Cool.

Hope there wont be too long In Queue submissions...Fix this..its ruining CF...

You should have post this announcement after today's QF Quals mirror is finished.

Is it ....emmmmm　geometrical?

please end... contests had be unrated for many times.

Is it rated?

Pls do something about the long queues. Yesterday's contest was just an example of it but in practice also these days the subs take too much time to judge.

is it rated?

hope this time in the questions the fullstops and multiplication signs will be different! last time they used '.' for both and it made the question A look way complicated XD lol!

Gl & Hf to everyone!

The queue problem isn't going away. Even in yesterday's NEERC Subregional, the queue was too long and it affected the contest. When you submit a code, and you get a WA after 7-8 minutes, it doesn't feel good. In a contest, every second counts. Also, you can't concentrate properly on another problem even if you want to while waiting for a queued verdict. It's frustrating. I hope this doesn't occur in today's round. If it does, then it'd be really disheartening. CF is the best, we want it to remain the same way.

Is it mathematical?

Hope the server will work effiently

My first contest in CODEFORCES is Round 396 which is conducted by you and I am very happy that again I am participating in your contest.

Thank you.

why the last contest was unrated?

Score Distribution ? Not Declared Yet , 1 & Half Hour Left

Auto comment: topic has been updated by mahmoudbadawy (previous revision, new revision, compare).

if repeted problem sucks

Scala support is completely broken and couple of users reported it for a while. See: http://mirror.codeforces.com/blog/entry/54283 http://mirror.codeforces.com/blog/entry/54245

I also messaged the admins about it and its still not fixed :(

Hope someone from CodeForces notices this!

Am I the only one that gets stuck on infinite loading when trying to submit? Have been trying to submit C for 15 min already with no success.

How to solve B ?

can someone help me figure out why my solution gives tle? I used a random algorithm, the expected running time should be small and it runs fast in my computer. I cant pass pretest 6.

How to solve C? Good problem in my opinion

Nice problems, how to solve C?

I had an idea that if a is odd number then aXOR(a + 1)XOR(a + 2)XOR(a + 3) = a - 1, so we generate first 4 * k numbers and find remaining ones with some bruteforce.

Can answer be "NO" in C except 2 0 case?

How to solve D?

And Don't lie that you didn't try to send random solve for C.

Will the random solution pass for problem C? (If not, how to hack it?)

Can anyone say this code bug?(Div 2/c)

#include<bits/stdc++.h>
#define int long long
using namespace std;
const int maxn=1e5+7;
int a[maxn];

int32_t main(){
    ios_base::sync_with_stdio(0);cin.tie(0);cout.tie(0);
	int n,x;cin>>n>>x;
	int sum=0;
	for(int i=0;i<n;i++)a[i]=i,sum^=i;
	int nwx=x^sum;
	a[0]=nwx;
	if(nwx>0&&nwx<n)a[nwx]+=(1<<15),a[(nwx==1?2:nwx-1)]+=(1<<15);
	for(int i=0;i<n;i++)cout<<a[i]<<" "
}

I wanna cry, it was 5 seconds left before I press submit and it tells me contest is over((( Pretty sure the code i was about to submit was correct

For problem E, I have O(N + M + (M — N) log (M — N) + Q log (M — N)). Is that right?

Here is a solution to F:

First of there is a basic observations which is important for the solution.

• lcp(s_l, s_l + 1, ..., s_r) = min(lcp(s_l, s_l + 1), lcp(s_l + 1, s_l + 2), ..., lcp(s_r - 1, s_r))

Now lets use a_i as lcp(i, i + 1) ( a_i = lcp(i, i + 1) ).

The problem becomes:

We have an array a_i. We have queries for finding the maximum value of min(a_i, a_i + 1, ..., a_j - 1) * (j - i + 1) for some i and j in a given range. We also have a query for updating a value at a given position.

We also have some additional constraints. There are at most different values a_i — it's easy to prove.

So the main idea is to store different segment trees (for every value of a_i). In the segment tree for a value X we will consider only elements with values  ≥ X. For simplicity let's assume that in the segment tree all positions with value  ≥ X are set to 1 and all other to 0. Then if we want to find the maximum length of an interval [i;j] inside our query interval such that min(a_i, a_i + 1, ..., a_j - 1) >  = X we can simply find the length of the longest interval in the segment tree for X with all elements equal to 1.

Then the query can be done by considering every segment tree, doing query for each of them and then getting max X * T_X.query(l, r).

The update can be done by simply preforming at most updates in segment trees.

The only problem with this solution is the memory. It will consume memory. Fortunately we can use a persistent segment tree instead of a regular one and this way the memory will be .

PS: I was to lazy to write it during the contest, but is should pass.

ignore*

found my mistake in div2c

How 2 solve D ? I do not even understand why there is mention of fflush

So fast system testing

System Testing Pending
Refresh Page after 10min
79% done

This times system testing is running at speed of thunder lmao :D

upvote for the very quick testing

![ ](
image upload
)

I could succeed to hack one's code ,if the website can run faster ,and my rank will be rise 200 and more!!! why my hack in the last minute can't do it's proper work?

By the way, how to solve the f**k problem C?

Easy solution for C. if n<3, brute force. otherwise, xor first n-3 numbers,say res.(res = 0^1^...^(n-4)) Then 0,1,...,(n-4),(1<<18),(1<<19),((1<<18)+(1<<19))^(x^res) is answer.

Easy solution for problem C : https://ideone.com/FZ6rfw

C with x=0 really made the problem trickier.

vintage_Vlad_Makeev, KAN, MikeMirzayanov, I have accidentally sent the same hack twice (the page did not reload and I sent it from a new tab). Now I have two unsuccessful attempts. Can it be fixed?

Looks like I'm finally going to div.1! I'm so happy, it was quite a long journey :)

Hello, actually my wrong solution for problem D passed main tests. Here is my solution — http://mirror.codeforces.com/contest/862/submission/30518750 and it runs into an infinite loop in the following case : "10"

Now my C solution failed for the stupidest reason, I forgot to print YES before the answer itself in one of the cases. Why is output so poorly formatted anyway? Isn't it the tradition of problem setting to print -1 when there is no solution?

Auto comment: topic has been updated by mahmoudbadawy (previous revision, new revision, compare).

After 17 fails and 1:49 my submission for D passed, but it failed again on 59th test. The worst feeling ever

nice problems, c is interesting for me and random_shuffle() solution hhe

How can one proove that random brute solution for C will not time-out ?

please don't make other rounds

can anyone please help me! This is my solution can someone please tell why is it a run time error one the codeforces compiler? it works in my system and had passed all the testcases! but then it gave a run time error?? thank you!

Does anyone knows what exactly compiler version and flags are used when you choose GNU C++11? I have a trouble with my solution http://mirror.codeforces.com/contest/862/submission/30520451. On my computer it answers 0 on first test using GCC 5.4.0, and on my friend's computer it returns the same answer using GCC 4.8. But test system thinks that answer is -273605408. What the problem?

Can someone give me some links to good random algorithms tutorial. I see a lot of people used it to solve C.

i have got 237 point int this round keep doing well mahmoudbadawy thank you so much for this round and problem :)

for problem C you can just select first n-1 elements randomly (from 0 to 2^19-1) and see if the number that you need to get the desired xor is available. try it 100 times, if none of those 100 times works then print NO. this works because the probability the nth number will be available is approximately 1/5 so the probability you get it wrong is 1/5^100. Of course this is just a heuristic, you need some sort of uniformity argument but w/e.