Function libnotcurses_sys::ffi::ncmetric

pub unsafe extern "C" fn ncmetric(
    val: uintmax_t, 
    decimal: uintmax_t, 
    buf: *mut c_char, 
    omitdec: c_int, 
    mult: uintmax_t, 
    uprefix: c_int
) -> *const c_char

Floating-point is never used, because an IEEE758 double can only losslessly represent integers through 2^53-1.

2^64-1 is 18446744073709551615, 18.45E(xa). KMGTPEZY thus suffice to handle an 89-bit uintmax_t. Beyond Z(etta) and Y(otta) lie lands unspecified by SI. 2^-63 is 0.000000000000000000108, 1.08a(tto). val: value to print decimal: scaling. ‘1’ if none has taken place. buf: buffer in which string will be generated omitdec: inhibit printing of all-0 decimal portions mult: base of suffix system (almost always 1000 or 1024) uprefix: character to print following suffix (‘i’ for kibibytes basically). only printed if suffix is actually printed (input >= mult).

You are encouraged to consult notcurses_metric(3).