Lucas Nussbaum <address@hidden> writes:
>> Come to think of it, if -h implied -g instead, you might have a
>> simpler implementation (albeit one with rounding errors and slower
>> performance).
>
> This would break with exabytes.
Why? -g uses floating point, and 1 EB (10**18) can be represented
exactly as an IEEE double. An IEEE double can even represent one ZB
(10**21) exactly. It does have a rounding error once you get to the
YB (10**24) range but like I said, there would be rounding errors.
And if you're talking about powers of 1024 not 1000, even one YiB
(2**80) can be represented exactly.
> My need for this is only to be able to sort output of commands like
> du -h.
Having -h imply -g would do this, no?