Re: how to implement sum of bits using template metaprogramming

From:

"Victor Bazarov" <v.Abazarov@comAcast.net>

Newsgroups:

comp.lang.c++

Date:

Tue, 25 Sep 2007 09:16:18 -0400

Message-ID:

<fdb1n2$8b9$1@news.datemas.de>

Alf P. Steinbach wrote:

* andrey.vul@gmail.com:

I'm tryin to convert this:
long sumOfBits(int n) {
    long x = 0;
    int i;
    for (i = n - 1; i >= 0; i--)
        x += (1 << i);
    return x;
}

into something like this:
template <int n> struct sumOfBits {
static const unsigned char u8value = (unsigned char)((1 << n) +
sumOfBits<n - 1>::u8value);
static const unsigned __int16 u16value = (unsigned __int16)((1 << n)
+ sumOfBits<n - 1>::u16value);
static const unsigned __int32 u32value = (unsigned __int32)((1 << n)
+ sumOfBits<n - 1>::u32value);
static const unsigned __int64 u64value = (unsigned __int64)((1 << n)
+ sumOfBits<n - 1>::u64value);
};
template <> struct sumOfBits <0> {
static const unsigned char u8value = (unsigned char)1;
static const unsigned __int16 u16value = (unsigned __int16)1;
static const unsigned __int32 u32value = (unsigned __int32)1;
static const unsigned __int64 u64value = (unsigned __int64)1;
};

and so far, all expansions, that have multiples of 8 (8, 16, 32, 64)
return 0 due to overflow.
Any way to correctly rewrite the template?

Off the cuff, to reproduce the behavior of the original non-tempalte
function:

template< int n >
struct SumOfBits
{
enum{ value = (1L << (n-1)) | ((1L << (n-1)) - 1) };
};

Well, the loop is uncalled for, so you're, right, it could just be

    template<unsigned n>
    struct SumOfBits
    {
        enum { value = (1 << n) - 1 }; // no need to OR anything
    };

V
--
Please remove capital 'A's when replying by e-mail
I do not respond to top-posted replies, please don't ask