assert, logging and testing C code

assert: checking your own assumptions

#include <assert.h>

double average(const int *marks, int n) {
    assert(marks != NULL);       /* a programming error if this is false */
    assert(n > 0);
    long sum = 0;
    for (int i = 0; i < n; i++) sum += marks[i];
    return (double) sum / n;
}
/* if violated: prog: report.c:5: average: Assertion 'n > 0' failed.  (then abort) */

An assertion documents and checks something that must ALWAYS be true if the code is correct. When it fails, the program stops immediately at the real cause, instead of crashing later in a confusing place. Compiling with -DNDEBUG removes every assert, so release builds pay nothing.

Error codes and errno

#include <errno.h>
#include <string.h>

FILE *f = fopen("config.ini", "r");
if (!f) {
    fprintf(stderr, "cannot open config.ini: %s\n", strerror(errno));   /* e.g. "No such file or directory" */
    return -1;
}

A common C convention: functions return 0 on success and a negative number on failure, and put results in output pointers. Document each function's error codes, and check them at every call.

A tiny test harness

#include <stdio.h>

static int failures = 0;
#define CHECK(cond) do { \
    if (!(cond)) { fprintf(stderr, "FAIL %s:%d: %s\n", __FILE__, __LINE__, #cond); failures++; } \
} while (0)

int gcd(int a, int b);

int main(void) {
    CHECK(gcd(84, 36) == 12);
    CHECK(gcd(17, 5) == 1);
    CHECK(gcd(10, 0) == 10);      /* edge case */
    CHECK(gcd(0, 0) == 0);        /* edge case: what SHOULD it return? */
    if (failures) { printf("%d test(s) failed\n", failures); return 1; }
    printf("all tests passed\n");
    return 0;
}

Test frameworks

Larger projects use a framework that organises tests, reports results and helps with mocks: Unity (popular in embedded work, a single C file), CMocka (with mocking support), Check, or GoogleTest (C++, but often used to test C libraries). The ideas are the same as in any language: arrange, act, assert, one behaviour per test, edge cases first.

What to test in C code

• Boundaries: 0, 1, the maximum size, exactly-full buffers, empty strings.
• Error paths: allocation failure (inject it), missing files, invalid input.
• Round trips: write then read a file; encode then decode.
• Run the whole test suite under sanitizers (topic 3).