support react's hydrate for enzyme-adapter-react-16

[brucewpaul]

It would be great to support testing react's hydrate method as opposed to just the render method. This could maybe be done via an options argument?

cc @ljharb

[ljharb]

With an empty DOM node, won't hydrate and render be identical?

[brucewpaul]

Ya I think so, but this would be able to specifically test server side rendering of a component.

cc @goatslacker

[ljharb]

Certainly mount invokes ReactDOM.render; an adapter option could be used in the v16+ adapter to specify ReactDOM.hydrate be used.

I'm just not sure what the point would be - all we'd be doing is testing ReactDOM itself, since hydrate and render theoretically do the same thing on an empty container.

[goatslacker]

There's currently a warning when you try to render() SSR'd markup, so it would at the very least silence that warning.

[ljharb]

In that case, maybe it'd be better to always use hydrate in the React 16 adapter here?

[goatslacker]

We should have both, for non server rendered content I think you want to use render

[ljharb]

OK, talked to @goatslacker about this - how about a mount option, parallel to attachTo, called hydrateIn? Both options are mutually exclusive; both take a DOM element; hydrateIn calls hydrate on the adapter if present and falls back to render if not?