Some tests (only in large cities):
5001{ 0.01, 0.02},{49.99,50.00},{50.00,50.01},{99.98,99.99}
3334{ 0.01, 0.03},{99.97,99.99}
2503{25.00,25.01},{33.32,33.33},{66.67,66.68},{74.99,75.00}
2501{ 0.01, 0.04},{24.99,25.00},{49.98,50.00},{50.00,50.02},
{75.00,75.01},{99.96,99.99}
2004{20.00,20.01},{79.99,80.00}
2003{33.34,33.35},{39.99,40.00},{60.00,60.01},{66.65,66.66}
2002{40.00,40.01},{59.99,60.00}
2001{ 0.01, 0.05},{19.99,20.00},{80.00,80.01},{99.95,99.99}
1667{ 0.01, 0.06},{49.97,50.00},{50.00,50.03},{99.94,99.99}
1429{ 0.01, 0.07},{33.31,33.33},{66.67,66.69},{99.93,99.99}
1257{12.49,12.50},{28.56,28.57},{71.43,71.44},{87.50,87.51}
1255{12.50,12.51},{87.49,87.50}
1253{16.67,16.68},{24.98,25.00},{33.34,33.36},{37.50,37.51},
{62.49,62.50},{66.64,66.66},{75.00,75.02},{83.32,83.33}
1251{ 0.01, 0.08},{25.00,25.02},{37.49,37.50},{49.96,50.00},
{50.00,50.04},{62.50,62.51},{74.98,75.00},{99.92,99.99}
1115{42.86,42.87},{57.13,57.14}
1112{ 0.01, 0.09},{99.91,99.99}
1009{10.00,10.01},{11.10,11.11},{88.89,88.90},{89.99,90.00}
1007{14.29,14.30},{29.99,30.00},{70.00,70.01},{85.70,85.71}
1004{20.00,20.02},{79.98,80.00}
The answers may be incorrect.

THANK YOU SO MUCH ! ! !!

Let me add some tests too:

1+
1

-10.0-
1

My prog passed all tests from topics below but failed these. Output obviously must be "Not a floating point number"

Представь вот мы сначала жарим две котлеты с одной стороны, первую меняем на третью, а вторую переворачиваем. Жарим еще минуту и теперь у нас есть одна готовая котлета и две наполовину готовые, и дожариваем их еще минуту. Вот и получается всего три минуты.

"It is guaranteed that there exists at least one a_i < 0"
So your test is incorrect

I also got answer 709227659 to this testcase.
Yes, seems like incorrect tests or incorrect problem statement.

Admins, please look at this issue. At least please check test 7.
Also please, clarify order of N and M values in input data (if they go in M,N order then correct statement please).

P.S.
  Also the problem is not so difficult as described in post-contest jury solutions. If problem statement is correct, it has got simple O(N*M) dp solution (based on https://en.wikipedia.org/wiki/Lindstr%C3%B6m%E2%80%93Gessel%E2%80%93Viennot_lemma principle).

We fixed wrong tests. Original author miswrote 10^9+7 as 10000007 in his code, that generated incorrect tests. We sincerely apologize for such an error

Edited by author 28.01.2025 02:06

For some old C++ compilers it was impossible to determine whether the exception is caught or not. At the same time there are very little practical use cases for using exceptions in C++ solutions. So, there was no reason in trying to fix the issue.

I don't know if the modern compilers already made it possible to determine. But even if they do I would still keep this feature disallowed because it takes reasonable effort to verify that a not so popular feature works properly on each new version of each of the C++ compilers.

Thanks for pointing this out. I added this information to the Guide.

Your test added. 4 authors lost accepted solutions after the rejudge. Thanks!

In fact, the commission rate is never 100 in the tests. I changed the problem statements to reflect that.

I decided not to add such tests because it would affect a significant number of accepted solutions with a particular algorithm where this matters, while the solutions using other algorithms would stay unaffected.

Because we need to find 2 separate indexes i < j :  |a[i] + a[j]| is minimal.
Not sum of segment |a[i]+a[i+1]+...+a[j]|, but only 2 numbers a[i] and a[j].