Largest All One Square in C++

#include <stdio.h>
#include <vector>
#include <algorithm>
#include <stdexcept>
#include <cstdint>
#include <limits>
// Largest All One Square in C++
//
// PURPOSE
// Each all-one square ending at a cell extends the smallest neighboring square from above,
// left, and upper-left.
//
// METHOD
// Use a single padded row and retain the previous diagonal. A zero cell resets its entry; a
// one cell adds one to the minimum of three neighbors.
//
// COMPLEXITY
// O(rows * cols) time, O(cols) auxiliary space.
//
// CONTRACT
// Flattened 0/1 matrix; rows,cols>=0. Returns SIDE LENGTH, not area.
// Public APIs infer array lengths except in C, which accepts pointer/length pairs. Vertex
// counts and matrix dimensions remain explicit. Arrays must contain signed 32-bit integers.
// Sorted order, correct tree topology, and caller-owned data-structure invariants are
// preconditions; checking them on every query would change query complexity.
// Scalar and shape checks are included. Value-range checks are included where needed to
// prevent invalid indexing or nontermination. Allocation sizes are capped at 4,000,000
// workspace entries. Arithmetic outside Python raises an error on 32-bit overflow; Python
// uses arbitrary precision. Native allocation failures propagate. In-place outputs may be
// partly updated after arithmetic failure. These are single-threaded, in-memory
// implementations.
//
// LANGUAGE DETAILS
// C++17: const vector references for inputs, mutable references for outputs, RAII workspace,
// standard lower_bound for LIS. Invalid inputs throw invalid_argument; overflow throws
// overflow_error; allocation failure propagates bad_alloc.
//
// EXAMPLE
// Expected return: 2.
// Run: g++ -std=c++17 -O2 example.cpp -o example && ./example
// Tests for this exact source artifact run before publication. The accompanying source
// archive includes unit cases, runtime results, and source hashes.

static int alg_min(int a, int b) { return a < b ? a : b; }
static int alg_max(int a, int b) { return a > b ? a : b; }
static void alg_require(bool valid) { if (!valid) throw std::invalid_argument("invalid algorithm input"); }
static int alg_checked(int64_t v) { if (v < INT32_MIN || v > INT32_MAX) throw std::overflow_error("32-bit result overflow"); return (int)v; }
static int alg_add(int a,int b) { return alg_checked((int64_t)a+b); }
static int alg_sub(int a,int b) { return alg_checked((int64_t)a-b); }
static int alg_mul(int a,int b) { return alg_checked((int64_t)a*b); }
static size_t alg_size(int n) { alg_require(n>=0 && n<=4000000); return (size_t)n; }

int largest_square(const std::vector<int>& grid, int rows, int cols) {
    alg_require(grid.size() <= 1000000);
    alg_require(((int)grid.size() == alg_mul(rows, cols)));
    alg_require(((rows >= 0) && (cols >= 0) && (rows <= 1000000) && (cols <= 1000000)));
    std::vector<int> dp(alg_size(alg_add(cols, 1)), 0);
    int r = 0;
    int c = 0;
    int diagonal = 0;
    int old = 0;
    int best = 0;
    while ((r < rows)) {
        diagonal = 0;
        c = 1;
        while ((c <= cols)) {
            old = dp[c];
            if ((grid[alg_sub(alg_add(alg_mul(r, cols), c), 1)] == 1)) {
                dp[c] = alg_add(1, alg_min(dp[c], alg_min(dp[alg_sub(c, 1)], diagonal)));
                best = alg_max(best, dp[c]);
            } else {
                dp[c] = 0;
            }
            diagonal = old;
            c = alg_add(c, 1);
        }
        r = alg_add(r, 1);
    }
    return best;
}

int main(void) {
    std::vector<int> grid = {1, 1, 0, 1, 1, 1, 0, 1, 1};
    int rows = 3;
    int cols = 3;
    int result = largest_square(grid, rows, cols);
    printf("%d\n", result);
    return 0;
}