C / C++ / MFC

16-Mar-08 16:16

Hello everyone,

I am looking for a good sample about how to implement C++ pre-condition and post condition check, but can not find a good sample code. Do you have any referred ones? Smile | :)

Since I can not find, I wrote the patterns in two ways, I am not sure which is correct and if both are not correct, how to implement this pattern?

Sample code 1,

#define MAX 1024

class Base
{
public:
	void foo(int i)
	{
	if (i > MAX)
	{
		// error handling
	}
	else
	{
		do_foo(i);
	}
}

private:
	virtual void do_foo(int i) = 0;
};

class Derived : public Base
{
private:
	virtual void do_foo(int i)
	{
	// i is never > MAX here
	}
};

int main()
{
	Derived d;
	d.foo (1000);

	return 0;
}

Sample 2,

#define MAX 1024

class Base
{
public:
	void foo(int i)
	{
	if (i > MAX)
	{
		// error handling
	}
	else
	{
		do_foo(i);
	}
}

private:
	virtual void do_foo(int i) = 0;
};

class Derived : public Base
{
public:
	virtual void do_foo(int i)
	{
		foo (i);
		// i is never > MAX here
	}
};

int main()
{
	Derived d;
	d.do_foo (1000);

	return 0;
}

thanks in advance,
George

Stephen Hewitt16-Mar-08 17:26

16-Mar-08 17:26

A word of advice not related to your problem: dump the macros! Use an enum instead, for example. One disadvantge of using macros is that they don't have scope and so can't be put into namespace and such.

enum { MAX = 1024 };

Steve

modified on Monday, March 17, 2008 1:49 AM

George_George16-Mar-08 19:48

16-Mar-08 19:48

Thanks Steve,

My bad, any comments or replies to my original question?

regards,
George

Stephen Hewitt16-Mar-08 19:57

16-Mar-08 19:57

I prefer the first; why make do_foo public? This seems to go against the pattern being used.

Steve

George_George16-Mar-08 20:01

16-Mar-08 20:01

Thanks Steve,

1.

Stephen Hewitt wrote:
against the pattern

You mean template method pattern? or something which hides implementation details?

2.

Do you have some more formal examples about pre and post condition implementation in C++? The samples I wrote are by myself. I want to learn some formal ones. Smile | :)

regards,
George

Stephen Hewitt16-Mar-08 20:05

16-Mar-08 20:05

See here[^].

Steve

George_George16-Mar-08 20:14

16-Mar-08 20:14

Thanks Steve,

I do not think it contains a sample of pre and post conditions and it only mentions the term pre and post condition check. I have read this article before.

If I am wrong, please feel free to correct me and point out the sample you refer in this article.

regards,
George

Stephen Hewitt16-Mar-08 20:23

16-Mar-08 20:23

The pre and post checks would obviously be specific to the function and the interface "contract".

Steve

George_George16-Mar-08 20:48

16-Mar-08 20:48

Thanks Steve,

My question is more about the pattern of pre- and post- condition check. Not about how to do a specific implementation. I have performend some search, but failed to find any definition.

Any recommended resources?

regards,
George

CPallini16-Mar-08 22:59

16-Mar-08 22:59

It is not public. Or did he just cheat changing it without notice?

If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.

This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong.
-- Iain Clarke

Stephen Hewitt17-Mar-08 13:05

17-Mar-08 13:05

It's public in class Derived in the second example.

Steve

CPallini17-Mar-08 22:07

17-Mar-08 22:07

You're right, I was blind Blush | :O

.
But we were both more than blind, because second example is completely wrong (recursive madness). Roll eyes | :rolleyes:

CPallini16-Mar-08 22:38

16-Mar-08 22:38

Well, if I state it is a precondition then I don't check it (i.e. I may check it in Debug build, not in Release one).

BTW approach (1) is possibly better because it frees Derived class methods from implementing checking logic (I mean, using approach (2) you must remember to call foo(i) and this is not good expecially if Derived class developer is not the same person who designed Base class).
Smile | :)

George_George16-Mar-08 22:48

16-Mar-08 22:48

Thanks CPallini,

Do you have some formal samples about pre- and post- condition pattern simple POC implementation? I have performed some search but can not find relevant stuff. Or you think my implementation is good enough.

Any ideas or recommendations?

regards,
George

CPallini16-Mar-08 22:50

16-Mar-08 22:50

No, I haven't.
I still stand in my original opinion: if I state a precondition then I haven't to check it.
Anyway, I like your first approach.
Smile | :)

George_George16-Mar-08 22:53

George, again in the recursive trap

16-Mar-08 22:53

Thanks CPallini,

With your agreement on my 1st approach of implementation, I think I am confident enough. Smile | :)

regards,
George

CPallini17-Mar-08 22:11

17-Mar-08 22:11

See here[^].
Poke tongue | ;-P

George_George17-Mar-08 22:23

17-Mar-08 22:23

So, you prefer implementation 1, I think, right, CPallini?

regards,
George

CPallini17-Mar-08 22:41

17-Mar-08 22:41

This is not matter of taste, George, example 2 is completely wrong.
if you fix it (exercise for you Poke tongue | ;-P

) then we can choose which of the two correnspond to a better design (...example 1 will win again Roll eyes | :rolleyes:

)

George_George17-Mar-08 23:29

17-Mar-08 23:29

Thanks CPallini,

Can you let me know what is the bug please? Smile | :)

regards,
George

CPallini18-Mar-08 0:39

18-Mar-08 0:39

OK, I do the exercise for you... Roll eyes | :rolleyes:

int main()
{
	Derived d;
	d.do_foo (1000);

	return 0;
}

in the above code you need to change d.do_foo(1000); to d.foo(1000); to avoid recursion. BTW if you followed Stephen Hewitt's suggestion, declaring private the Derived class do_foo member, then the compiler was able to stop you from committing recursion madness (it's called OOP nemesis).
Smile | :)

George_George18-Mar-08 1:02

18-Mar-08 1:02

Thanks CPallini,

You are correct. My bad. Actually, I have not tested it comprehensively. I just show my ideas about what the pattern looks like.

1.

So, if I fix the bug, you think both could be called pre- and post- condition pattern?

2.

If yes, what do you think the pros and cons of each implementation?

regards,
George

CPallini18-Mar-08 1:34