Proposal: Test Cleanup

The test/ directory is rather difficult to follow at the moment. When adding new tests it is not always clear where the test should be added. I had some difficulties when working on #437.

I’m willing to progressively fix up the test/ directory in a series of PRs, but since it would be a fair bit of effort, I do not want to start the process if there is not much of a chance that the PRs would be merged.

This would go nicely with the coverage report that will be added with #468.

Tests

• algorithm ECDSA
• algorithm HMAC
• algorithm invalid
• algorithm none
• algorithm RSA
• claim audience/aud
• claim expiresIn/exp (ignoreExpiration, clockTolerance, clockTimestamp)
• claim iat/maxAge (clockTolerance, clockTimestamp)
• claim invalid
• claim issuer/iss
• claim jwtid/jti
• claim notBefore/nbf (ignoreNotBefore, clockTolerance, clockTimestamp)
• claim private
• claim subject/sub
• header keyId/kid
• option clockTimestamp
• option clockTolerance
• option encoding
• option header
• option ignoreExpiration
• option ignoreNotBefore
• option mutatePayload
• option noTimestamp

[ziluvatar]

It sounds really good!

[ziluvatar]

Certain options operate on several claims (ex. clockTimestamp). In those cases I have been placing a test for that option with the claim test (ex. ‘should verify “nbf” using “clockTimestamp”’) in nbf.test.js.

I like this pattern ^ If an option will modify how a claim is handled I think that is under the responsibility of that feature / claim support. One thing you can do is to prepare a group of utility functions for testing regarding that option that can be reused in every claim.

Even if I like more that option, it is not like I hate the other one it is a matter of decide what we think it is the best.