Changelog

New in 3.11

v3.11 focuses on transition timing and runtime efficiency.

Screen lifecycle state now stays synchronized with the motion that is actually visible, transformed parent layouts no longer distort destination measurements, and live handoff behaves more reliably across longer navigation flows.

Synchronized transition lifecycle

entering, closing, animating, settled, and willAnimate now describe one coherent transition lifecycle across programmatic navigation and gestures.

willAnimate is a one-frame signal emitted before a new idle gesture or programmatic transition begins. It is intended for work that must be ready before the first animated frame, such as refreshing a Boundary source measurement.

1screenStyleInterpolator: ({ current }) => {
2  if (current.willAnimate) {
3    // Prepare transition-owned state before motion starts.
4  }
5
6  return {
7    content: {
8      style: {
9        opacity: current.progress,
10      },
11    },
12  };
13}

Re-grabbing a gesture while its values are still settling does not create another lifecycle start. Closing state is committed before its animation-start pulse, and interrupted entrances continue toward the correct active target.

gesture.initiator exposes the accepted gesture identity for the current attempt. Prefer it over gesture.active, which is now deprecated.

More accurate Boundary measurement

Boundary measurement no longer relies on the screen origin provider. The origin provider could over-correct a Boundary inside a parent layout with a transform such as translateY, causing the destination to be measured from an offset position.

Destinations now initially render without transition styles so Bounds can measure their actual layout. Once the required source and destination measurements are ready, the library applies the transition styles and starts the animation.

This keeps transformed parent layouts from shifting the measured destination and does not require any configuration changes.

Receiver-driven live handoff

Live handoff now treats every matching Boundary as a potential receiver rather than assigning permanent ownership of the payload to the first screen.

The active visual route selects the receiver during pushes, pops, interrupted navigation, and longer A → B → C chains. Closing transitions keep the payload attached to the closing receiver until the visual close completes, then return it directly to the previous valid receiver.

1<Transition.Boundary id="video" handoff>
2  <Transition.Boundary.Target style={styles.video} />
3</Transition.Boundary>

The public handoff contract is unchanged: every receiving Boundary uses the same id and opts into handoff.

On Android, native-backed payloads still depend on their surface implementation. For expo-video, surfaceType="surfaceView" avoids texture-composition flashes that can otherwise resemble a handoff timing failure.

Upgrade notes

The deprecated sharedBoundTag and remeasureOnFocus props have been removed. Use Transition.Boundary and its id instead:

1<Transition.Boundary id="profile-avatar">
2  <Transition.Boundary.Target>
3    <Avatar />
4  </Transition.Boundary.Target>
5</Transition.Boundary>

The deprecated shared-image preset argument has likewise been renamed from sharedBoundTag to id.

For lifecycle-driven custom code:

• use willAnimate for one-frame pre-animation work
• use closing for committed dismissal state
• use animating or settled for public motion state
• replace gesture.active with gesture.initiator when you need to know which gesture triggered the recognizer
• replace gesture.dismissing with closing
• replace gesture.settling with animating or settled

For the full APIs, see Custom Animations, Shared Elements, and Inactive Screen Behavior.