Proper mouse multi-button workarounds

[jonathanolson]

So phetsims/resistance-in-a-wire#147 brings up an interesting case where we can actually NOT receive a pointerup event for a pointerdown when pressing left/right mouse buttons (probably due to the context menu).

We CAN get the lostpointercapture event (which also triggers after usual ups). So we could theoretically:

• When we get lostpointercapture while it's a touch/pen, cancel it.
• When we get lostpointercapture while it's a mouse, do mouseup for the pressed buttons (has complications)
• When we get a mouse up, do NOT fire the event if we got the lostpointercapture earlier (i.e. it's recorded as up on the Mouse object)

Besides being likely to break things:

• We can't just call mouseUp(), since we need to control the mouse button than we're forcing up. That's stored on the PointerEvent, so we'd need to factor that out.
• We'd need to call it for potentially multiple buttons (say the left and right buttons were down at the same time, as in the example from the linked issue).
• We'd need to consider passing through the button for events in multiple places.

Luckily we should be able to put a button field on our custom Event type. It turns out I'm the only person who has ever referenced this in code (it's used in canStartPress), so I can localize the changes to Scenery.

Assigning to @ariel-phet for prioritization (as the reward/work ratio for this seems somewhat low).

[jonathanolson]

Also, we'd want to have an array of "buttons" pressed on the Mouse, instead of the three options we have now. (buttons being by number).

[zepumph]

Note that this issue is the only thing holding up a new RC (spot check) of RIAW before publication.

[jonathanolson]

I'm also deprecating Pointer.isDownProperty, it will need more fine-grained control (since that doesn't mention which mouse button is down).

[ariel-phet]

This might be useful to handle more cleanly in the future, but given the very "edge-case" nature of the bugs reported, not worth putting in any effort at the moment

[jessegreenberg]

This was also noticed in phetsims/balloons-and-static-electricity#91 (comment).