https://bugs.kde.org/show_bug.cgi?id=460232
--- Comment #2 from [email protected] --- Turns out string to knumber's float variant conversion uses GNU MPFR's old init function, thus lost the precision from user input to internal representation. I changed the init function to "mpfr_init2" with 1024 bit precision. Should be fixed now. A similar change will be applied to KCalc, since that's where Kalk's knumber from. :) -- You are receiving this mail because: You are watching all bug changes.
