Implement raw mouse movement for Mac

[swiftcoder]

Also hide the cursor when grabbed, to prevent the cursor being in an inconsistent state after both hiding and grabbing.

[stephank]

Looks like this needs merging with master, if we actually want to match on the axis, because at this point the axis ID was still private.

There are no conflicts, but should apply this on top:

diff --git a/src/platform/macos/events_loop.rs b/src/platform/macos/events_loop.rs
index 37ee535e..8973ab5e 100644
--- a/src/platform/macos/events_loop.rs
+++ b/src/platform/macos/events_loop.rs
@@ -7,7 +7,6 @@ use std::sync::{Arc, Mutex, Weak};
 use super::window::Window;
 use std;
 use super::DeviceId;
-use AxisId;
 
 
 pub struct EventsLoop {
@@ -481,14 +480,14 @@ impl EventsLoop {
 
                 let delta_x = (scale_factor * ns_event.deltaX() as f32) as f64;
                 if delta_x != 0.0 {
-                    let motion_event = DeviceEvent::Motion { axis: AxisId(0), value: delta_x };
+                    let motion_event = DeviceEvent::Motion { axis: 0, value: delta_x };
                     let event = Event::DeviceEvent{ device_id: DEVICE_ID, event: motion_event };
                     events.push_back(event);
                 }
 
                 let delta_y = (scale_factor * ns_event.deltaY() as f32) as f64;
                 if delta_y != 0.0 {
-                    let motion_event = DeviceEvent::Motion { axis: AxisId(1), value: delta_y };
+                    let motion_event = DeviceEvent::Motion { axis: 1, value: delta_y };
                     let event = Event::DeviceEvent{ device_id: DEVICE_ID, event: motion_event };
                     events.push_back(event);
                 }

[swiftcoder]

I've rebased these against mainline and added the requested changes

[mitchmindtree]

I just tested the window and grabbing examples and the changes seem to work nicely. The code also looks fine, although it still uses the DEVICE_ID placeholder rather than using something provided by the system, but this is probably fine for now.

I guess auto hiding the cursor makes sense for most cases when grabbed, @tomaka @vberger is this what windows and wayland do? Maybe there should be a separate API for manually hiding or showing the cursor, though I guess that discussion would be best left to a separate issue.

Either way I think it's fine to merge this as is 👍

[tomaka]

I guess auto hiding the cursor makes sense for most cases when grabbed, @tomaka @vberger is this what windows and wayland do?

That's what Windows does.

Anyway, thanks for the PR.

[elinorbgr]

I guess auto hiding the cursor makes sense for most cases when grabbed, @tomaka @vberger is this what windows and wayland do?

On wayland, it is really up to the client to decide, there is no automatic behavior. So yes, auto-hiding the cursor seems legit to me.