Summary:
Changelog: [Internal]
This diff add types to some of the common globals so uses of
them through `global` are now typed.
All the globals are marked as read-only for their intented uses.
However, some of them do have write cites (mostly are in tests to
deliberately set up a special test environment). Those write cites
are considered as "necessary evil" and annotated as `FlowFixMe`.
Reviewed By: yungsters
Differential Revision: D30158145
fbshipit-source-id: 93a99063361a4b7a1e33d9fc97a661be30a4d8f9
Summary:
Changelog: [internal]
Make sure the order of call is preserved for `NativeAnimatedModule`. The order of calls to NativeAnimatedModule needs to be preserved because its internals depend on it. For example, if you `getValue` is called before `createAnimatedNode`, it causes a crash.
To resolve it, we need to enqueue `getValue` onto operationQueue.
Reviewed By: JoshuaGross
Differential Revision: D30035911
fbshipit-source-id: bbd698a96cada5d2b1312a1a689ca99b04a07cdc
Summary:
Add animation queuing back into createAnimatedComponent_EXPERIMENTAL.js, which is a concurrent-safe version of createAnimatedComponent.
T93269035 for details on why this is needed.
# How does this work?
In the old createAnimatedComponent, Animations were queued by calling `setWaitingForIdentifier` before render, and then calling `unsetWaitingForIdentifier` after render.
In this diff, instead we are calling `setWaitingForIdentifier` in an `useLayoutEffect` before calling `useAnimatedProps`, and we are calling `unsetWaitingForIdentifier` in a `useEffect` after `useAnimatedProps`. So the ordering for the effects are:
1. `useLayoutEffect` with `setWaiting`
2. `useLayoutEffect`s in `useAnimatedProps`
3. `useEffect`s in `useAnimatedProps`
4. `useEffect` with `unsetWaiting`.
There's a React guarantee that **if one effect is called, all of them will be called**, so we don't have a concern about the queue getting locked.
## **Main concerns:**
1. This works in my test cases, but it's not the same behavior as the old createAnimatedComponent (which is wait before and unset wait after render). This may still be ok because the relevant side effects in render from that component have been moved to `useLayoutEffect` or `useEffect` in `useAnimatedProps` (so the ordering is still the same?).
2. I'm not sure about the ordering of `onLayoutEffect`, `onLayout` callbacks, and `useEffect`. createAnimatedComponent_EXPERIMENTAL doesn't use `onLayout`, but with this new method of queuing, **`onLayout` calls will likely be called before the animation queue has been flushed**. It's not clear to me whether this is bad.
Changelog: [Internal]
Reviewed By: yungsters
Differential Revision: D29467458
fbshipit-source-id: 2be23a8968404526d0fa394a7879fda8b5ffbfdc
Summary: Changelog: [Internal] - Animated.event from my limited understanding attaches Animated.Values to native views. In terms of animation, Animated.Values are already handled by AnimatedMock to skip interpolation.
Reviewed By: GijsWeterings
Differential Revision: D29241590
fbshipit-source-id: 98650883dad53d6f77dbc398ba0cac903194deca
Summary:
This pre-suppresses the 154 error diff ahead of its release, since it is large.
Changelog: [Internal]
Reviewed By: samwgoldman
Differential Revision: D29065246
fbshipit-source-id: f418041305a46df410dcbe3d9a4db81a61ac7014
Summary:
Creates a `useAnimatedProps` hook that is compatible with concurrent mode, and uses this in `createAnimatedComponent_EXPERIMENTAL`.
Changelog:
[Internal]
Reviewed By: lunaleaps
Differential Revision: D28390393
fbshipit-source-id: d8aa471507a5d096832b93caf63a24f71733deda
Summary:
Creates an experimental stub implementation of `createAnimatedComponent` to be implemented in future commits.
Changelog:
[Internal]
Reviewed By: lunaleaps, kacieb
Differential Revision: D28799738
fbshipit-source-id: dc3fbee557db353de6807bd87561f8f372d7cab5
Summary:
Sets up an injection mechanism for experimenting in production with an alternate implementation of `createAnimatedComponent`.
This will be used to implement and refine a new `createAnimatedComponent` that is compatible with concurrent rendering.
Changelog:
[Internal]
Reviewed By: lunaleaps
Differential Revision: D28799739
fbshipit-source-id: 46ba2dd6137f7bf73ce8a659698533ed3985516f
Summary:
This pre-suppresses the 153 error diff ahead of its release, since it is large.
Changelog: [Internal]
Reviewed By: mroch
Differential Revision: D28754374
fbshipit-source-id: 1806f53bc7d804644d434583a2dcd6da63d00328
Summary:
Fixes https://github.com/facebook/react-native/issues/28517
Animated.loop needs to reset the animation after each iteration but currently, natively driven animations are not getting reset properly.
## Changelog
<!-- Help reviewers and the release process by writing your own changelog entry. For an example, see:
https://github.com/facebook/react-native/wiki/Changelog
-->
[General] [Fixed] - Fix natively driven animations not getting reset properly
Pull Request resolved: https://github.com/facebook/react-native/pull/29585
Test Plan: Tested within RNTester
Reviewed By: kacieb
Differential Revision: D28383538
Pulled By: yungsters
fbshipit-source-id: 40790ad8f825c402afc3515ef6afc00570704109
Summary:
This check will never pass because `this._propsAnimated` contains an instance of `AnimatedProps`, but `nextProps` contains an object literal containing the new props.
Changelog:
[Internal]
Reviewed By: JoshuaGross, TheSavior, kacieb
Differential Revision: D28271627
fbshipit-source-id: c563eec1eeaee5eb84bb01525313b46db502225a
Summary:
Decouples `__attach` from the constructor in `AnimatedProps`.
This change will enable the instantiation of `AnimatedProps` (and subsequent invocation of `__getValue()`) without having to trigger side effects until after mount or update. This is important in order for `Animated` to ever become safe for Concurrent Mode.
Changelog:
[Internal]
Reviewed By: lunaleaps
Differential Revision: D28271628
fbshipit-source-id: 6ccfed6de79382cecdfa6939c7dad3134e1ecaaa
Summary:
NativeAnimationModule on Android is using batches to combine related updates together. In Fabric, this process is controlled by JS, which schedules starts batching when Animated component props are about to be updated.
In rare cases when there's no prop update (like calls to `setValue` on `AnimatedValue`), changes weren't applied until the next batch was scheduled. This change ensures that such changes get applied directly.
This problem doesn't happen on iOS or non-Fabric surfaces, because they use per-frame scheduling instead.
Changelog: [Internal]
Reviewed By: JoshuaGross
Differential Revision: D28100641
fbshipit-source-id: 87b3acb3a8de131cdcb2104dc121d40eccfbc29d
Summary:
This change caused crashes in animations on some surfaces.
Changelog: [Internal]
Reviewed By: JoshuaGross
Differential Revision: D28013969
fbshipit-source-id: 95845c69d6e67d59582ea14ad08cbf42fd3e2f8f
Summary:
In 0.62, `createAnimatedComponent` was changed to use `forwardRef` instead of requiring callers to use `ref.getNode()`. In order to preserve backward-compatibility, `ref.getNode()` was monkey-patched onto the returned ref, if possible, to return the `ref` and output a console warning.
Three major (well, technically, minor) releases later, we are dropping support for `getNode()`. Calling it on the `ref` of an animated component will begin to fail after this (unless the underlying component itself actually implements a `getNode()` method on its imperative handle).
Changelog:
[General][Removed] - Removed `getNode()` from animated component refs.
Reviewed By: nadiia
Differential Revision: D27979882
fbshipit-source-id: 885c3dbf4f2749c994fc2662dd6f16ff3dd887c7
Summary:
D27682424 (https://github.com/facebook/react-native/commit/ea1ff374de2ece7d1698b14d4e1aa8075df22cdd) updated how animated node batches are executed in Fabric. On Paper, these batches were controlled by native module in some places (batch was executed ~every 2 frames), but some animations were switching animation batching control to JS globally there as well.
This change updates two things:
- If batching is controlled by native, it makes sure batches are calculated correctly.
- At the same time, this change switches control for animation node batching to JS, aligning it with Fabric.
Changelog: [Internal]
Reviewed By: JoshuaGross, mdvacca
Differential Revision: D27939659
fbshipit-source-id: d6251bce2a303a4a007dc10297edc0175cc4b348
Summary:
"The instance should not stick around after unmount..." - Tim Yung, 2021
I have a hypothesis that, if a component instance of an animated component sticks around after unmount, it could cause memory leaks due to references to Fabric ShadowNodes across the JSI (this would not impact non-Fabric... in theory).
Wild guess. If OOMs disappear then maybe this hypothesis is correct, but it's a long shot. I figure there's ~no harm in doing this cleanup here anyway.
Changelog: [Internal]
Reviewed By: sammy-SC
Differential Revision: D26650348
fbshipit-source-id: 90633db650b65755cacfb52344e7b53e46c9b125
Summary:
There's logic in Animated JS that prevents flattening of animated views in Fabric. However, we cannot actually detect Fabric vs non-Fabric in the first render pass; in the past we defaulted to assuming non-Fabric. Now we assume Fabric for View flattening purposes.
Changelog: [Internal]
Reviewed By: shergin
Differential Revision: D26647393
fbshipit-source-id: c91b51aeeb4f352cc502bc018f086e36fd1ffd85
Summary:
ES Modules implicitly enable strict mode. Adding the "use strict" directive is, therefore, not required.
This diff removes all "use strict" directives from ES modules.
Changelog:
[Internal]
Reviewed By: motiz88
Differential Revision: D26172715
fbshipit-source-id: 57957bcbb672c4c3e62b1db633cf425c1c9d6430
Summary:
This change contains the suppressions for the up coming v0.142.0 Flow release.
The new suppressions are a result the following changes:
* Disallow flowing functions or inexact objects to indexed objects to improve object soundness. This can cause errors if you are passing a function or inexact objects when an indexed object is expected.
* Flow now processes imports before checking the body of a file. In some rare cases this can expose previously skipped errors due to the processing order.
Reviewed By: mroch
Differential Revision: D25820434
fbshipit-source-id: 59cc1d852ffc8cc39f0d5112ce485fb33f05c092
Summary:
The `isFabric` method used in createAnimatedComponent is unreliable (another reason in a long list of reasons why you should not duplicate this code elsewhere, and why we want to delete it ASAP).
In particular, during the first render, the ref component has not been set yet, so we /cannot/ detect if the component is Fabric or non-Fabric and assume it's non-Fabric.
In Fabric, this causes `collapsable` and `nativeID` values to change after the first render.
To reduce this re-rendering, but not eliminate it for all components, I've introduced a flag that indicates if a component will /never/ be flattened. In particular, Image components, ScrollViews, Text cannot ever be flattened,
so we should always pass `collapsable:false` and the same nativeID prop for those components. For Animated <View>s and other components, the re-rendering issue is still a problem in Fabric for now.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25720322
fbshipit-source-id: fe3234d8ae974911a2b5f82e4f6a093216f43d4b
Summary:
In D24521951 (https://github.com/facebook/react-native/commit/2ff1d4c041ee662871a84363a3f85a8bc9e857ce) I refactored the JS-side queueing for NativeAnimated API calls, and used randomized IDs for queueing. This could cause bugs or unexpected behavior, and potentially crashes, if there's ever a collision in random numbers generated or
a collision between random number and one of the deterministic numbers generated in createAnimatedComponent.
In this diff I make both of them namespaced with a string, and deterministic, to eliminate any potential collisions. This could also be slightly faster (but not meaningfully) since we're not relying on Math.random.
Changelog: [Internal]
Reviewed By: yungsters
Differential Revision: D24553557
fbshipit-source-id: 8b765e21597ad4f8e641453c1f9f90bdf1ee022f
Summary:
This small PR fixes few "no-unused-var" issues and and removes two old entries for no longer existing files from the `.eslintignore`.
## Changelog
[Internal] [Fixed] - Lint: fix few "no-unused-var" warnings for imports
Pull Request resolved: https://github.com/facebook/react-native/pull/30157
Test Plan: Successfully run of `yarn lint` script. Warnings count has been reduced from `61` to `58`.
Reviewed By: yungsters
Differential Revision: D24288226
Pulled By: appden
fbshipit-source-id: 06e4ef015a331e3f2eac3b9aa6f757a3764e3ed9
Summary:
There's no reason for us to have lint ignores for `react-native/codegen/react-native-modules`. This diff removes all such ignores. I'll address any actual problems with the specs in subsequent diffs.
Changelog: [Internal]
Reviewed By: fkgozali
Differential Revision: D24529238
fbshipit-source-id: bbd2f4fb5dace65d803a8f93bd0d9a1c5a1cfb34
Summary:
I believe the old method of queueing these operations for Fabric is causing crashes because "connectNode" is on a separate JS queue from setting up nodes.
In hindsight, this seems silly. We must ensure that nodes are created before they're connected, and we weren't doing that?
Using a single queue is conceptually simpler, should be easier to reason about, and should fix some crashes.
Changelog: [Internal]
Reviewed By: fkgozali
Differential Revision: D24521951
fbshipit-source-id: f6c38ac0023faa28414063d8b96daf32ba95524d
Summary:
Same as D20969087 (https://github.com/facebook/react-native/commit/be7867375580ed391bb10c50b768d998087e848d) but a bit more sophisticated.
We currently see a lot of errors happens because of division by zero in AnimatedDivision module. We already have a check for that in the module but it happens during the animation tick where the context of execution is already lost and it's hard to find why exactly it happens.
Adding an additional check to the constructor should trigger an error right inside render function which should make the error actionable.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: fkgozali
Differential Revision: D23908993
fbshipit-source-id: d21be9a72ec04fe4ff0740777d9ff49cf1bcde73
Summary:
Open source this ESLint rule so that we can lint our open source NativeModule specs.
Changelog: [Internal]
Reviewed By: shergin, cpojer
Differential Revision: D23791748
fbshipit-source-id: e44444bc87eaa9dc9b7f2b3ed03151798a35e8a5
Summary:
We disabled animations on Android because native driver animations weren't working yet. After recent changes in NativeAnimatedModule, animations are now working as expected. This diff re-enables animations for bridgeless mode on Android.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D22605612
fbshipit-source-id: 7a7078bf23bb88e7745fc899e1a5ff551425dbad
Summary:
Deletes `Animated/polyfills` which is no longer necessary (and only creates extra configuration burden).
Changelog:
[Internal]
Reviewed By: cpojer
Differential Revision: D22451963
fbshipit-source-id: 7a9a48b96b3783f2f6340226bdafd2eaa43f32e7
Summary:
Removes the legacy `react-animated` package configuration and collapses the `Animated/src/` directory into `Animated/`.
Also, reconfigures all references to `Animated/src/` to just be `Animated/`.
Changelog:
[Internal]
Reviewed By: cpojer
Differential Revision: D22450848
fbshipit-source-id: 9fd4861e9f357d817d82e9fec71967a2936a3830
Summary:
This is the JS side of D23010844 (https://github.com/facebook/react-native/commit/73242b45a9f7f9ecce8f5124ac3d1665f575be74).
In the past I tried a few heuristics to guess when a batch of Animated Operations were ready, and none of these were super reliable. But it turns out we can safely allow JS to manage that explicitly.
Non-Fabric still uses the old behavior which seems fine.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D23010937
fbshipit-source-id: 4a513672c166ceaa8b6daa387e46a2a95b60ce53
Summary:
In D22801173 (https://github.com/facebook/react-native/commit/9e6ba9ddb8608d4e3a4a80d0138600130b766d4c) I introduced a mechanism to queue Animated operations on the JS side until all mounting effects were finished. I also added an unnecessary check that the NativeAnimated module is loaded, and it seems like if it's called too early, before the NM is initialized, it will crash.
It's not necessary in these two methods, so we just delete the invariant check.
Changelog: [Internal]
Reviewed By: lunaleaps
Differential Revision: D23040331
fbshipit-source-id: e70ec887695b9abefa0f01408a4a1c2749946d1c
Summary:
In the past I tried a few heuristics to guess when a batch of Animated Operations were ready, and none of these were super reliable. But it turns out we can safely allow JS to manage that explicitly.
Non-Fabric still uses the old behavior which seems fine.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D23010844
fbshipit-source-id: 4c688d3a61460118557a4971e549ec7457f3eb8f
Summary:
Unfortunately, the new codegen doesn't allow us to import types from other files. Therefore, I've inlined the interface specification of NativeAnimatedModule into NativeAnimatedTurboModule.
Changelog: [Internal]
Reviewed By: JoshuaGross
Differential Revision: D22858790
fbshipit-source-id: 759bb960240afaba6b70c2730c3359b7e8c46c83
Summary:
In Fabric, there are two subtle differences in Animated:
1) On the native side, queued NativeAnimated operations are executed more frequently
2) On the JS side there seems to be a bigger gap between AnimatedNodes being set up, and the final `componentDidMount`/`componentDidUpdate` that results in `_attachNativeEvents` being called.
This results in some visual glitching for certain features like StickyScrollViewHeader and other animations that are frequently rerendered. Rerendering an animated component relies on sync-like behavior
(new animations must be set up in the same frame that old animations are torn down, and the UI must flush any updates in the same frame). This sync-like behavior is trickier to achieve in Fabric, so
we introduce this mechanism, in Fabric only, to queued operations more aggressively on the JS side and only flush them when `componentDidMount`/`componentDidUpdate` have been called on all relevant nodes.
This seems to resolve longstanding issues with animations in Fabric on Android.
There are unrelated issues on iOS that make this difficult to test currently, so I'll keep the new path Android-only for now.
Changelog: [Internal]
Reviewed By: sammy-SC
Differential Revision: D22801173
fbshipit-source-id: 3db2235483636c9074df941cfac0a30f3f97545a
Summary:
The Animated native module cannot be converted to a TM because it has perf regressions when used in conjunction with Paper renderer. Instead of fixing these complicated perf issues (which will disappear when Fabric ships this half), temporarily fork the native module. The new fork is converted to a TM, and only used to unblock bridgeless mode testing.
Changelog:[Internal]
Reviewed By: RSNara
Differential Revision: D22651010
fbshipit-source-id: 912123ef38ac8c66025b7bba34a65ec6d98f330d
Summary:
This appears to be the only usage of whitelist/blacklist in react-native-github, let's use some more inclusive language.
Changelog: [JS][Internal]
Reviewed By: kacieb
Differential Revision: D22539431
fbshipit-source-id: 21d4cd54a5a2a676996ccec7b02ef15c421efee1
Pieter De Baets
Xuan Huang
Samuel Susla
Kacie Bawiec
Luna Wei
Daniel Sainati
Tim Yung
Tien Pham
Andrei Shikov
Rubén Norte
Nadiia D
Joshua Gross
Micha Reiser
Pieter Vanderwerff
simek
Ramanpreet Nara
Valentin Shergin
Emily Janzer
Peter Argany