alphanumdun

Alphanumeric + - (dash) and _ (underscore) is what this W3C specification is after:

> A _valid path id_ is a string where each character is ASCII alphanumeric or "_" or "-".

I have no idea why W3C would define such limitation/spec around what could be used as unique identifier for what is going to be used as path, where / or even :\\ are common and valid chars, among all others, but sure thing I am pragmatic enough that "I don't care", I just want users of my libraries to be able to vfs.mount('/tmp') to define /tmp as unique idenfitier for their mounted folder into any virtual FileSystem so this is what this library does:

encode(str) to return a "safe to be used*" string to define a unique mounting path
* decode(encoded_str) to return the original value of that string used to mount such path

``js
import * as id from 'https://esm.run/alphanumdun';

console.assert(id.decode(id.encode('/test')) === '/test');
// that's it, that /test will work when you use it
// as 'id for the FileSystemHandle API
``

Implementation Details

This 2 (effective) LOC utility simply excludes - as valid char and converts anything that is not a-z, A-Z or 0-9 into its base36 counterpart surrounded by - accepted chars to guarantee every intent is unique but penalizing strings that heavily use - within their value.

That's it, you can convert back and forward anything you like and use that as id in any FileSystemHandle related operation that accepts such id as optional hint.



Reasons behind - as penalized char



First of all, the easy way to convert any string to an alphanumeric compatible value is base36, which is also the top conversion utility we have for numbers in JS so that (1234567890).toString(36) would produce kf12oi which is fully compatible with the alphanumeric ID the specs expect.

Now, because it's possible to use also _ or - as identifier, the choice to avoid collision around users' defined entries and actual transformation is left between those two chars.

As a both JS and Python developer, I know that imports from Python accept _ as module name, due syntax restrictions, but not - so that import my-module would fail but import my_module wouldn't.

And that's it, since most people out there developing and using the FileSystemHandle primitive will need to deal with syntax constraints and/or Python, I've felt like it was natural to keep the _ char that is widely supported in various PLs syntaxes, and exclude - from the equation so that - or the explicit -19-` input would produce two entirely different unique identifiers as output 🥳

- - -

> "wait ... WUT? Python on the Web?"

Sure thing: PyScript