CODEWITHSUNDEEP
c++ctemplates
Dexter
Dexter

Reputation: 1337

setting default value depending upon data type in c++

I have a sample code like 

template <typename BIT_LENGTH>
class mod {
public:
mod(BIT_LENGTH val) : m_val (powl(2,sizeof(BIT_LENGTH) * 8) - 1), _val(val) {
}

void value() {
    std::cout << " sizeof -- " << sizeof(BIT_LENGTH) << std::endl;
    std::cout << " value is : ---- " << m_val << std::endl;
    std::cout << " value is : ---- " << _val << std::endl;
}
private:
BIT_LENGTH m_val;
BIT_LENGTH _val;

};

int main(){
mod<uint64_t> mod1(10);
mod1.value();
}

Now i want to initialize my member variable to the max value of the templatized type.

What can be the best possible way to do it ?.

Anything cleaner than using powl function ?

Upvotes: 0

Views: 123

Answers (2)

Cheers and hth. - Alf
Cheers and hth. - Alf

Reputation: 145279

check out std::numeric_limits.

powl is not a standard C++ function: it's from Linux. Oh, it's there in C99. Oh well! In C++ essentially the same function is available with just the name pow, an overload (curiously powl is not mentioned in the C++11 standard).

By the way, you can (statistically) avoid annoying unintended text substitutions by reserving ALL UPPERCASE for macro names. It's also much easier on the eye. In general.

Also, just in case you're not aware, in C and C++ a byte is not guaranteed to be 8 bits. On some platforms, notably some Texas Instruments digital signal processors, a byte (e.g. a char) is 16 bits. Historically it has also been other sizes. It has to be at least 8 bits but that's it. The number of bits per byte is available as CHAR_BIT from <limits.h>.

Finally, if you always use systematic indentation you will find that you get a much better understanding of the code.

And that also others will much easier understand the code.

And almost magically, that results in a great reduction of the number of bugs you have to spend time on hunting down and correcting.

Upvotes: 2

jogojapan
jogojapan

Reputation: 69977

 #include <limits>

 std::numeric_limits<BIT_LENGTH>::max()

will do this if BIT_LENGTH is a built-in type. That is a very misleading name for a data type, by the way.

Upvotes: 1

Related Questions