Exponent detection with precision not working properly

[elfkuzco]

While using the library to format filesizes, it tends to show numbers in a weird way with fixed precision (most likely) because it is not computing the proper exponent.
Example:

Without precison

> filesize.filesize(1058223158, {base: 2, standard: 'iec'})
'1009.2 MiB'
> filesize.filesize(529825922, {base: 2, standard: 'iec'})
'505.28 MiB'

With precison

> filesize.filesize(1058223158, {base: 2, standard: 'iec', precision: 3})
'1.01e+3 MiB'
> filesize.filesize(529825922, {base: 2, standard: 'iec', precision: 3})
'505 MiB'

While not using precision seems to give a more readable output, I would like to use a defined precision and have '1.01e+3 MiB' shown as '0.990 GiB' instead.

[elfkuzco]

Thank you very much for responding fast but I think you might have introduced an error.

Current Behaviour

filesize.filesize(1058223158, {base: 2, standard: 'iec', precision: 3})
'1.01 MiB'

But this is wrong, this filesize should be close to Gib.

Without precision

filesize.filesize(1058223158, {base: 2, standard: 'iec'})
'1009.2 MiB'

I am re-opening the issue. Feel free to close if there's something I am missing. Thanks again.

[elfkuzco]

Apparently, I can't re-open the issue. Would wait a bit and re-open in a newer one

[avoidwork]

Yeah, it's not actually reasonable to try to implement as it changes the meaning of the calculation. Since the exponent can't be used for calculating the unit.

if the number is 1009.2 MiB, it's still 1009.2 MiB if you want to see it with precision... it doesn't become the number you'd like to become.

you're asking for "next major unit - current number as new unit".

[elfkuzco]

you're asking for "next major unit - current number as new unit".

yeah

Not entirely sure I understand what you mean by "current number as new unit". But if it shows as 1.01 GiB or 0.990 GiB, then, yeah (although 0.990 seems more accurate)

[avoidwork]

I mean that if there's an exponent in the result of the toPrecision() you'd like to have the number processed from the next unit up, e.g. (1009/1024).toPrecision(3). <-- this is not the number you're providing though, it's just the calculation of the delta to the next unit.

[elfkuzco]

I mean that if there's an exponent in the result of the toPrecision() you'd like to have the number processed from the next unit up, e.g. (1009/1024).toPrecision(3). <-- this is not the number you're providing though, it's just the calculation of the delta to the next unit.

Yes, indeed

[avoidwork]

it's a simple op to code, but i don't think it's correct for all cases

[avoidwork]

i'm prep'n the version of what you'd like to see; lets see how it's received.