Correct way to implement singleton with shared pointer? - c++

I'm attempting my first bigger C++ object and I'm running into a bit of a problem when it comes to implementing a singleton using a shared pointer.
When I try to compile the following, I Visual Studio gives me this error:
"Error C2248 'PaletteManager::PaletteManager': cannot access private
member declared in class 'PaletteManager' {omitted}\xmemory line 228"
I'm guessing the issue is because I have a private constructor/destructor and make_shared is trying to call the constructor. This access problem makes sense to me, but if I want to use a shared pointer as the way to access my singleton object, what am I supposed to do? The code works just fine with a raw pointer, but I wanted to try and do things the clean way with a smart pointer.
Here is the header file for the code in question:
class PaletteManager
{
private:
// array representing palette colors
uint* paletteColors;
// private constructor/destructor because singleton
PaletteManager();
~PaletteManager();
// load palette from file TODO: not implemented
void loadPallette();
static std::shared_ptr<PaletteManager> instance;
public:
const uint PALETTE_MAX_COLORS = 4;
uint getPaletteColor(uint idx);
// singleton accessor
static std::shared_ptr<PaletteManager> getInstance();
};
And here is the function at issue in the cpp file:
std::shared_ptr<PaletteManager> PaletteManager::instance = nullptr;
std::shared_ptr<PaletteManager> PaletteManager::getInstance()
{
if (!PaletteManager::instance)
{
PaletteManager::instance = std::make_shared<PaletteManager>();
}
return PaletteManager::instance;
}

PaletteManager::instance = std::make_shared<PaletteManager>();
This results in std::make_shared attempting to new a PalletteManager object, and then construct a std::shared_ptr from it. This is what std::make_shared does, that's how it works.
This is not going to work here: that's because PalletteManager has a private constructor, and std::make_shared, the template in the C++ library, is not it's friend.
You will have to explicitly new the object in getInstance, a member of this class, which can use the private constructor, and then manually construct the std::shared_ptr from the pointer to this new object.
TLDR: you cannot use std::make_shared for an object with a private constructor. C++'s rules do not allow you to do that.

You can make std::make_unique<>() a friend function of your class with the private constructor. Presumably this would work with std::make_shared<>() too? (But I haven't tested it.)
For example, see:
How to make std::make_unique a friend of my class

Related

Passkey idiom with std::make_shared. xmemory cannot access private Key constructor

In my program I need a factory function that provides instances of separate class because I need control over the details of each instance and to be aware of how many instances are in existence at a time. In particular returning a std::shared_ptr is ideal, but this is initially impossible due to a known issue with the "make" fucntions of the std::pointer types as they would need to be friends with my Widget class as well, which isn't portable since it relies on the current implementation of those methods that may change.
To get around this, I want to employ the Passkey idiom, which was directly recommend for this situation as described at the bottom of this: https://abseil.io/tips/134. I also based my implementation off the lessons learned here: https://arne-mertz.de/2016/10/passkey-idiom/
This is a sample project that uses my same setup as my full project:
#include <iostream>
class Widget
{
public:
class Key
{
friend class Factory;
private:
Key() {};
Key(const Key&) = default;
};
int mTest;
explicit Widget(Key, int test) { mTest = test; }
int getTestVar() { return mTest; }
};
class Factory
{
public:
int mTestPass;
Factory(int input) { mTestPass = input; }
std::shared_ptr<Widget> factoryMake() { return std::make_shared<Widget>(Widget::Key{}, mTestPass); }
};
int main()
{
Factory testFactory(10);
std::shared_ptr<Widget> testWidget = testFactory.factoryMake();
std::cout << testWidget->getTestVar();
return 0;
}
However, I get
Error C2248 'Widget::Key::Key': cannot access private member declared in class 'Widget::Key' TestProject ...\include\xmemory 204
This has me completely lost, since the error coming from xmemory.cpp indicates that std::make_shared is sill trying to access a private constructor. As far as I'm aware, the construction of the Key instance occurs within the factoryMake() function, which belongs to Factory, and then that instance is passed into the std::make_shared function; therefore, std::make_shared should not need access to the Key constructor since an already constructed instance is being passed to it, which is the entire point of using this idiom in this context. The class itself is public so it should have no issues interacting with the type Key, only the constructor should be inaccessible.
In the end I can just skip using std::make_shared and instead use the shared_ptr(*T) constructor with a raw pointer, but this is slightly less efficient due to the extra allocation it requires, as noted in my first link. It isn't a big deal as I'm not making many widgets but I'd ultimately prefer to get the more ideal implementation working.
What am I missing here?
The problem is that the compiler needs to copy your Widget::Key when you call std::make_shared, and you have declared the copy constructor private. You can solve this in one of two ways:
Make the copy constructor of Widget::Key public.
Change the Widget constructor to take the Widget::Key by const reference:
explicit Widget(const Key&, ...

Calling a non-static variable from a static function

I encountered an issue while trying to do something in the process of learning C++ and I am not sure how to handle the situation:
class Command
{
public:
const char * Name;
uint32 Permission;
bool (*Handler)(EmpH*, const char* args); // I do not want to change this by adding more arguments
};
class MyClass : public CommandScript
{
public:
MyClass() : CommandScript("listscript") { }
bool isActive = false;
Command* GetCommands() const
{
static Command commandtable[] =
{
{ "showlist", 3, &DoShowlistCommand } // Maybe handle that differently to fix the problem I've mentioned below?
};
return commandtable;
}
static bool DoShowlistCommand(EmpH * handler, const char * args)
{
// I need to use isActive here for IF statements but I cannot because
// DoShowlistCommand is static and isActive is not static.
// I cannot pass it as a parameter either because I do not want to
// change the structure of class Command at all
// Is there a way to do it?
}
};
Any help would be greatly appreciated! :)
// Is there a way to do it?
No.
Either pass it as parameter, make it static, or make DoShowlistCommand non-static.
There are two potential answers here:
1. about use of non static items in a static functions:
As said in our previous question/answer, this is not possible, unless you'd have in the static function a specific MyClass object (and use object.isActive). Unfortunately, you can't do this here :
your code comments clearly show that you can't add a MyClass parameter to the function call;
the existing parameters don't suggest that you have already a pointer to parent class object;
it would not be adivsable to use global objects in such a context.
2. about what your're trying to do:
It seems that you want to have the function static, because you want to provide it in a table that maps script-commands to function pointers.
Alternative A
If all the function pointers used in commandtable are members of MyClass, you could think of using a pointer to a member function instead of a pointer to a function. The outside object/function that sets isActive on an object, could then refer the pointer to the member function, on the MyClass object it knows.
Alternative B
Revise the design of your code to implement your script engine by using the command design pattern: it's ideally suited for this kind of problems. It will require some refactoring of your code, but it will be so much more maintenable and extensible afterwards !
I don't think there is any way to do it. Here is why:
A static member function is not attached to any particular object, which means it cannot access other members that are not static, since they are attached to an object.
It doesn't look like you need to make it a static member. If you are sure you do - then pass it as a parameter. For example, make a
bool isActive();
function, and pass an argument from it to that function somewhere when you call this 'problematic' one.
You also could change your member variable to static, but it looks like you need it for EACH object, not one-for-all

std::map::insert exception

This is another, "My code isn't working and i don't know why, " question i'm afraid. I just don't have enough knowledge of the stl to know why std::map::insert would throw an exception. If you know what cases it throws an exception, you can probably skip this wall of text and just answer. If you just desperately need some background on the issue, then have at it. I'll post my code and explain what is done, and i would be very grateful if all you with a better knowledge of the stl could explain what could be wrong with my call to insert.
I wrote an object awhile ago that i use occasionally as my go to factory object. It's main purpose is basically to take a string and store both the string and a "create new object function" pointer, so that in the end, you can call a function, pass a string, and if there is a valid registration for it, it returns a new instance of a derived object. Less talk, more code, here's what i got:
factory.h
#ifndef FACTORY_H
#define FACTORY_H
// library tools
#include <map>
#include <string>
// Simplified registration macros
#define DECLARE_DERIVED(T, base) static Factory<base>::DerivedRegister<T> reg;
#define DEFINE_DERIVED(T, base, s) Factory<base>::DerivedRegister<T> T::reg(s);
template<class base>
class Factory
{
protected:
template<class T>
static base * createT() { return new T;}
public:
typedef std::map<std::string, base*(*)()> map_type;
virtual ~Factory(){ }
static base * createInstance(const std::string & s)
{
if(!m_Map.count(s))
return nullptr;
std::map<std::string, base*(*)()>::iterator it = m_Map.find(s);
return it->second();
}
template <class T>
struct DerivedRegister;
protected:
static map_type m_Map;
};
template<class base>
template<class T>
struct Factory<base>::DerivedRegister : public Factory<base>
{
DerivedRegister(std::string const & s)
{
m_Map.insert(std::pair<std::string, base*(*)()>(s, &createT<T>));
}
};
#endif
here's a better explanation of what it does real quick. Let's say you have a base class, class A . and then you have any number of derived classes. I make a factory object somewhere templated to A, and then either create a derived register object manually, or use the macro at the top within the derived classes declaration to create a static registry object. Then you define it in the implementation and call it's constructor, passing in a string to be used to identify the object. using the factory member createInstance you can pass in a string identifier and have a derived object returned, pointed to by an A *.
example:
A.h
class A
{
};
A.cpp
// the map for this factory template has to be defined somewhere, as it is static
Factory<A>::map_type Factory<A>::m_Map;
b.h
#include <A.h>
class B : public A
{
// anywhere in declaration of derived B
DECLARE_DERIVED(A, B)
};
b.cpp
// just somewhere in cpp file
DEFINE_DERIVED(A, B, "B")
main.cpp
int main()
{
A * ptr;
Factory<A> factory;
ptr = factory.createInstance("B");
}
This object has worked for me in the past, mostly without a hitch. Now i'm doing a project a little more complicated. I've taken a liking to the data organization/ api design involved with game engines, and i'm just trying to implement a solution of cataloging, (but not instantiated) shaders, so that you have a whole list of the shaders you've programmed, but they will not be instantiated at run-time unless needed. That aside, this question actually has nothing to do with d3d11, or at least i hope not.
So here is what's going on. I have an object that represents a graphics-shader abstract class. All the shaders you wish to write must derive from this object. The you derive from and implement it's functions differently for all your different shaders.
let's call the base object "SYNC::D3D11Shader" in namespace sync and the derived shaders "ColorShader" "LightShader" and "TextureShader". Since i do not simply want to make an std::map of instances of these shaders within the rendering object, i make a factory within the rendering object like this.
D3D11Renderer.h
class D3D11Renderer
{
// many other members...
Factory<D3D11Shader> m_ShaderFactory;
// many other member...
};
D3D11Renderer.cpp
// define this templated classes map or you'll get undefined errors
Factory<SYNC::D3D11Shader>::map_type Factory<SYNC::D3D11Shader>::m_Map;
and then in the ColorShader i use the macros like so
D3D11ColorShader.h
class D3D11ColorShader : public SYNC::D3D11Shader
{
// ...lotsa members
DECLARE_DERIVED(D3D11ColorShader, SYNC::D3D11Shader)
// lotsa member...
};
D3D11ColorShader.cpp
// define the registery object with it's key here
DEFINE_DERIVED(D3D11ColorShader, SYNC::D3D11Shader, "ColorShader")
this all compiles fine, and where it throws it's exception is where i first call the registryObjects constructor in D3D11ColorShader.cpp, spefically at the insert call. the exception error is this:
Unhandled exception at 0x772315de in Syncopate.exe: 0xC0000005: Access
violation reading location 0x00000004.
So in reality, the question boils down to, when does std::map::insert throw an exception and why. I just knew everyone would be asking for some background on what i'm doing. Low and behold, a giant wall of text has appeared! All i really need is a hunch.
also should i or should i not tag d3d11, because the question doesn't really pertain to it?
Here's a problem:
std::map<std::string, base*(*)()>::iterator it = m_Map.find(s);
return it->second();
if the call to find fails (i,e. it can't find 's' in the map), then it will return m_Map.end(). Dereferencing that is a no-no.
My guess would be that this is due to the order of initialization of static variables. There is no way to control this order. So you are not guaranteed that your initialization:
Factory<A>::map_type Factory<A>::m_Map;
gets called before this initialization:
DEFINE_DERIVED(A, B, "B")
In this case the latter statement must be getting initialized first and so you map has not been allocated.
An alternative design pattern would control the initialization of the singleton factories. If you have an explicit Initialize function on each which creates the factory object then you can call this at the start of your main. E.g.
Factory.h
class Factory {
private:
static Factory* instance_;
public:
static Initialize(){instance_=new Factory;}
Factory* instance(){return instance_;}
}
Factory.cpp
static Factory* Factory::instance_ = NULL;
If you have a lot of factories you will probably want a single initialize function that initializes them all, and you will have to remember to add in the new factories as you create them.
Okay, i have actually been laboring over this error for about a day, and only now do i realize what is wrong.
problem 1:
the derived shaders header was never actually included anywhere throughout the project, and despite the fact that it never needs to be directly instantiated, it still has to be included somewhere so it can be linked and included in build.
problem 2:
interesting enough, just like combinatorial said, the initialization order was not done one after the other, but then looking over my old code, it seemed to initialize correctly before. what the difference here was, i put the factory of the derived objects within a different object then the base class. what i used to do was declare a static function and static factory within the base class so that you could instantiate any of it's registered derived classes from the base class itself. When the factory is included within the base class instead, and instantiation is done through a static function, the initialization order of all the statics seems to be constently in order ( not sure if this is always true). It runs fine now after changing this.
so now, my answer, you can get operating system exceptions like this for trying to use references to objects that were never actually included anywhere in your project. I don't have a very good knowledge of compilers or linkers to tell you why it seemed to compile fine, despite this object never being included. If someone wants to extend my answer, please.
I use MSVC++ 2010 express if that pertains to this predicament.

Member CComPtr in a C++/CLR class

I'm writing a C++/CLI class that needs to hold on to CComPtr for the duration of its lifefime, e.g.
public ref class MyClass
{
public:
MyClass()
{
CComPtr<ISomeType> pSomeType;
// init someType;
m_pSomeType = pSomeType;
}
private:
CComPtr<ISomeType> m_pSomeType;
void DoSomething()
{
m_pSomeType->DoSomething();
}
}
However this doesn't compile as mixed types are not supported - the solution is to use an AutoPtr. I still need reference counting and so this is what I came up with.
public ref class MyClass
{
public:
MyClass()
{
CComPtr<ISomeType> pSomeType;
// init someType;
m_pSomeType = new CComPtr<ISomeType>(pSomeType);
}
private:
AutoPtr<CComPtr<ISomeType>> m_pSomeType;
void DoSomething()
{
CComPtr<ISomeType> pSomeType = *m_pSomeType.GetPointer();
pSomeType->DoSomething();
}
}
This looks nasty to me, and I also suspect that its wrong in some way (I come from a C# background so I'm kind of learning a lot of this as I go).
How should I "store" a CComPtr as a member of a C++/CLR class?
I've made a smart pointer that might help you. It can't be used to manage COM pointers directly, but it can be used to hold a CComPtr inside a managed class.
scoped_ptr for C++/CLI (ensure managed object properly frees owned native object)
Please respect the license requirements if you choose to use it (commercial use is not prohibited, but you must give credit).
Also, you can write
(*m_pSomeType)->DoSomething();
instead of copying the CComPtr (which has to update the reference count).
Besides that, I would use the ctor-initializer-list instead of assignment in the constructor.
MyClass()
: m_pSomeType(new CComPtr<ISomeType>)
{
}
If you want to compile managed and unmanaged code seperately then here is a link.
Using this you can retain flavour of both.

Best way to start a thread as a member of a C++ class?

I'm wondering the best way to start a pthread that is a member of a C++ class? My own approach follows as an answer...
This can be simply done by using the boost library, like this:
#include <boost/thread.hpp>
// define class to model or control a particular kind of widget
class cWidget
{
public:
void Run();
}
// construct an instance of the widget modeller or controller
cWidget theWidget;
// start new thread by invoking method run on theWidget instance
boost::thread* pThread = new boost::thread(
&cWidget::Run, // pointer to member function to execute in thread
&theWidget); // pointer to instance of class
Notes:
This uses an ordinary class member function. There is no need to add extra, static members which confuse your class interface
Just include boost/thread.hpp in the source file where you start the thread. If you are just starting with boost, all the rest of that large and intimidating package can be ignored.
In C++11 you can do the same but without boost
// define class to model or control a particular kind of widget
class cWidget
{
public:
void Run();
}
// construct an instance of the widget modeller or controller
cWidget theWidget;
// start new thread by invoking method run on theWidget instance
std::thread * pThread = new std::thread(
&cWidget::Run, // pointer to member function to execute in thread
&theWidget); // pointer to instance of class
I usually use a static member function of the class, and use a pointer to the class as the void * parameter. That function can then either perform thread processing, or call another non-static member function with the class reference. That function can then reference all class members without awkward syntax.
You have to bootstrap it using the void* parameter:
class A
{
static void* StaticThreadProc(void *arg)
{
return reinterpret_cast<A*>(arg)->ThreadProc();
}
void* ThreadProc(void)
{
// do stuff
}
};
...
pthread_t theThread;
pthread_create(&theThread, NULL, &A::StaticThreadProc, this);
I have used three of the methods outlined above.
When I first used threading in c++ I used static member functions, then friend functions and finally the BOOST libraries. Currently I prefer BOOST. Over the past several years I've become quite the BOOST bigot.
BOOST is to C++ as CPAN is to Perl. :)
The boost library provides a copy mechanism, which helps to transfer object information
to the new thread. In the other boost example boost::bind will be copied with a pointer, which is also just copied. So you'll have to take care for the validity of your object to prevent a dangling pointer. If you implement the operator() and provide a copy constructor instead and pass the object directly, you don't have to care about it.
A much nicer solution, which prevents a lot of trouble:
#include <boost/thread.hpp>
class MyClass {
public:
MyClass(int i);
MyClass(const MyClass& myClass); // Copy-Constructor
void operator()() const; // entry point for the new thread
virtual void doSomething(); // Now you can use virtual functions
private:
int i; // and also fields very easily
};
MyClass clazz(1);
// Passing the object directly will create a copy internally
// Now you don't have to worry about the validity of the clazz object above
// after starting the other thread
// The operator() will be executed for the new thread.
boost::thread thread(clazz); // create the object on the stack
The other boost example creates the thread object on the heap, although there is no sense to do it.