C: rational-numbers

c/rational-numbers/src/rational_numbers.c

@@ -0,0 +1,115 @@
+#include "rational_numbers.h"
+#include <stdlib.h>
+#include <math.h>
+
+#define N(r)  ((r).numerator)
+#define D(r)  ((r).denominator)
+
+/* Note. We should probably check for possible overflow in all
+ * functions below (not covered in exercise).
+ * To do this, a simple solution could be to make operations with
+ * long long (or some other method), and add a field in rational_t
+ * to express such overflow, division by zero, etc...
+ */
+
+/* Euclidean algorithm (by Donald Knuth) */
+rational_t reduce(rational_t r)
+{
+    int16_t a=abs(N(r)), b=abs(D(r)), t;
+
+    while (b != 0) {
+        t = b;
+        b = a % b;
+        a = t;
+    }
+    if (D(r) < 0)
+        a=-a;
+    return (rational_t) { N(r)/a, D(r)/a };
+}
+
+/* to avoid pow() for integers
+ * BUG: does not check for overflow
+ */
+static inline int power(int n, int p)
+{
+    int res=n;
+
+    if (p==0)
+        return 1;
+    while (--p)
+        res*=n;
+    return res;
+}
+
+/* All formulas below come from https://en.wikipedia.org/wiki/Rational_number
+ */
+rational_t add(rational_t r1, rational_t r2)
+{
+    return reduce((rational_t) {
+            N(r1) * D(r2) + N(r2) * D(r1),
+            D(r1) * D(r2)
+        });
+}
+
+rational_t subtract(rational_t r1, rational_t r2)
+{
+    return reduce((rational_t) {
+            N(r1) * D(r2) - N(r2) * D(r1),
+            D(r1) * D(r2)
+        });
+}
+
+rational_t multiply(rational_t r1, rational_t r2)
+{
+    return reduce((rational_t) {
+            N(r1) * N(r2),
+            D(r1) * D(r2)
+        });
+}
+
+rational_t divide(rational_t r1, rational_t r2)
+{
+    return reduce((rational_t) {
+            N(r1) * D(r2),
+            D(r1) * N(r2)
+        });
+}
+
+rational_t absolute(rational_t r)
+{
+    return (rational_t) {
+        abs(N(r)),
+        abs(D(r))
+    };
+}
+
+rational_t exp_rational(rational_t r, uint16_t n)
+{
+    return reduce((rational_t) {
+            power(N(r), n),
+            power(D(r), n)
+        });
+}
+
+float exp_real(uint16_t x, rational_t r)
+{
+    return powf((float)x, (float)N(r)/D(r));
+}
+
+/* See GNUmakefile below for explanation
+ * https://github.com/braoult/exercism/blob/master/c/templates/GNUmakefile
+ */
+#ifdef UNIT_TEST
+int main(int ac, char **av)
+{
+    int arg=1;
+    rational_t r1, r2;
+
+    for (; arg<ac-1; ++arg, ++arg) {
+        r1.numerator=atoi(av[arg]);;
+        r1.denominator=atoi(av[arg+1]);;
+        r2=reduce(r1);
+        printf("reduce(%d, %d)=(%d, %d)\n", N(r1), D(r1), N(r2), D(r2));
+    }
+}
+#endif

c/rational-numbers/src/rational_numbers.h

@@ -0,0 +1,33 @@
+#ifndef RATIONAL_NUMBERS
+#define RATIONAL_NUMBERS
+
+#include <stdint.h>
+
+typedef struct {
+    int16_t numerator;
+    int16_t denominator;
+} rational_t;
+
+rational_t add(rational_t r1, rational_t r2);
+rational_t subtract(rational_t r1, rational_t r2);
+rational_t multiply(rational_t r1, rational_t r2);
+rational_t divide(rational_t r1, rational_t r2);
+rational_t absolute(rational_t r);
+rational_t exp_rational(rational_t r, uint16_t n);
+float exp_real(uint16_t x, rational_t r);
+rational_t reduce(rational_t r);
+
+/* See GNUmakefile below for explanation
+ * https://github.com/braoult/exercism/blob/master/c/templates/GNUmakefile
+ */
+#if defined UNIT_TEST || defined DEBUG
+#include <stdio.h>
+#include <stdlib.h>
+#endif
+
+#ifdef  TESTALL
+#undef  TEST_IGNORE
+#define TEST_IGNORE() {}
+#endif
+
+#endif