FlxMouseEventManager: correct the mouseChildren parameter, closes #2102

[MSGhero]

Closes #2102

Fixes a bug with mouseChildren in FlxMouseEventManager where items with the parameter set to false still allowed overlapping objects to receive events. E.g. the children of a parent sprite group with mouseChildren = false still received events if they were added to FlxMEM after.

I forgot I made this change I've been using this change since the issue and haven't had any problems. It does break the order of the objects if that's a concern, which I could change but would be more involved.

[Gama11]

I would also like to have a unit test for this, if possible.

[MSGhero]

reorder also needs to be changed since it disrupts the add order.

I'd like for sprite groups with mouseChildren = false to block their children's potential mouseChildren = false, drawing from how Flash does it. So this needs a bit more imo.

[MSGhero]

Essentially, the mouseChildren false and true objects should be in two separate arrays, or one array where the boundary between the two halves is managed. Adding an object or changing the mouseChildren param should place that object at the front of its proper subarray, which preserves the user's order.

@Gama11 my breaking question is if Flx(Sprite)Groups should block their children from receiving events like AS3's mouseChildren, which is not the case currently. That would have to go into reorder because I'm not really sure how to accomplish that in add without just calling reorder.

[MSGhero]

Will unit test, but the most recent commit keeps mouseChildren = true consistent with = false in terms of the add order.

[Gama11]

What's the status here? Is that Flx(Sprite)Group question still relevant?

[MSGhero]

Ehhhhh if you want it to be. Finding some time this and next weekend to code.

[MSGhero]

Ok

[Gama11]

This is really hacky, a unit test shouldn't need to rely on @:access(), only public API.

the parameter set to false still allowed overlapping objects to receive events.

You should be able to test for that not happening anymore?

[Gama11]

(you should be able to dispatch mouse events manually, similar to what we do here:
https://github.com/HaxeFlixel/flixel/blob/dev/tests/unit/src/flixel/input/gamepad/FlxGamepadManagerTest.hx)

[MSGhero]

Ohhh I didn't think of that. That's a much better test, yeah.

[MSGhero]

Ok, again.

[Gama11]

Nice, that's much cleaner. 👍