Communities

Writing
Writing
Codidact Meta
Codidact Meta
The Great Outdoors
The Great Outdoors
Photography & Video
Photography & Video
Scientific Speculation
Scientific Speculation
Cooking
Cooking
Electrical Engineering
Electrical Engineering
Judaism
Judaism
Languages & Linguistics
Languages & Linguistics
Software Development
Software Development
Mathematics
Mathematics
Christianity
Christianity
Code Golf
Code Golf
Music
Music
Physics
Physics
Linux Systems
Linux Systems
Power Users
Power Users
Tabletop RPGs
Tabletop RPGs
Community Proposals
Community Proposals
tag:snake search within a tag
answers:0 unanswered questions
user:xxxx search by author id
score:0.5 posts with 0.5+ score
"snake oil" exact phrase
votes:4 posts with 4+ votes
created:<1w created < 1 week ago
post_type:xxxx type of post
Search help
Notifications
Mark all as read See all your notifications »
Q&A

Welcome to Software Development on Codidact!

Will you help us build our independent community of developers helping developers? We're small and trying to grow. We welcome questions about all aspects of software development, from design to code to QA and more. Got questions? Got answers? Got code you'd like someone to review? Please join us.

Comments on Why can't a derived class add a const qualifier to a method?

Parent

Why can't a derived class add a const qualifier to a method?

+5
−0

Say we have an abstract class Foo declared as so:

class Foo {
  public:
    virtual void test() = 0;
};

Let's say that we make a derived concrete class FooDerived, and decide to mark it's version of test as const as it doesn't modify its state.

class FooDerived : public Foo {
  public:
    void test() const override { std::cout << "FooDerived!" << std::endl; }
};
main.cpp:12:10: error: ‘void FooDerived::test() const’ marked ‘override’, but does not override

Why is this not allowed? I thought that it would be, given that the only real function of the qualification is to be able to be called on const instances. It shouldn't be any more restrictive on the caller than the non-const version. Also, is there a good way to achieve this?

History
Why does this post require attention from curators or moderators?
You might want to add some details to your flag.
Why should this post be closed?

0 comment threads

Post
+4
−1

@InfiniteDissent‭ gives an excellent explanation why such a feature would be problematic from a programming language design perspective. As a consequence, C++ is defined such that the code you wanted to write is no valid C++ code and thus compilers will reject it.

Although not directly addressed by the technical nature of your question, I would like to add an aspect about semantic aspects of such overriding scenarios: You expected this feature to be supported because, from your position, adding constness in a derived method is just an additional guarantee to the caller, and thus it seems safe to add this guarantee when overriding.

But, not modifying the object may in some cases actually mean a violation of the base class interface contract: Assume the base class method is defined such that calling it will bring the object to a defined state (for example, re-initializing it). Then a client holding an object of the base class type will, when calling that method, assume that the object will be brought to that state. If the overriding function, however, will not touch the object, this can mean a breach of this contract.

I just mention this to make it clear that, also from an interface design point of view, adding constness for a derived method is not always "safe" but can result in a broken class hierarchy.

History
Why does this post require attention from curators or moderators?
You might want to add some details to your flag.

1 comment thread

LSP (1 comment)
LSP
Derek Elkins‭ wrote over 2 years ago

What you're describing is simply a violation of the LSP (Liskov Substitution Principle) and has nothing to do with constness. The scenarios you describe (both specific and generic) are ones where there is no possible const implementation that doesn't violate the LSP. The problem is C++ provides no way to differentiate a declaration that allows self mutation and one that requires self-mutation per its spec. In the former case, having a adding const modifier on an overriding method would be fine, while in the latter it would never make sense. (Technically, it would still be safe vacuously.) We could imagine adding a nonconst modifier to represent the "requires" case and treating no modifier as a union of const and nonconst. The scenarios you describe would then use the nonconst modifier and const overrides would be disallowed. However, methods with no modifier could be overridden with either const or nonconst methods.