Why does C++ not have reflection?

Question

This is a somewhat bizarre question. My objectives are to understand the language design decision and to identify the possibilities of reflection in C++.

1. Why C++ language committee did not go towards implementing reflection in the language? Is reflection too difficult in a language that does not run on a virtual machine (like java)?

2. If one were to implement reflection for C++, what will be the challenges?

I guess the uses of reflection are well-known: editors can be more easily written, program code will be smaller, mocks can be generated for unit tests and so on. But it would be great if you could comment on uses of reflection too.

Answer 1

There are several problems with reflection in C++.

• It's a lot of work to add, and the C++ committee is fairly conservative, and don't spend time on radical new features unless they're sure it'll pay off. (A suggestion for adding a module system similar to .NET assemblies has been made, and while I think there's general consensus that it'd be nice to have, it's not their top priority at the moment, and has been pushed back until well after C++0x. The motivation for this feature is to get rid of the #include system, but it would also enable at least some metadata).

• You don't pay for what you don't use. That's one of the must basic design philosophies underlying C++. Why should my code carry around metadata if I may never need it? Moreover, the addition of metadata may inhibit the compiler from optimizing. Why should I pay that cost in my code if I may never need that metadata?

• Which leads us to another big point: C++ makes very few guarantees about the compiled code. The compiler is allowed to do pretty much anything it likes, as long as the resulting functionality is what is expected. For example, your classes aren't required to actually be there. The compiler can optimize them away, inline everything they do, and it frequently does just that, because even simple template code tends to create quite a few template instantiations. The C++ standard library relies on this aggressive optimization. Functors are only performant if the overhead of instantiating and destructing the object can be optimized away. operator[] on a vector is only comparable to raw array indexing in performance because the entire operator can be inlined and thus removed entirely from the compiled code. C# and Java make a lot of guarantees about the output of the compiler. If I define a class in C#, then that class will exist in the resulting assembly. Even if I never use it. Even if all calls to its member functions could be inlined. The class has to be there, so that reflection can find it. Part of this is alleviated by C# compiling to bytecode, which means that the JIT compiler can remove class definitions and inline functions if it likes, even if the initial C# compiler can't. In C++, you only have one compiler, and it has to output efficient code. If you were allowed to inspect the metadata of a C++ executable, you'd expect to see every class it defined, which means that the compiler would have to preserve all the defined classes, even if they're not necessary.

• And then there are templates. Templates in C++ are nothing like generics in other languages. Every template instantiation creates a new type. std::vector<int> is a completely separate class from std::vector<float>. That adds up to a lot of different types in a entire program. What should our reflection see? The template std::vector? But how can it, since that's a source-code construct, which has no meaning at runtime? It'd have to see the separate classes std::vector<int> and std::vector<float>. And std::vector<int>::iterator and std::vector<float>::iterator, same for const_iterator and so on. And once you step into template metaprogramming, you quickly end up instantiating hundreds of templates, all of which get inlined and removed again by the compiler. They have no meaning, except as part of a compile-time metaprogram. Should all these hundreds of classes be visible to reflection? They'd have to, because otherwise our reflection would be useless, if it doesn't even guarantee that the classes I defined will actually be there. And a side problem is that the template class doesn't exist until it is instantiated. Imagine a program which uses std::vector<int>. Should our reflection system be able to see std::vector<int>::iterator? On one hand, you'd certainly expect so. It's an important class, and it's defined in terms of std::vector<int>, which does exist in the metadata. On the other hand, if the program never actually uses this iterator class template, its type will never have been instantiated, and so the compiler won't have generated the class in the first place. And it's too late to create it at runtime, since it requires access to the source code.

• And finally, reflection isn't quite as vital in C++ as it is in C#. The reason is again, template metaprogramming. It can't solve everything, but for many cases where you'd otherwise resort to reflection, it's possible to write a metaprogram which does the same thing at compile-time. boost::type_traits is a simple example. You want to know about type T? Check its type_traits. In C#, you'd have to fish around after its type using reflection. Reflection would still be useful for some things (the main use I can see, which metaprogramming can't easily replace, is for autogenerated serialization code), but it would carry some significant costs for C++, and it's just not necessary as often as it is in other languages.

Edit: In response to comments:

cdleary: Yes, debug symbols do something similar, in that they store metadata about the types used in the executable. But they also suffer from the problems I described. If you've ever tried debugging a release build, you'll know what I mean. There are large logical gaps where you created a class in the source code, which has gotten inlined away in the final code. If you were to use reflection for anything useful, you'd need it to be more reliable and consistent. As it is, types would be vanishing and disappearing almost every time you compile. You change a tiny little detail, and the compiler decides to change which types get inlined and which ones don't, as a response. How do you extract anything useful from that, when you're not even guaranteed that the most relevant types will be represented in your metadata? The type you were looking for may have been there in the last build, but now it's gone. And tomorrow, someone will check in a small innocent change to a small innocent function, which makes the type just big enough that it won't get completely inlined, so it'll be back again. That's still useful for debug symbols, but not much more than that. I'd hate trying to generate serialization code for a class under those terms.

Evan Teran: Of course these issues could be resolved. But that falls back to my point #1. It'd take a lot of work, and the C++ committee has plenty of things they feel is more important. Is the benefit of getting some limited reflection (and it would be limited) in C++ really big enough to justify focusing on that at the expense of other features? Is there really a huge benefit in adding features the core language which can already (mostly) be done through libraries and preprocessors like QT's? Perhaps, but the need is a lot less urgent than if such libraries didn't exist. For your specific suggestions though, I believe disallowing it on templates would make it completely useless. You'd be unable to use reflection on the standard library, for example. What kind of reflection wouldn't let you see a std::vector? Templates are a huge part of C++. A feature that doesn't work on templates is basically useless.

But you're right, some form of reflection could be implemented. But it'd be a major change in the language. As it is now, types are exclusively a compile-time construct. They exist for the benefit of the compiler, and nothing else. Once the code has been compiled, there are no classes. If you stretch yourself, you could argue that functions still exist, but really, all there is is a bunch of jump assembler instructions, and a lot of stack push/pop's. There's not much to go on, when adding such metadata.

But like I said, there is a proposal for changes to the compilation model, adding self-contained modules, storing metadata for select types, allowing other modules to reference them without having to mess with #includes. That's a good start, and to be honest, I'm surprised the standard committee didn't just throw the proposal out for being too big a change. So perhaps in 5-10 years? :)