The code below compiles correctly on clang 3.8.1-1 on ArchLinux.
Is this clang bug?
gcc issue correct warning/error on this.
template <class T>
struct BugReproducer{
using size_type = typename T::size_type;
int bug1(size_type count);
int bug2(size_type count) const;
static int bug3(size_type count);
};
template <class T>
int BugReproducer<T>::bug1(size_type const count){
// this is a bug. must be not allowed
count = 5;
// return is to use the result...
return count;
}
template <class T>
int BugReproducer<T>::bug2(size_type const count) const{
// same for const method
count = 5;
return count;
}
template <class T>
int BugReproducer<T>::bug3(size_type const count){
// same for static method
count = 5;
return count;
}
struct DummyVector{
using size_type = int;
};
int main(){
using BugRepr = BugReproducer<DummyVector>;
BugRepr reproducer;
auto a = reproducer.bug1(1);
auto b = reproducer.bug2(1);
auto c = BugRepr::bug3(1);
// return is to use the result...
return a + b + c;
}
Here how I compile:
[nmmm#zenbook HM3]$ clang x.cc -std=c++11 -lstdc++ -Wall -Wpedantic -Wconversion
clang and c++14 - same result.
[nmmm#zenbook HM3]$ clang x.cc -std=c++14 -lstdc++ -Wall -Wpedantic -Wconversion
Here is gcc output:
[nmmm#zenbook HM3]$ gcc x.cc -std=c++11 -lstdc++ -Wall -Wpedantic -Wconversion
x.cc: In instantiation of ‘int BugReproducer<T>::bug1(BugReproducer<T>::size_type) [with T = DummyVector; BugReproducer<T>::size_type = int]’:
x.cc:46:28: required from here
x.cc:13:8: error: assignment of read-only parameter ‘count’
count = 5;
~~~~~~^~~
x.cc: In instantiation of ‘int BugReproducer<T>::bug2(BugReproducer<T>::size_type) const [with T = DummyVector; BugReproducer<T>::size_type = int]’:
x.cc:47:28: required from here
x.cc:22:8: error: assignment of read-only parameter ‘count’
count = 5;
~~~~~~^~~
x.cc: In instantiation of ‘static int BugReproducer<T>::bug3(BugReproducer<T>::size_type) [with T = DummyVector; BugReproducer<T>::size_type = int]’:
x.cc:48:20: required from here
x.cc:29:8: error: assignment of read-only parameter ‘count’
count = 5;
~~~~~~^~~
Yes, this is a bug in clang; filed at https://llvm.org/bugs/show_bug.cgi?id=30365.
The nature of the bug is that in a class template member function definition appearing outside ([class.mfct]/1) the class template, with the type of a parameter dependent on the class template parameters, clang uses the parameter type of the declaration rather than the parameter type of the definition where they differ in topmost cv-qualification. Simplified example:
template<class T> struct A { void f(typename T::U); };
template<class T> void A<T>::f(typename T::U const i) { i = 1; }
struct X { using U = int; };
int main() { A<X>{}.f(0); }
Per [dcl.fct]/5 the type of i within the definition of A<X>::f is int const (Use of 'const' for function parameters):
5 - [...] After producing the list of parameter types, any top-level cv-qualifiers modifying a parameter
type are deleted when forming the function type. [...]
[ Note: This transformation does not affect the types of the parameters. [...] — end note ]
Related
I am trying to implement in a C++ class a constexpr member function which returns a template parameter. The code is supposed to be c++11 compatible. However I encounter compilation issues when the templated class also contains STL containers as data members such as std::vector (which are untouched by the constexpr member function).
A minimal example is given by the following code:
#include <vector>
#include <iostream>
#include <array>
template<size_t n>
struct A
{
constexpr size_t dimensions() const
{
return n;
}
private:
std::vector<double> a;
};
int main(int argc,char ** argv)
{
auto a=A<3>();
std::array<double,a.dimensions()> arr;
}
The code compiles correctly with the commands
g++ -std=c++14 -O3 quickTest.cpp -o test -Wall
clang++ -std=c++11 -O3 quickTest.cpp -o test -Wall
but fails when I use
g++ -std=c++11 -O3 quickTest.cpp -o test -Wall
with the error
quickTest.cpp:22:33: error: call to non-‘constexpr’ function ‘size_t A<n>::dimensions() const [with long unsigned int n = 3; size_t = long unsigned int]’
std::array<double,a.dimensions()> arr;
~~~~~~~~~~~~^~
quickTest.cpp:10:20: note: ‘size_t A<n>::dimensions() const [with long unsigned int n = 3; size_t = long unsigned int]’ is not usable as a ‘constexpr’ function because:
constexpr size_t dimensions() const
^~~~~~~~~~
quickTest.cpp:22:33: error: call to non-‘constexpr’ function ‘size_t A<n>::dimensions() const [with long unsigned int n = 3; size_t = long unsigned int]’
std::array<double,a.dimensions()> arr;
~~~~~~~~~~~~^~
quickTest.cpp:22:33: note: in template argument for type ‘long unsigned int’
Why does the code not compile with gcc -std=c++11 but does compile with clang++ -std=c++11 ?
How could one make this code snippet work with older versions of gcc which many not support c++14/17, but only c++11 ?
I am using gcc 8.1.1 and clang 6.0.1
C++11 had a rule [dcl.constexpr]/8:
... The class of which that function is a member shall be a literal type ([basic.types]).
struct A is not a literal type because of the vector, hence its non-static member functions cannot be constexpr.
So GCC is right to reject the code in C++11 mode.
C++14 removed that restriction.
The solution for C++11 is to declare dimensions() static:
static constexpr size_t dimensions()
{
return n;
}
Live demo
The following code does not compile with GCC 5.2 (C++14). It does compile with clang 3.6 (C++14). (original code can be found here)
#include <cstddef>
#include <algorithm>
#include <type_traits>
#include <utility>
template <typename T>
class aggregate_wrapper;
template <typename T, std::size_t n>
class aggregate_wrapper<T[n]> {
public:
using array = T[n];
template <typename... Ts, typename = decltype(array{std::declval<Ts>()...})>
aggregate_wrapper(Ts&&... xs)
: arr_{std::forward<Ts>(xs)...} {
// nop
}
aggregate_wrapper(const array& arr) {
std::copy(arr, arr + n, arr_);
}
aggregate_wrapper(array&& arr) {
std::move(arr, arr + n, arr_);
}
operator T* () {
return arr_;
}
operator const T* () const {
return arr_;
}
constexpr std::size_t size() const {
return n;
}
private:
array arr_;
};
int main() {
aggregate_wrapper<int[3]> arr;
static_assert(arr.size() == 3, "");
}
The error message produced is
main.cpp: In function 'int main()':
main.cpp:44:3: error: non-constant condition for static assertion
static_assert(arr.size() == 3, "");
^
main.cpp:44:25: error: call to non-constexpr function 'constexpr std::size_t aggregate_wrapper<T [n]>::size() const [with T = int; long unsigned int n = 3ul; std::size_t = long unsigned int]'
static_assert(arr.size() == 3, "");
^
main.cpp:34:25: note: 'constexpr std::size_t aggregate_wrapper<T [n]>::size() const [with T = int; long unsigned int n = 3ul; std::size_t = long unsigned int]' is not usable as a constexpr function because:
constexpr std::size_t size() const {
^
main.cpp:34:25: error: enclosing class of constexpr non-static member function 'constexpr std::size_t aggregate_wrapper<T [n]>::size() const [with T = int; long unsigned int n = 3ul; std::size_t = long unsigned int]' is not a literal type
main.cpp:10:7: note: 'aggregate_wrapper<int [3]>' is not literal because:
class aggregate_wrapper<T[n]> {
^
main.cpp:10:7: note: 'aggregate_wrapper<int [3]>' is not an aggregate, does not have a trivial default constructor, and has no constexpr constructor that is not a copy or move constructor
Any ideas? Should the code compile according to the standard?
Or you could just make your existing variadic constructor serve as your constexpr constructor to perform the default construction:
template <typename... Ts, typename = decltype(array{std::declval<Ts>()...})>
constexpr // <---- ADD THIS
aggregate_wrapper(Ts&&... xs)
: arr_{std::forward<Ts>(xs)...} {
// nop
}
In order for g++ to get it compiled you will need to add a default constructor:
aggregate_wrapper() = default;
please see it in action at: http://coliru.stacked-crooked.com/a/df1ac057960bebc7
I have the feeling that clang under the hood added it, but I am not 100% sure ...
GCC is wrong. Its diagnostic, in part, says:
main.cpp:34:25: note: '<...>' is not usable as a constexpr function because:
main.cpp:34:25: error: enclosing class of constexpr non-static member function '<...>' is not a literal type
... but there is no such rule. See [dcl.constexpr]/3 for the list of constraints that apply here.
You can work around the bogus GCC diagnostic by adding a dummy constexpr constructor (it's fine for that constructor to be private and/or deleted if you don't want any of your real constructors to be constexpr) or by making size be static.
I have the following piece of code which represents an actual bigger piece of code:
#include <iostream>
using namespace std;
template<size_t N> class A {
public:
static constexpr size_t getN() {return N;}
};
template<size_t N> class B {
public:
void print() { cout << "B created: " << N << '\n';}
};
template <class T> class C {
public:
void set(T* a) {
t_ptr = a;
}
void create() {
constexpr int m = t_ptr->getN();
B<m> b;
b.print();
}
private:
T* t_ptr;
};
int main() {
constexpr int n = 2;
A<n> a;
C<A<n> > c;
c.set(&a);
c.create();
}
Compiling with g++ -o main main.cpp -std=c++11 and GCC/G++ 4.8.3 the expected output is received:
B created: 2
However, with GCC/G++ 4.9.1 the code does not compile, output:
main.cpp: In member function ‘void C<T>::create()’:
main.cpp:27:15: error: the value of ‘m’ is not usable in a constant expression
B<m> b;
^
main.cpp:26:27: note: ‘m’ used in its own initializer
constexpr int m = t_ptr->getN();
^
main.cpp:27:16: error: the value of ‘m’ is not usable in a constant expression
B<m> b;
^
main.cpp:26:27: note: ‘m’ used in its own initializer
constexpr int m = t_ptr->getN();
^
main.cpp:27:19: error: invalid type in declaration before ‘;’ token
B<m> b;
^
main.cpp:28:15: error: request for member ‘print’ in ‘b’, which is of non-class type ‘int’
b.print();
^
This is caused by a known bug in GCC 4.9: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=59937 and in this older thread https://gcc.gnu.org/ml/gcc-bugs/2013-11/msg00067.html the usage of extern is proposed as a workaround. However, I am not able to get this workaround working.
Could you guys help me to make this code compile in GCC 4.9? Thank you!
Since this is not constexpr the access to this->t_ptr is not either.
clang's error is a bit more helpful
implicit use of 'this' pointer is only allowed within the
evaluation of a call to a 'constexpr' member function
Referring to:
N3690 5.19/2 (emphasis added)
A conditional-expression e is a core constant expression unless the evaluation of e, following the rules of the abstract machine, would evaluate one of the following expressions:
— this, except in a constexpr function or a constexpr constructor that is being evaluated as
part of e;
Calling the static member function via the typename works
constexpr int m = T::getN();
Here's a sample program:
#include <type_traits>
#include <stdio.h>
template <typename X>
struct test
{
operator int() const { puts("?"); return 0; }
template <typename T, typename = typename std::enable_if<std::is_same<X, void*>::value, T>::type>
operator T() const { puts("T"); return 0; }
};
int main()
{
test<void*> t;
char* c = (char*)t;
switch (t)
{
case 0: break;
}
return 0;
}
And this is the error that g++-4.7 gives
user#user:~$ g++-4.7 -std=c++0x test.cpp
test.cpp: In function ‘int main()’:
test.cpp:13:14: error: ambiguous default typeconversion from ‘test<void*>’
test.cpp:13:14: error: candidate conversions include ‘template<class T, class> test::operator void*() const [with T = T; <template-parameter-2-2> = <template-parameter-1-2>; X = void*]’
g++ 4.6 compiles it without errors and different operators are actually called.
Is there a way to make this work under g++ 4.7?
UPDATE: actually it works in 4.6 without any enable_if at all... so the question still applies but I'm now not sure if enable_if will help.
If you add an explicit cast to int here:
switch ((int)t)
Then it should compile.
I think it's complaining about the conversion being ambiguous since there exists more than one type that can hold a 0 value.
I'm using g++ 4.8 though.
I found at least one "acceptable" solution:
#define switch(x) \
switch( (typename switch_type<__typeof(x)>::type)(x) )
which switch_type trait can be extended to resolve ambiguity for specific app-related types (property types).
I have the following template function:
template <std::size_t first, std::size_t last, typename T>
bool in_range(T& in)
{
for(auto i = in.begin(); i!=in.end(); ++i)
if(*i<first || *i>last)
return false;
return true;
}
but when I try to use it as such:
std::vector<int> test;
test.push_back(1);
test.push_back(5);
test.push_back(6);
std::cout<<in_range<4,7>(test);
I get this weird error:
main.cpp: In instantiation of 'bool in_range(T&) [with long long unsigned int first = 4ull; long long unsigned int last = 7ull; T = std::vector<int>]':
main.cpp:31:34: required from here
What am I doing wrong?
EDIT: full build log: http://pastebin.com/Cwemq2Hk
If I build that with C++11 support enabled, then it compiles. Here is a demonstation.
Before C++11, auto had a different meaning, and so auto i = ... was invalid - it declared a variable with no type.
I guess you're using GCC; depending on the version, you'll need to specify either -std=c++0x or -std=c++11 as a command-line option.