- From: François REMY <francois.remy.dev@outlook.com>
- Date: Wed, 20 Aug 2014 23:08:45 +0200
- To: "Arthur Stolyar" <nekr.fabula@gmail.com>, Scott González <scott.gonzalez@gmail.com>
- Cc: "Rick Byers" <rbyers@chromium.org>, "Olli Pettay" <olli@pettay.fi>, <public-pointer-events@w3.org>
>> So the assumption here is that developers will continue to
>> choose native over web because of the .3ms spent on hit
>> testing during pointer movement? I think that's really far
>> fetched. I've met many many developers who prefer
>> Pointer Events over Touch Events, but less than a handful
>> of developers who prefer Touch Events over Pointer Events.
>> I can't imagine that many developers are concerned about
>> the performance hit in this area over developer ergonomics.
>
> It would be good to have
> element.setPointerCapture({ trackBoundaries: false });
>
> I understand position of Blink here, but PointerEvents are
> already here and they have this 3ms problem. I think it
> absolutely should be fixed.
The hit-testing cost is already removed for pointer-events if you do set a
'normal' pointer capture. The hit-testing cost is only incurred if you don't
set a pointer capture (because it has to compute the element owning the
event at every single pointer move, like it currently happens for the mouse
events in all desktop browsers).
Additonnally, any sane gesture library will definitely set a pointer capture
for you, so this isn't something you really need to worry about.
------
An interesting idea to ponder would be to allow auto-capture to be
expressable in CSS using "touch-action: trap-pointer", so that you don't
even have to register an event handler to disable hit-testing on swipe.
Another good thing would be to specify the hit-testing should be limited to
the current node and its children, like "touch-action: trap-pointer-inside",
to specify you only intend your drag to span a reduced set of your document.
To make this system complete, we could add another property like
"touch-action: trap-pointer-outside" which would make sure the hit-testing
during pointer move event do not consider all the children of an element
independently.
With this, it would be super-efficient to create a drag and drop between two
columns:
- the container would have "touch-action: none, trap-pointer-inside"
- the column containers would have "touch-action: trap-pointer-outside"
- the column containers would have pointerenter and pointerleave registered
- the items in the columns would have pointerdown registered (and it would
register pointerup on the global container)
Received on Wednesday, 20 August 2014 21:09:12 UTC