Imagine you have a class with 100 or more member variables:
class DataContainer { int a1; int a2; ... double a100; };
and another class which accesses and analysis parts of the variables in the first class:
class Analysis {
protected:
DataContainer *myData;
void myVar1, myVar2;
public:
Analysis() {...}
void myChoice(void a, void b) { myVar1 = a; myVar2 = b }
void Analyse(DataContainer data) {
myData = data;
/* do something with myVar1 and myVar2 */
}
};
but before analysing you have to choose which parameters to use. I would like to know how to write a call for this myChoice-function which is not simply passing the content of a variable, but the choice of the membervariable itself.
In order to ask more generally: I think the position in the memory of a class-member relatively to the class itself is fixed (is it?) Independet if it is a new instance of that class or not. So how can I tell the combiler "Please use the membervariable a3 always". Instead of the content of a3.
This is exactly what pointers-to-data-members are for.
int DataContainer::* myVar1 = &DataContainer::a1;
double DataContainer::* myVar2 = &DataContainer::a100;
I think the position in the memory of a class-member relatively to the class itself is fixed (is it?)
It is indeed. The relative position does not change at run time. This is why pointers-to-data-members work. They can be implemented as a memory offset.
PS. Considering how massive DataContainer is, I would not recommend passing it by value if you can avoid it.
PPS. Consider grouping the member variables into sub-structures.
If the purpose is to have a way to select specific fields in known ways, based on the type of analysis you are doing, then make Analyze virtual pure, and have specific types of analysis pick their fields directly.
From the question, I get the idea that you want to perform the same underlying computation, but use different data points based on the myChoice.
So the whole issue is how do you create the selection metadata? If there are a few specific, well documented ways to do it, then providing each in a subclass will be a very clear, concise and maintainable approach.
class Analysis {
public:
virtual void Analyze(const DataContainer &data) = 0;
virtual ~Analysis();
protected:
void special_computation(int x, int y, double z);
};
class SpecificKindOfAnalysis : public Analysis {
public:
void Analyze(const DataContainer &data){
int x = data.a27;
int y = data.a12;
double z = data.a99;
special_computation(x,y,z);
}
};
This actually looks pretty horrible with variable names like x = data.a27, but I assume we're trying to be general here, and the actual program has meaningful names.
Related
I want to have a Collider interface class in which will have a overloaded -> operator to have access directy to the BoxCollider derived class. I want to have access to the members of box collider through the interface and chnage the type of collider at run-time.
So I thought of using templates:
template<typename T>
class ColliderV2 {
public:
virtual T* operator ->() = 0;
};
class BoxColliderV2 : public ColliderV2<BoxColliderV2> {
public:
float width;
float height;
BoxColliderV2* operator ->() {
return this;
}
};
int main()
{
ColliderV2<BoxColliderV2>* col = new BoxColliderV2;
(*col)->width = 1;
}
This works. But templates , as far as I know, will generate a brand new Collider class in compile-time filling T with Box Collider, correct? Thats why it worked. But later it prevents me from changing the collider type. I also thought of just making a virtual Collider class with Collider* operator->() ; overload in the derived class BoxCollider* operator->() ;
But if I tried :
Collider<BoxCollider>* col = new BoxCollider;
(*col)->width = 1; // won't work
doesn't work since Collider is not BoxCollider. And I don't want to dynamic_cast every possible collider type I could have. So, what can be done here?
As you've already found out, this doesn't work. Templates and runtime behavior are kind of contradicting mechanics. You can't create a common base class and let it act like a generic pointer to give you access to its derived types' members.
An interface specifies a contract against which you can code. You don't code against a specific implementation but the interface, so the interface has to provide all the members that you'd like to access. In your case this would result in width and height beeing part of ColliderV2 instead of BoxColliderV2. However this defeates the logic you are trying to mimic.
There are a few approaches that you can take:
Either make your collider type a variant, like
using ColliderType = std::variant<BoxColliderV2, MyOtherCollider, ...>;
and check for the actual type when you want to access the member
ColliderType myCollider = /* generate */;
if (auto boxCollider = std::get_if<BoxColliderV2>(&myCollider); boxCollider)
boxCollider->width = 0;
Or, keep the base class that you have, remove the operator-> and the template and do a dynamic cast on it:
ColliderV2* col = new BoxColliderV2;
if (auto boxCollider = dynamic_cast<BoxColliderV2*>(col); boxCollider)
boxCollider->width = 0;
You can also hide details like width or height behind more generic functions that are part of the interface. For example:
class ColliderV2 {
public:
virtual void setBounds(float width, float height) = 0;
};
class BoxColliderV2 : public ColliderV2 {
public:
void setBounds(float width, float height) override {
this->width = width;
this->height = height;
}
private:
float width;
float height;
};
int main()
{
ColliderV2* col = new BoxColliderV2;
col->setBounds(1, 1);
}
What you are trying to do is discouraged by C++. What you are trying to do is to change the type of something based on the return value of a function. The type system is designed to stop you from writing code like this.
One important restriction of a function is that can only return one type-of-thing. You can return one of a list of things if you wrap those possibilities in a class, and return that. In C++17, a ready-made class for this is std::variant. The restriction on this is that the list of things must be fixed (or a closed-set). If you want an arbitrary set of return values (open-set), you must use a different approach. You must restate your problem in terms a function that is done on the return value.
class BoxColliderV2 : public MyBaseCollider {
public:
void SetWidth(float new_width) override;
};
You may find this video useful. The bit of interest starts at around 40 minutes (but watch the whole video if you can). If you are interested in advice, I would suggest starting with std::variant, and if it works, move to virtual functions. Problems like collision detection get really complicated really quickly, and you will almost certainly require double dispatch at some stage. Start simple, because it's only going to get more complicated.
These excerpts from the ISO-Guidelines may help
1. When you change the semantic meaning of an operator, you make it
harder for other programmers to understand you code. guideline.
2. Dynamic casting is verbose and ugly, but deliberately so, because dynamic casting is dangerous, and should stand out. guideline
I think you are approaching the problem from the wrong direction. The purpose of an interface is that you don't have to know about the exact type or the implementation.
For example: You are using Axis-Aligned Bounding Boxes for collision detection. So, even if your CircleCollider uses a radius, you are still able to calculate its width and height from it. Now, you don't have to worry about if you are dealing with a BoxCollider or a CircleCollider, you have everything to make a Bounding Box.
class Collider
{
public:
virtual float x() const = 0;
virtual float y() const = 0;
virtual float width() const = 0;
virtual float height() const = 0;
};
class BoxCollider : public Collider
{
// Implementation...
};
class CircleCollider : public Collider
{
// Implementation...
};
Of course, you are maybe using something else, and not AABBs. I just wanted to demonstrate how you can use interfaces effectively.
After adopting Prototype Pattern into a game, it pleasantly improved maintainability of my code.
However, I have started to fear that when the actual object become more detail-customized,
I tend to code the corresponding prototype to become more like a copy of the actual object.
(notice #v.2 in code)
class Prototype{ //aka. "fake object"
public: Vec3 position;
//bool isShootable; //(#v.2) //#X
//float delayBetweenShoot; //(#v.2)
//float hp; //(#v.2)
};
class Actual{ //aka. "actual object"
public: PhysicObject* rigidBody=nullptr;
//bool isShootable; //(#v.2) //#X
//float delayBetweenShoot; //(#v.2)
//float hp; //(#v.2)
};
int main(){ // simplify
Prototype prototype;
prototype.position = Vec3(1,2,3);
//^ end prototype creation
//v these lines are inside a function, but it simply works like this
PhysicObject* phy=new PhysicObject(prototype.position);
Actual actual;
actual.rigidBody=phy;
//actual.isShootable =prototype.isShootable; //(#v.2) #Y
//actual.delayBetweenShoot=prototype.delayBetweenShoot; //(#v.2)
//actual.hp =prototype.hp; //(#v.2)
gameLogic.add(actual); //roughly speaking
}
There are two bad signals (#v.2):-
1. repetitive code in Prototype vs Actual (#X)
2. tedious copying of fields. (#Y)
Thus, I think something start to go wrong.
This pattern may naturally cause new maintainability issues.
In the real situation, the actual2 contains another actual1.
To adopt Prototype Pattern, I use a corresponding prototype2 inside another corresponding prototype1 :-
class Actual1{
//... some fields e.g. A1 a1; A2 a2; A3 a3;
};
class Actual2{
//... some other field e.g. B1 B2 B3
Actual1 actual1;
};
class Prototype1{
//... some fields e.g. very similar to A1 A2 A3
};
class Prototype2{
//... some other field e.g. very similar to B1 B2 B3
Prototype1 prototype1;
};
Question
(1) Is it common that Prototype Pattern create new maintainability issues?
(2) If yes, how to avoid it?
(3) If no, where am I wrong (especially the style of coding)? ... or this trend (of repetitive code) is not really an issue at all (i.e. I just panic.)?
My poor solution
I think it might be a good idea to encapsulate the repetitive part into a single structure named Settings.
class Settings{
bool isShootable;
float delayBetweenShoot;
float hp;
};
class Prototype{ //aka. "fake object"
public: Vec3 position;
Settings settings; //(#v.2)
};
class Actual{ //aka. "real object"
public: PhysicObject* rigidBody=nullptr;
Settings settings; //(#v.2)
};
However, it might increase unfavorable cohesion (i.e. glue or strong relation) between Prototype and Actual. Thus, it may leader to another new maintainability problem again.
You can avoid the unnecessary duplication by subclassing your 'actuals' from the prototype. For example:
struct Prototype {
bool isShootable = false;
float delayBetweenShoot = DEFAULT_DELAY;
float hp = DEFAULT_HP;
Vec3 position = STARTING_POSITION;
...
};
struct Actual : Prototype {
PhysicObject* rigidBody;
Actual() : Prototype(), rigidBody(new PhysicObject(position)) {}
}
int main() {
Actual actual;
// now you can access these methods
if (actual.isShootable) {
...
}
...
}
Your intuition is correct in that by grouping 'common' fields together, you increase the coupling between those fields. In some sense there is a tradeoff between coupling and code duplication. It is up to you to determine what is an acceptable compromise that best suits your application.
Probably using different types (with the ensuing double book keeping) is not the best interpretation of that design pattern.
See below for an approach which avoids the double book keeping and still holds the benefits of basic idea - to preconfigure a sample or template object once, then use that instance to initialize many other instances.
class A {
int32_t id;
bool shootable;
bool moveable;
bool destructable;
public:
// A template instance specific constructor.
A(bool shoot, bool move, bool destruct)
: id(-1)
, shootable(shoot)
, moveable(move)
, destructable(destruct)
{
}
// One or more "real" instance constructors.
A(int32_t idval, const A& source)
: id(idval)
, shootable(source.shootable)
, moveable(source.moveable)
, destructable(source.destructable)
{
}
// ...
};
int main(int argc, const char *argv[])
{
A kind(true,false,true);
A instance0(1,kind);
A instance1(2,kind);
return 0;
}
As a variation of the above idea, you could as well store a reference to the template instance and indeed use 2 types.
class UnitType
{
int32_t hp;
bool ranged;
//...
};
class Unit
{
int32_t id;
const UnitType *type;
// more data
public:
Unit(int32_t idval, const UnitType* unitType)
: id(idval)
, type(unitType)
{
}
//...
};
Of course, then the UnitType instance should not be writable by instances. And once, there is also e.g. currentHp, you have another form of duplication to handle. Plus, you need to ensure the life time for the UnitType template instance exceeds that of each Unit instance using it.
I have this class:
class Phone {
private:
string producer, color;
int weight, dimension;
public:
Phone(string &producer, string &color, int &weight, int &dimension):
producer(producer), color(color), weight(weight), dimension(dimension) {};
Phone():
producer(""), color(""), weight(0), dimension(0) {};
virtual ~Phone() {};
string getProducer(void) const;
string getColor(void) const;
int getWeight(void) const;
int getDimension(void) const;
virtual void displayInfo(void) const;
};
The problem is here caused by the fact that I expose the internal implementation of the object via getters.
But how can I prevent this?
Because usually in my code, I need to know some private data from my object (for comparision is one example), and that's why I use getters.
So then I rewrite the class to something like this:
class Phone {
private:
string producer, color;
int weight, dimension;
public:
Phone(string &producer, string &color, int &weight, int &dimension):
producer(producer), color(color), weight(weight), dimension(dimension) {};
Phone():
producer(""), color(""), weight(0), dimension(0) {};
virtual ~Phone() {};
bool isTheProducer(string& producer) const { return this->producer == producer };
bool hasWeight(int& weight) const { return this->weight == weight };
bool hasDimension(int& dimension) const { return this->dimension == dimension };
virtual void displayInfo(void) const;
};
Is this a better design (by the fact that I don't get the actual private value)?
As you might have seen from the other answers and comments, the answer is: It depends.
In fact, it depends mainly on the usecases where your class is used. Let's stick first to the example given in the question, the comparison of objects. Since it is not clearly visible from the question if we want to compare two phone objects or just a specific data member I will discuss both situations here.
Comparing a data member to out-of-class data
Let's take this usecase where we search for all phones with a weight bigger than x(just pseudocode):
for (Phone& p in phoneList) {
if (p.getWeight() > x) {
cout << "Found";
}
}
Then the first class example is perfectly fine, since this is not an intrinsic feature of the phone, and thus the phone class is not responsible for handling it. In addition, the result does not expose more than absolutely required for the task.
Comparing two phone objects
In this case both code examples are equally good (or in this case equally bad). In both cases the user has to know a lot of details about how phones are represented to compare all necessary members. If in a later revision a new member is added to the class, every code segment that compares two phones has to be adapted. To overcome this, one can add a function to the class that does exactly the comparison.
class Phone {
private:
string producer, color;
int weight, dimension;
public:
bool IsEqualTo(const Phone& other)
{
return (producer == other.producer && color == other.color &&....);
}
Non comparitive usecase
But let's go to a more advanced example. Let's assume the following task: A user enters the pin to a phone and if it is the correct one, the phone should unlock. Let's assume a very naive approach:
class Phone
{
private:
int pin;
bool unlocked;
public:
int getPin() { return pin; }
void unlock() { unlocked = true; }
};
and the corresponding call
if (phone.getPin() == enteredPin)
phone.unlock();
In this case we have a totally different situation. Here we need to consider the "tell, don't ask" rule, which basically says that it is a bad design to query the state of an object first, make a decision and then tell the object what to do. Instead we should only tell the object what we want, and let it do the work for us. In this usecase this is obvious, since unlocking the phone only when the pin is correct is a responsibility of the phone, not of the user that uses the phone class. But in more complex scenarious many programmers will do exactly what I described here.
Back to the problem: A good solution here would be for example
class Phone
{
private:
int pin;
bool unlocked;
public:
void CheckPin(int enteredPin) {
if (pin == enteredPin)
unlocked = true;
}
};
with the code
phone.CheckPin(enteredPin);
Hope this helps, and thanks to #KonradRudolph for pointing to the "tell, don't ask rule". Feel free to help me to improve the answer per commenting on it :)
The first one, even with getter, is encapsulated. Consider the color() method, which returns a string. Even if you change the implementation of Phone such that you store the color as an enum rather than a string, your method can still return a string if you do some sort of conversion first. The important part is that you can change the implementation of color() and the underlying storage without users of the class needing to change.
Compare to a class that stores color as a publicly accessible string. If you later change the data member to an enum, you need to modify every location that uses the color. This is less of a property of encapsulation and more a property of separating interface from implementation.
Encapsulation allows controlling of attributes exclusively via methods within the class. Both examples are encapsulated.
I have a very complicated code structure, but the important bits are:
typical setup: I have a base class and two classes that derive from this base class and each has own members, and which don't have a standard constructor
class BaseSolver{
...
};
class SolverA : BaseSolver{
public:
std::string a;
SolverA(TypeA objectA);
};
class SolverB : BaseSolver{
public:
int b;
SolverB(TypeB objectB);
};
Now I have a config xml file from which I read whether I have to use SolverA or SolverB. Therefore I have an IOService:
template<class T>
class IOService
{
BaseSolver* getSolver()
{
std::string variableThatIReadFromXML;
/* here I have to perform many actions before I can create a solver object
* to retrieve the data needed for the constructors */
TypeA variableIConstrucedWithDataFromXML;
TypeB anotherVariableIConstrucedWithDataFromXML;
if (variableThatIReadFromXML == "a")
return new SolverA(variableIConstrucedWithDataFromXML); // I know that this can leak memory
else if (variableThatIReadFromXML == "b")
return new SolverB(anotherVariableIConstrucedWithDataFromXML);
}
};
And somewhere in my application (for simplicity let's say it's the main.cpp):
int main(){
IOService ioService;
BaseSolver* mySolver = ioService.getSolver();
}
That is absolutely fine.
But now, in the main I have to access the members of the derived classes a and b respectively.
How can I do this?
I thought of retreving only the type of the Solver from the IOService:
class IOService
{
decltype getSolverType()
{
std::string variableThatIReadFromXML;
/* here I have to perform many actions before I can create a solver object
* to retrieve the data needed for the constructors */
TypeA variableIConstrucedWithDataFromXML;
TypeB anotherVariableIConstrucedWithDataFromXML;
if (variableThatIReadFromXML == "a")
return new SolverA(variableIConstrucedWithDataFromXML); // I know that this can leak memory
else if (variableThatIReadFromXML == "b")
return new SolverB(anotherVariableIConstrucedWithDataFromXML);
}
TypeA getConstructorDataForSolverA()
{
/* here I have to perform many actions before I can create a solver object
* to retrieve the data needed for the constructors */
return variableIConstrucedWithDataFromXML;
}
TypeB getConstructorDataForSolverB()
{
/* here I have to perform many actions before I can create a solver object
* to retrieve the data needed for the constructors */
return anotherVariableIConstrucedWithDataFromXML;
}
};
But of course I can't specify decltype as return value.
I'm really helpless. I would appreciate any hint into the right direction, or even a solution for this problem.
[Edit]: The derived solver classes need more than only the information from the xml file to work properly. That means, that I have to set some more properties which come from a mesh file. So I could give the meshfile to the IOService, so that the IOService could set the appropriate members this way:
class IOService
{
BaseSolver* getSolver(MeshType myMesh)
{
std::string variableThatIReadFromXML;
/* here I have to perform many actions before I can create a solver object
* to retrieve the data needed for the constructors */
TypeA variableIConstrucedWithDataFromXML;
TypeB anotherVariableIConstrucedWithDataFromXML;
if (variableThatIReadFromXML == "a")
{
auto solverA = new SolverA(variableIConstrucedWithDataFromXML); // I know that this can leak memory
solverA.a = mesh.a;
}
else if (variableThatIReadFromXML == "b")
{
auto solverB = new SolverB(anotherVariableIConstrucedWithDataFromXML);
solverB.b = mesh.b;
}
}
};
But then the IOService needs to know the class MeshType, what I want to avoid, because I think that it breaks encapsulation.
So I wanted to set the member a and b, respectively, in another part of my program (here for simplicity in the main).
Taking this into account, only the answer from Daniel Daranas seems like a solution for me. But I wanted to avoid dynamic casts.
So a reformulated question could be: How should I change my design to ensure encapsulation and avoid dynamic casts? [/Edit]
I am using clang 3.4 ob ubuntu 12.04 lts.
Use dynamic_cast to try to cast a pointer-to-base-class to pointer-to-derived-class. It will return NULL if the pointed-to object of the base class does not exist (NULL value of the base pointer), or is not actually a derived class object. If the result, instead, is not NULL, you have a valid pointer-to-derived-class.
int main(){
IOService ioService;
BaseSolver* mySolver = ioService.getSolver();
SolverB* bSolver = dynamic_cast<SolverB*>(mySolver);
if (bSolver != NULL)
{
int finallyIGotB = bSolver->b;
cout << finallyIGotB;
}
}
Note that there may be some better design solutions than using dynamic_cast. But at least this is one possibility.
The funny thing about polymorphism is that it points out to you when you are not using it.
Inheriting a base class in the way you are serves 1 purpose: to expose a uniform interface for objects with different behaviors. Basically, you want the child classes to look the same. If I have classes B and C that inherit from A, I want to say "do foo" to the class, and it'll do foob or fooc.
Essentially, you're flipping it around: I have a B and C of type A, and if it is B i want to do foob and if it is C I want to do fooc. While this may seem scary, usually the best way to solve the problem is to rephrase the question.
So to your example, you are currently saying "OK, so I have an XML file, and I will read data from it one way if I'm making an A, or another way if I'm making a B." But the polymorphic way would be "I have an XML file. It tells me to make an A or a B, and then I tell the instance to parse the XML file".
So one of the ways to solve this to change your solver interface:
class BaseSolver
{
public:
virtual void ReadXMLFile(string xml) = 0;
...
};
While this does rephrase the problem in a way that uses polymorphism, and removes the need for you to see what you've created, you probably don't like that for the same reason I don't: you'd have to supply a default constructor, which leaves the class in an unknown state.
So rather than enforce it at the interface level, you could enforce it at the constructor level, and make both SolverA and SolverB have to take in the XML string as part of the constructor.
But what if the XML string is bad? Then you'd get an error state in the constructor, which is also a no-no. So I'd deal with this using the factory pattern:
class SolverFactory;
class BaseSolver
{
public:
virtual void solve() = 0;
protected:
virtual int ReadXML(std::string xml) = 0;
friend class SolverFactory;
};
class A : public BaseSolver
{
public:
virtual void solve() {std::cout << "A" << std::endl;}
protected:
A(){}
virtual int ReadXML(std::string xml) {return 0;}
friend class SolverFactory;
};
class B : public BaseSolver
{
public:
virtual void solve() {std::cout << "B" << std::endl;}
protected:
B(){}
virtual int ReadXML(std::string xml) {return 0;}
friend class SolverFactory;
};
class SolverFactory
{
public:
static BaseSolver* MakeSolver(std::string xml)
{
BaseSolver* ret = NULL;
if (xml=="A")
{
ret = new A();
}
else if (xml=="B")
{
ret = new B();
}
else
{
return ret;
}
int err = ret->ReadXML(xml);
if (err)
{
delete ret;
ret = NULL;
}
return ret;
}
};
I didn't put any actual XML processing in here because I am lazy, but you could have the factory get the type from the main tag and then pass the rest of the node in. This method ensures great encapsulation, can catch errors in the xml file, and safely separates the behaviors you are trying to get. It also only exposes the dangerous functions (the default constructor and ReadXMLFile) to the SolverFactory, where you (supposedly) know what you are doing.
Edit: in response to the question
The problem you've stated is "I have a B and C of type A, and if is B i want to set "b" settings and if it is C i want to set "c" settings".
Taking advantage of polymorphism, you say "I have a B and C of type A. I tell them to get their settings."
There a couple of ways to do this. If you don't mind mangling your IO with the class, you can simply expose the method:
class BaseSolver
{
public:
virtual void GetSettingsFromCommandLine() = 0;
};
And then create the individual methods for each class.
If you do want to create them separate, then what you want is polymorphism in the io. So expose it that way:
class PolymorphicIO
{
public:
virtual const BaseSolver& get_base_solver() const = 0;
virtual void DoSettingIO() = 0;
};
an example implmentation
class BaseSolverBIO : PolymorphicIO
{
public:
virtual const BaseSolver& get_base_solver() const {return b;}
virtual void DoSettingIO() { char setting = get_char(); b.set_b(setting);}
private:
BaseSolverB b;
};
At first glance this seems like a lot of code (we've doubled the number of classes, and probably need to supply a factory class for both BaseSolver and the IO interface). Why do it?
It is the issue of scaleability/maintainability. Lets say you have figured out a new solver you want to add (D). If you are using dynamic cast, you have to find all the places in your top level and add a new case statement. If there is only 1 place, then this is pretty easy, but if it is 10 places, you could easily forget one and it would be hard to track down. Instead, with this method you have a separate class that has all the specific IO functionality for the solver.
Lets also think of what happens to those dynamic_cast checks as the number of solvers grows. You've been maintaining this software for years now with a large team, and lets say you've come up with solvers up to the letter Z. Each of those if-else statements are hundreds-a tousand of lines long now: if you have an error in O you have to scroll through A-M just to find the bug. Also, the overhead for using the polymorphism is constant, while reflection just grows and grows and grows.
The final benefit for doing it this way is if you have a class BB : public B. You probably have all the old settings from B, and want to keep them, just make it a little bigger. Using this model, you can extend the IO class as well for the io for BB and reuse that code.
One way to achieve this is to add an interface method into the base class:
class BaseSolver{
virtual void SolverMethodToCallFromMain() = 0;
...
};
class SolverA : BaseSolver{
public:
std::string a;
SolverA(TypeA objectA);
virtual void SolverMethodToCallFromMain() {/*SolverA stuff here*/};
};
class SolverB : BaseSolver{
public:
int b;
SolverB(TypeB objectB);
virtual void SolverMethodToCallFromMain() {/*SolverB stuff here*/};
};
And in main:
int main(){
IOService ioService;
BaseSolver* mySolver = ioService.getSolver();
mySolver->SolverMethodToCallFromMain();
}
Consider the following code:
class Rectangle
{
public:
// Constructors
Rectangle(){ init(0,0); }
Rectangle(int h, int w){ init(h,w); }
// Methods
void init(int h, int w)
{
_h = h;
_w = w;
}
// Getters / Setters
double get_h(void){ return _h; }
double get_w(void){ return _w; }
void set_h(double h){ _h = h; }
void set_w(double w){ _w = w; }
std::string get_name(void){ return _name; }
void set_name(std::string name){ _name = name; }
private:
// Private Members
int _h, _w;
std::string _name;
};
class House
{
public:
// <BEGIN PASSTHROUGHS>
std::string get_b_name(void){ return _base.get_name() };
std::string get_r_name(void){ return _roof.get_name() };
void set_b_name(std::string name){ _base.set_name(name); }
void set_r_name(std::string name){ _roof.set_name(name); }
// </END PASSTHROUGHS>
private:
// Private Members
Rectangle _base;
Triangle _roof;
};
This code works fine.
My question deals with the "passthrough" functions in the House class, enclosed by the PASSTHROUGHS tags. Is this the best way to do this? The arguments and return types will always match and there is no "intelligence" in these passthrough functions other than to make things cleaner and more straightforward.
My instinct would be something like one of the following:
get_b_name = _base.get_name;
// OR
std::string get_b_name(void) = _base.get_name;
... but neither seem to work unfortunately and it was only wishful thinking in the first place. If there are no easier options, telling me that is fine too. Thanks!
The problem, I think, is conceptual. Your design is quite un-object oriented in that the house does not represent an entity, but rather provides a bit of glue around the components. From that standpoint, it would make more sense to provide accessors to the elements, rather than pass-through functions:
class House {
Rectangle _base;
Triangle _roof;
public:
const Rectangle& base() const {
return _base;
}
const Triangle& roof() const {
return _roof;
}
};
I imagine that this is just a toy example, but the same reasoning applies: a class should represent an entity on which a set of operations are preformed, in some cases those operations might be implemented in terms of internal subobjects, but they are still operations on the type, and how they are gathered is an implementation detail.
Consider:
class House {
Thermostat t;
public:
int temperature() const {
return t.temperature();
}
};
From the user point of view the house has a temperature that can be read, and in this particular implementation, it is read from a thermostat that is a member. But that is an implementation detail. You might want to later install more thermostats in the house and substitute the single reading by an average of the readings, but that will not change the fact that the entity House (in this model) has a temperature.
That is, you should not be thinking in implementing pass-through functions, but rather on implementing features of the type. If the implementation happens to be a single forwarding to an internal method, that is fine.
But if the type contains internal members and it makes sense to access properties of the members, consider that it might be that you actual type should just provide access to its internal members. Consider that you want to move a piano inside the house, then you might just provide access to the door member and let the user check:
class House {
Door d;
public:
Door const & door() const {
return d;
}
};
bool can_enter_piano( House const & h, Piano const & p ) {
return h.door().width() > p.size();
}
There is no need to provide House::get_door_width(), and House::get_door_color() so that you can describe the entrance to a friend, and House::get_door_handle() so that they can know when they arrive...
That's possibly because your design is contradictory. Why on earth would you make a public member variable, then write a function that just forwards to one of that variable's functions? As a user of your class, I'd just call the function on the public variable myself. You're just confusing me by providing two ways to do the same thing. Or write getters and setters for a Rectangle class? That thing is just a bunch of variables, and doesn't need any getters and setters. You're not exactly going to inherit from it, and you can't really change the internal logic and maintain the same semantics, so it's very meaningless to not just make the variables public.
The Rectangle class needs a very healthy dose of YAGNI, and the House class just needs to look at itself again. The fact that there's no intelligence in the "passthrough" methods should be a huge alarm bell telling you that they are quite probably redundant and not helpful- especially since you can't change the public variables without breaking your interface anyway, it's not like the getters and setters are decreasing coupling or anything like that.
Methods should perform logic, or in the very least case, exist where logic might have to be done.