https://bugs.documentfoundation.org/show_bug.cgi?id=154792
--- Comment #7 from Eike Rathke <[email protected]> --- (In reply to NOYB from comment #6) > Oh but I do understand it. I doubt that. You probably don't know why all this not-straight-forward rounding is implemented. Just take a look at the example of bug 124286 =ROUNDDOWN(8.94-8;2) that for 8.94-8 with the floating point inaccuracy already produces 0.9399999999999995 which following the "rounding a number to nth digit is exact" would produce 0.93 instead of the expected 0.94, hence we try to account for inaccuracy which works for many but not all values and needs to be refined. > Within the > precision limits of the underlying system e.g. 19 digits etc. There is no 19 digits precision. IEEE 754 double has a decimal digits precision of 15 to 17. -- You are receiving this mail because: You are the assignee for the bug.
