Stress Testing Kit

Catches wrong answers by comparing your optimised solution against a brute force on random inputs.

---

Files

File	Purpose	Edit per problem?
gen.cpp	Random input generator	✅ make_test() only
solution.cpp	Your optimised solution	✅
brute.cpp	Your correct brute force	✅
stress.bat	Test runner	❌

After running, these folders are created automatically:

build\     <- compiled binaries
dump\      <- temporary I/O files (auto-cleaned after each run)
result\    <- saved inputs and outputs from failing tests

---

Usage

stress.bat              :: 1000 tests, compiles everything
stress.bat 500          :: 500 tests
stress.bat 500 0        :: 500 tests, skip recompile

---

On a Mismatch

The input, your output, and the expected output are printed directly in cmd. After pressing any key they are saved to:

result\fail_input.txt
result\fail_output.txt
result\fail_expected.txt
result\fail_seed.txt

To replay the exact failing input at any time:

build\gen <seed> > dump\input.txt
build\solution  < dump\input.txt

---

On a Crash

If solution.exe or brute.exe exits with a non-zero code (segfault, failed assert, out of bounds), the run stops immediately with the input printed and saved to result\.

---

Editing gen.cpp

The only part you edit per problem is make_test(). Everything else is reusable across problems.

void make_test()
{
    int n = get(1, 100);
    auto a = get_array<int>(n, 1, 1000);

    cout << n << "\n";
    for (int i = 0; i < n; i++)
        cout << a[i] << " \n"[i == n - 1];
}

For multi-test problems, uncomment and configure t in main():

t = get(1, 10);
cout << t << "\n";

Available Helpers

Function	Description
get(l, r)	Random integer in [l, r], no modulo bias
get_string(n)	Random lowercase string of length n
get_array<T>(n, l, r)	Random array of n values in [l, r]
get_permutation(n)	Random permutation of 1..n
get_tree(n)	Random tree, n nodes, 1-indexed, returns edge list
get_graph(n, m)	Random simple graph, n nodes, m edges

Reproducibility

stress.bat passes the loop index as the seed to gen automatically. Every test is fully reproducible — seed 47 always produces the same input regardless of when you run it.

---

Recommended Workflow

1. Write solution.cpp and brute.cpp
2. Edit gen.cpp -> make_test() to match the problem's input format
3. Run stress.bat
4. On mismatch: debug solution.cpp using the saved result\ files
5. Fix and re-run until all tests pass