Continuation of JWKS support from PR #71

[davidallendj]

This PR continues the work from PR #71 to add JWKS support, but removes the dynamic verifier as requested.

[davidallendj]

@pkieltyka Can I get a review for this?

[alexlovelltroy]

I'm working on rebasing and adding more documentation. I don't understand why encode isn't supported in this case and I think we need a better error message for it. Can you explain what this is for?

[davidallendj]

I think that's the wrong error message probably from some copy-pasta. I think the message should just reflect that the JWKS isn't set so nothing can be encoded here.

Edit: This may have been something added by the original author, so I'm not 100% sure of the intent here, so I'm speculating.

[alexlovelltroy]

Based on my understanding, I'm considering updating Encode like this:

// Encode generates a JWT token string with the provided claims.
// It returns the encoded token as a string, along with the token object and any error encountered.
// If the JWTAuth instance has a key set, encoding is not supported and an error is returned.
func (ja *JWTAuth) Encode(claims map[string]interface{}) (t jwt.Token, tokenString string, err error) {
	if ja.keySet != nil {
		return nil, "", fmt.Errorf("encoding is not supported with key set")
	}

	t = jwt.New()
	for k, v := range claims {
		if err := t.Set(k, v); err != nil {
			return nil, "", err
		}
	}
	payload, err := ja.sign(t)
	if err != nil {
		return nil, "", err
	}
	tokenString = string(payload)
	return
}

[davidallendj]

Looks fine to me. I was reading the code wrong and was thinking the ja.keySet was an error. I'll have to go back and look at the entire PR to recall the context why this was done.

[alexlovelltroy]

I'm refactoring it a bit now that I understand it better.