Currently we leak the internal representation of them as two GValues that
contain a fraction. Without this we could store fraction ranges as
data[0] = (min_n << 32) | (min_d)
data[1] = (max_n << 32) | (max_d)
and wouldn't require an additional allocation per range.
/* g_value_unset (&end); */
}
+/* FIXME 2.0: Don't leak the internal representation of fraction
+ * ranges but instead return the numerator and denominator
+ * separately.
+ * This would allow to store fraction ranges as
+ * data[0] = (min_n << 32) | (min_d)
+ * data[1] = (max_n << 32) | (max_d)
+ * without requiring an additional allocation for each value.
+ */
+
/**
* gst_value_get_fraction_range_min:
* @value: a GValue initialized to GST_TYPE_FRACTION_RANGE