hyung-hwan/hcl
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

enhanced the multiplication function to truncate the resulting fixed-point decimal to the equal precision to the operands

Browse Source
This commit is contained in:
hyung-hwan
2018-03-30 18:11:11 +00:00
parent e466c6b68a
commit 8df5087b9d
2 changed files with 39 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

41
lib/number.c
View File

@ -96,6 +96,26 @@ static hcl_ooi_t equalize_scale (hcl_t* hcl, hcl_oop_t* x, hcl_oop_t* y)
return xs;
}
static hcl_oop_t truncate (hcl_t* hcl, hcl_oop_t iv, hcl_ooi_t cs, hcl_ooi_t ns)
{
/* this function truncates an existing fixed-point decimal.
* it doesn't create a new object */
if (cs > ns)
{
do
{
/* TODO: optimizatino... less divisions */
iv = hcl_divints(hcl, iv, HCL_SMOOI_TO_OOP(10), 0, HCL_NULL);
if (!iv) return HCL_NULL;
cs--;
}
while (cs > ns);
}
return iv;
}
hcl_oop_t hcl_addnums (hcl_t* hcl, hcl_oop_t x, hcl_oop_t y)
{
if (!HCL_IS_FPDEC(hcl, x) && !HCL_IS_FPDEC(hcl, y))
@ -156,7 +176,7 @@ hcl_oop_t hcl_subnums (hcl_t* hcl, hcl_oop_t x, hcl_oop_t y)
hcl_oop_t hcl_mulnums (hcl_t* hcl, hcl_oop_t x, hcl_oop_t y)
{
hcl_ooi_t xs, ys, scale;
hcl_ooi_t xs, ys, cs, ns;
hcl_oop_t nv;
hcl_oop_t xv, yv;
@ -189,16 +209,21 @@ hcl_oop_t hcl_mulnums (hcl_t* hcl, hcl_oop_t x, hcl_oop_t y)
nv = hcl_mulints(hcl, xv, yv);
if (!nv) return HCL_NULL;
scale = xs + ys;
if (scale > HCL_SMOOI_MAX)
{
/* TODO: limit scale */
}
cs = xs + ys;
if (cs <= 0) return nv; /* the result must be an integer */
return hcl_makefpdec(hcl, nv, scale);
ns = (xs > ys)? xs: ys;
/* cs may be larger than HCL_SMOOI_MAX. but ns is guaranteed to be
* equal to or less than HCL_SMOOI_MAX */
HCL_ASSERT (hcl, ns > 0 && ns <= HCL_SMOOI_MAX);
nv = truncate(hcl, nv, cs, ns);
if (!nv) return HCL_NULL;
return hcl_makefpdec(hcl, nv, ns);
}
hcl_oop_t hcl_divnums (hcl_t* hcl, hcl_oop_t x, hcl_oop_t y)
{
hcl_ooi_t xs, ys, i;