Need help understanding Templates

I went through my book's chapter about function templates and I'm trying to make sure I understand this correctly.
There’s an example in the book so I tried to do a program using templates to create a program on absolute values to see If I get it, but I’m getting a bunch of errors so obviously I didn’t.

Here’s what I did:

Code:

#include<iostream>

using namespace std;



template <class Value>
void printAbsValue()
{
	if (Value<0)
	{
		cout<<"The absolute value of "<<Value<<" is "<<(-Value)<<endl;
	}
	else
	{
		cout<<"The absolute value of "<<Value<<" is "<<Value<<endl;
	}
}


int main()
{
	const int aValue=10;
	const float bValue=10.4;
	const double cValue=1.5;

	int a[aValue]=-23;
    float b[bValue]=-23.55;
	double c[cValue]=-2.4;

	printAbsValue(a, aValue);
	printAbsValue(b, bValue);
	printAbsValue(c, cValue);
	
	return 0;
}

Here are the errors I’m getting:

Code:

.cpp(28) : warning C4305: 'initializing' : truncation from 'double' to 'const float'
.cpp(31) : error C2440: 'initializing' : cannot convert from 'int' to 'int [10]'
        There are no conversions to array types, although there are conversions to references or pointers to arrays
.cpp(32) : error C2057: expected constant expression
.cpp(32) : error C2466: cannot allocate an array of constant size 0
.cpp(32) : error C2440: 'initializing' : cannot convert from 'double' to 'float []'
        There are no conversions to array types, although there are conversions to references or pointers to arrays
.cpp(33) : error C2057: expected constant expression
.cpp(33) : error C2466: cannot allocate an array of constant size 0
.cpp(33) : error C2440: 'initializing' : cannot convert from 'double' to 'double []'
        There are no conversions to array types, although there are conversions to references or pointers to arrays
.cpp(35) : error C2780: 'void printAbsValue(void)' : expects 0 arguments - 2 provided
.cpp(12) : see declaration of 'printAbsValue'
.cpp(36) : error C2780: 'void printAbsValue(void)' : expects 0 arguments - 2 provided
.cpp(12) : see declaration of 'printAbsValue'
.cpp(37) : error C2780: 'void printAbsValue(void)' : expects 0 arguments - 2 provided
.cpp(12) : see declaration of 'printAbsValue'

Can anyone tell me if I understand this concept correctly and point me in the right direction to understand this.
Thanks.

you're close enough. I think you have the general idea right, but just not the syntax (which can be pretty scary at first!!)

ok in your function you have

Code:

template <class Value>
void printAbsValue()
{
//snip
}

which should be

Code:

template <typename Value> // 1
void printAbsValue(Value v) // 2
{
// 3
	if (v<0)
	{
		cout<<"The absolute value of "<<v<<" is "<<(-v)<<endl;
	}
	else
	{
		cout<<"The absolute value of "<<v<<" is "<<v<<endl;
	}
}

1. this is a nit, but you should really use typename instead of class.
2. ok you've declared the template parameter, but you still need to pass an instance of the templated type to the function
3. use the instance of the templated type in the body of the function

ok now for you client code

Code:

int a[aValue]=-23;
float b[bValue]=-23.55;
double c[cValue]=-2.4;

no idea what you thought you were doing here but it's not what you want. (you're actually declaring an array here)

Code:

printAbsValue(a, aValue);
printAbsValue(b, bValue);
printAbsValue(c, cValue);

ok again you're close here, but not quite
syntax is actually

Code:

printAbsValue<int>(a);
printAbsValue<float>(b);
printAbsValue<double>(c);

// but a good compiler will deduce the arguements so it's just
printAbsValue(a);
printAbsValue(b);
printAbsValue(c);

hope that helps

ChaosEngine
Thanks It should help big time.
I'll work on this and let you know how it goes.

Thanks again.

Hi,

You are treating the template parameter incorrectly: you are using it as if it were a quantity. For instance, you are trying to cout Value as if it were a number. That is somewhat understandable since you chose the misleading name Value. However, the template parameter is a type, e.g. int, double, float--not a value; and you can't display a type. You wouldn't do this:

cout<<int<<endl; //error

If you want an easy way to write a template function, try this: write the function for an int type, e.g.

Code:

void printAbsValue(int theNumber)
{
	if (theNumber <0)
	{
		cout<<"The absolute value of "<<theNumber<<" is "<<-theNumber<<endl;
	}
	else
	{
		cout<<"The absolute value of "<<theNumber<<" is "<<theNumber<<endl;
	}
}

Next, the whole reason you would want a template function is so that it will work for different types--not just ints. So, make the type a variable by templating the function. To do that, replace int with T (which stands for Type) everywhere you want the int type to be a variable:

Code:

template<typename T>
void printAbsValue(T theNumber)
{
	if (theNumber <0)
	{
		cout<<"The absolute value of "<<theNumber<<" is "<<-theNumber<<endl;
	}
	else
	{
		cout<<"The absolute value of "<<theNumber<<" is "<<theNumber<<endl;
	}
}

When you call the template function with a certain type, e.g.

printAbsValue(10.35);

the compiler deduces what type is required depending on what type of argument you use--in this case a double--and it creates a version of the function by substituting double everywhere T appears.

Note that generally you can't just use the "search and replace" facilitites of a text editor to replace int with T everywhere int appears in your function. There are some int types in a function you may not want to vary, like a loop variable:

Code:

for(int i = 0; i < 10; i++)
{
         ...
         ...
}

You wouldn't want i's type to vary with the template parameter, so in that case don't replace it with T when converting the funtion into a template function.

Thank to both of you I Think I've got it. I think I'm understanding the template concept much better.

I'll try to do another "little" program using template to make sure.

Thanks so much.

wats the difference with using template<typename T> instead of template<class T> because this is what i am used to using

There is no difference. class is the old-style way of doing it.

Originally Posted by dwks

There is no difference. class is the old-style way of doing it.

yep, it's purely a style-thing, but I prefer typename as class implies it must be a class (as opposed to a pod).

Some very old compilers only support class. I generally use typename, too, though.