josh
/
Teacup_Firmware
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

gcode_parse.c: review decfloat_to_int() yet again. 

This time, raise range at the cost of some precision for the
newly explored areas.
This commit is contained in:
Markus Hitter 2012-03-06 12:08:43 +01:00
parent 88236ba698
commit 5a9a134ae4
1 changed files with 9 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
gcode_parse.c
View File

@ -57,9 +57,9 @@ extern const uint32_t powers[]; // defined in sermsg.c
/// convert a floating point input value into an integer with appropriate scaling.
/// \param *df pointer to floating point structure that holds fp value to convert
/// \param multiplicand multiply by this amount during conversion to integer
/// \param divide_by_1000 divide by 1000 during conversion to integer
///
/// lots of work has been done in exploring this function's limitations in terms of overflow and rounding
/// Tested for up to 42'000 mm (accurate), 420'000 mm (precision 10 um) and
/// 4'200'000 mm (precision 100 um).
static int32_t decfloat_to_int(decfloat *df, uint32_t multiplicand) {
uint32_t r = df->mantissa;
uint8_t e = df->exponent;
@ -68,6 +68,13 @@ static int32_t decfloat_to_int(decfloat *df, uint32_t multiplicand) {
if (e)
e--;
// This raises range for mm by factor 1000 and for inches by factor 100.
// It's a bit expensive, but we should have the time while parsing.
while (e && multiplicand % 10 == 0) {
multiplicand /= 10;
e--;
}
r *= multiplicand;
if (e)
r = (r + powers[e] / 2) / powers[e];