Using partial_sum(), I am trying to get the position from previous during the recurrence. I don't understand the result. I was expecting to get :
recurrence i=0
recurrence i=1
...
As the recurrence unfolds. But I am getting :
recurrence i=-48115006
recurrence i=-48115006
...
What am doing wrong ?
#include <vector>
#include <algorithm>
#include <numeric>
#include <stdio.h>
using namespace std;
int main()
{
const int n=15;
vector<vector<int> > vv(n 1);
vv[0]={42};
auto next=[&](const vector<int>& previous, const vector<int>&){
const int i = &previous - &vv[0];
printf("recurrence i=%d\n", i);
fflush(stdout);
vector<int> v;
return v;
};
partial_sum(vv.begin(), vv.end(), vv.begin(), next);
}
CodePudding user response:
The implementation for partial_sum
that you are using is probably similar to the one described in cppreference, Possible implementation, Second version:
template<class InputIt, class OutputIt, class BinaryOperation>
constexpr // since C 20
OutputIt partial_sum(InputIt first, InputIt last,
OutputIt d_first, BinaryOperation op)
{
if (first == last)
return d_first;
typename std::iterator_traits<InputIt>::value_type sum = *first;
*d_first = sum;
while ( first != last)
{
sum = op(std::move(sum), *first); // std::move since C 20
* d_first = sum;
}
return d_first;
}
That being the case, the first argument that partial_sum
passes to op
(the binary operation, next
in your code) is fix: it's a reference to an accumulator. It is the second argument the one that varies: it's a reference to the current element in each iteration (starting at the second element). But, in your example, you don't use this second argument to set i
. Instead, you always use &vv[0]
, which is also a fix value (unless vv
changes in size, which is not the case). So it's expected that you are always printing the same value: the difference of two fix addresses.
You can see it working here with some more debugging output.