Support single-character units for parsing

[bartoncasey]

Adds support for parsing single-character units (k, m, g, t, p).

[dougwilson]

Thank you for your pull request. Isn't this already possible like the following?

input = '45k'
console.log(bytes(`{input}b`))

Can you elaborate on the purpose of this feature and how it would be used, etc.?

[bartoncasey]

Thank you for your pull request. Isn't this already possible like the following?

Sure, but then you're just putting extra parsing logic into your application code.

Can you elaborate on the purpose of this feature and how it would be used, etc.?

It just makes the parsing a little more flexible when working with input from different sources. The meaning of the single-character unit is unambiguous, so I don't see why bytes shouldn't accept it gracefully.

[dougwilson]

The meaning of the single-character unit is unambiguous, so I don't see why bytes shouldn't accept it gracefully.

How will it interact with the addition of kib, mib, and similar units (#26)?

[bartoncasey]

How will it interact with the addition of kib, mib, and similar units?

I suspect the least surprising thing to do would be to map it to classic binary units.

That would be a breaking change due to the current units switching to metric anyway, so it would probably be part of a bigger refactoring and api change.

[dougwilson]

Gotcha. So we should wait on this to land support for the classic binary units first, then add the single letter alias as a mapping to those? This is likely a breaking change already, as folks may be relying on the existing parsing to reject '100k', for example.

[bartoncasey]

/shrug. I have no idea about the current roadmap. I just hacked this into my local copy as part of my current project, and thought it would be generally useful.