[POC] end-user-programming "REPL" exploration

[szymonkaliski]

This PR is proof-of-concept for REPL for connecting from laptop sidecar machine to running capstone instance.

Basic running: install, restart capstone, and then (on sidecar machine):

$ cd src/apps/repl-cli
$ ./capstone-cli "capstone://Workspace/..." # grab the workspace from running capstone instance

Welcome to Capstone CLI [hash]
>>>

You can now execute one-off commands within the Capstone env:

# not too exciting
>>> 2 + 2
4

# dump the current store state
>>> Content.store
{ ... }

# snap all cards to grid
>>> Content.once(Content.rootBoardUrl, change => change(doc => {
    Object.keys(doc.cards).forEach(key => {
      const snap = 100
      doc.cards[key].x = Math.round(doc.cards[key].x / snap) * snap
      doc.cards[key].y = Math.round(doc.cards[key].y / snap) * snap
    })

    return doc
  })
)

I'm also experimenting with idea of hooks to attach behaviours, the only existing hook right now is hooks.onCardDragEnded but this allows us to have snap-to-grid happen after each drag is completed (hook is execute within doc.change context so we can update card directly):

>>> hooks.onCardDragEnded = (card) => {
  const snap = 100
  card.x = Math.round(card.x / snap) * snap
  card.y = Math.round(card.y / snap) * snap
}

Putting this out so you can play with it before Friday's show&tell where I hope to demo this functionality and share some notes.

This probably shouldn't be merged.