In the following simple code fragment:
#include <cstddef>
struct B
{
virtual ~B() = default;
static void operator delete(void *, int);
static void * operator new(size_t, int);
};
struct C : B
{
virtual ~C() = default;
};
clang 3.7 complains that "non-deleted function '~C' cannot override a deleted function":
http://goo.gl/Ax6oth
Neither Visual Studio nor GCC report an error in this code. Is it a clang defect or what?
static void operator delete(void *, int);
No, it's
static void operator delete(void *, std::size_t);
and that type difference causes an ambiguity that gets relevant:
cppreference.com has
The implicitly-declared or defaulted destructor for class T is
undefined (until C++11)defined as deleted (since C++11) if any of the
following is true:
[...]
The implicitly-declared destructor is virtual (because the base class
has a virtual destructor) and the lookup for the deallocation
function (operator delete() results in a call to ambiguous, deleted,
or inaccessible function.
And in the standard (draft n4140) §12.4 that is
5 A defaulted destructor for a class X is defined as deleted if:
[...]
(5.3) or, for a virtual destructor, lookup of the non-array
deallocation function results in an ambiguity or in a function that is
deleted or inaccessible from the defaulted destructor.
I encountered same problem. #decltype_auto said it's a C++11 feature. So I use option "-std=c++98" to bypass the problem.
Related
Consider the following code:
class Base {
public:
#ifdef __VIRTUAL__
virtual ~Base() {}
#else
~Base() {}
#endif
};
class Derived : public Base {
public:
~Derived() {}
private:
static void operator delete(void*) = delete;
};
int main() {
Derived d;
}
It'll compiled successfully with cmd
g++ -std=c++11 main.cpp
but failed with cmd
g++ -std=c++11 -D__VIRTUAL__ main.cpp
The output shows the operator delete is required
main.cpp: In destructor ‘virtual Derived::~Derived()’:
main.cpp:12:17: error: use of deleted function ‘static void Derived::operator delete(void*)’
~Derived() {}
^
main.cpp:14:17: error: declared here
static void operator delete(void*) = delete;
^
main.cpp: In destructor ‘virtual Derived::~Derived()’:
main.cpp:12:17: error: use of deleted function ‘static void Derived::operator delete(void*)’
~Derived() {}
^
main.cpp:14:17: error: declared here
static void operator delete(void*) = delete;
^
It means that if I use virtual destructor function, I cannnot delete operator delete .
Why is this happened, why virtual destructor required global operator delete even if created on stack.
The operator delete function is only called during deletion of an object with dynamic storage duration. It's not called by your program.
However, the standard requires that this function be available if there is a virtual destructor, even if the object is never actually used dynamically.
In the C++17 standard this is [class.dtor]/13:
At the point of definition of a virtual destructor (including an implicit definition), the non-array deallocation function is determined as if for the expression delete this appearing in a non-virtual destructor of the destructor’s class. If the lookup fails or if the deallocation function has a deleted definition, the program is ill-formed. [Note: This assures that a deallocation function corresponding to the dynamic type of an object is available for the delete-expression. —end note ]
Why does the standard require this? I have no idea but you're going to have to find some other approach to solve your problem. Maybe this thread will be useful: Any way to prevent dynamic allocation of a class?
If I have a pure virtual class InterfaceA that consists solely of a pure virtual destructor, why do I have to define the destructor as inline? I I don't I get an error when I try to link it.
Below is an admittedly contrived example, however it illustrates the point. The point does not compile for me using cmake and g++. However, if I change the InterfaceA destructor definition as follows - inline InterfaceA::~InterfaceA(){}; then it compiles.
Why is this? What does the inline keyword do?
// InterfaceA.h, include guards ommitted for clarity
class InterfaceA
{
public:
virtual ~InterfaceA() = 0;
};
InterfaceA::~InterfaceA(){};
// A.h, include guards ommitted for clarity
#include "InterfaceA.h"
class A : public InterfaceA
{
public:
A(int val)
: myVal(val){};
~A(){};
int myVal;
};
// AUser.h, include guards ommitted for clarity
#include "InterfaceA.h"
class AUser
{
public:
AUser(InterfaceA& anA)
: myA(anA){};
~AUser(){};
int getVal() const;
private:
InterfaceA& myA;
};
// AUser.cpp
#include "AUser.h"
#include "A.h"
int AUser::getVal() const
{
A& anA = static_cast<A&>(myA);
return anA.myVal;
}
// main.cpp
#include "AUser.h"
#include "A.h"
#include <iostream>
int main(){
A anA(1);
AUser user(anA);
std::cout << "value = " << user.getVal() << std::endl;
return 0;
}
You have to use the inline keyword when defining functions in header files. If you do not, and the file is included in more than one translation unit, the function will be defined twice (or more times).
The linker error is probably something like "Symbol ... is multiply defined" right?
If you defined the member function in the body of the class, it would be implicitly inline and it would also work.
See this answer
To answer the question "What does the inline keyword do?":
In the old days it would be used to ask the compiler to inline functions i.e. insert the code whenever the function is used instead of adding a function call. Eventually it turned into a simple suggestion since compiler optimizers became more knowledgeable about which functions were inline candidates. These days it is used almost exclusively to define functions in header files that must have external linkage.
inline means that compiler is allowed to add code directly to where the function was called. It also removes function from external linkage, so both your compile units would have local version of.. pure destructor.
// InterfaceA.h, include guards ommitted for clarity
class InterfaceA
{
public:
virtual ~InterfaceA() = 0;
};
You declare destructor virtual, so compiler almost never would make it inline. Why? because virtual functions are called through vtable - a internal working of virtual functions system, vtable most likely implemented as an array of pointers to member functions. If function is inlined, it would have no address, no legal pointer. If attempt to get address of function is taken, then compiler silently disregards inline keyword. The other effect will be still in place: inlined destructor stops to be visible to linker.
It may look like declaring pure virtual destructor looks like oxymoron , but it isn't. The pure destructor is kind of destructor that would be always called without causing UB. Its presence would make class abstract, but the implicit call in sequence of destructor calls would still happen. If you didn't declare destructor body, it would lead to an UB, e.g. purecall exception on Windows.
If you don't need an abstract base class, then inline definition will suffice:
class InterfaceA
{
public:
virtual ~InterfaceA() {}
};
that is treated by compiler as inline as well, but mixing inline definition and pure member declaration is not allowed.
The following code will not compile on gcc 4.8.2.
The problem is that this code will attempt to copy construct an std::pair<int, A> which can't happen due to struct A missing copy and move constructors.
Is gcc failing here or am I missing something?
#include <map>
struct A
{
int bla;
A(int blub):bla(blub){}
A(A&&) = delete;
A(const A&) = delete;
A& operator=(A&&) = delete;
A& operator=(const A&) = delete;
};
int main()
{
std::map<int, A> map;
map.emplace(1, 2); // doesn't work
map.emplace(std::piecewise_construct,
std::forward_as_tuple(1),
std::forward_as_tuple(2)
); // works like a charm
return 0;
}
As far as I can tell, the issue isn't caused by map::emplace, but by pair's constructors:
#include <map>
struct A
{
A(int) {}
A(A&&) = delete;
A(A const&) = delete;
};
int main()
{
std::pair<int, A> x(1, 4); // error
}
This code example doesn't compile, neither with coliru's g++4.8.1 nor with clang++3.5, which are both using libstdc++, as far as I can tell.
The issue is rooted in the fact that although we can construct
A t(4);
that is, std::is_constructible<A, int>::value == true, we cannot implicitly convert an int to an A [conv]/3
An expression e can be implicitly converted to a type T if and only if the declaration T t=e; is well-formed,
for some invented temporary variable t.
Note the copy-initialization (the =). This creates a temporary A and initializes t from this temporary, [dcl.init]/17. This initialization from a temporary tries to call the deleted move ctor of A, which makes the conversion ill-formed.
As we cannot convert from an int to an A, the constructor of pair that one would expect to be called is rejected by SFINAE. This behaviour is surprising, N4387 - Improving pair and tuple analyses and tries to improve the situation, by making the constructor explicit instead of rejecting it. N4387 has been voted into C++1z at the Lenexa meeting.
The following describes the C++11 rules.
The constructor I had expected to be called is described in [pairs.pair]/7-9
template<class U, class V> constexpr pair(U&& x, V&& y);
7 Requires: is_constructible<first_type, U&&>::value is true and
is_constructible<second_type, V&&>::value is true.
8 Effects: The
constructor initializes first with std::forward<U>(x) and second with
std::forward<V>(y).
9 Remarks: If U is not implicitly convertible to
first_type or V is not implicitly convertible to second_type this
constructor shall not participate in overload resolution.
Note the difference between is_constructible in the Requires section, and "is not implicitly convertible" in the Remarks section. The requirements are fulfilled to call this constructor, but it may not participate in overload resolution (= has to be rejected via SFINAE).
Therefore, overload resolution needs to select a "worse match", namely one whose second parameter is a A const&. A temporary is created from the int argument and bound to this reference, and the reference is used to initialize the pair data member (.second). The initialization tries to call the deleted copy ctor of A, and the construction of the pair is ill-formed.
libstdc++ has (as an extension) some nonstandard ctors. In the latest doxygen (and in 4.8.2), the constructor of pair that I had expected to be called (being surprised by the rules required by the Standard) is:
template<class _U1, class _U2,
class = typename enable_if<__and_<is_convertible<_U1, _T1>,
is_convertible<_U2, _T2>
>::value
>::type>
constexpr pair(_U1&& __x, _U2&& __y)
: first(std::forward<_U1>(__x)), second(std::forward<_U2>(__y)) { }
and the one that is actually called is the non-standard:
// DR 811.
template<class _U1,
class = typename enable_if<is_convertible<_U1, _T1>::value>::type>
constexpr pair(_U1&& __x, const _T2& __y)
: first(std::forward<_U1>(__x)), second(__y) { }
The program is ill-formed according to the Standard, it is not merely rejected by this non-standard ctor.
As a final remark, here's the specification of is_constructible and is_convertible.
is_constructible [meta.rel]/4
Given the following function prototype:
template <class T>
typename add_rvalue_reference<T>::type create();
the predicate condition for a template specialization is_constructible<T, Args...> shall be satisfied if and only if the following variable definition would be well-formed for some invented variable t:
T t(create<Args>()...);
[Note: These tokens are never interpreted as a function declaration. — end note] Access checking is performed as if in a context unrelated to T and any of the Args. Only the validity of the immediate context of the variable initialization is considered.
is_convertible [meta.unary.prop]/6:
Given the following function prototype:
template <class T>
typename add_rvalue_reference<T>::type create();
the predicate condition for a template specialization is_convertible<From, To> shall be satisfied if and
only if the return expression in the following code would be well-formed, including any implicit conversions
to the return type of the function:
To test() {
return create<From>();
}
[Note: This requirement gives well defined results for reference types, void types, array types, and function types. — end note] Access checking is performed as if in a context unrelated to To and From. Only
the validity of the immediate context of the expression of the return-statement (including conversions to
the return type) is considered.
For your type A,
A t(create<int>());
is well-formed; however
A test() {
return create<int>();
}
creates a temporary of type A and tries to move that into the return-value (copy-initialization). That selects the deleted ctor A(A&&) and is therefore ill-formed.
I have recently wrapped my mind around the C++0x's concepts of glvalues, xvalues and prvalues, as well as the rvalue references. However, there's one thing which still eludes me:
What is "an rvalue reference to function type"? It is literally mentioned many times in the drafts. Why was such a concept introduced? What are the uses for it?
I hate to be circular, but an rvalue reference to function type is an rvalue reference to function type. There is such a thing as a function type, e.g. void (). And you can form an rvalue reference to it.
In terms of the classification system introduced by N3055, it is an xvalue.
Its uses are rare and obscure, but it is not useless. Consider for example:
void f() {}
...
auto x = std::ref(f);
x has type:
std::reference_wrapper<void ()>
And if you look at the synopsis for reference_wrapper it includes:
reference_wrapper(T&) noexcept;
reference_wrapper(T&&) = delete; // do not bind to temporary objects
In this example T is the function type void (). And so the second declaration forms an rvalue reference to function type for the purpose of ensuring that reference_wrapper can't be constructed with an rvalue argument. Not even if T is const.
If it were not legal to form an rvalue reference to function, then this protection would result in a compile time error even if we did not pass an rvalue T to the constructor.
In the old c++ standard the following is forbidden:
int foo();
void bar(int& value);
int main()
{
bar(foo());
}
because the return type of foo() is an rvalue and is passed by reference to bar().
This was allowed though with Microsoft extensions enabled in visual c++ since (i think) 2005.
Possible workarounds without c++0x (or msvc) would be declaring
void bar(const int& value);
or using a temp-variable, storing the return-value of foo() and passing the variable (as reference) to bar():
int main()
{
int temp = foo();
bar(temp);
}
Clang's own diagnostics propaganda contains this exerpt:
Since Clang has range highlighting, it never needs to pretty print your code back out to you. This is particularly bad in G++ (which often emits errors containing lowered vtable references), but even GCC can produce inscrutible error messages in some cases when it tries to do this.
Googling this phrase doesn't give anything very helpful, and the subsequent example is completely unrelated.
Can someone please post an example of what it's talking about?
Thanks.
Here is an example:
struct a {
virtual int bar();
};
struct foo : public virtual a {
};
void test(foo *P) {
return P->bar()+*P;
}
Clang produces:
t.cc:9:18: error: invalid operands to binary expression ('int' and 'foo')
return P->bar()+*P;
~~~~~~~~^~~
GCC 4.2 produces:
t.cc: In function ‘void test(foo*)’:
t.cc:9: error: no match for ‘operator+’ in ‘(((a*)P) + (*(long int*)(P->foo::<anonymous>.a::_vptr$a + -0x00000000000000020)))->a::bar() + * P’
t.cc:9: error: return-statement with a value, in function returning 'void'
GCC does this because its C++ frontend is bolted on top of the C frontend in many cases. Instead of building C++-specific Abstract Syntax Trees (ASTs) for various C++ operations, the parser just lowers them immediately to their C equivalent. In this case, GCC synthesizes an struct to contain the vtable, and the pointer dereference to bar is then lowered into a series of C pointer dereferences, casts, pointer arithmetic etc.
Clang does not have this problem, because it has a very clean AST that directly represents the source code. If you change the example to:
struct a {
virtual int bar();
};
struct foo : public virtual a {
};
void test(foo *P) {
P->bar();
}
.. so that the code is valid, then ask clang to dump its ast with "clang -cc1 -ast-dump t.cc", you get:
...
void test(foo *P)
(CompoundStmt 0x10683cae8 <t.cc:8:19, line:10:1>
(CXXMemberCallExpr 0x10683ca78 <line:9:3, col:10> 'int'
(MemberExpr 0x10683ca40 <col:3, col:6> '<bound member function type>' ->bar 0x10683bef0
(ImplicitCastExpr 0x10683cac8 <col:3> 'struct a *' <UncheckedDerivedToBase (virtual a)>
(ImplicitCastExpr 0x10683ca28 <col:3> 'struct foo *' <LValueToRValue>
(DeclRefExpr 0x10683ca00 <col:3> 'struct foo *' lvalue ParmVar 0x10683c8a0 'P' 'struct foo *'))))))
-Chris