In C++, a coder doesn't know whether other coders will inherit his class. Should he make every function in that class virtual? Are there any drawbacks? Or is it just not acceptable at all?
In C++, you should only make a class inheritable from if you intend for it to be used polymorphically. The way that you treat polymorphic objects in C++ is very different from how you treat other objects. You don't tend to put polymorphic classes on the stack, or pass them by or return them from functions by value, since this can lead to slicing. Polymorphic objects tend to be heap-allocated, be passed around and returns by pointer or by reference, etc.
If you design a class to not be inherited from and then inherit from it, you cause all sorts of problems. If the destructor isn't marked virtual, you can't delete the object through a base class pointer without causing undefined behavior. Without the member functions marked virtual, they can't be overridden in a derived class.
As a general rule in C++, when you design the class, determine whether you want it be inherited from. If you do, mark the appropriate functions virtual and give it a virtual destructor. You might also disable the copy assignment operator to avoid slicing. Similarly, if you want the class not to be inheritable, don't give it any of these functions. In most cases it's a logic error to inherit from a class that wasn't designed to be inherited from, and most of the times you'd want to do this you can often use composition instead of inheritance to achieve this effect.
No, not usually.
A non-virtual function enforces class-invariant behavior. A virtual function doesn't. As such, the person writing the base class should think about whether the behavior of a particular function is/should be class invariant or not.
While it's possible for a design to allow all behaviors to vary in derived classes, it's fairly unusual. It's usually a pretty good clue that the person who wrote the class either didn't think much about its design, lacked the resolve to make a decision.
In C++ you design your class to be used either as a value type or a polymorphic type. See, for example, C++ FAQ.
If you are making a class to be used by other people, you should put a lot of thought into your interface and try to work out how your class will be used. Then make the decisions like which functions should be virtual.
Or better yet write a test case for your class, using it how you expect it to be used, and then make the interface work for that. You might be surprised what you find out doing it. Things you thought were absolutely necessary might turn out to be rarely needed and things that you thought were not going to be used might turn out to be the most useful methods. Doing it this way around will save you time not doing unnecessary work in the long run and end up with solid designs.
Jerry Coffin and Dominic McDonnell have already covered the most important points.
I'll just add an observation, that in the time of MFC (middle 1990s) I was very annoyed with the lack of ways hook into things. For example, the documentation suggested copying MFC's source code for printing and modifying, instead of overriding behavior. Because nothing was virtual there.
There are of course a zillion+1 ways to provide "hooks", but virtual methods are one easy way. They're needed in badly designed classes, so that the client code can fix things, but in those badly designed classes the methods are not virtual. For classes with better design there is not so much need to override behavior, and so for those classes making methods virtual by default (and non-virtual only as active choice) can be counter-productive; as Jerry remarked, virtuals provide opportunites for derived classes to screw up.
There are design patterns that can be employed to minimize the possibilities of screw-ups.
For example, wrapping internal virtuals in exposed non-virtual methods with sanity checks, and, for example, using decoupled event handling (where appropriate) instead of virtuals.
Cheers & hth.,
When you create a class, and you want that class to be used polymorphically you have to consider that the class has two different interfaces. The user interface is defined by the set of public functions that are available in your base class, and that should pretty much cover all operations that users want to perform on objects of your class. This interface is defined by the access qualifiers, and in particular the public qualifier.
There is a second interface, that defines how your class is to be extended. At that level you have to think on what behavior you want to be overridden by extending classes, and what elements of your object you want to provide to extending classes. You offer access to derived classes by means of the protected qualifier, and you offer extension points by means of virtual functions.
You should try to follow the Non-Virtual Interface idiom whenever possible. That idiom (google for it) basically tries to fully separate the two interfaces by not having public virtual functions. Users call non-virtual functions, and those in turn call on configurable functionalities by means of protected/private virtual functions. This clearly separates extension points from the class interface.
There is a single case, where virtual has to be part of the user interface: the destructor. If you want to offer your users the ability to destroy derived objects through pointers to the base, then you have to provide a virtual destructor. Else you just provide a protected non-virtual one.
He should code the functions as it is, he shouldn't make them virtual at all, as in the circumstances specified by you.
The reasons being
1> The CLASS CODER would obviously have certain use of functions he is using.
2> The inherited class may or may not make use of these functions as per requirement.
3> Any function may be overwritten in derived class without any errors.
Related
Is there any real reason not to make a member function virtual in C++? Of course, there's always the performance argument, but that doesn't seem to stick in most situations since the overhead of virtual functions is fairly low.
On the other hand, I've been bitten a couple of times with forgetting to make a function virtual that should be virtual. And that seems to be a bigger argument than the performance one. So is there any reason not to make member functions virtual by default?
One way to read your questions is "Why doesn't C++ make every function virtual by default, unless the programmer overrides that default." Without consulting my copy of "Design and Evolution of C++": this would add extra storage to every class unless every member function is made non-virtual. Seems to me this would have required more effort in the compiler implementation, and slowed down the adoption of C++ by providing fodder to the performance obsessed (I count myself in that group.)
Another way to read your questions is "Why do C++ programmers do not make every function virtual unless they have very good reasons not to?" The performance excuse is probably the reason. Depending on your application and domain, this might be a good reason or not. For example, part of my team works in market data ticker plants. At 100,000+ messages/second on a single stream, the virtual function overhead would be unacceptable. Other parts of my team work in complex trading infrastructure. Making most functions virtual is probably a good idea in that context, as the extra flexibility beats the micro-optimization.
Stroustrup, the designer of the language, says:
Because many classes are not designed to be used as base classes. For example, see class complex.
Also, objects of a class with a virtual function require space needed by the virtual function call mechanism - typically one word per object. This overhead can be significant, and can get in the way of layout compatibility with data from other languages (e.g. C and Fortran).
See The Design and Evolution of C++ for more design rationale.
There are several reasons.
First, performance: Yes, the overhead of a virtual function is relatively low seen in isolation. But it also prevents the compiler from inlining, and that is a huge source of optimization in C++. The C++ standard library performs as well as it does because it can inline the dozens and dozens of small one-liners it consists of. Additionally, a class with virtual methods is not a POD datatype, and so a lot of restrictions apply to it. It can't be copied just by memcpy'ing, it becomes more expensive to construct, and takes up more space. There are a lot of things that suddenly become illegal or less efficient once a non-POD type is involved.
And second, good OOP practice. The point in a class is that it makes some kind of abstraction, hides its internal details, and provides a guarantee that "this class will behave so and so, and will always maintain these invariants. It will never end up in an invalid state".
That is pretty hard to live up to if you allow others to override any member function. The member functions you defined in the class are there to ensure that the invariant is maintained. If we didn't care about that, we could just make the internal data members public, and let people manipulate them at will. But we want our class to be consistent. And that means we have to specify the behavior of its public interface. That may involve specific customizability points, by making individual functions virtual, but it almost always also involves making most methods non-virtual, so that they can do the job of ensuring that the invariant is maintained. The non-virtual interface idiom is a good example of this:
http://www.gotw.ca/publications/mill18.htm
Third, inheritance isn't often needed, especially not in C++. Templates and generic programming (static polymorphism) can in many cases do a better job than inheritance (runtime polymorphism). Yes, you sometimes still need virtual methods and inheritance, but it is certainly not the default. If it is, you're Doing It Wrong. Work with the language, rather than try to pretend it was something else. It's not Java, and unlike Java, in C++ inheritance is the exception, not the rule.
I'll ignore performance and memory cost, because I have no way to measure them for the "in general" case...
Classes with virtual member functions are non-POD. So if you want to use your class in low-level code which relies on it being POD, then (among other restrictions) any member functions must be non-virtual.
Examples of things you can portably do with an instance of a POD class:
copy it with memcpy (provided the target address has sufficient alignment).
access fields with offsetof()
in general, treat it as a sequence of char
... um
that's about it. I'm sure I've forgotten something.
Other things people have mentioned that I agree with:
Many classes are not designed for inheritance. Making their methods virtual would be misleading, since it implies child classes might want to override the method, and there shouldn't be any child classes.
Many methods are not designed to be overridden: same thing.
Also, even when things are intended to be subclassed / overridden, they aren't necessarily intended for run-time polymorphism. Very occasionally, despite what OO best practice says, what you want inheritance for is code reuse. For example if you're using CRTP for simulated dynamic binding. So again you don't want to imply your class will play nicely with runtime polymorphism by making its methods virtual, when they should never be called that way.
In summary, things which are intended to be overridden for runtime polymorphism should be marked virtual, and things which don't, shouldn't. If you find that almost all your member functions are intended to be virtual, then mark them virtual unless there's a reason not to. If you find that most of your member functions are not intended to be virtual, then don't mark them virtual unless there's a reason to do so.
It's a tricky issue when designing a public API, because flipping a method from one to the other is a breaking change, so you have to get it right first time. But you don't necessarily know before you have any users, whether your users are going to want to "polymorph" your classes. Ho hum. The STL container approach, of defining abstract interfaces and banning inheritance entirely, is safe but sometimes requires users to do more typing.
The following post is mostly opinion, but here goes:
Object oriented design is three things, and encapsulation (information hiding) is the first of these things. If a class design is not solid on this, then the rest doesn't really matter very much.
It has been said before that "inheritance breaks encapsulation" (Alan Snyder '86) A good discussion of this is present in the group of four design pattern book. A class should be designed to support inheritance in a very specific manner. Otherwise, you open the possibility of misuse by inheritors.
I would make the analogy that making all of your methods virtual is akin to making all your members public. A bit of a stretch, I know, but that's why I used the word 'analogy'
As you are designing your class hierarchy, it may make sense to write a function that should not be overridden. One example is if you are doing the "template method" pattern, where you have a public method that calls several private virtual methods. You would not want derived classes to override that; everyone should use the base definition.
There is no "final" keyword, so the best way to communicate to other developers that a method should not be overridden is to make it non-virtual. (other than easily ignored comments)
At the class level, making the destructor non-virtual communicates that the class should not be used as a base class, such as the STL containers.
Making a method non-virtual communicates how it should be used.
The Non-Virtual Interface idiom makes use of non-virtual methods. For more information please refer to Herb Sutter "Virtuality" article.
http://www.gotw.ca/publications/mill18.htm
And comments on the NVI idiom:
http://www.parashift.com/c++-faq-lite/strange-inheritance.html#faq-23.3
http://accu.org/index.php/journals/269 [See sub-section]
Is there a real-world example where non-virtual multiple inheritance is being used? I'd like to have one mostly for didactic reasons. Slapping around classes named A, B, C, and D, where B and C inherit from A and D inherits from B and C is perfectly fine for explaining the question "Does/Should a D object have one or two A sub-objects?", but bears no weight about why we even have both options. Many examples care about why we do want virtual inheritance, but why would we not want virtual inheritance?
I know what virtual base classes are and how to express that stuff in code. I know about diamond inheritance and examples of multiple inheritance with a virtual base class are abundant.
The best I could find is vehicles. The base class is Vehicle which is inherited by Car and Boat. Among other things, a Vehicle has occupants() and a max_speed(). So an Amphibian that inherits from both Car and Boat inherits different max_speed() on land and water – and that makes sense –, but also different occupants() – and that does not make sense. So the Vehicle sub-objects aren't really independent; that is another problem which might be interesting to solve, but this is not the question.
Is there an example, that makes sense as a real-world model, where the two sub-objects are really independent?
You're thinking like an OOP programmer, trying to design abstract models of things. C++ multiple inheritance, like many things in C++, is a tool that has a particular effect. Whether it maps onto some OOP model is irrelevant next to the utility of the tool itself. To put it another way, you don't need a "real-world model" to justify non-virtual inheritance; you just need a real-world use case.
Because a derived class inherits the members of a base class, inheritance often is used in C++ as a means of collecting a set of common functionality together, sometimes with minimal interaction from the derived class, and injecting this functionality directly into the derived class.
The Curiously Recurring Template Pattern and other mixin-like constructs are mechanisms for doing this. The idea is that you have a base class that is a template, and its template parameter is the derived class that uses it. This allows the base class to have some access to the derived class itself without virtual functions.
The simplest example I can think of in C++ is enable_shared_from_this, which allows an object whose lifetime is currently managed by a shared_ptr to actually retrieve a shared_ptr to that object just from a pointer/reference to that object. That uses CRTP to add the various members and interfaces needed to make shared_from_this possible to the derived class. And since the inheritance is public, it also allows shared_ptr's various functions that "enable shared_from_this" to to detect that a particular type has the shared_from_this stuff in it and to properly initialize it.
enable_shared_from_this doesn't need virtual inheritance, and indeed would probably not work very well with it.
Now imagine that I have some other CRTP class that injects some other functionality into an object. This functionality has nothing to do with shared_ptr, but it uses CRTP and inheritance.
Well, if I now write some type that wants to inherit from both enable_shared_from_this and this other functionality, well, that works just fine. There is no need for virtual inheritance, and in fact doing so would only make composition that much harder.
Virtual inheritance is not free. It fundamentally changes a bunch of things about how a type relates to its base classes. If you inherit from such a type, your constructors have to initialize any virtual base classes directly. The layout of such a type is very odd and is highly unlikely to be standardized. And various other things. C++ tries not to make programmers pay for functionality they don't use, so if you don't need the special properties of virtual inheritance, you shouldn't be using it.
Its the same reason C++ has non-virtual methods -- because the implementation is simpler and more efficient if you use non-virtual inheritance, so you need to explicitly ask for virtual inheritance if you want it. Since you don't need it if your classes never use multiple inheritance, that is the default.
Quoting the Effective C++ (Scott Meyers), third edition, item 7:
Declare destructors virtual in polymorphic base classes.
This implies that classes intended to be inherited from should, at a minimum, have the destructor virtual.
When writing some applications/libraries, some classes (I would say not few) are not designed with the intention of being inherited from. We usually rely on some conventions, in which one is not supposed to inherit classes that are not his own or without checking if it's safe to.
Now, a coding standard may require to write and design classes such that inheritance is always safe. I feel like this might be too much. C++11 has added the final keyword that makes sure classes are not inherited from. Would you recommend marking all classes that are not designed for inheritance as final by default?
This would make the compiler enforce what we have been doing by convention for ages. But that convention might be considered well enough understood and easy to follow (especially since inheritance is usually avoided, with composition being preferred) such that it might be just another thing to care of while writing the code.
Just because a class does not have any virtual functions does not mean that inheriting from it does not make sense. It just means it was probably not intended and that you have to be careful not to delete any instances of the derived objects through pointers to the base class. private inheritance would often be used in such a case, as an alternative to composition, to signal that inheriting from the base is purely an implementation detail and the derived class is not to be considered as having a "is-a" relationship with the base.
Marking a class final is a great way to express that a class is not intended to be derived from, at all, by making it impossible. So if that's what you want, slap a final on the class.
In some cases, marking a (derived) class final to prevent any further derivatives can also have the added benefit of making it easier for the compiler to devirtualize function calls, thus leading to better performance in some cases.
In my opinion, final should (usually) be the default and you can then remove it later if you find that you need/want to derive from the class after all.
But that's just opinion.
I need to finish others developer work but problem is that he started in different way...
So now I found in situation to use existing code where he chooses to inherit a non-abstract class (very big class, without any virtual functions) that already implements bunch of interfaces or to dismiss that code (which shouldn't be to much work) and to write another class that implements interfaces I need.
What are the pros and cons that would help me to choose the better approach.
p.s. please note that I don't have to much experience
Many Thanks
Although it is very tempting to say write it from scratch again, don't do it! The existing code may be ugly, but it looks like it does work. Since the class is big, I assume there is fair bit of history behind it as well. It might have solutions for some very obscure cases which you might not have imagined till now. What I suggest is, if possible first talk to the person who developed that class, understand how it works, then derive from it (after making its destructor virtual of course) and complete your work. Then as and when time permits slowly refactor the parts of the class into smaller more manageable classes. Also, don't forget to write a good unit-tester before you start so that you can validate the new behavior against the existing class's behavior. One more thing, there is nothing wrong in inheriting from a non-abstract base class as long as it makes sense and the base class destructor is virtual.
If the other developer has written a base-class with no virtual functions, then those functions do not need to be overridden, and it is correct to define them in a non-abstract base class.
If those functions define functionality that all the child-classes require then it would be a mistake to get rid of the base class, as you would then need to implement those functions individually in each of the child classes.
I've seen a lot of developers go 'interface-mad' in the last couple of years, but base classes still serve a function over interfaces - to provide a concrete implementation that is common to all child classes. It would be a mistake to get rid of the base class and have seperate implementations of these functions in each of the child classes.
HOWEVER, if the child classes are inheriting functionality that they do not require, or require a separate implementation of, then the Base class is a mistake and interfaces would seem like the better option to divide the functionality between the child classes.
Despite this, I would agree with Naveen that its probably not worth the extra work this will give you, it may seem simple, but if this is a big class with a lot of inheritors then it could turn out to be a nightmare. Quite often in Software Engineering you have to deal with another developer's code that you might have implemented differently. If you re-implemented it ever time you will be a very unproductive developer. I say work with what you've got and get the project finished on time.
Is there anything at all you want to use from the base class or would you end up overriding everything?
Does it define some sort of type that you want to use for an "is-a" relationship?
(for example, base class is "animal" and you want to make "cat", but if it doesn't add any behavior to its interface, that doesn't seem likely)
Is the base class used in other interfaces you need to use? (like if someone is passing objects through a reference/pointer to the base class)
If not, I'd say there's no advantage in inheriting from that class over implementing the interface(s) yourself.
What are the pros and cons that would help me to choose the better approach.
It's legal to derive from a class with no virtual functions, but that doesn't make it a good idea. When you derive from a class with virtual functions, you often use that class through pointers (eg., a class Derived that inherits from Base is often manipulated through Base*s). That doesn't work when you don't use virtual functions. Also, if you have a pointer to the base class, delete-ing it can lead to a memory leak.
However, it sounds more like these classes aren't being used through pointers-to-the-base. Instead the base class is simply used to get a lot of built in functionality, although the classes aren't related in the normal sense. Inversion of control (and has-a relationships) is a more common way to do that nowadays (split the functionality of the base class into a number of interfaces -- pure virtual base classes -- and then have the objects that currently derive from the base class instead have member variables of those interfaces).
At the very least, you'll want to split the big base class into well-defined smaller classes and use those (like mixins), which sounds like your second option.
However, that doesn't mean rewrite all the other code that uses the blob base class all in one go. That's a big undertaking and you're likely to make small typos and similar mistakes. Instead, buy yourself copies of Working Effectively With Legacy Code and Large-Scale C++ Software Design, and do the work piecemeal.
From you question it is not too clear what the problem is - looking at the title (Using non-abstract class as base) I can tell you that using an abstract class (non pure virtual - when you talk about interfaces in C++ I am assuming pure virtual abstract classes) as base makes sense only if there is common functionality you can share between subclasses - meaning that a number of classes extend the same abstract class inheriting the common implementation. If that's not the case (and you're pretty confident it's never gonna happen) then it doesn't make sense to use an abstract class.
If you can extract out some of the functionality in you big class in such a way that leads to (even potential) code reuse then it could make sense - otherwise I wouldn't see the point.
Should be a newbie question...
I have existing code in an existing class, A, that I want to extend in order to override an existing method, A::f().
So now I want to create class B to override f(), since I don't want to just change A::f() because other code depends on it.
To do this, I need to change A::f() to a virtual method, I believe.
My question is besides allowing a method to be dynamically invoked (to use B's implementation and not A's) are there any other implications to making a method virtual? Am I breaking some kind of good programming practice? Will this affect any other code trying to use A::f()?
Please let me know.
Thanks,
jbu
edit: my question was more along the lines of is there anything wrong with making someone else's method virtual? even though you're not changing someone else's implementation, you're still having to go into someone's existing code and make changes to the declaration.
If you make the function virtual inside of the base class, anything that derives from it will also have it virtual.
Once virtual, if you create an instance of A, then it will still call A::f.
If you create an instance of B and store it in a pointer of type A*. And then you call A*::->f, then it will call B's B::f.
As for side effects, there probably won't be any side effects, other than a slight (unnoticeable) performance loss.
There is a very small side effect as well, there could be a class C that also derives from A, and it may implement C::f, and expect that if A*::->f was called, then it expects A::f to be called. But this is not very common.
But more than likely, if C exists, then it does not implement C::f at all, and in which case everything is fine.
Be careful though, if you are using an already compiled library and you are modifying it's header files, what you are expecting to work probably will not. You will need to recompile the header and source files.
You could consider doing the following to avoid side effects:
Create a type A2 that derives from A and make it's f virtual
Use pointers of type A2 instead of A
Derive B from type A2.
In this way anything that used A will work in the same way guaranteed
Depending on what you need you may also be able to use a has-a relationship instead of a is-a.
There is a small implied performance penalty of a vtable lookup every time a virtual function is called. If it were not virtual, function calls are direct, since the code location is known at compile time. Wheras at runtime, a virtual function address must be referenced from the vtable of the object you're calling upon.
To do this, I need to change A::f() to
a virtual method, I believe.
Nope, you do not need to change it to a virtual method in order to override it. However, if you are using polymorphism you need to, i.e. if you have a lot of different classes deriving from A but stored as pointers to A.
There's also a memory overhead for virtual functions because of the vtable (apart from what spoulson mentioned)
There are other ways of accomplishing your goal. Does it make sense for B to be an A? For example, it makes sense for a Cat to be an Animal, but not for a Cat to be a Dog. Perhaps both A and B should derive from a base class, if they are related.
Is there just common functionality you can factor out? It sounds to me like you'll never be using these classes polymorphically, and just want the functionality. I would suggest you take that common functionality out and then make your two separate classes.
As for cost, if you're using A ad B directly, the compile will by-pass any virtual dispatching and just go straight to the functions calls, as if they were never virtual. If you pass a B into a place expecting `A1 (as a reference or pointer), then it will have to dispatch.
There are 2 performance hits when speaking about virtual methods.
vtable dispatching, its nothing to really worry about
virtual functions are never inlined, this can be much worse than the previous one, function inlining is something that can really speed things in some situations, it can never happen with a virtual function.
How kosher it is to change somebody else's code depends entirely on the local mores and customs. It isn't something we can answer for you.
The next question is whether the class was designed to be inherited from. In many cases, classes are not, and changing them to be useful base classes, without changing other aspects, can be tricky. A non-base class is likely to have everything private except the public functions, so if you need to access more of the internals in B you'll have to make more modifications to A.
If you're going to use class B instead of class A, then you can just override the function without making it virtual. If you're going to create objects of class B and refer to them as pointers to A, then you do need to make f() virtual. You also should make the destructor virtual.
It is good programming practise to use virtual methods where they are deserved. Virtual methods have many implications as to how sensible your C++ Class is.
Without virtual functions you cannot create interfaces in C++. A interface is a class with all undefined virtual functions.
However sometimes using virtual methods is not good. It doesn't always make sense to use a virtual methods to change the functionality of an object, since it implies sub-classing. Often you can just change the functionality using function objects or function pointers.
As mentioned a virtual function creates a table which a running program will reference to check what function to use.
C++ has many gotchas which is why one needs to be very aware of what they want to do and what the best way of doing it is. There aren't as many ways of doing something as it seems when compared to runtime dynamic OO programming languages such as Java or C#. Some ways will be either outright wrong, or will eventually lead to undefined behavior as your code evolves.
Since you have asked a very good question :D, I suggest you buy Scott Myer's Book: Effective C++, and Bjarne Stroustrup's book: The C++ Programming Language. These will teach you the subtleties of OO in C++ particularly when to use what feature.
If thats the first virtual method the class is going to have, you're making it no longer a POD. This can break things, although the chances for that are slim.
POD: http://en.wikipedia.org/wiki/Plain_old_data_structures