I am trying to accomplish the following: I have 4 classes(lets call one primary, other 3 secondary ones) within one namespace, I want to store instances of secondary classes as private members of primary, to do this I need to call secondary constructors from primary's one and then store instances. But Unfortunately I do not completely understand how to do it (not really experienced with c++): here is what I have in header file:
class secondary_one
{
private:
int number1;
public:
secondary_one(int);
int get_number1() const;
};
class secondary_two
{
private:
int number2;
public:
secondary_two(int);
int get_number2() const;
};
class secondary_three
{
private:
int number3;
public:
secondary_three(int);
int get_number3() const;
};
And 'primary' class is:
class primary
{
private:
secondary_one one;
secondary_two two;
secondary_three three;
public:
primary(int,int,int);
};
Upon calling primary constructor I want first argument to be send to constructor of secondary_one, second argument to constructor of secondary_two and so on.
And then store instances as private members. Is it even possible or I am just wasting time? If it is, can you give a short example what should I have in header and source file?
Use the constructor initialization list:
class primary
{
public:
primary(int a, int b, int c) : one(a), two(b), three(c) {}
private:
secondary_one one;
secondary_two two;
secondary_three three;
};
Related
I am not sure is my question is right or not? But let me still try to ask once.
I have a Class with have few member variables defined. As per OO concepts, every member function can access , all member variables of its class.
But I want these member variable to be accessed via specific methods (Lets say Getters) , even within same class member functions.
It there any way to do it?
class A {
public:
void func1();
void func2();
B getB();
private:
B b;
}
void A::func1() {
b.functionFromB(); // function uses member variable b directly
}
void A::func2() {
B b1=getB(); // function ask for B from a function and then uses it. // I need something like this... And ensure each function uses same way otherwise there should be warning...
b1.functionFromB();
}
Thanks,
Kailas
No, there is not. You can do it via encapsulation and inheritance like:
class shape
{
private:
int angles;
protected:
shape(int angles_):angles(angles_){};
int getAngles() const;
}
class square : private shape
{
public:
square():shape(4){}
void doSth()
{
\\ you can access angles only via getAngles();
}
}
Any private members of the class can be accessed from within the class, but not by users of the class. So it looks like you need private members and public methods that allow access to them.
class A
{
private:
int a;
public:
int getA() {return a;}
};
int main()
{
A inst;
int t;
inst.a =5; // error member a is private
t = inst.getA(); //OK
}
The concept extends fine to nested class declarations in case you only want to allow instance of a class to be created from another class; details here
As others have said - you have to add an additional layer.
If you want to give access to specific methods then you can use the friend keyword. E.g.
// Public.h
#pragma once
class Public
{
public:
Public();
int GetI() const;
float GetF() const;
private:
std::unique_ptr<Private> p_;
};
//Public.cpp
#include "Public.h"
Public::Public()
: p_(new Private)
{
}
int Public::GetI() const
{
return p_->i_;
}
float Public::GetF() const
{
return p_->f_;
}
// Private.h
#pragma once
class Private
{
friend int Public::GetI() const;
friend float Public::GetF() const;
int i_;
float f_;
};
Keep in mind that every friend method can access ALL private members.
If you really really want to limit which methods can access which members then you can wrap each member in a separate class/struct and make only the getter/setter of that member a friend of that class/struct but I would not recommend this approach.
If I have these two sets of .cpp and .h files set up like this and I l want to put an add() function in a.cpp that adds ten integers and returns sum. Then I have class b that will return the average but I want to use the sum of the ten integers that I got from class a. How can I call function add() in function average() so it can return the sum? I don't know if this is getting my question across well, I just made this example up to try it illustrate it.
class a {
private:
int a[10];
public:
a( );
int add( int);
};
class b {
private:
int d;
public:
int average(int );
}
You can call member functions on an object of a class, not on the class itself. Creating an object is almost as creating a variable, for example a my_a;. Then you can use the add member function of the my_a object as my_a.add(42).
If you do not need objects (but do need classes for some reason), use static member functions and variables as follows.
class MyClass {
private:
static int variable;
public:
static int accessor() { return variable; }
};
In this case, you can call the static member function without creating an instance as MyClass::accessor().
class A {
public:
static int add(int);
};
class B {
public:
void average(int val){
A::add(val);
// Some logic
}
};
So in my current understanding of how to use encapsulation correctly it is a good rule of thumb to make variables private and access them through member functions of the class like this:
class Aclass{
private:
int avar;
public:
void ch_avar(int val){avar = val};
int get_avar() {return avar;}
};
My question is how would I access a private member of a class instance which is its self a private member of another class. Here is an example of the way I have been trying to do it (not retyping the example above for brevity)
class LargerClass{
private:
int other_var;
Aclass A; //has two instances of class "Aclass"
Aclass B;
public:
void ch_other_var(int val){other_var = val;}
int get_other_var() {return other_var;}
// this is the important line for the question
int get_avar(Aclass X){return X.get_avar();}
};
Now In my real program there are a few more levels of this and I keep getting the compilation error that "Aclass" is an unknown type. Even though I have included the header file for Aclass in the Larger class. Since I am stuck I thought It would be good to find out if this is even the correct (or an acceptable way) of doing what I want. I am new to OOP and this feels sloppy to me.
To access the private member of a class instance which is its self a private member of another class, can be done this way. You don't need to pass a Aclass X. It is unneccessary. You can call it by the instance name you have given..
class LargerClass{
private:
int other_var;
Aclass A; //has two instances of class "Aclass"
Aclass B;
public:
void ch_other_var(int val){other_var = val;}
int get_other_var() {return other_var;}
// this is the important line for the question
int get_avar_A(){return A.get_avar();}
};
If you have 20 instances of Aclass, rather you create a vector of Aclass instances.
class LargerClass{
private:
int other_var;
Aclass A[20];
public:
void ch_other_var(int val){other_var = val;}
int get_other_var() {return other_var;}
// this is the important line for the question
int[] get_avar_A()
{
int other_var[20];
for(int i= 0; i<20; i++)
{
other_var[i] = A[i].get_avar();
}
return other_var;
}
};
I have seen multiple public and private keywords in a class definition, like the example below:
class myClass
{
public:
void func1(){}
public:
void func2(){}
private:
int x;
int y;
private:
int z;
int baz;
};
What is the practical use of this (if any)? Is there any situation in which this could be useful?
Is there any situation in which this could be useful?
I can think of a situation where it would be very problematic otherwise:
class myClass
{
public:
void func1(){}
public:
void func2(){}
COORDINATES; // <-- Note the macro here!
private:
int z;
int baz;
};
which, after the expansion of the COORDINATES macro, becomes:
class myClass
{
public:
void func1(){}
public:
void func2(){}
private:
int x;
int y;
private:
int z;
int baz;
};
If multiple private / public keywords weren't allowed, it would be very painful to do it. Although using macros isn't good practice; neither is introducing access modifiers silently for all the members appearing after the macro. Nevertheless, it could be useful for RAD tools generating C++ source code.
I can only guess why you see that in human written classes. My guess is that it is poor coding; the writer probably wanted to express that a chunk of data belongs together and / or wanted to be able to move up and down those chunks within the class definition, together with the corresponding access modifier (private/protected/public).
I'll go one step farther from my comment for this answer, with a snippet of code.
class myClass {
// initializers etc
public:
myClass();
~myClass();
// signal processing
public:
void modifyClass();
private:
float signalValue;
// other class responsibilities
public:
void doWork();
private:
void workHelper();
};
and so on. I wouldn't say this is a solid DESIGN for the class, but it's a good way to show the different capabilities of a class.
Another use is when you want to have a specific destruction order. Lets consider two cases:
Bad case
// Auxiliary class
Class StorageCleaner {
public:
StorageCleaner(ExternalStorage* storage) : storage_(storage);
~StorageCleaner() { storage_->DeleteEverything(); }
private:
ExternalStorage* const storage_; // Not owned
}
// Class with bad declaration order
Class ExternalStorageWrapper {
public:
ExternalStorageWrapper(ExternalStorage* storage) : cleaner_(storage) {
pointer_to_storage_.reset(storage);
}
const StorageCleaner cleaner_;
private:
std::unique_ptr<ExternalStorage> pointer_to_storage_;
// Other data which should be private
int other_int;
string other_string;
....
};
What happens when existing ExternalStorageWrapper object goes out of scope?
The ExternalStorageWrapper destructor will first
destroy other data
Then it will destroy external storage pointed by pointer_to_external_storage_ (because fields are destroyed in reversed declaration order).
Then it will attempt to destroy cleaner_
But cleaner inside its own destructor attempts to manipulate external storage, which has already been destroyed!
Good case
// Class StorageCleaner same as before
...
// Class with better declaration order
Class ExternalStorageWrapper {
private:
std::unique_ptr<ExternalStorage> pointer_to_storage_;
public:
ExternalStorageWrapper(ExternalStorage* storage) : cleaner_(storage) {
pointer_to_storage_.reset(storage);
}
const StorageCleaner cleaner_;
private:
// Other data which should be private
int other_int;
string other_string;
....
};
What happens in this case when existing ExternalStorageWrapper object goes out of scope?
The ExternalStorageWrapper destructor will first destroy other data.
Then it will attempt to destroy cleaner_. Cleaner will DeleteEverything() from storage using storage_ pointer to still existing storage.
Finally, storage gets destroyed through pointer_to_storage_.
I actually had to debug such problem in a company, so although rare, this peculiar case is possible to occur.
One use case I encounter is for the clarity. As shown in the example below, class1 inherits base class which contains a pure virtual function that class1 needs to implement. To indicate that method1 in class1 is coming from some other class class1 inherits from (e.g. base) instead of class1's own function, we use another public to make this point clear:
class base {
public:
virtual void method1() = 0;
}
class class1: base {
public:
myOwnMethod1();
myOwnMethod2();
public: /* base interface */
method1();
Just finished reading those examples. Hopefully this one could contribute on how useful multiple keyword definition is.
We don't quite need to assume much since this is my current problem as of now but consume the following:
class IOHandler {
public:
enum COLOR_DEFINITIONS : unsigned char
{
BLACK,
DARK_ER_BLUE,
DARK_GREEN,
DARK_SKY_BLUE,
DARK_RED,
DARK_PINK,
DARK_YELLOW,
DARK_WHITE,
DARK_GREY,
DARK_BLUE,
BRIGHT_GREEN,
BRIGHT_SKY_BLUE,
BRIGHT_RED,
BRIGHT_PINK,
BRIGHT_YELLOW,
BRIGHT_WHITE
};
template <typename dtEX>
void showOutputEx(dtEX literalOutput, _COLOR_OUTPUT textColorSet = {COLOR_DEFINITIONS::BRIGHT_WHITE , COLOR_DEFINITIONS::BRIGHT_YELLOW}, bool appendOutputType = true, OUTPUT_TYPE levelOutput = OUTPUT_TYPE::OUTPUT_NORMAL, EMBRACE_TYPE embraceOutputType = EMBRACE_TYPE::EMBRACE_OUTPUT_LEVEL, ...);
// ! ^^^^^^^^^^^^^^ Doesn't detect the struct being referenced at which is at below.
private:
typedef struct colorProps // This is the struct that the public function where trying to get referenced at but failed to do so.
{
unsigned char C_BG = 0,
C_FG = 0;
} _COLOR_OUTPUT;
};
The error code.
identifier "_COLOR_OUTPUT" is undefined.
At this point, my intelliSense keep complaining that _COLOR_OUTPUT is undefined within public class scope.
The most probable solution is to put the struct inside public scope instead of private scope.
But I don't want to.
The reason this happens was due to the compiler reads the file from top to bottom. Anything that is declared
that requires reference should be on top and that should resolve the issue. Since I don't want to make things messed up by putting all private class functions and variables
on top. I should just declare another specifier on the top so that any public function requiring some reference might see it ahead.
So the solution is the following: (Is to move all referrable variables and struct on top of the class so that any private and public function argument reference is being recognized.)
class IOHandler {
// Private Class Scope for Variables and Structure
private:
typedef struct colorProps // This is the struct we move at the top for recognizing references.
{
unsigned char C_BG = 0,
C_FG = 0;
} _COLOR_OUTPUT;
public:
enum COLOR_DEFINITIONS : unsigned char
{
BLACK,
DARK_ER_BLUE,
DARK_GREEN,
DARK_SKY_BLUE,
DARK_RED,
DARK_PINK,
DARK_YELLOW,
DARK_WHITE,
DARK_GREY,
DARK_BLUE,
BRIGHT_GREEN,
BRIGHT_SKY_BLUE,
BRIGHT_RED,
BRIGHT_PINK,
BRIGHT_YELLOW,
BRIGHT_WHITE
};
template <typename dtEX>
void showOutputEx(dtEX literalOutput, _COLOR_OUTPUT textColorSet = {COLOR_DEFINITIONS::BRIGHT_WHITE , COLOR_DEFINITIONS::BRIGHT_YELLOW}, bool appendOutputType = true, OUTPUT_TYPE levelOutput = OUTPUT_TYPE::OUTPUT_NORMAL, EMBRACE_TYPE embraceOutputType = EMBRACE_TYPE::EMBRACE_OUTPUT_LEVEL, ...);
// ! ^^^^^^^^^^^^^^ Now recognizable reference.
private:
... // Any other functions, excerpt.
};
I think I have coded everything correctly in this program but still getting errors.
The object si it says it's not accessible.
#include<conio.h>
#include<iostream.h>
class s_interest
{
int p,t;
float r;
s_interest(int a, int b, float c)
{
p=a;
r=b;
t=c;
}
void method()
{
int result=(float)(p*t*r)/100;
cout<<"Simple interest:"<<result;
}
};
void main()
{
int pr,ti;
float ra;
cout<<"\n Enter the principle, rate of interest and time to calculate Simple Interest:";
cin>>pr>>ti>>ra;
s_interest si(pr,ti,ra);
si.method();
}
When the compiler tells you that something is not accessible, it's talking about the public: vs. protected: vs. private: access control. By default, all members of a class are private:, so you cannot access any of them from main(), including the constructor and the method.
To make the constructor public, add a public: section to your class, and put the constructor and the method there:
class s_interest
{
int p,t;
float r;
public: // <<== Add this
s_interest(int a, int b, float c)
{
...
}
void method()
{
...
}
};
Default member access for a class is private (whereas the default for struct is public). You need to make the constructor and method() public:
class s_interest
{
int p,t; // private
float r; // also private
public: // everything that follows has public access
s_interest(int a, int b, float c) { .... }
void method() { ....}
};
Note also that void main() is not standard C++. The return type needs to be int, so you need
int main()
{
...
}
And finally, iostream.h is not a standard C++ header. You need to include <iostream> if you are using a standards compliant C++ implementation.
Following High Integrity C++ Coding Standard guidelines, always declare first public, then protected and private members. See Rule 3.1.1 of hicpp-manual-version-3-3.pdf
All the variables & functions in your class are private. This is the default when access is not specified with the private: , protected: and public: specifiers. I suggest you have a good read of a tutorial - google C++ classes.
also it is int main() and never void main()
The problem is due to access specifiers. By default class methods and data members are private. Make your data members private and methods public. so you can set the private data members value using public methods.
class{
private:
int a;
int b;
int c;
public:
void method();
void print_sum();
};