PointerEvents always adds hit testing overhead to touchmove events about pep HOT 2 CLOSED

Ah yep, that'd be a pretty easy optimization.

I'm trying to think of a simpler hit testing heuristic for the common case, maybe using the bounding box of the current target and its children? I'd love to hear other ideas.

from pep.

Jacob Rossi from the IE team argued for why the spec'd behavior was good here: http://lists.w3.org/Archives/Public/public-pointer-events/2014JanMar/0060.html

But we can still avoid doing a hit test on move when:

• the pointer has been captured,
• the capturing element has no pointerenter/pointerleave handler, AND
• the capturing element and it's ancestors have no pointerover/pointerout handlers (which are discouraged anyway)

I suggest we change polymer to do this now. If it's difficult we can always start with a simpler version that just avoids the hit test whenever the pointer has capture and leave it as a lower-priority bug that you don't get pointerover/out/enter/leave when crossing the boundary of the captured element.

from pep.