:
derived() : base{}
, theFSM_{}
{ startFSM(); }
/// FSM interface actions
private:
virtual void startFSM()
{ theFSM_.start(); }
private:
SomeFSMType theFSM_;
}
In this case class derived is marked as final so no o further sub-objects can exist. Ergo the virtual call will resolve correctly (to the most derived type).
Is it still considered bad practice?
This would still be considered bad practice as this sort of this almost always indicates bad design. You'd have to comment the heck out of the code to explain why this works in this one case.
T.C.'s comment above reinforces one of the reasons why this is considered bad practice.
What happens if, a year down the line, you decide that derived
shouldn't be final after all?
That said, in the example above, the pattern will work without issue. This is because the constructor of the most derived type is the one calling the virtual function. This problem manifests itself when a base class's constructor calls a virtual function that resolves to a subtype's implementation. In C++, such a function won't get called, because during base class construction, such calls will never go to a more derived class than that of the currently executing constructor or destructor. In essence, you end up with behavior you didn't expect.
Edit:
All (correct/non-buggy) C++ implementations have to call the version of the function defined at the level of the hierarchy in the current constructor and no further.
The C++ FAQ Lite covers this in section 23.7 in pretty good detail.
Scott Meyers also weighs in on the general issue of calling virtual functions from constructors and destructors in Effective C++ Item 9
Regarding
” Normally calling virtual functions from constructors is considered bad practice, because overridden functions in sub-objects will not be called as the objects have not been constructed yet.
That is not the case. Among competent C++ programmers it’s normally not regarded as bad practice to call virtual functions (except pure virtual ones) from constructors, because C++ is designed to handle that well. In contrast to languages like Java and C#, where it might result in a call to a method on an as yet uninitialized derived class sub-object.
Note that the dynamic adjustment of dynamic type has a runtime cost.
In a language oriented towards ultimate efficiency, with "you don't pay for what you don't use" as a main guiding principle, that means that it's an important and very much intentional feature, not an arbitrary choice. It's there for one purpose only. Namely to support those calls.
Regarding
” In this case class derived is marked as final so no o further sub-objects can exist. Ergo the virtual call will resolve correctly (to the most derived type).
The C++ standard guarantees that at the time of construction execution for a class T, the dynamic type is T.
Thus there was no problem about resolving to incorrect type, in the first place.
Regarding
” Is it still considered bad practice?
It is indeed bad practice to declare a member function virtual in a final class, because that’s meaningless. The “still” is not very meaningful either.
Sorry, I didn't see that the virtual member function was inherited as such.
Best practice for marking a member function as an override or implementation of pure virtual, is to use the keyword override, not to mark it as virtual.
Thus:
void startFSM() override
{ theFSM_.start(); }
This ensures a compilation error if it is not an override/implementation.
It can work, but why does startFSM() need to be virtual? In no case do you actually want to actually call anything but derived::startFSM(), so why have any dynamic binding at all? If you want to have it call the same thing as a dynamically binded method, make another non-virtual function called startFSM_impl() and have both the constructor and startFSM() call it instead.
Always prefer non-virtual to virtual if you can help it.
A pure virtual function isn't supposed to have a body, but I just noticed that the following code is accepted by the compiler:
class foo
{
virtual void dummy() = 0
{
cout << "hello";
}
};
So, why are pure virtual functions allowed to have a body? Also, even when the function has a body, the class still can't be instantiated, why is that?
Pure virtual function can have a body, but the fact that you declare them pure virtual is exactly to say that a derived implementation is required.
You can execute the pure virtual method from a derived method (using an explicit BaseClass::method()) but still you have to provide an implementation.
Not being able to instantiate a class with a pure virtual method that has not been overriden is the main point of the pure virtual declaration. In other words the idea of declaring a method pure virtual is to ensure that the programmer will not forget about providing its implementation.
Generally speaking, abstract class is used to define interface and/or partial implementation, and is intended to be inherited by concrete classes. It is a way of enforcing a contract between class designer and users of that class, whatever they are.
If you don't need this contract and want instantiation, then do not enforce it to be abstract class, i.e. remove = 0.
If your concern is why would pure virtual functions have body in the first place, then one good example could be this quote from the Effective C++ book by Scott Meyers:
Derived classes that implement this pure virtual function may call
this implementation somewhere in their code. If part of the code of
two different derived classes is similar then it makes sense to move
it up in the hierarchy, even if the function should be pure virtual.
Another example is pure virtual destructor (which is also a function by the way):
...
virtual ~foo() = 0;
...
foo::~foo() {} // required!
when you are required to provide implementation for it, but at the same time want to make it pure virtual in order to make the foo class abstract. So, in fact, the core of the language itself relies on this feature.
Diving back into the world of C++, and experimenting with interfaces. I can find plenty of online examples, but 99% of them are trivial 'all-in-one-file'.
Interfaces can be classified into 3 basic types- interfaces where (all|some|zero) functions must be overridden in a concrete subclass (all|some|zero pure virtual functions in c++ parlance).
Is it possible to implement any of the 3 in a single header file? (no .cpp twin) Why/why not? How? If not, what are my options for the .h/.cpp pair?
Some of the options I have seen are: virtual destructor, protected destructor, inline destructor, pure virtual destructor with an implementation,... my head is spinning!
EDIT: meant pure virtual, not virtual void
If you make your interface pure virtual, the compiler will give an error if you forget to implement a method in your concrete class.
If there are methods that have some sane default implementation you can provide it, but this is just a convenience.
If you make the destructor virtual, you can delete the object via its interface pointer rather than requiring the concrete object pointer.
Herb Sutter has some interesting thoughts on interfaces which are at direct odds with the way most of us implement an interface: http://www.gotw.ca/publications/mill18.htm
Read this awesome piece of internet wisdom: http://www.parashift.com/c++-faq-lite/dtors.html
1) You never need a .cpp file, you can code all your programs in .h, but doing this is terrible style and will inevitabely slow your compilation.
MyClass : public MyInterface {
public:
virtual void myFunction() {
// override and implemenht
}
};
2) When inheriting classes you don't need to worry about declaring your destructor to be virtual, but you need to make sure that classes you are inheriting from have virtual destructors. Otherwise you might be faced with memory leaks. In so far as I know this is not necessary for interface classes with only pure virtual functions.
Destructor being inline makes no difference to you.
Destructor being declared as void is something that should give you a compiler error :).
Edit:
About the 3 types of interfaces, I am getting the sense that you are a little bit confused and this is hampering your googling.
An interface is usually a class that has no functions defined and you must inherit and implement all of them.
A class that has some functionality but still has some pure virtuals is usually called abstract.
When 'zero' functions must be overridden this is neither an interface nor an abstract class, it is just a plain old class that you can inherit and do whatever.
Check out the C++ faq.
Interface where all functions must be overridden by a concrete subclass is called interface
Interface where some functions must be overridden by a concrete subclass is called abstract class
Interface where zero functions must be overridden by a concrete subclass is called class.
You can implement virtual member functions inside header file. They will not inline though, because they have to have address to be placed into virtual table.
... my head is spinning!
To get it spinning more you can also have pure virtual destructor. This is useful trick when you want to have abstract class, but no real useful method to put pure virtual:
class A
{
public:
virtual ~A() = 0;
};
A::~A() {}
class B : public A
{
public:
~B() {}
};
Class A has destructor implemented, but it still can not be instantiated directly, because it is interface. Yet, when it comes to destroying B compiler has to obey to protocol and call ~A(). That is why you have to implement it
I read different opinions about this question. Let's say I have an interface class with a bunch of pure virtual methods. I implement those methods in a class that implements the interface and I do not expect to derive from the implementation.
Is there a need for declaring the methods in the implementation as virtual as well? If yes, why?
No - every function method declared virtual in the base class will be virtual in all derived classes.
But good coding practices are telling to declare those methods virtual.
virtual is optional in the derived class override declaration, but for clarity I personally include it.
Real need - no. Once a method is declared as virtual in the base class, it stays virtual for all derived classes. But it's good to know which method is virtual and which - not, instead of checking this in the base class. Also, in most cases, you cannot be sure, if your code will be derived or not (for example, if you're developing some software for some firm). As I said, it's not a problem, as once declared as virtual, it stays virtual, but just in case .. (:
There is no requirement to mark them virtual.
I'd start by arguing that virtual advertises to readers that you expect derived classes to override the virtual to do something useful. If you are implementing the virtual to do something, then the virtual method might have nothing to do with the kind of thing your class is: in which case marking it virtual is silly. consider:
class CommsObject {
virtual OnConnect();
virtual OnRawBytesIn();
};
class XMLStream : public CommsObject {
virtual OnConnect();
OnRawBytesIn();
virtual OnXMLData();
};
In that example, OnConnect is documented as virtual in both classes because it makes sense that a descendent would always want to know. OnRawBytesIn doesn't make sense to "Export" from XMLStream as it uses that to handle raw bytes, and generate parsed data - which it notifies via OnXMLData().
Having done all that, I'd then argue that the maintainer of a 3rd class, looking at XMLStream, might think that it would be "safe" to create their own OnRawBytes function and expect it to work as a normal overloaded function - i.e. the base class would call the internal correct one, and the outer one would mask the internal OnRawBytes.
So omitting the virtual has hidden important detail from consumers of the class and made the code behave in unexpected ways.
So ive gone full circle: Don't try to use it as a hint about the intended purpose of a function - DO use it as a hint about the behaviour of the function: mark functions virtual consistently so downstream programmers have to read less files to know how a function is going to behave when overridden.
No, it is not needed and it doesn't prevent any coding errors although many coders prefer to put it.
Once C++0x becomes mainstream you'll be able to use the override specifier instead.
Once 'virtual', it's virtual all the way down to the last child. Afaik, that's the feature of the c++.
If you never derive from a class then there is no point making its methods virtual.