New option for precision

[binarykitchen]

Would be good to have an option to allow one digit after the floating-point.

Especially for scientific based apps.

[sindresorhus]

Can you provide an example of input, the output you currently get, and the output you'd like to get?

[binarykitchen]

What I currently get:

prettyBytes(183715);
//=> '184 kB'

What I'm after:

prettyBytes(183715, { digits: 1 }); // new option
//=> '183.7 kB'

[sindresorhus]

prettyBytes(1337) results in 1.34 kB, so with the current behavior, I'm honestly not sure how it decides when to round or not...

How about a option, that by default, does the human readable thing and rounds when it makes sense? Like the current behavior was intended to do. And if you turn it off, you get the behavior you want with fraction digits. Like: https://developer.apple.com/documentation/foundation/bytecountformatter/1417887-isadaptive

[binarykitchen]

I hear you. Probably sounds right.

The thing is, we are porting old code to a newer system (migrations) and have to ensure behaviour stays the same. And there is always one fraction digit in our bytes, can't question this. Hence, as long as I can force that with a new option, then I'm happy with it.

[sindresorhus]

How I'm imagining it, turning off the isAdaptive option would result in all units having one fractional digit.