read this: Precision Error Problems

This is totally unfair. when there is like 30 min left in the contest the pages didn't even load let alone submitting .at first I thought its my connection problem, later I found its codeforces problem. even what surprising to me, I got those pop up like 30 min after the contest end.very pathetic.

In the problem, E.I'm stuck in bs implementation of @GreenGrape.

my question is,for smallest t we got count_le(t) == k.how can I prove that t is in the set of some combination of(D(A)*D(B))?

Direction array:

int dx[4] = {-1 ,1 ,0 , 0} ;

int dy[4] = {0 ,0 -1 , 1} ; ==> gave wrong answer. but then

changing it to ,

int dx[4] = {-1 ,0 ,0 , 1} ;

int dy[4] = {0 ,-1 , 1 , 0} ; ==> the solution got accepted!

Suppose the smallest number( let's call it P ) in gcd set S divides every other numbers in S.That means the problem has a solution. now,

Case 1:

when i == j .

Then we know that gcd( a[i] , a[j] ) = a[i] / a[j] ( belongs S ).so every elements of S can be in the answer set A.

Case 2:

when i != j

now no matter how you chose your range it always includes P, if you put n-1 Ps alternatively in between n numbers.

Combining the two cases above produce the correct answer.

please make the training camp materials available with lots of related problems.because there is almost no resources on this for people who don't have the opportunity to attend the training camp.