Expose assert on sandbox instances

[ljharb]

Per https://github.com/wealthfront/sinon-sandbox/issues/7:

In our codebase, we use import sinon from 'sinon-sandbox', and have a linter rule that forbids importing sinon directly. However, this means we can't use sinon.assert.

Would you be willing to accept a PR that attaches sinon.assert to sandbox instances?

[mantoni]

Maybe you can use sinon.assert.expose to attach it yourself? See http://sinonjs.org/docs/#assertions

[ljharb]

@mantoni according to http://sinonjs.org/docs/#assert-expose, that means that the actual sinon.assert API won't be there - ie, sinon.assert.called will become assertCalled - the whole idea is that sinonSandbox.assert has identical methods as sinon.assert :-/

[mantoni]

Well, then you can add the following here:

sandbox.assert = sinon.assert;

Or is there more to it?

[ljharb]

Indeed! Per https://github.com/wealthfront/sinon-sandbox/issues/7#issuecomment-228912571 however, I think it would be better for the upstream to do it - since the sandbox already exposes match, the precedent is there.

Is there a reason that this semver-minor change wouldn't be a good thing?

[mantoni]

Fine with me, was just curious. We're going to review the API surface for the 3.0 release and I wouldn't want to introduce anything that doesn't fit with that. Maybe @mroderick or @fatso83 have any feelings?

[fatso83]

I just reviewed #1000 (offtopic: we should try and get this ball rolling again) and I could not really see much preventing us from doing this. It is a very minor change AFAIK, and seemingly only aligns the sandbox api with that of the sinon object.

So I am pro.

[ljharb]

Submitted #1076

[ljharb]

Yay, thanks everyone! Looking forward to the release.