Initialize static const multidimensional array with inferred dimensions inside class definition - c++

Since C++11, one can initialize static const built-in types inside a class definition, like so:
class A {
public:
static const unsigned int val = 0; //allowed
};
However, doing this in Visual C++ 2013 with an array gives me an error telling me that this is not allowed:
class B {
public:
static const unsigned int val[][2] = { { 0, 1 } }; //not allowed
};
The error message simply reads "a member of type const unsigned int [][2] cannot have an in-class initializer." Instead, I'm forced to do the following:
class C {
public:
static const unsigned int val[][2];
};
const unsigned int C::val[][2] = { { 0, 1 } };
This is unfortunate because I have code which relies on the size of val, and I want to be able to change the contents of val without having to remember to go back and change a constant somewhere. Is there a different way of doing this that allows me to use sizeof on val from any point in the file below the declaration?

Your array must be a constexpr (clang and gcc specify it in their error messages) :
class B {
public:
static constexpr const unsigned int val[][2] = { { 0, 1 } };
// ^^^^^^^^
};
See it working here.
Visual Studio CTP 2013 (a "beta" version, that should be avoided for production) provides support for constepxr, which should be available in future releases, too.
EDIT :
If your compiler does not support constexpr (hopefully for you, not for too long), then you cant do in-class initialization of your static array, and must do the old way :
class C {
public:
static const unsigned int val[][2];
};
const unsigned int C::val[][2] = { { 0, 1 } };
If your array type is complete (if you declare all the array dimensions), then sizeof can be applied (the compiler knows how many elements to expect) :
;
class C {
public:
static const unsigned int val[2][2]; // Specify all dimensions.
void foo() { cout << sizeof(C::val); } // OK
};
const unsigned int C::val[][2] = { { 0, 1 } , { 2, 3 } };
int main() {
C c;
c.foo();
return 0;
}

Related

Is it possible to initialize a data member as const based on a bool passed in as argument in the constructor?

I have a class that takes in a boolean called fixed as an argument. I want it to initialize the data member position as const if fixed is true. Is this possible at all?
class PhysicsVertex
{
public:
PhysicsVertex(olc::vf2d position, const bool fixed = false) :
position(position), fixed(fixed)
{
}
olc::vf2d position; //make this const if fixed is true.
const bool fixed = false;
};
int main()
{
PhysicsVertex v1{ {0, 0}, true }; //initialize v1.position as const
PhysicsVertex v1{ {0, 0} }; //initialize v1.position as non-const
}
If your information about fixed parameter is needed at compile time then you can use templates
#include <type_traits>
struct vf2d
{
};
template <bool fixed>
class PhysicsVertex
{
public:
PhysicsVertex() :
position()
{
}
typename std::conditional<fixed, const vf2d, vf2d>::type position;
};
void foo()
{
auto vertex = PhysicsVertex<true>();
// vertex.position = vf2d(); - compile time error, you cannot change const member
auto vertex2 = PhysicsVertex<false>();
vertex2.position = vf2d(); // It works fine, you can change non const member
}
const class members are problematic, because they inhibit a whole bunch of compiler generated methods. For example PhysicsVertex cannot be copied. Often the better alternative is to not make it const but instead provide no means to modify it from outside the class:
class foo {
int x = 42;
public:
int get() { return x; }
};
The member is not modifyable.
Now to your quesiton. You cannot make it const or non-const based on a runtime value (at least not easily), but following the suggestion above you can conditionally allow to modify it or not:
class foo {
int x = 42;
bool modifyable = false;
public:
foo(bool modifyable) : modifyable(modifyable) {}
void set(int y) {
if (!modifyable) throw "cannot modify";
x = y;
}
};
You should reconsider whether you can already decide to make it const or non-const at compile time, because then you need not rely on runtime checks and excpetions but can use the approach described in this answer.

Initializing C struct with unions declared in C++17

This question is specifically about options in C++17. Assuming following declaration in C library (I cannot change them):
typedef enum {
TYPEA = 0,
TYPEB = 2,
TYPEC = 4
} SPECIFIC_TYPE_t;
typedef struct {
uint16_t Method : 1;
uint16_t Access : 2;
uint16_t VendorSpecific : 1;
uint16_t Direction : 1;
uint16_t Persistent : 1;
uint16_t Internal : 4;
uint16_t Reserved : 6;
} PROPERTY_t;
typedef RESULT_t (*CB_DataPointRead_t) (void *Service, uint8_t Pinpoint, bool VendorSpecific,
uint16_t GroupID, uint16_t ElementID, void *Data,
uint8_t *DataLengthInOut);
typedef RESULT_t (*CB_DataPointWrite_t) (void *service, uint8_t Pinpoint, bool VendorSpecific,
uint16_t GroupID, uint16_t ElementID, void *Data,
uint8_t DataLength);
typedef struct {
uint16_t GroupID;
uint16_t ElementID;
uint8_t Pinpoint;
SPECIFIC_TYPE_t Type;
uint8_t Size;
PROPERTY_t Property;
union {
struct {
CB_DataPointRead_t Read;
CB_DataPointWrite_t Write;
} Callback;
struct {
void *Data;
} DirectAccess;
} AccessType;
} DataPoint_t;
Status in pure C
To initialize the last element designator initializers work well in C:
uint16t dataPointValue = 0;
const DataPoint_t firstDatapointConfig = {
/*...*/,
.DirectAccess = { (void*)&dataPointValue; }
};
Designator initializers in C++
They appear in C++20 and aren't compatible with C in many aspects.
Problem
I'd like to initialize variable like firstDatapointConfig as const qualified in C++17. So far I don't see a way other than write a function in C (compiled as C code) and return the initialized structure to a variable before use. I tried various ways, including gnuc++17 which handles designator initializers, except it tells me:
error: 'const DataPoint_t' has no non-static data member named 'DirectAccess'
and MSVC don't digest this method of initialization at all without C++20 enabled.
Addressing the last element outside initializer, don't work either:
datapoint.DirectAccess = { &value };
results in the following error:
error: 'struct DataPoint_t' has no member named 'DirectAccess'
Comment
It was much easier to use these structures in Rust after processing them through bindgen :-)
Question
Is there a way to initialize the variable of DataPoint_t type in C++17 with DirectAccess element filled with the right value?
Is there a way to initialize the variable of DataPoint_t type in C++17 with DirectAccess element filled with the right value?
Yes, even though my way of doing it is a little cumbersome. There may be easier ways:
// Make a usable type of that anonymous entity:
using AccessType_t = decltype(DataPoint_t::AccessType);
// Use the new type and prepare what you need:
AccessType_t at;
at.DirectAccess.Data = nullptr;
// initialize your DataPoint_t
const DataPoint_t firstDatapointConfig{1,2,3,TYPEA,4, PROPERTY_t{}, at};
If you do this a lot you could make a helper function:
using AccessType_t = decltype(DataPoint_t::AccessType);
using Callback_t = decltype(AccessType_t::Callback);
using DirectAccess_t = decltype(AccessType_t::DirectAccess);
template<class U>
constexpr auto init_AccessType(U u) {
AccessType_t at;
if constexpr (std::is_same_v<U,Callback_t>) {
at.Callback = u;
} else if constexpr (std::is_same_v<U,DirectAccess_t>) {
at.DirectAccess = u;
} else {
// uninitialized
}
return at;
}
const DataPoint_t firstDatapointConfig{1,2,3,TYPEA,4, PROPERTY_t{},
init_AccessType(DirectAccess_t{nullptr})};
I'd like to initialize variable like firstDatapointConfig as const qualified in C++17
With all that C++ has to offer:
constexpr DataPoint_t create_DataPoint_t(uint16_t *v) {
DataPoint_t r{};
r.AccessType.DirectAccess.Data = v;
return r;
}
const DataPoint_t firstDatapointConfig = create_DataPoint_t(&dataPointValue);
with designated initializers:
const DataPoint_t firstDatapointConfig2 = {
.AccessType = {
.DirectAccess = {
.Data = &dataPointValue
}
}
};
Addressing the last element outside initializer, don't work either:
datapoint.DirectAccess = { &value };
Because there is no such element, there is datapoint.AccessType.DirectAccess.Data.
To initialize the last element designator initializers work well in C:
uint16t dataPointValue = 0;
const DataPoint_t firstDatapointConfig = {
/*...*/,
.DirectAccess = { (void*)&dataPointValue; }
};
The presented code is invalid - uint16t is meant to be uint16_t and ; is a typo. And still after fixing the typos, no, the presented code is invalid in "pure C" and "does not work well" godbolt link. There is no such thing as DirectAccess in DataPoint_t - there is such member in the unnamed union declared inside DataPoint_t. You can do in C:
const DataPoint_t firstDatapointConfig3_in_C = {
.AccessType = {
.DirectAccess = { (void*)&dataPointValue }
}
};
or
const DataPoint_t firstDatapointConfig4_in_C = {
.AccessType.DirectAccess = { (void*)&dataPointValue }
};
or
const DataPoint_t firstDatapointConfig4_in_C = {
.AccessType.DirectAccess.Data = (void*)&dataPointValue
};
The cast to void* is superfluous - all pointers are implicitly converted to void*. Note that the following:
const DataPoint_t firstDatapointConfig5_in_C = {
.AccessType = { (void*)&dataPointValue }
};
would be equal to:
const DataPoint_t firstDatapointConfig6_in_C = {
.AccessType.Callback.Read = (void*)&dataPointValue
};
Most probably you are coding under -fms-extensions with GNU gcc or with MSVC, in which case you should be aware that you are using an extension that imports unnamed structure members to parent structure. The code is invalid in "pure C", it's using an extension to C.

Passing const variable between member functions to be index of array

I have tried to solve this, but I can't. I have a class definition and I want a member function (siz) to return a constant value to another member function (abc). This value is used as maximum index in an array declaration in that function. But this doesn't seems to work. Here is a simplified version:
class bin {
constexpr int siz();
public:
void abc();
};
constexpr int bin::siz() {
const int sizz = sizeof(int) * 8;
}
void bin::abc() {
char arr[siz()]; // compiler: this expression didn't evaluate as constant (¿?)
};
However, this other very similar code (but using simple functions) does compile...
constexpr int siz() {
const int sizz = sizeof(int) * 8;
return sizz;
}
int main() {
char arr[siz()];
return 0;
}
I am not entirely sure but I think the problem is that in bin::abc, the object can be anything at run time. Hence, bin::siz() cannot be evaluated at compile time.
The following works fine
int main()
{
bin b;
char arr[b.siz()];
}
after changing bin to:
class bin {
public:
constexpr int siz();
};
constexpr int bin::siz() {
return sizeof(int) * 8;
}
If siz does not depend on the state of the object, as in your posted code, I suggest making it a static member function.
The following works fine for me.
class bin {
public:
static constexpr int siz();
void abc() const;
};
constexpr int bin::siz() {
return sizeof(int) * 8;
}
void bin::abc() const {
char arr[siz()];
}
int main()
{
bin b;
char arr[b.siz()];
}

Can't a class have static constexpr member instances of itself?

This code is giving me incomplete type error.
What is the problem? Isn't allowed for a class to have static member instances of itself?
Is there a way to achieve the same result?
struct Size
{
const unsigned int width;
const unsigned int height;
static constexpr Size big = { 480, 240 };
static constexpr Size small = { 210, 170 };
private:
Size( ) = default;
};
A class is allowed to have a static member of the same type. However, a class is incomplete until the end of its definition, and an object cannot be defined with incomplete type. You can declare an object with incomplete type, and define it later where it is complete (outside the class).
struct Size
{
const unsigned int width;
const unsigned int height;
static const Size big;
static const Size small;
private:
Size( ) = default;
};
const Size Size::big = { 480, 240 };
const Size Size::small = { 210, 170 };
see this here: http://coliru.stacked-crooked.com/a/f43395e5d08a3952
This doesn't work for constexpr members, however.
Is there a way to achieve the same result?
By "the same result", do you specifically intend the constexpr-ness of
Size::big and Size::small? In that case maybe this would be close enough:
struct Size
{
const unsigned int width = 0;
const unsigned int height = 0;
static constexpr Size big() {
return Size { 480, 240 };
}
static constexpr Size small() {
return Size { 210, 170 };
}
private:
constexpr Size() = default;
constexpr Size(int w, int h )
: width(w),height(h){}
};
static_assert(Size::big().width == 480,"");
static_assert(Size::small().height == 170,"");
As a workaround you can use a separate base class which definition is complete when defining the constants in the derived class.
struct size_impl
{
//data members and functions here
unsigned int width;
unsigned int height;
};
struct size: public size_impl
{
//create the constants as instantiations of size_impl
static constexpr size_impl big{480,240};
static constexpr size_impl small{210,170};
//provide implicit conversion constructor and assignment operator
constexpr size(const size_impl& s):size_impl(s){}
using size_impl::operator=;
//put all other constructors here
};
//test:
constexpr size a = size::big;
You can put the base class in a separate namespace to hide its definition if you want to.
The code compiles with clang and gcc

initialize a const array in a class initializer in C++

I have the following class in C++:
class a {
const int b[2];
// other stuff follows
// and here's the constructor
a(void);
}
The question is, how do I initialize b in the initialization list, given that I can't initialize it inside the body of the function of the constructor, because b is const?
This doesn't work:
a::a(void) :
b([2,3])
{
// other initialization stuff
}
Edit: The case in point is when I can have different values for b for different instances, but the values are known to be constant for the lifetime of the instance.
With C++11 the answer to this question has now changed and you can in fact do:
struct a {
const int b[2];
// other bits follow
// and here's the constructor
a();
};
a::a() :
b{2,3}
{
// other constructor work
}
int main() {
a a;
}
Like the others said, ISO C++ doesn't support that. But you can workaround it. Just use std::vector instead.
int* a = new int[N];
// fill a
class C {
const std::vector<int> v;
public:
C():v(a, a+N) {}
};
It is not possible in the current standard. I believe you'll be able to do this in C++0x using initializer lists (see A Brief Look at C++0x, by Bjarne Stroustrup, for more information about initializer lists and other nice C++0x features).
std::vector uses the heap. Geez, what a waste that would be just for the sake of a const sanity-check. The point of std::vector is dynamic growth at run-time, not any old syntax checking that should be done at compile-time. If you're not going to grow then create a class to wrap a normal array.
#include <stdio.h>
template <class Type, size_t MaxLength>
class ConstFixedSizeArrayFiller {
private:
size_t length;
public:
ConstFixedSizeArrayFiller() : length(0) {
}
virtual ~ConstFixedSizeArrayFiller() {
}
virtual void Fill(Type *array) = 0;
protected:
void add_element(Type *array, const Type & element)
{
if(length >= MaxLength) {
// todo: throw more appropriate out-of-bounds exception
throw 0;
}
array[length] = element;
length++;
}
};
template <class Type, size_t Length>
class ConstFixedSizeArray {
private:
Type array[Length];
public:
explicit ConstFixedSizeArray(
ConstFixedSizeArrayFiller<Type, Length> & filler
) {
filler.Fill(array);
}
const Type *Array() const {
return array;
}
size_t ArrayLength() const {
return Length;
}
};
class a {
private:
class b_filler : public ConstFixedSizeArrayFiller<int, 2> {
public:
virtual ~b_filler() {
}
virtual void Fill(int *array) {
add_element(array, 87);
add_element(array, 96);
}
};
const ConstFixedSizeArray<int, 2> b;
public:
a(void) : b(b_filler()) {
}
void print_items() {
size_t i;
for(i = 0; i < b.ArrayLength(); i++)
{
printf("%d\n", b.Array()[i]);
}
}
};
int main()
{
a x;
x.print_items();
return 0;
}
ConstFixedSizeArrayFiller and ConstFixedSizeArray are reusable.
The first allows run-time bounds checking while initializing the array (same as a vector might), which can later become const after this initialization.
The second allows the array to be allocated inside another object, which could be on the heap or simply the stack if that's where the object is. There's no waste of time allocating from the heap. It also performs compile-time const checking on the array.
b_filler is a tiny private class to provide the initialization values. The size of the array is checked at compile-time with the template arguments, so there's no chance of going out of bounds.
I'm sure there are more exotic ways to modify this. This is an initial stab. I think you can pretty much make up for any of the compiler's shortcoming with classes.
ISO standard C++ doesn't let you do this. If it did, the syntax would probably be:
a::a(void) :
b({2,3})
{
// other initialization stuff
}
Or something along those lines. From your question it actually sounds like what you want is a constant class (aka static) member that is the array. C++ does let you do this. Like so:
#include <iostream>
class A
{
public:
A();
static const int a[2];
};
const int A::a[2] = {0, 1};
A::A()
{
}
int main (int argc, char * const argv[])
{
std::cout << "A::a => " << A::a[0] << ", " << A::a[1] << "\n";
return 0;
}
The output being:
A::a => 0, 1
Now of course since this is a static class member it is the same for every instance of class A. If that is not what you want, ie you want each instance of A to have different element values in the array a then you're making the mistake of trying to make the array const to begin with. You should just be doing this:
#include <iostream>
class A
{
public:
A();
int a[2];
};
A::A()
{
a[0] = 9; // or some calculation
a[1] = 10; // or some calculation
}
int main (int argc, char * const argv[])
{
A v;
std::cout << "v.a => " << v.a[0] << ", " << v.a[1] << "\n";
return 0;
}
Where I've a constant array, it's always been done as static. If you can accept that, this code should compile and run.
#include <stdio.h>
#include <stdlib.h>
class a {
static const int b[2];
public:
a(void) {
for(int i = 0; i < 2; i++) {
printf("b[%d] = [%d]\n", i, b[i]);
}
}
};
const int a::b[2] = { 4, 2 };
int main(int argc, char **argv)
{
a foo;
return 0;
}
You can't do that from the initialization list,
Have a look at this:
http://www.cprogramming.com/tutorial/initialization-lists-c++.html
:)
A solution without using the heap with std::vector is to use boost::array, though you can't initialize array members directly in the constructor.
#include <boost/array.hpp>
const boost::array<int, 2> aa={ { 2, 3} };
class A {
const boost::array<int, 2> b;
A():b(aa){};
};
How about emulating a const array via an accessor function? It's non-static (as you requested), and it doesn't require stl or any other library:
class a {
int privateB[2];
public:
a(int b0,b1) { privateB[0]=b0; privateB[1]=b1; }
int b(const int idx) { return privateB[idx]; }
}
Because a::privateB is private, it is effectively constant outside a::, and you can access it similar to an array, e.g.
a aobj(2,3); // initialize "constant array" b[]
n = aobj.b(1); // read b[1] (write impossible from here)
If you are willing to use a pair of classes, you could additionally protect privateB from member functions. This could be done by inheriting a; but I think I prefer John Harrison's comp.lang.c++ post using a const class.
interestingly, in C# you have the keyword const that translates to C++'s static const, as opposed to readonly which can be only set at constructors and initializations, even by non-constants, ex:
readonly DateTime a = DateTime.Now;
I agree, if you have a const pre-defined array you might as well make it static.
At that point you can use this interesting syntax:
//in header file
class a{
static const int SIZE;
static const char array[][10];
};
//in cpp file:
const int a::SIZE = 5;
const char array[SIZE][10] = {"hello", "cruel","world","goodbye", "!"};
however, I did not find a way around the constant '10'. The reason is clear though, it needs it to know how to perform accessing to the array. A possible alternative is to use #define, but I dislike that method and I #undef at the end of the header, with a comment to edit there at CPP as well in case if a change.