PanResponder
PanResponder
reconciles several touches into a single gesture. It makes single-touch gestures resilient to extra touches, and can be used to recognize basic multi-touch gestures.
By default, PanResponder
holds an InteractionManager
handle to block long-running JS events from interrupting active gestures.
It provides a predictable wrapper of the responder handlers provided by the gesture responder system. For each handler, it provides a new gestureState
object alongside the native event object:
A native event is a synthetic touch event with the following form:
nativeEvent
changedTouches
- Array of all touch events that have changed since the last eventidentifier
- The ID of the touchlocationX
- The X position of the touch, relative to the elementlocationY
- The Y position of the touch, relative to the elementpageX
- The X position of the touch, relative to the root elementpageY
- The Y position of the touch, relative to the root elementtarget
- The node id of the element receiving the touch eventtimestamp
- A time identifier for the touch, useful for velocity calculationtouches
- Array of all current touches on the screen
A gestureState
object has the following:
stateID
- ID of the gestureState- persisted as long as there at least one touch on screenmoveX
- the latest screen coordinates of the recently-moved touchmoveY
- the latest screen coordinates of the recently-moved touchx0
- the screen coordinates of the responder granty0
- the screen coordinates of the responder grantdx
- accumulated distance of the gesture since the touch starteddy
- accumulated distance of the gesture since the touch startedvx
- current velocity of the gesturevy
- current velocity of the gesturenumberActiveTouches
- Number of touches currently on screen
Basic Usage
Working Example
To see it in action, try the PanResponder example in RNTester
Reference
Methods
create()
@param {object} config Enhanced versions of all of the responder callbacks that provide not only the typical ResponderSyntheticEvent
, but also the PanResponder
gesture state, by replacing the word Responder
with PanResponder
in each of the typical onResponder*
callbacks. For example, the config
object would look like:
onMoveShouldSetPanResponder: (e, gestureState) => {...}
onMoveShouldSetPanResponderCapture: (e, gestureState) => {...}
onStartShouldSetPanResponder: (e, gestureState) => {...}
onStartShouldSetPanResponderCapture: (e, gestureState) => {...}
onPanResponderReject: (e, gestureState) => {...}
onPanResponderGrant: (e, gestureState) => {...}
onPanResponderStart: (e, gestureState) => {...}
onPanResponderEnd: (e, gestureState) => {...}
onPanResponderRelease: (e, gestureState) => {...}
onPanResponderMove: (e, gestureState) => {...}
onPanResponderTerminate: (e, gestureState) => {...}
onPanResponderTerminationRequest: (e, gestureState) => {...}
onShouldBlockNativeResponder: (e, gestureState) => {...}
In general, for events that have capture equivalents, we update the gestureState once in the capture phase and can use it in the bubble phase as well.
Be careful with onStartShould*
callbacks. They only reflect updated gestureState
for start/end events that bubble/capture to the Node. Once the node is the responder, you can rely on every start/end event being processed by the gesture and gestureState
being updated accordingly. (numberActiveTouches) may not be totally accurate unless you are the responder.