为什么g++5在自动类型推导中推导对象而不是initializer_list

Question

I recently came upon this code:

struct Foo{};

int main() 
{
    Foo a;
    // clang++ deduces std::initializer_list
    // g++5.1 deduces Foo
    auto b{a}; 
    a = b;
}

It compiles fine with g++5.1, but fails in clang++ (used both -std=c++11 and -std=c++14 , same results). The reason is that clang++ deduces the type of b as std::initializer_list<Foo> , whereas g++5.1 deduces as Foo . AFAIK, the type should indeed be (counter-intuitive indeed) std::initializer_list here . Why does g++5 deduces the type as Foo ?

Answer 1

There is a proposal for C++1z that implements new type deduction rules for brace initialization ( N3922 ), and I guess gcc implemented them:

For direct list-initialization:
1. For a braced-init-list with only a single element, auto deduction will deduce from that entry;
2. For a braced-init-list with more than one element, auto deduction will be ill-formed.

[Example:

 auto x1 = { 1, 2 }; // decltype(x1) is std::initializer_list<int> auto x2 = { 1, 2.0 }; // error: cannot deduce element type auto x3{ 1, 2 }; // error: not a single element auto x4 = { 3 }; // decltype(x4) is std::initializer_list<int> auto x5{ 3 }; // decltype(x5) is int.

-- end example]

Here is the gcc patch concerning the new changes with regards to "Unicorn initialization."