Summary:
Refactor EventEmitters to take an optional SurfaceId that Fabric will use, and non-Fabric will not.
Migrating touches is a good proof-of-concept for how this could be used generally, and as it turns out, TouchEvent's API is more flexible than most other event APIs (because it uses a dictionary to pass data around, so we can just stuff SurfaceId into it - not efficient, but flexible).
All new APIs are backwards-compatible and designed to work with old-style events, with Fabric and non-Fabric. Native Views that migrate to the new API will be backwards-compatible and get an efficiency boost in Fabric.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D26025135
fbshipit-source-id: 5b418951e9d0a3882f2d67398f2aaadac8a3a556
Summary:
Create V2 EventEmitter that surfaceId can be passed into, with a backwards-compat layer, and some debug-only logging to help assist with migration.
Changelog: [Changed][Android] RCTEventEmitter has been deprecated in favor of RCTModernEventEmitter for optimal Fabric support; RCTEventEmitter will continue to work in Fabric, but there are minor perf implications.
Reviewed By: mdvacca
Differential Revision: D26027104
fbshipit-source-id: 784ca092bbc88d19c82f6c42537c34460d96de86
Summary:
There's a field called `surfaceID` in a couple of classes that isn't the same as the integer `surfaceId` in Fabric.
For consistency, I've deprecated a couple of them, or renamed when appropriate.
In addition, now we're passing the actual integer surfaceId into the ThemedReactContext. This means that every single View created in Fabric gets annotated with the surfaceId it's in. Currently this isn't used, but the idea is that now each View has a mapping back to its surface, which could be used to simplify / optimize operations with SurfaceMountingManager. In particular, we might be able to use this in the future to optimize animations and/or event emitters.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D26021571
fbshipit-source-id: b7db7de123db07fa928a6f815be86bdbb030e62c
Summary:
This refactors MountingManager into a minimal API that shims into a more fully-featured SurfaceMountingManager. The SurfaceMountingManager keeps track of surface start/stop, surface ID, and surface Context.
This solves a number of issues around (1) race conditions around StopSurface/StartSurface, (2) memory management of Views, (3)
Concrete improvements:
1. Simpler to reason about race conditions around StopSurface/StartSurface.
2. 1:1 relationship between SurfaceId and all Views/tags.
3. When surface is stopped, all descendent Views can be GC'd immediately.
4. Fixed separation of concerns and leaky abstractions: surfaceId/rootTag and Surface Context are now stored and manipulated *only* in SurfaceMountingManager.
5. Simpler StopSurface flow: we simply remove references to all Views, and the Fragment (outside of the scope of this code) removes the RootView. This will trigger GC and we do ~0 work. Previously, we ran a REMOVE and DELETE instruction and kept track of each View in a HashMap. Now we can simply delete the map and move on.
The caveat: NativeAnimated (or other native modules that go through UIManager). APIs like `updateProps` currently uses only the ReactTag and does not store SurfaceId. This is a good argument for moving away from ReactTag, at least in its current incarnation, but: for now this requires that you do a lookup of a ReactTag across N surfaces (worst-case) to determine which Surface a ReactTag is in.
So, to summarize, the "con" of this approach is that now `getSurfaceManagerForViewEnforced` could be slower. It is used in:
* NativeAnimatedModule calls `updateProps` through UIManager
* FabricEventEmitter calls `receiveEvent` on FabricUIManager directly
* On audit, I could find zero native callsites to `sendAccessibilityEvent` through UIManager
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D26000781
fbshipit-source-id: 386ae40c4333f8c584e05818c404868dbee6ce73
Summary:
MountingManager keeps a map of tags to Views, and attempts to clean it up by (1) deleting tags when they're explicitly deleted, (2) recursively deleting all Views when the View hierarchy is torn down.
However, there appear to be.... substantial gaps here. In tests, when navigating between screen X -> screen Y -> back to X (triggering a StopSurface), each "StopSurface" resulted in 200-600 Views being leaked (!).
What is causing these leaks? Well, for one, the "dropView" mechanism isn't perfect, so it might be missed Views. Second, Views don't always guarantee that `reactTag == getId()`, so that could result in leaks. Third, View preallocation on Android complicates things: Views can be preallocated and then never even inserted into the View hierarchy, so DELETE mutations could never be issued. Fourth, StopSurface is also complicated on Android (largely because of View preallocatioAndroid (largely because of View preallocation).
So, I introduce a new mechanism: keep a list of all tags for a surface, and remove all tags for a surface when the surface is torn down. This should be fool-proof: it handles preallocation and normal creation; it can handle deletes; and we're guaranteed that tags cannot be added after a surface is stopped.
Is this overly complicating things? Well, hopefully we can simplify all of this in the longterm. But until we get rid of View Preallocation, it seems like we need this mechanism - and View Preallocation might be around for a while, or forever.
Other thoughts: it's possible that using other data-structures could be more efficient, but I'm guessing the perf implications here are marginal (compared to the insane amount of memory leaks we're fixing). It could also simplify things to have a SurfaceMountingManager interface that implies all actions happen on a specific surface, including teardown.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25985409
fbshipit-source-id: f55b533770b1630c6c2a9b7a694d953aa3324428
Summary:
Original commit changeset: 3ed8e78e31b0
Backing-out D25938851 (https://github.com/facebook/react-native/commit/69b3016171bb2f994dd4a62c34c2c4645b5a7d56) and D25935785 (https://github.com/facebook/react-native/commit/bdea479a1faa0f1f7d7c9d9162212cce94bc9720). Based on analysis documented in T83141606, I believe this issue should be fixed in JS.
Additionally, this crash actually has nothing to do with (un)flattening or the differ; it is a side-effect of stale ShadowNodes being cloned, which I believe is either UB or a contract violation. Either way, it should probably be fixed either in JS, or in node cloning. So this isn't the right solution for this issue and should be reverted.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25949569
fbshipit-source-id: 8cf1094a767da98fff4430da60d223412e029545
Summary:
This diff changes the type of the SwipeRefreshLayoutManager.size prop from Int to String in Fabric.
The current implementation of this prop allows JS developers to use "int" type when fabric is enables and "int or string" types when using Fabric is disabled.
Since long term we want to only support "string" type for this prop, I'm changing the type of the prop to be String.
After my diff Fabric will start supporting only "string" types, non fabric screens will keep supporting "int or string" values.
**Will this break production?**
No, because there are no usages of RefreshControl.Size prop in fbsource
**What about if someone start using this prop next week?**
IMO It's very unlikely because of the nature of this prop, I will be monitoring next week and if there's an usage it will be detected by flow when trying to land D25933457.
Changelog: [Android][Changed] - RefreshControl.size prop changed its type to string, the valid values are: 'default' and 'large'
Reviewed By: JoshuaGross
Differential Revision: D25933458
fbshipit-source-id: 55067d7405b063f1e8d9bb7a5fd7731f5f168960
Summary:
Android has some optimizations around view allocation and pre-allocation that, in the case of View Unflattening, can cause "Create" mutations to be skipped.
To make sure that doesn't happen, we add a flag to ShadowViewMutation (in the core) that any platform can consume, that indicates if the mutation is a "recreation" mutation.
It is still a bit unclear why this is needed, in the sense that I would expect props revision to increment if a view is unflattened. However, there is at least one documented reproduction where that is *not* the case. So for now, we'll have a hack pending further investigation.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25935785
fbshipit-source-id: 6fb4f0a6dedba0fe46ba3cd558ac1daa70f671f5
Summary:
This diff refactors the intialization of Fabric in order to avoid loading UIManagerModule as part of the creation of FabricJSIModuleProvider.
One caveat is that now we are not taking into consideration the flag mLazyViewManagersEnabled
master/xplat/js/react-native-github/ReactAndroid/src/main/java/com/facebook/react/CoreModulesPackage.java177
if (mLazyViewManagersEnabled) {
As a side effect of this diff view managers will be initialized twice if the user has fabric and paper enabled
This diff was originally backed out in D25739854 (https://github.com/facebook/react-native/commit/4984c1e525e310f15c7d89230fdb2fa8fea91f05) because it produced a couple of bugs:
https://fb.workplace.com/groups/rn.support/permalink/4917641074951135/https://fb.workplace.com/groups/rn.support/permalink/4918163014898941/
These bugs are fixed by D25667987 (https://github.com/facebook/react-native/commit/2e631471092090e743245377742166ecae1d7e26).
This diff was reverted a couple of times because of the change in the registration of eventDispatcher. That's why I'm gating that behavior change as part of the "StaticViewConfig" QE.
changelog: [internal] internal
Reviewed By: JoshuaGross
Differential Revision: D25858934
fbshipit-source-id: a632799ccac728d4efca44ee685519713b4a7cbb
Summary:
These methods can all throw exceptions that get caught and reported by JS. The logviews aren't currently very helpful; hopefully adding additional information will make batch debugging a little easier.
Changelog: [Internal]
Reviewed By: fkgozali
Differential Revision: D25870788
fbshipit-source-id: a1cab225b11a3d2868f098d4575e475ee4064e65
Summary:
This diff migrates all the lookups of EventDispatcher to not depend on UIManagerModule anymore.
This refactor is necessary because:
- Users running in Fabric / Venice should not load on the UIManagerModule class
- D25858934 will introduce a change that will break all of these callsites
In the migration I'm relying on the method UIManagerHelper.getEventDispatcherFromReactTag() that returns the correct EventDispatcher for a reactTag.
I'm planning to land this change early in the week (to catch potential errors in alpha / beta versions)
As a followup we need to deprecate and prevent developers to continue using getNativeModule(UIManagerModule.class) moving forward. That will be part of another diff
changelog: [internal] internal
Reviewed By: JoshuaGross
Differential Revision: D25858933
fbshipit-source-id: e26c99759307517b5bef483274fe0e0d71bb4c6c
Summary:
Hermes has a way to set up a callback that is invoked when a fatal error
such as Out of Memory occurs. It is a static API that should be called at
most once, so it uses `std::call_once` to achieve that.
The fatal error handler is simple, it just uses glog to log an error message
to logcat, then aborts (using `__android_log_assert`).
The reason is typically very helpful for understanding why `hermes_fatal` was called.
Changelog:
[Android][Internal] - Print a logcat message when Hermes has a fatal error
Reviewed By: mhorowitz
Differential Revision: D25792805
fbshipit-source-id: 45de70d71e9bd8eaa880526d8835b4e32aab8fe3
Summary:
This diff logs a SoftError when there is not EventDispatcher associated to UIManager
The app will crash in Debug mode, this will not affect production users
changelog: [internal]
Reviewed By: JoshuaGross
Differential Revision: D25859546
fbshipit-source-id: 8045bcd67f613ea6286f30fe6f3c66113c700b0b
Summary:
The purpose of this diff is to ensure that visibility changes are handled correctly when the value of "display" for a View changes from 'flex' to 'none'.
RNTester is nesting several Views with different kind of visibilities. When the user tap on an item there's a state update that changes the visibility styles for some of these views. Fabric does not reflect the right changes of visibility on the screen.
changelog: internal
Reviewed By: shergin
Differential Revision: D25841763
fbshipit-source-id: 769b97afb72939d346a4c6f2669ff938b35596bc
Summary:
This is the Android native implementation of sendAccessibilityEvent for Fabric.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25852418
fbshipit-source-id: cb51e667a7f673da6b9c9e539770225b02bdc902
Summary:
This diff refactors the createViewInstance method in order to ensure that viewID is set before props are updated in the view.
This is necessary because there are components that deliver events at the same time their props are set. This means that some components might not have their viewId set correctly when events are delivered.
Since viewId is used to determine if a view belongs to Fabric or Paper, there are cases when the events are not delivered to the right renderer
changelog: [internal]
Reviewed By: JoshuaGross
Differential Revision: D25667987
fbshipit-source-id: 4acfa8f80d66e9e59514354481957d7d3b571248
Summary:
This experiment has been successfully running for several weeks and show small but statsig perf improvements. Delete the old code and ship this 100% in code to simplify Fabric code.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25775668
fbshipit-source-id: d2b41dfe691775e52b1e89c2fb6790a6500e560e
Summary:
Right now we assume in the Image component that any prop changes requires a redraw of the image, even if the props set are identical.
Noop prop updates can be caused in Fabric by LayoutAnimations. This may go away in the future, but only when we have a new animations system.
I don't think most other components need to be concerned with this, and many other components already guard against unnecessary redraws. Since the image "flashes"
when it is loaded, unlike most other native components, this issue is more noticeable for images.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25727482
fbshipit-source-id: 75ffa456bddc1208900733140ce4ff19f7e2c11e
Summary:
## Description
Suppose this was the codegen declaration before:
```
rn_library(
name = "FooModule",
native_module_spec_name = "FBReactNativeSpec",
codegen_modules = True,
# ...
)
```
Previously, this would generate the following BUCK targets:
- generated_objcpp_modules-FooModuleApple
- generated_java_modules-FooModuleAndroid
- generated_java_modules-FooModule-jniAndroid
## Changes
We will now generate:
- FBReactNativeSpecApple
- FBReactNativeSpecAndroid
- FBReactNativeSpec-jniAndroid
This matches the naming scheme of the old codegen.
Changelog: [Internal]
Reviewed By: fkgozali
Differential Revision: D25680224
fbshipit-source-id: 617ac18fd915f3277f6bd98072d147f20fb193e5
Summary:
This replaces the internal core implementation of `setState` with the new `updateStateWithAutorepeat` which is now the only option.
In short, `updateStateWithAutorepeat` works as `setState` with the following features:
* The state update might be performed several times until it succeeds.
* The callback is being called on every retry with actual previous data provided (can be different on every call).
* In case of a static value is provided (simple case, not lambda, the only case on Android for now), the same *new*/provided value will be used for all state updates. In this case, the state update cannot fail.
* If a callback is provided, the update operation can be canceled via returning `nullptr` from the callback.
This diff removes all mentions of the previous state update approach from the core; some other leftovers will be removed separatly.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D25695600
fbshipit-source-id: 14b3d4bad7ee69e024a9b0b9fc018f7d58bf060c
Summary:
Generate Fabric C++ files along side TM spec files for RNTester. The combined .so then has both TM and Fabric files.
This commit also removed the checked-in JNI files.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25674313
fbshipit-source-id: 8091d5a00f42849a74cab50e8d24f4010d500e5b
Summary:
For core components, we can start using the codegen output during build time instead of the checked in files in: https://github.com/facebook/react-native/tree/master/ReactAndroid/src/main/java/com/facebook/react/viewmanagers
Note: Some files seemed to be handwritten, so this only removed those that use codegen.
Changelog: [Internal]
Reviewed By: JoshuaGross
Differential Revision: D25453157
fbshipit-source-id: f7eabddfd3fd668bef0c4aef3fddcb38c8b046a0
Summary:
This diff fixes a racecondition in registration of event listeners.
mPostEventDispatchListeners is accessed from different threads, most of the times this variable is used to executed the listeners. It is only written during initialization and turn down of the renderer.
changelog: [internal]
Reviewed By: PeteTheHeat
Differential Revision: D25667988
fbshipit-source-id: 1bf95f5193d55a737bad9403206cc3320185b8cb
Summary:
## Context
The legacy NativeModule infra implements promise methods using [async NativeModule method calls](https://fburl.com/diffusion/tpkff6vg). In the TurboModule infra, promise methods [treat as sync methods](https://fburl.com/diffusion/yde7xw71), and executed directly on the JavaScript thread. This experiment makes TurboModule promise methods async, and dispatches them to the NativeModules thread, when they're executed from JavaScript.
Reviewed By: fkgozali
Differential Revision: D25623192
fbshipit-source-id: 2b50d771c5272af3b6edf150054bb3e80cab0040
Summary:
## Context
Every time we require a NativeModule in Java, we [first try to create it with the TurboModuleManager](https://fburl.com/diffusion/3nkjwea2). In the TurboModule infra, when a NativeModule is requested, [we first create it](https://fburl.com/diffusion/d2c6iout), then [if it's not a TurboModule, we discard the newly created object](https://fburl.com/diffusion/44gjlo6y). This is extremely wasteful, especially when a NativeModule is requested frequently and periodically, like UIManagerModule.
Therefore, in D24811838 (https://github.com/facebook/react-native/commit/803a26cb003e6b790e3a1ab31beb0c95795fff0c) fkgozali launched a fix to the infra that would avoid creating the non-TurboModule object in the first place. Today, we're launching this optimization.
Reviewed By: fkgozali
Differential Revision: D25621570
fbshipit-source-id: dedba4d5ac6fcf2ec3c31e7163a6a226065c708b
Summary:
This diff fixes a race condition in the execution of FabricUIManager.StartSurface method.
The rootcause is that startSurface is executing getViewportOffset from a background thread.
changelog: [internal]
Reviewed By: shergin
Differential Revision: D25617154
fbshipit-source-id: 9351201088164e74bb0b9454e30651e1de0da912
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/30595
Changelog: [Internal]
Add support for loading HBC bundles from Metro in Twilight.
* adds `runtimeBytecodeVersion` to the bundleURL
* adds logic to check chunk hermes bytecode bundles
* adds support for running the bytecode bundles
Reviewed By: cpojer
Differential Revision: D24966992
fbshipit-source-id: acdd03a2e9e2b3e4c29c99c35a7c9136a3a7ef01
Summary:
Ship LayoutAnimations to 100% of users by removing feature-flag gating.
The `collapseDeleteCreateMountingInstructions_` stuff is always disabled for LayoutAnimations, so we can get rid of that too.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25510740
fbshipit-source-id: 71bac44f829530458e4906ecd1e7e68e766de2ec
Summary:
In preparation for flipping the default, marking autoglob as False in places where it isn't explicitly specified.
Changelog: [Internal]
Reviewed By: strulovich
Differential Revision: D25497305
fbshipit-source-id: 142e5caca2d67efcb3c25067a36934f7f6dd4b21
Joshua Gross
David Vacca
generatedunixname89002005325676
generatedunixname89002005325674
Ramanpreet Nara
Riley Dulin
Valentin Shergin
Igor Klemenski
Jayme Deffenbaugh
Kacie Bawiec
Kevin Gozali
Ron Edelstein
Keion Anvaripour
