C++ Constructors vs. Init() Functions

3,927

February 26, 2014 04:12 PM

Well, constructor's don't have any kind of return value, so, many people use an Init() type function that returns an error code of some sort.

The standard way would be to throw an exception from the constructor. (I dont do that either.)

I would never do anything "complicated" inside the constructor, like loading images, setting up textures, etc. I would just initialise member variables to a known state and leave the other stuff for an init function/method or using getters/setters, depending on the situation. I wouldn't do anything that could throw exceptions inside a constructor. Throwing an exception inside a constructor sounds like a bad thing to me. Is this really the standard way?

Yes, throwing an exception is the standard way of handling a constructor that fails.

http://www.parashift.com/c++-faq-lite/ctors-can-throw.html

I feel offloading work into an init() method is somewhat "hackish" and goes against the RAII principle, not to mention everyone who uses your class has to remember to call init().

"I would try to find halo source code by bungie best fps engine ever created, u see why call of duty loses speed due to its detail." -- GettingNifty

frob

46,223

February 26, 2014 05:58 PM

I feel offloading work into an init() method is somewhat "hackish" and goes against the RAII principle, not to mention everyone who uses your class has to remember to call init().

If you come from a background of older languages, such as C, pre-90's BASIC, cobol, pascal, ada, or even machine code, RAII means something very different.

In the older languages, allocation means you get a blob of memory with completely unspecified data inside. It might be zeros. It might be white noise. It might be the contents from a previous (perfectly valid) object. The next step after allocating is nearly always to initialize the data to known values. It could be bzero() or memset() to initialize the values to zero. It could be to initialize them to otherwise known values. Or sometimes, people would follow the less-than-ideal practice of leaving the memory buffer around with garbage data until they get around to putting real data in.

Most modern languages (under two decades old) have strict enforcement of RAII. For example, in Java or C# if you write int i; the compiler automatically applies RAII and initializes it to a value of zero on your behalf. If you write bool b; the language performs RAII and gives you a value of false.

In contrast with older languages (from two decades to nearly seven decades old) in the past when you created an integer it contained whatever happened to be in a register or memory address at the time. It might be zero, 36, 72, or any other value. Unless you initialized it you had no guarantees about the contents until you performed a separate initialization step.

So really, the RAII cries you hear are mostly an old vs. young interpretation. Those who have only worked in young languages (Java is still a teenager) and not studied history often interpret RAII as requiring multiple resource allocations and performing costly work. Those who have worked in older languages or who have studied history interpret RAII as just doing something as simple as bzero() after a successful malloc().

DividedByZero

1,261

February 26, 2014 08:47 PM

Wouldn't throwning an exception inside a constructor cause a memory leak, as the destructor would never get called?

This seems hackish to me.

What about occasions where the error is 'not the end of the world' and the fallback option might not even have anything to do with that class (i.e. you don't need that class if the init() fails).

Seems funny to throw an exception from the constructor in that case. I'd still go for an init() function.

Brother Bob

10,350

February 26, 2014 09:04 PM

Wouldn't throwning an exception inside a constructor cause a memory leak, as the destructor would never get called?

This seems hackish to me.

What about occasions where the error is 'not the end of the world' and the fallback option might not even have anything to do with that class (i.e. you don't need that class if the init() fails).

Seems funny to throw an exception from the constructor in that case. I'd still go for an init() function.

The destructor of base classes and already constructed members will be called. If your class doesn't handle resources manually but via smart objects, as it should do, then resources will be correctly released if they were allocated before constructor failed.

cozzie

5,063

February 26, 2014 09:08 PM

- if you define your own constructor, the 'default' constructor, destructor and assignment operator are no longer applied from the compiler, meaning you need to write them yourself to (look up "the rule of three")

Unless somebody made really weird changes to the language, that's simply not true. Creating your own constructor will never prevent default implementations of anything but the default constructor.

The "Rule of Three " doesn't say "you must implement all of them, because the compiler doesn't create them", it says "if you need one, you probably need all three". In fact, the big problem IS that the default implementations don't do everything they should in that case.

Obvious example:

class Demo()

{

Demo() : ptr(new Thing) {}

~Demo() { delete ptr; }

Thing* ptr;

}

And you probably don't want to know how often I'm seeing this kind of thing, always hand-waved away with "nah, it's never going to be copied anyway". Unless it suddenly is and "someone" (typically me) gets to spend hours debugging and tracking down the double delete introduced months ago.

If the compiler actually would stop creating defaults for assignment and _copy_ constructor (which is the one relevant to the Rule of Three) the code would have the decency to stop compiling. You'd also have an army of C++ coders lynching the person that came up with it.

Another common rule is that the constructor should only do minimal work to get the object in a valid state. Any heavy lifting that isn't absolutely needed would then go into an init() function.

Thanks, I've misunderstood it and looked it up:

The rule of three (also known as the Law of The Big Three or The Big Three) is a rule of thumb in C++ that claims that if a class defines one of the following it should probably explicitly define all three:[1]

These three functions are special member functions. If one of these functions is used without first being declared by the programmer it will be implicitly implemented by the compiler with the default semantics of performing the said operation on all the members of the class. The default semantics are:

Destructor - Call the destructors of all the object's class-type members
Copy constructor - Construct all the object's members from the corresponding members of the copy constructor's argument, calling the copy constructors of the object's class-type members, and doing a plain assignment of all non-class type (e.g., int or pointer) data members
Copy assignment operator - Assign all the object's members from the corresponding members of the assignment operator's argument, calling the copy assignment operators of the object's class-type members, and doing a plain assignment of all non-class type (e.g., int or pointer) data members.

The Rule of Three claims that if one of these had to be defined by the programmer, it means that the compiler-generated version does not fit the needs of the class in one case and it will probably not fit in the other cases either.

Crealysm game & engine development: http://www.crealysm.com

Looking for a passionate, disciplined and structured producer? PM me

frob

46,223

February 26, 2014 11:24 PM

The rule of three is updated for C++11, to a rule of five.

It now also also includes a move constructor and a move assignment operator.

SunDog

232

February 26, 2014 11:40 PM

Trivial constructors (basically constructors that might initialize variables to a defined state, but don't do anything like dynamically allocate memory), along with trivial copy constructors and destructors make RAIII much easier, without having to resort to shared/smart/auto pointers. The problem with C++ is that constructors and destructors are often called implciitly without the programmer realizing it, in addition the compiler generates default constructors in any case. Using a complicated constructor leads one to use pointers to objects and call new(), when really standard simple scoped objects would be a much better solution.

Example:


class A {
private:
       int* intPtr;
       int intPtrSize;
public: 
       A()  {  intPtr = nullptr; inttPtrSize = 0; }
       A(const &A other) { intPtr = nullptr; }       // no copy constructor
       virtual ~A  { }                                          // empty destructor
       void init(int size)   { intPtr = new int[size]; inttPtrSize = size; }
       void dispose()  { delete[] intPtr; intPtr = nullptr;}  
};
 
class B {
private:
      A a;
      float *floatPtr;
      int floatPtrSize;
public:
      B() {  floatPtr = nullptr; floatPtrSize = 0; }   
      B(const &B other) { floatPtr = nullptr; }   
      virtual ~B() { } // empty destructor
      void init(int aSize, int bSize)  { a.init(aSize);  floatPtr = new float[bSize]; floatPtrSize = bSize; }
      void dispose()  { a.dispose(); delete[] floatPtr; floatPtr = nullptr; }  
}

In the above example, I am using RAII with respect to A's lifetime in B. Theres no complicated constructor chaining, and no possibility of an exception being thrown during A's constructor, which would leave B's floatPtr in an undefined state. On the destructor side, I don't do anything and have to call an explicit dispose() method. This makes the code much clearer as to whats actually going on, rather than implicit destrucotrs being called At least, if I try to do something with "unitilaized B", the program will crash trying to dereference a null pointer. This is a much easier bug to track down, than a memory leak.

Bottom line for me, is that the only two sane choices are to use something lie the above pattern, or wrap everything around smart pointers to avoid shooting yourself in the foot.

Hodgman

52,717

February 27, 2014 12:24 AM

Well, constructor's don't have any kind of return value, so, many people use an Init() type function that returns an error code of some sort.

The standard way would be to throw an exception from the constructor. (I dont do that either.)

I would never do anything "complicated" inside the constructor, like loading images, setting up textures, etc. I would just initialise member variables to a known state and leave the other stuff for an init function/method or using getters/setters, depending on the situation. I wouldn't do anything that could throw exceptions inside a constructor. Throwing an exception inside a constructor sounds like a bad thing to me. Is this really the standard way?
Yes, throwing an exception is the standard way of handling a constructor that fails.
http://www.parashift.com/c++-faq-lite/ctors-can-throw.html

Yes, but no, maybe.
In theory it's the standard mechanism (or using an out-parameter or a zombie object state)... however, I've never actually seen exceptions used in the professional games industry.

Over the past 10 years I've used about 7 different professional C++ game engines (on about a dozen different products), and they all avoided the use of C++ exceptions completely.
I don't want to turn this into a "Are exceptions good/bad" thread, as that's a different topic -- but C++'s exceptions should not be compared to C#/Java/Python/etc's exceptions -- they are a completely different beast. Also C++ is a very complex language, to the point where almost every project leader will define an acceptable sub-set of the language to be used.

For large/complex professional game engines, this commonly includes:

don't use the standard new/delete/malloc/free (use some engine-specific replacement / wrapper),
don't use std containers (as they will call new, and custom allocator mechanisms are broken),
don't use exceptions (as writing to satisfy the "strict exception safety" guarantee is hard in C++, there's a performance impact, and some gaming platforms may not even support them),
don't use RTTI or dynamic_cast,

and sometimes includes (these used to be common 10 years ago, but not so much today).

don't use anything in std,
don't use templates.

What about occasions where the error is 'not the end of the world' and the fallback option might not even have anything to do with that class (i.e. you don't need that class if the init() fails).

Can you give an example of where this would happen?

In OO, it should be extremely rare to find a valid case for a constructor to fail.

To go off on a rant for a moment --

OO here doesn't mean that you're using an OOP language and you're using keywords like class... It means that you're making use of the large body of software design and engineering knowledge that's been collected under that moniker.

To use a straw-man example of what's wrong here, let's say that we've got a Texture class, responsible for managing the lifetime of pixel-data inside the GPU, who's constructor loads an image file from disk. This is a problem because errors can occur during file loading, such as FILE_NOT_FOUND. If that occurs, you'd have to abort from inside the constructor!

Bzzt. You just broke the SRP (Texture is responsible for GPU-resource lifetime management AND disk IO logic), so you're actually using your own methodology here, you're not using OO!

While we're reading up on SRP, we also decide to read about DI and IoC.

Now, we end up with a TextureLoader class, who opens a file, handles FILE_NOT_FOUND errors, and then once it's actually able to load the pixel data from disk, only then is a Texture object constructed and passed that data. Wow, after actually using OO, this whole aborting-construction "problem" went away, look at that...

So, I would treat "I need to throw an exception from this constructor" as a code-smell, indicating that you probably need some DI and IoC up in yo code.

. 22 Racing Series .

frob

46,223

February 27, 2014 12:35 AM

So, I would treat "I need to throw an exception from this constructor" as a code-smell, indicating that you probably need some DI and IoC up in yo code.

Which brings us right back to items covered on page one of the discussion...

RAII means to initialize things to be ready to use. "Empty" and "Disconnected" are perfectly valid definitions of ready to use, and for non-trivial objects are usually the best default.

I would never do anything "complicated" inside the constructor, like loading images, setting up textures, etc. I would just initialise member variables to a known state and leave the other stuff for an init function/method or using getters/setters, depending on the situation. I wouldn't do anything that could throw exceptions inside a constructor.

The best constructors are the ones that instantly init to an empty or blank object, deferring the heavy processing work so the developer can schedule the work to a time and place that is appropriate.

If you must do work, and especially if that work can fail, it seems like you are doing more than constructing an object.

Paradigm Shifter

5,832

February 27, 2014 12:37 AM

Init is a terrible name though, since it doesn't really describe anything.

Bind, Open, Connect, etc. much better.

EDIT: Init is ok for "call this first, don't do anything else beforehand" but that's probably better in a constructor unless reasons already posted here (lots of work, fails a lot, etc.) apply.

"Most people think, great God will come from the sky, take away everything, and make everybody feel high" - Bob Marley

C++ Constructors vs. Init() Functions

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

C++ Constructors vs. Init() Functions

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Reticulating splines