Prev: rvalue-reference result premature destruction
Next: draft standard n3090.pdf, std::unique() algorithm
From: PGK on 10 Apr 2010 15:01
Is it possible to find the maximum value of a C++1x tuple, and then
update that value? I could almost imagine calculating the maximum
value in a tuple, but the index corresponding to this value would not
be a compile-time constant literal, as required by the tuple "get"
function.
Graham.

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

From: Mathias Gaunard on 11 Apr 2010 05:01
On 11 avr, 07:01, PGK <graham.k...(a)gmail.com> wrote:
> Is it possible to find the maximum value of a C++1x tuple, and then
> update that value?

Yes.

> I could almost imagine calculating the maximum
> value in a tuple, but the index corresponding to this value would not
> be a compile-time constant literal, as required by the tuple "get"
> function.

Well obviously, you need a meta-programming loop, not a normal loop.

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

From: Pete Becker on 11 Apr 2010 05:01
PGK wrote:
> Is it possible to find the maximum value of a C++1x tuple,

It's impossible to say. There is no work being done on C++1x.

--
Pete
Roundhouse Consulting, Ltd. (www.versatilecoding.com) Author of
"The Standard C++ Library Extensions: a Tutorial and Reference"
(www.petebecker.com/tr1book)

[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

From: PGK on 12 Apr 2010 06:41
> Is it possible to find the maximum value of a C++1x tuple, and then
> update that value?

The code below will find the index of the max value of a tuple of
elements supporting the lessthan operator. The problem is, how do I
use this to update that tuple value; say set it to zero? (The const
int I obtain cannot be used in the tuple "get" function as it cannot
appear in a constant expression.)

template <typename T_Tuple,size_t size>
struct max_helper {
static const int tup_max2( T_Tuple &t, double cur_max, int &index )
{
if (get<size-1>(t) > cur_max) {
cur_max = get<size-1>(t);
index = size-1;
}
max_helper<T_Tuple,size-1>::tup_max2(t,cur_max,index);
return index;
}
};

template <typename T_Tuple>
struct max_helper<T_Tuple,0> {
static const int tup_max2( T_Tuple &t, double cur_max, int &index )
{ return index; }
};

int main(int argc, char *argv[]) {
int i = -1;
double cur = -100.0;
tuple<int,double,long,int,float> t(3,5,2,14,7); // max is at index 3
const int index = max_helper<decltype(t),
tuple_size<decltype(t)>::value>
::tup_max2(t,cur,i);
std::cout << "max value is at index: " << i << std::endl;
// get<index>(t); // error
return 0;
}

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

From: SG on 12 Apr 2010 06:40
On 11 Apr., 08:01, PGK wrote:
> Is it possible to find the maximum value of a C++1x tuple,

You mean C++0x.

> and then update that value?

Yes. You could unroll the loop via metaprogramming/recursion and
return a reference to the max element -- provided that the element
types are all the same. You chould start like this (untested):

#include <tuple>
#include <utility>

template<int Remaining>
struct tuple_for_each_helper
{
template<typename Tuple, typename Func>
static void doit(Tuple & tup, Func & f)
{
const int ts = std::tuple_size<Tuple>::value;
f( std::get<(ts-Remaining)>(tup) );
tuple_for_each_helper<(Remaining-1)>::doit(tup,f);
}
};

template<>
struct tuple_for_each_helper<0>
{
template<typename Tuple, typename Func>
static void doit(Tuple & tup, Func & f) {}
};

template<typename Tuple, typename Func>
Func tuple_for_each(Tuple & tup, Func f)
{
const int ts = std::tuple_size<Tuple>::value;
tuple_for_each_helper<ts>::doit(tup,f);
return std::move(f);
}

The rest follows from that. For example, you might want to use
tuple_for_each with a lambda expression. Alternativly, you could use
std::array instead of std::tuple. std::array allows indexing with
"runtime values".

Cheers,
SG

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

 |  Next  |  Last
Pages: 1 2
Prev: rvalue-reference result premature destruction
Next: draft standard n3090.pdf, std::unique() algorithm