Define pointermove and pointerrawupdate may have delta from previous corresponding pointer event as movementX and movementY

[masayuki-nakano]

MouseEvent.movementX and MouseEvent.movementY are extensions of Pointer Lock 2.0 and the spec of Pointer Events does not refer these attributes. As of current definition, those values of pointermove and pointerrawupdate should be 0. However, Chromium explicitly sets them.

• Computing the value
• Retrieving the last position
• Storing the last position at dispatching a pointer event

I think that the Chromium's behavior is reasonable and Mozilla got a bug report about this. So, I suggest that the spec of Pointer Events should define pointermove and pointerrawupdate may have non-zero values of movementX and movementY.

• If the event type is pointermove, the source event of the last position should be pointerdown, pointermove or pointerup which had same pointerId
• If the event type is pointerrawupdate, the source event of the last position should be pointerrawupdate which had same pointerId
• At same time of Pointer Lock definition, the last position should be reset.

Then, it would be better to redesign the type of movementX and movementY are double, but it must be better browsers to keep rounding the values of MouseEvent to integer for backward compatibility.

[smaug----]

Looks like this is basically a duplicate of w3c/pointerlock#100

[patrickhlauke]

Discussed at meeting today: we plan to keep this issue open until the pointerlock issue is closed, but group fundamentally thinks this should be handled there rather than in pointer events.

[smaug----]

@masayuki-nakano do browsers currently use non-fractional values with mousemove?

[masayuki-nakano]

@masayuki-nakano do browsers currently use non-fractional values with mousemove?

First, Firefox always use integer coordinates, and also as far as I've checked, Chromium always use std::floor in the getter methods of MouseEvent.

I'm not sure about Safari, but looks like WebKit's MouseEvent is always initialized with integer coordinates. And anyway, it seems that they don't support fractional coordinates even with PointerEvent.

• https://wpt.fyi/results/pointerevents/pointerevent_fractional_coordinates.html%3Fmouse?label=experimental&label=master&aligned
• https://wpt.fyi/results/pointerevents/pointerevent_fractional_coordinates_untrusted.html?label=experimental&label=master&aligned