Would this library support generating namespaced UUIDs without a namespace?

Is your feature request related to a problem? Please describe.

I am working with some other microservices implemented in Java, their acceptance tests use UUIDs which are generated from string names, and no namespaces, using this library: docs, source.

Describe the solution you’d like

I think it would be possible to include a similar option, which omits the namespace if it is not given.

Describe alternatives you’ve considered

I am considering just using the URL namespace, but i was hoping to raise as little eyebrows as possible for future workers who have to read my code. Tests are like documentation after all, and I was hoping to reflect the logic as closely as possible as the Java side.

It is also possible to give { length: 16 } as the argument instead of a uuid string, which achieves the desired result, but would be preferable to just give null as the argument, or something along those lines.

Additional context

Here is the logic which i was hoping to get from this library, in Java code (please forgive me, i couldnt get a job doing nodejs in my city):

    @Override
    public UUID generate(byte[] nameBytes)
    {
        byte[] digest;
        synchronized (_digester) {
            _digester.reset();
            if (_namespace != null) {
                _digester.update(UUIDUtil.asByteArray(_namespace));
            }
            _digester.update(nameBytes);
            digest = _digester.digest();
        }
        return UUIDUtil.constructUUID(_type, digest);
    }

[Alan-Liang]

I came across this when trying to implement the Yggdrasil protocol for Minecraft; if you really need to generate the same UUID as UUID.nameUUIDFromBytes(bytes) in Java does, here is a hack to trick the library into doing this:

const uuid = require('uuid')
// copied from https://github.com/uuidjs/uuid/blob/7feb2c3e854964999afa9aad026058b400ed1388/src/v35.js#L4-L14
function stringToBytes (str) {
  str = unescape(encodeURIComponent(str)) // UTF8 escape
  const bytes = []
  for (let i = 0; i < str.length; ++i) bytes.push(str.charCodeAt(i))
  return bytes
}

const str = 'hello'
const bytes = stringToBytes(str)
const name = {
  ...bytes,
  get length () {
    return name.ok
      ? bytes.length
      : (name.ok = true, bytes.length - 16)
  },
}
const namespace = {
  get length () {
    return namespace.ok
      ? 0
      : (namespace.ok = true, 16)
  },
}
uuid.v5(name, namespace) // => 'aaf4c61d-dcc5-58a2-9abe-de0f3b482cd9'

It’s a hack anyway (so pin the npm package version), and I really hope we can achieve this in an easier way.

[ctavan]

@alexanderankin yes, this is totally expected. The two different UUIDs you are generating with the JAVA library are:

hash(name) != hash(NIL + name)

where NIL are just 16 0-bytes.