Add validator

[aljones15]

This adds a validator for both Identifiers and Documents
that should conform to v0.13 of the DID Spec.
This is being used for the spec tests.

[dmitrizagidulin]

My main question about this PR is - why not use the did-veres-one driver's VeresOneDidDoc.validateDid() (or at least extract the validation logic there to a static method, and use it here)? I'm not sure it makes sense to duplicate a lot of that logic..

[aljones15]

My main question about this PR is - why not use the did-veres-one driver's VeresOneDidDoc.validateDid() (or at least extract the validation logic there to a static method, and use it here)? I'm not sure it makes sense to duplicate a lot of that logic..

was not aware that existed (I did look) so will look into that in a moment.

EDIT: actually did see VeresOneDoc validator, but I guess it seemed out of date or to specific.

[aljones15]

@dmitrizagidulin looks like the validation there is:

A. much more specific than this one:

https://github.com/veres-one/did-veres-one/blob/master/lib/VeresOneDidDoc.js

const VERES_DID_REGEX = /^(did:v1:)(test:)?(uuid|nym):(.+)/;

• only accepts method-names uuid|test|nym (other implementations might use other method names)
• has a version number on did (is that correct? EDIT: sorry v1 is the methodName)

    if(this.id && this.id.startsWith('urn:uuid:')) {
      return {valid: true}; // short-circuit, UUID URNs are fine as is
    }

• returns true if the id starts with urn (no further validation)

if(mode === 'test' && didMode !== 'test') {
      return {
        error: new Error(`DID is invalid for test mode: "${this.id}".`),
        valid: false
      };
    }

• has a mode

honestly I think that validator is alright for veres, but definitely not general enough for all implementors of the spec. Perhaps we should run this validator on the document then do the rest of the existing VeresOneDidDoc validation?

Others:

• throws on uppercase did or methodName which is correct, but validator in this PR throws more specific errors.
• does not validate for the actual document i.e. ['@context'] being there etc.
• does not validate authentication, publicKeys, etc.

I guess the did identifier validation is good, but this seems more through or am I missing something?

[aljones15]

@dmitrizagidulin

Can the existing VeresOneDid validator pass the tests here:

w3c-ccg/did-test-suite#1

I think it would throw a few false positives etc.

p.s. thanks for reviewing this.