New features

[wltrup]

This PR adds the following changes:

Release 0.2.0 breaks backward compatibility with the earlier release, since one of the protocols has disappeared and a new one has been added. Specific changes and new features are as follows:

• The presenting view controller is no longer required to conform to DrawerPresenting. In fact, DrawerPresenting no longer exists. Instead, a new protocol was created to take its place, DrawerCoordinating, so that any object can conform to it and be responsible for vending the drawer display controller. Of course, the presenting view controller can still fulfil that responsibility but it no longer must do so.

• Added tap-to-expand. Now, if the corresponding configuration boolean is turned on, tapping on the drawer when it's in its partially expanded state will trigger it to transition to its fully expanded state.

• Added some under-the-hood improvements in preparation for adding upcoming new features.

[sergdort]

So compiler now can synthesize equatable?

[RuiAAPeres]

For enums without associated values I think it had for a while.

Oh wait, missed that CGFloat

[wltrup]

No. The implementation is internal because it depends on a custom equal method that is also internal (actually, private but it will soon have to be internal). See the PresentationController file.

extension DrawerState {
    static func ==(lhs: DrawerState, rhs: DrawerState) -> Bool {
        switch (lhs, rhs) {
        case (.collapsed, .collapsed),
             (.partiallyExpanded, .partiallyExpanded),
             (.fullyExpanded, .fullyExpanded):
            return true
        case let (.transitioning(lhsCurPosY), .transitioning(rhsCurPosY)):
            return equal(lhsCurPosY, rhsCurPosY)
        default:
            return false
        }
    }
}

[wltrup]

The issue is that equality of floating point numbers is a tricky thing, so I have

private func equal(_ lhs: CGFloat, _ rhs: CGFloat) -> Bool {
    let epsilon: CGFloat = 1e-6
    return abs(lhs - rhs) <= epsilon
}

[diegopetrucci]

guard maximumCornerRadius != 0
    && drawerPartialY != 0
    && drawerPartialY != containerViewH
else { return 0 }

I think it's more aligned to our conventions.

[wltrup]

Fixed in 4d92109.

[diegopetrucci]

These pharentesis :S

[wltrup]

Sorry. Unless more people complain about them, I'm leaving them as is. I find that they improve readability.

[diegopetrucci]

here I like them :D

[wltrup]

Here they are needed, else the boolean logic would have a different result.

[diegopetrucci]

indentation per two comments above

[wltrup]

Fixed in c83e689.

[diegopetrucci]

What does // recall that Y-axis points down mean?

[wltrup]

It's somewhat counter-intuitive that a negative speed means moving up rather than down, and the reason is that the vertical axis is positive pointing down, hence the comment.

[diegopetrucci]

super minor stuff but lgtm

[zzcgumn]

case transitioning(currentY: CGFloat) is, at least in my opinion, clearer than adding a documenting comment.