C 二维数组损坏

Question

我遇到了一个问题，我的二维数组正在损坏。考虑这段代码:

int **generate_clique_adjacency_matrix(const formula f) {
    int num_terms = f.num_clauses * kVarsPerClause;
    printf("\nnum_terms: %d\n", num_terms);

    // construct a 2D array of appropriate size
    int **compatible = malloc(sizeof(int) * num_terms);
    for (int i = 0; i < num_terms; i++) {
        compatible[i] = malloc(sizeof(int) * num_terms);
        // initialize every cell to 0
        for (int j = 0; j < num_terms; j++) {
            printf("writing 0 to (%d, %d)\n", i, j);
            compatible[i][j] = 0;
        }
    }

    printf("num_terms: %d\n", num_terms);
    printf("matrix:\n");
    for (int i = 0; i < num_terms; i++) {
        printf("row %d: ", i);
        for (int j = 0; j < num_terms; j++) {
            printf("%d ", compatible[i][j]);
        }
        printf("\n");
    }
    exit(0);
}

哪个正在产生令人震惊的输出:

num_terms: 6
writing 0 to (0, 0)
writing 0 to (0, 1)
writing 0 to (0, 2)
writing 0 to (0, 3)
writing 0 to (0, 4)
writing 0 to (0, 5)
writing 0 to (1, 0)
writing 0 to (1, 1)
writing 0 to (1, 2)
writing 0 to (1, 3)
writing 0 to (1, 4)
writing 0 to (1, 5)
writing 0 to (2, 0)
writing 0 to (2, 1)
writing 0 to (2, 2)
writing 0 to (2, 3)
writing 0 to (2, 4)
writing 0 to (2, 5)
writing 0 to (3, 0)
writing 0 to (3, 1)
writing 0 to (3, 2)
writing 0 to (3, 3)
writing 0 to (3, 4)
writing 0 to (3, 5)
writing 0 to (4, 0)
writing 0 to (4, 1)
writing 0 to (4, 2)
writing 0 to (4, 3)
writing 0 to (4, 4)
writing 0 to (4, 5)
writing 0 to (5, 0)
writing 0 to (5, 1)
writing 0 to (5, 2)
writing 0 to (5, 3)
writing 0 to (5, 4)
writing 0 to (5, 5)
num_terms: 6
matrix:
row 0: 22661104 0 22661136 0 0 0 
row 1: 0 0 0 0 0 0 
row 2: 0 0 0 0 0 0 
row 3: 0 0 0 0 0 0 
row 4: 0 0 0 0 0 0 
row 5: 0 0 0 0 0 0 

我的阵列是如何损坏的？显然，第0行的数据应该全为0。我什至尝试添加断言以确保 malloc 成功。这让我抓狂。

Answer 1

这是错误的:

int **compatible = malloc(sizeof(int) * num_terms);

您需要为 n 个指向 int 的指针预留空间(不是为 n 个 int)

更改为

int **compatible = malloc(sizeof(int *) * num_terms);

或者更好

int **compatible = malloc(sizeof(*compatible) * num_terms);