Failing test for <StrictMode>

[timdorr]

Doesn't fix #897. But lets us know what's going to break in future versions of React:

Warning: Unsafe lifecycle methods were found within a strict-mode tree:

    componentWillReceiveProps: Please update the following components to use static getDerivedStateFromProps instead: Connect(Container)

    componentWillUpdate: Please update the following components to use componentDidUpdate instead: Connect(Container)

    Learn more about this warning here:
    https://fb.me/react-strict-mode-warnings

If others want to fix this, you can use this branch as a base. Feel free to try it out!

[theKashey]

In terms of React-Hot-Loader componentWillUpdate could not be replaced by componentDidRender, as long RHL needs to being able render "a new" Connected component, not a cached version :(
Something major should be changed.

[timdorr]

@theKashey We may be able to use the new getSnapshotBeforeUpdate and cDU while keeping oldListeners around between the two (or going up to the top module level).

[theKashey]

Or just not cache rendered component :)
If a new state will be calculated in getDerivedStateFromProps, then it is possible just to cancel the unnecessary render. (and react-hot-loader will bypass this cancel)

Trying to find that cached component I am talking about, and could not find anything....

Anyway - using componentDidUpdate should be ok. And why should component reconnect to the store, not just update selector?