Division problem about cc65-floatlib HOT 5 OPEN

Did you use a recent checkout, ie after i changed the binary format to IEEE754? That might be more accurate, before that it used just two bytes for the mantissa.

That said, there may well be other problems with the rounding and int/float conversion - i did not do any extensive tests.

from cc65-floatlib.

Looks like I grabbed it 3 days before the last update. I'll pull and try again.

from cc65-floatlib.

If you are adventurous, you could also try https://github.com/mrdudz/cc65/tree/fptest (see cc65/cc65#1777 )

from cc65-floatlib.

I love a good adventure. The latest update did not change the result, so I'll give fptest a try.

from cc65-floatlib.

The problem is likely the crude way i am converting from int to float, basically just truncating the mantissa instead of rounding it.

from cc65-floatlib.