You are given an array of _n_ integer numbers. Let _sum_(_l_, _r_) be the sum of all numbers on positions from _l_ to _r_ non-inclusive (_l_\-th element is counted, _r_\-th element is not counted). For indices _l_ and _r_ holds 0 ≤ _l_ ≤ _r_ ≤ _n_. Indices in array are numbered from 0.
For example, if _a_ = \[ - 5, 3, 9, 4\], then _sum_(0, 1) = - 5, _sum_(0, 2) = - 2, _sum_(1, 4) = 16 and _sum_(_i_, _i_) = 0 for each _i_ from 0 to 4.
Choose the indices of three delimiters _delim_0, _delim_1, _delim_2 (0 ≤ _delim_0 ≤ _delim_1 ≤ _delim_2 ≤ _n_) and divide the array in such a way that the value of _res_ = _sum_(0, _delim_0) - _sum_(_delim_0, _delim_1) + _sum_(_delim_1, _delim_2) - _sum_(_delim_2, _n_) is maximal.
Note that some of the expressions _sum_(_l_, _r_) can correspond to empty segments (if _l_ = _r_ for some segment).
## Input
The first line contains one integer number _n_ (1 ≤ _n_ ≤ 5000).
The second line contains _n_ numbers _a_0, _a_1, ..., _a__n_ - 1 ( - 109 ≤ _a__i_ ≤ 109).
## Output
Choose three indices so that the value of _res_ is maximal. If there are multiple answers, print any of them.
[samples]