Add NotImplements and variants

[hslatman]

Summary

Add assert.NotImplements (and variants)

Changes

Added assert.NotImplements (and variants)

Motivation

There was no way to test an object did not implement a specific interface yet. Its counterpart does exist, so it was somewhat surprising to not be able to test the inverse.

import (
	"crypto/ecdsa"
	"crypto/elliptic"
	"crypto/rand"
	
	"github.com/stretchr/testify/assert"
        "github.com/stretchr/testify/require"
)


func Test_NotASigner(t *testing.T) {
	e, err := ecdsa.GenerateKey(elliptic.P256(), rand.Reader)
        require.NoError(t, err)
        assert.NotImplements(t, (*crypto.Signer)(nil), e.Public()) // public key only isn't a signer
}

Related issues

Closes #1065

[coditamar]

@CodiumAI-Agent please review

[CodiumAI-Agent]

PR Analysis

• 🎯 Main theme: Adding a new assertion method
• 🔍 Description and title: Yes
• 📌 Type of PR: Enhancement
• 🧪 Relevant tests added: Yes
• ✨ Minimal and focused: Yes, because the PR is focused on adding a new assertion method and all changes are relevant to this addition.
• 🔒 Security concerns: No, because the PR is focused on adding a new assertion method and does not involve any security-sensitive operations.

PR Feedback

• 💡 General PR suggestions: The PR is well-structured and follows the existing code style. The new method is added in all the necessary places and the tests cover the main use cases.

How to use

Tag me in a comment '@CodiumAI-Agent' to ask for a new review after you update the PR.
You can also tag me and ask any question, for example '@CodiumAI-Agent is the PR ready for merge?'

[dolmen]

Patch is good.

However I'm not sure we want to extend the API. At least the must wait for 1.9.0.

[brackendawson]

Personally I would be hesitant to include this kind of function in the API because it is somewhat a static assertion. I would prefer to encourage testing for state in the case used to justify this.

However in such cases I will defer to our predecessors, the previous acceptance of Implements and all the existing negative assertions makes this a good fit and the implementation looks sound.

At the time of writing there have been no functional changes since v1.8.5. Either we cut v1.8.6 soon and then merge this, or we merge this now and cut v1.9.0.