sys/fmt: make fmt_s32_dfp() string based

sys/fmt/fmt.c

@@ -308,61 +308,62 @@ size_t fmt_s16_dec(char *out, int16_t val)
     return fmt_s32_dec(out, val);
 }
 
-size_t fmt_s16_dfp(char *out, int16_t val, int fp_digits)
+size_t fmt_s16_dfp(char *out, int16_t val, int scale)
 {
-    return fmt_s32_dfp(out, val, fp_digits);
+    return fmt_s32_dfp(out, val, scale);
 }
 
-size_t fmt_s32_dfp(char *out, int32_t val, int fp_digits)
+size_t fmt_s32_dfp(char *out, int32_t val, int scale)
 {
-    assert(fp_digits > -(int)TENMAP_SIZE);
+    unsigned pos = 0;
 
-    unsigned  pos = 0;
-
-    if (fp_digits == 0) {
+    if (scale == 0) {
         pos = fmt_s32_dec(out, val);
     }
-    else if (fp_digits > 0) {
+    else if (scale > 0) {
         pos = fmt_s32_dec(out, val);
         if (out) {
-            memset(&out[pos], '0', fp_digits);
+            memset(&out[pos], '0', scale);
         }
-        pos += fp_digits;
+        pos += scale;
     }
     else {
-        fp_digits *= -1;
-        uint32_t e = _tenmap[fp_digits];
-        int32_t abs = (val / (int32_t)e);
-        int32_t div = val - (abs * e);
-
-        /* the divisor should never be negative */
-        if (div < 0) {
-            div *= -1;
-        }
-        /* handle special case for negative number with zero as absolute value */
-        if ((abs == 0) && (val < 0)) {
+        scale = -scale;
+        char buf[10]; /* "2147483648" */
+        int negative = val < 0;
+        uint32_t uval = negative ? -val : val;
+        int len = fmt_u32_dec(buf, uval);
+        if (negative) {
             if (out) {
                 out[pos] = '-';
             }
             pos++;
         }
+        if (len <= scale) {
+            int zeroes = scale - len + 1;
+            if (out) {
+                memset(&out[pos], '0', zeroes);
+            }
+            pos += zeroes;
+        }
+        if (out) {
+            memcpy(&out[pos], buf, len);
+        }
 
-        if (!out) {
-            /* compensate for the decimal point character... */
-            pos += fmt_s32_dec(NULL, abs) + 1;
+        pos += len;
+
+        if (out) {
+            unsigned dot_pos = pos - scale;
+            for (unsigned i = pos; i >= dot_pos; i--) {
+                out[i] = out[i - 1];
+            }
+            out[dot_pos] = '.';
         }
-        else {
-            pos += fmt_s32_dec(&out[pos], abs);
-            out[pos++] = '.';
-            unsigned div_len = fmt_s32_dec(&out[pos], div);
-            fmt_lpad(&out[pos], div_len, (size_t)fp_digits, '0');
-        }
-        pos += fp_digits;
+        pos += 1;
     }
 
     return pos;
 }
-
 /* this is very probably not the most efficient implementation, as it at least
  * pulls in floating point math.  But it works, and it's always nice to have
  * low hanging fruits when optimizing. (Kaspar)

sys/include/fmt.h

@@ -287,39 +287,30 @@ size_t fmt_s16_dec(char *out, int16_t val);
  *
  * @param[out] out          Pointer to the output buffer, or NULL
  * @param[in]  val          Fixed point value
- * @param[in]  fp_digits    Number of digits after the decimal point, MUST be
- *                          >= -7
+ * @param[in]  scale        Scale value
  *
  * @return      Length of the resulting string
  */
-size_t fmt_s16_dfp(char *out, int16_t val, int fp_digits);
+size_t fmt_s16_dfp(char *out, int16_t val, int scale);
 
 /**
  * @brief Convert 32-bit fixed point number to a decimal string
  *
- * The input for this function is a signed 32-bit integer holding the fixed
- * point value as well as an integer defining the position of the decimal point.
- * This value is used to shift the decimal point to the right (positive value
- * of @p fp_digits) or to the left (negative value of @p fp_digits).
+ * This multiplies a 32bit signed number by 10^(scale) before formatting.
  *
- * Will add a leading "-" if @p val is negative.
+ * The resulting string will always be padded with zeros after the decimal point.
  *
- * The resulting string will always be patted with zeros after the decimal point.
- *
- * For example: if @p val is -3548 and @p fp_digits is -2, the resulting string
- * will be "-35.48". The same value for @p val with @p fp_digits of 2 will
- * result in "-354800".
- *
- * @pre fp_digits > -8 (TENMAP_SIZE)
+ * For example: if @p val is -35648 and @p scale is -2, the resulting
+ * string will be "-352.48"( -35648*10^-2). The same value for @p val with
+ * @p scale of 2 will result in "-3524800" (-35648*10^2).
  *
  * @param[out] out          Pointer to the output buffer, or NULL
  * @param[in]  val          Fixed point value
- * @param[in]  fp_digits    Number of digits after the decimal point, MUST be
- *                          >= -7
+ * @param[in]  scale        Scale value
  *
  * @return      Length of the resulting string
  */
-size_t fmt_s32_dfp(char *out, int32_t val, int fp_digits);
+size_t fmt_s32_dfp(char *out, int32_t val, int scale);
 
 /**
  * @brief Format float to string