Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43985
This is just personal preference.
The name "OnJsError" makes the intent of the abstraction clear: an instance of OnJsError is a function that gets called when a js error is caught.
The name "JsErrorHandlingFunc" is not as good.
Changelog: [General][Breaking] - JsErrorHandler: Rename JsErrorHandlingFunc to OnJsError
Reviewed By: christophpurrer
Differential Revision: D55563580
fbshipit-source-id: 4d20bc984e6633aeac6193b9276a88d76961df2c
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43957
Right now, JsErrorHandler is only used to handle fatal exceptions.
So, let's just scope handleJsError down to handleFatalError.
Changelog: [General][Breaking] - JsErrorHandler: Rename handleJsError to handleFatalError
Reviewed By: cortinico
Differential Revision: D55547901
fbshipit-source-id: 261e0c8fea2852bc95e53c688d90d012d4abea34
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43956
I think we should try to centralize all things js error handling related inside JsErrorHandler. So, I moved this bool into JsErrorHandler.
This makes ReactInstance easier to understand: it removes one member variable from ReactInstance.
Changelog: [Internal]
Reviewed By: cipolleschi
Differential Revision: D55547897
fbshipit-source-id: 73d1e0eedf3896c42cda4ce1013863960585da2c
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43955
Just makes it easier to pass around JsErrorHandler.
We'll need this in D55547897, when we start storing the "has fataled" boolean inside the JsErrorHandler.
Changelog: [internal]
Reviewed By: cipolleschi
Differential Revision: D55547898
fbshipit-source-id: 162faaeff43bada0301de29111b2c17f7ef878c6
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43952
getRuntimeScheduler() allows things to schedule work on the js thread by bypassing main bundle buffering.
This is unsafe: almost everything should be using the buffered runtime executor, unless it sets up bindings used in the main bundle.
I filed a task for the investigation to see if there's any problems. And added it to the code in this diff.
Changelog: [Internal]
Reviewed By: cipolleschi
Differential Revision: D55547899
fbshipit-source-id: 7785b9777e93f36ea0278993332662ed45a20cf2
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43967
Following up https://github.com/facebook/react-native/issues/43943, the metro loading banner is presented twice in Bridgeless mode.
This happens because both the RCTInstance and the RCTHost are listening to the Reload Command and issuing the instructions to refetch the JSBundle and to present the banner.
The RCTInstance should not concern itself with lifecycle events, owned by the RCTHost.
## Changelog:
[iOS][Fixed] - Avoid to show Metro Loading banner twice.
Reviewed By: cortinico
Differential Revision: D55870640
fbshipit-source-id: addb67d3226f7d7db20736309172a42fc15f3aa3
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43920
React native is shipped as a whole, so it makes no sense for individual pods to specify which version of boost they support.
With this change we let the `react_native_pods` and the `boost.podspec` file to decide which version of boost is supported and all the other podspecs will follow.
## Changelog:
[Internal] - Remove explicit boost version from other podspecs
Reviewed By: NickGerleman
Differential Revision: D55801708
fbshipit-source-id: 3dcbbfb25010d2ee615afc4acfd5232fdc0c2a14
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43951
## Context
The **early js** error reporting pipeline catches javascript exceptions.
After errors are caught, the pipeline parses the exception, and puts the data into into a map buffer. 🤨
## Problems
We don't need to use a mapbuffer here: The structure of this exception data is known and never changes. (A map buffer is a type-unsafe bag of key/value pairs).
Instead, we could just use lower-level type-safe language primitives: regular C++ struct, and java class w/ fbjni.
## Changes
Migrate the **early js** error handling infra to C++ structs/fbjni.
## Impact
Now, there is no mapbuffer usage on iOS. We could re-introduce it when there is a need.
Changelog: [Internal]
Reviewed By: fkgozali
Differential Revision: D55265170
fbshipit-source-id: cda97633d4c6ccaad541e5d416067390fe6f61b2
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43913
changelog: [internal]
We have this turned on in ParagraphShadowNode, let's roll it out for TextInput as well.
I came across this while profiling scroll performance when TextInput is part of the view hierarchy.
Reviewed By: javache
Differential Revision: D55751341
fbshipit-source-id: 2af20ddb5a4fb9b0ccd33217e60e8b9e8a95b920
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43859
This frees up the `reactnative` CMake target so we could use it as single .so
for the CMake build.
Changelog:
[Internal] [Changed] - Rename reactnative.a -> react_cxxreact.a
Reviewed By: javache
Differential Revision: D55745640
fbshipit-source-id: 3cad512cc07a277af2a0cea696863c85a17dabc1
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43824
Changelog: [Internal]
# Feature flag cleanup/migration
This migration is part of the new Feature Flag system built by rubennorte. The goal of this migration is to clean up our Feature Flags and consolidate them into a single place, accessible by all platforms.
# In this diff
Replaced `RCTSetUseNativeViewConfigsInBridgelessMode` and `RCTGetUseNativeViewConfigsInBridgelessMode` with `ReactNativeFeatureFlags.useNativeViewConfigsInBridgelessMode()` and ReactNativeFeatureFlags::override
Reviewed By: javache
Differential Revision: D55705805
fbshipit-source-id: 861675a1a94da0fcef8d8a02ccbd8ecdd97ec700
Summary:
With 0.74.rc-5 one bug which related to errors was fixed (https://github.com/facebook/react-native/issues/41950). However, the fix introduced another one: the shape of Error objects that come from native modules has changed. This PR attempts to fix that, though it's not (yet) doing it in a way that would be 100% compatible with the old arch.
The problem was observed on iOS, not sure what the situation is on Android but believe it's okay there.
edit: on Android, there are no issues but the `message` field is enumerable, so that part is different from ios (see logs below).
Consider this code, where `error` is produced from a promise rejection inside of a native module.
```ts
console.log(
'own properties: ',
JSON.stringify(Object.getOwnPropertyNames(error), null, 2),
);
console.log(
'own enumerable properties: ',
JSON.stringify(Object.entries(error), null, 2),
);
```
These are the results for
<details>
<summary>Old architecture</summary>
```
LOG Running "google-one-tap-example" with {"rootTag":1,"initialProps":{}}
LOG own properties: [
"stack",
"code",
"message",
"domain",
"userInfo",
"nativeStackIOS"
]
LOG own enumerable properties: [
[
"code",
"-5"
],
[
"message",
"RNGoogleSignIn: The user canceled the sign in request., Error Domain=com.google.GIDSignIn Code=-5 \"The user canceled the sign-in flow.\" UserInfo={NSLocalizedDescription=The user canceled the sign-in flow.}"
],
[
"domain",
"com.google.GIDSignIn"
],
[
"userInfo",
{
"NSLocalizedDescription": "The user canceled the sign-in flow."
}
],
[
"nativeStackIOS",
[
"0 ReactTestApp 0x0000000102f4a6d8 RCTJSErrorFromCodeMessageAndNSError + 112",
"1 ReactTestApp 0x0000000102eeedd0 __41-[RCTModuleMethod processMethodSignature]_block_invoke_2.73 + 152",
"2 ReactTestApp 0x0000000102e2ae24 +[RNGoogleSignin rejectWithSigninError:withRejector:] + 548",
"3 ReactTestApp 0x0000000102e2aa8c -[RNGoogleSignin handleCompletion:serverAuthCode:withError:withResolver:withRejector:fromCallsite:] + 184",
"4 ReactTestApp 0x0000000102e2a8e0 -[RNGoogleSignin handleCompletion:withError:withResolver:withRejector:fromCallsite:] + 236",
"5 ReactTestApp 0x0000000102e28628 __40-[RNGoogleSignin signIn:resolve:reject:]_block_invoke_2 + 100",
"6 ReactTestApp 0x0000000102dc9d80 __35-[GIDSignIn addCompletionCallback:]_block_invoke_2 + 132",
...
]
]
]
```
</details>
<details>
<summary>RN 74 rc-5 (with bridgeless on)</summary>
```
(NOBRIDGE) LOG Bridgeless mode is enabled
(NOBRIDGE) LOG Running "google-one-tap-example" with {"rootTag":1,"initialProps":{"concurrentRoot":true},"fabric":true}
(NOBRIDGE) LOG own properties: [
"stack",
"message",
"cause"
]
(NOBRIDGE) LOG own enumerable properties: [
[
"cause",
{
"code": "-5",
"message": "RNGoogleSignIn: The user canceled the sign in request., Error Domain=com.google.GIDSignIn Code=-5 \"The user canceled the sign-in flow.\" UserInfo={NSLocalizedDescription=The user canceled the sign-in flow.}",
"nativeStackIOS": [
"0 ReactTestApp 0x00000001023a7b38 RCTJSErrorFromCodeMessageAndNSError + 112",
"1 ReactTestApp 0x00000001026cf774 ___ZZN8facebook5react15ObjCTurboModule13createPromiseERNS_3jsi7RuntimeENSt3__112basic_stringIcNS5_11char_traitsIcEENS5_9allocatorIcEEEEU13block_pointerFvU13block_pointerFvP11objc_objectEU13block_pointerFvP8NSStringSH_P7NSErrorEEENK3$_0clES4_RKNS2_5ValueEPSQ_m_block_invoke.57 + 332",
"2 ReactTestApp 0x0000000102270958 +[RNGoogleSignin rejectWithSigninError:withRejector:] + 548",
"3 ReactTestApp 0x00000001022705c0 -[RNGoogleSignin handleCompletion:serverAuthCode:withError:withResolver:withRejector:fromCallsite:] + 184",
"4 ReactTestApp 0x0000000102270414 -[RNGoogleSignin handleCompletion:withError:withResolver:withRejector:fromCallsite:] + 236",
"5 ReactTestApp 0x000000010226e15c __40-[RNGoogleSignin signIn:resolve:reject:]_block_invoke_2 + 100",
"6 ReactTestApp 0x000000010220f328 __35-[GIDSignIn addCompletionCallback:]_block_invoke_2 + 132",
...
],
"domain": "com.google.GIDSignIn",
"userInfo": {
"NSLocalizedDescription": "The user canceled the sign-in flow."
}
}
]
]
```
</details>
<details>
<summary>with the diff from this PR</summary>
```
(NOBRIDGE) LOG own properties: [
"stack",
"message",
"code",
"nativeStackIOS",
"domain",
"userInfo"
]
(NOBRIDGE) LOG own enumerable properties: [
[
"code",
"-5"
],
[
"nativeStackIOS",
[
"0 ReactTestApp 0x000000010083b8f8 RCTJSErrorFromCodeMessageAndNSError + 112",
"1 ReactTestApp 0x0000000100b63534 ___ZZN8facebook5react15ObjCTurboModule13createPromiseERNS_3jsi7RuntimeENSt3__112basic_stringIcNS5_11char_traitsIcEENS5_9allocatorIcEEEEU13block_pointerFvU13block_pointerFvP11objc_objectEU13block_pointerFvP8NSStringSH_P7NSErrorEEENK3$_0clES4_RKNS2_5ValueEPSQ_m_block_invoke.57 + 332",
"2 ReactTestApp 0x0000000100704718 +[RNGoogleSignin rejectWithSigninError:withRejector:] + 548",
"3 ReactTestApp 0x0000000100704380 -[RNGoogleSignin handleCompletion:serverAuthCode:withError:withResolver:withRejector:fromCallsite:] + 184",
"4 ReactTestApp 0x00000001007041d4 -[RNGoogleSignin handleCompletion:withError:withResolver:withRejector:fromCallsite:] + 236",
"5 ReactTestApp 0x0000000100701f1c __40-[RNGoogleSignin signIn:resolve:reject:]_block_invoke_2 + 100",
"6 ReactTestApp 0x00000001006a30e8 __35-[GIDSignIn addCompletionCallback:]_block_invoke_2 + 132",
...
]
],
[
"domain",
"com.google.GIDSignIn"
],
[
"userInfo",
{
"NSLocalizedDescription": "The user canceled the sign-in flow."
}
]
]
```
</details>
You see there is a change compared to old arch because `message` is no longer own enumerable property. If that needs to change (I guess it should), it'd be nice if someone more familiar with JSI pointed me in the right direction. Even with this inconsistency, the PR is an improvement and would be nice to have this fix included in the next RC.
This is output from Chrome's console for completeness, just to have something to compare to:
```
let err = new Error('hello')
undefined
Object.getOwnPropertyNames(err)
> ['stack', 'message']
Object.entries(err)
> []
```
bypass-github-export-checks
## Changelog:
<!-- Help reviewers and the release process by writing your own changelog entry.
Pick one each for the category and type tags:
[IOS] [FIXED] - add missing fields to native errors in new arch
For more details, see:
https://reactnative.dev/contributing/changelogs-in-pull-requests
Pull Request resolved: https://github.com/facebook/react-native/pull/43649
Test Plan: Tested locally with an example app running RN 74-rc 5
Reviewed By: cortinico
Differential Revision: D55690184
Pulled By: cipolleschi
fbshipit-source-id: 60a857b9871af888dcd526782b5e6b73c07c051a
Summary:
Working with gabrieldonadel, we realized that static frameworks of the React-RendererRuntime are not following the proper folder structure.
When a user tries to import `ReactCommon/RCTHost` in the app delegate, for example, the user ends up with an error and they can't find the files.
These changes fixes this by establishing the right folder structure in the static frameworks
## Changelog:
[Internal] - Make sure that React-RuntimeCore and JSErrorHandler are created with the proper structure for static frameworks
Pull Request resolved: https://github.com/facebook/react-native/pull/43846
Test Plan:
Tested locally on an app with 0.74.
Before: it failed to build.
After: it build successfully.
Reviewed By: cortinico
Differential Revision: D55741581
Pulled By: cipolleschi
fbshipit-source-id: 11ac0882d3feea05ef8904d55856ba5704b7a3b8
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43777
It's best to avoid mutex reentrancy, even when using "read" locks (i.e., std::shared_lock). Since we already have a read lock on the ShadowTree in `UIManager::setNativeProps_DEPRECATED`, we can avoid the reentrancy by grabbing ancestor node from the shadow tree instead of relying on a recursive call to ShadowTreeRegistry::visit
## Changelog
[GENERAL][FIXED] - Avoid ShadowTreeRegistry::mutex_ read lock reentrancy
Reviewed By: rubennorte
Differential Revision: D55640201
fbshipit-source-id: 7a5c6674d290ea280ab584ae734733f12a65f6f8
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43701
Changelog: [internal]
We were defining an `UNDEFINED` type of entry that should never happen in practice and we were using unnecessarily to signal "no type" where an optional type would be more suitable. Most importantly, **we were incorrectly allocating a buffer for entries of this type**.
This removes that type and the unnecessary buffer.
Reviewed By: rshest
Differential Revision: D55478890
fbshipit-source-id: 145210a9c4e2614a342f2d913b9eb6b3d62f676f
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43702
Changelog: [internal]
Just a minor refactor to use C++20 designated initializers in `PerformanceEntryReporter` and its tests, while removing unnecessary initialization for optional fields.
Reviewed By: rshest
Differential Revision: D55477745
fbshipit-source-id: a643adf7ae48df23c5c383420fd4c4dd550e1322
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43703
Changelog: [internal]
(internal because this API isn't available in OSS yet)
I found a bug in the current implementation of `performance.measure` where the API would use the first `mark` reported under a specific name instead of the last one (found it in the new example in RNTester in D55477746 that re-logs the marks every time we click on a button).
The root cause for this problem is that we were using `insert` from `std::unordered_set` to update the value, but `insert` doesn't modify the value if it's already present.
This fixes the issue by doing a lookup and removing the value prior to inserting it.
Reviewed By: rshest
Differential Revision: D55477743
fbshipit-source-id: e72aa784a936828db64b572988fe0acb2ad78214
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43752
Gentest tests started failing because Chrome changed behavior of overflowed align-content container. Spec says should fallback to "safe center", which is really just "start", instead of previous "center" behavior. This changes behavior accordingly.
There is one bit where I think we are doing the wrong thing wrt alignment of flex start vs start (which we don't support yet), but couldn't repro a failing chrome test.
Changelog: [Internal]
Reviewed By: joevilches
Differential Revision: D55617689
fbshipit-source-id: 08f23d198c75f2c2f51ccaa8795289e6e4a92cb8
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43664
Changelog: [internal]
Now that we have the event loop, we can modify the implementation of `MutationObserver` (which is still not enabled by default) to dispatch the notifications as microtasks, making the API more spec-compliant.
Reviewed By: javache
Differential Revision: D55380178
fbshipit-source-id: f876ffba49f9744f6603053f1485e7c2f43cb230
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43672
Changelog: [internal]
Small cleanup to move the last C++ native module defined in a JS directory to the new directory in `react-native/ReactCommon/react/nativemodule`.
Reviewed By: javache
Differential Revision: D55384106
fbshipit-source-id: 3bf477c2aceab6838f7f8131174b6eb74e890a23
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43663
Changelog: [internal]
We have a new directory for built-in C++ native modules, but the native modules for `MutationObserver` and `IntersectionObserver` were created before we had it.
This moves the native module for `MutationObserver` to `react-native/ReactCommon/react/nativemodule/mutationobserver` to follow the convention.
Reviewed By: javache
Differential Revision: D55380179
fbshipit-source-id: 0c64acbec973f2e5b57a0e38a0992bba49a01a45
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43653
Changelog: [internal]
This will allow us to clean up some code in `UIManager` (using methods in the native module instead) and prepare to use the DOM APIs in OSS behind a feature flag.
This doesn't enable the DOM APIs in OSS, only the native module.
Reviewed By: javache
Differential Revision: D55365252
fbshipit-source-id: 70ec0eb022df586ad554c5b8ce6915b8ceddef5f
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43659
Changelog: [internal]
This adds an implementation for the legacy layout measurement methods in React Native (`measure`, `measureInWindow` and `measureLayout`) in the DOM native module, so we can clean up the API from the `nativeFabricUIManager` binding.
Reviewed By: javache
Differential Revision: D55368141
fbshipit-source-id: 196d4d29be3b78ffc22fdc136be6e0cf5ab9dd26
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43679
The current implementation of clear-with-predicate first copies unconsumed
elements and then all others. This works correctly when the buffer is full
(wraps around), but fails if size() < maxSize: add() may no longer insert
an element in the correct position (after the last unconsumed entry; see
new unit test).
Replace it with a loop that iterates over all entries in order, and adjusts
cursorStart and cursorEnd to point to the last numToConsume elements of the
vector.
Changelog: [Internal]
Reviewed By: rshest
Differential Revision: D55273402
fbshipit-source-id: 647dc35faeb35c7fa99b8113cf85ce7f02f073e5
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43538
The Hermes RuntimeConfig for bridgeless accidentally force-disabled ES6Proxy, resulting in https://github.com/facebook/react-native/issues/43523
Let's remove the incorrect override.
To test using RNTester, add the following change:
```
diff --git a/packages/rn-tester/js/RNTesterAppShared.js b/packages/rn-tester/js/RNTesterAppShared.js
index 87cb6b69dfe..f2512d09c5a 100644
--- a/packages/rn-tester/js/RNTesterAppShared.js
+++ b/packages/rn-tester/js/RNTesterAppShared.js
@@ -50,6 +50,8 @@ const RNTesterApp = ({
);
const colorScheme = useColorScheme();
+ new Proxy({}, {});
+
const {
activeModuleKey,
activeModuleTitle,
```
Before this change, RNTester will get an error at start-up. After, the app loads correctly.
Changelog: [General][Fixed] Correctly keep ES6Proxy for bridgeless mode
Reviewed By: cortinico
Differential Revision: D55045780
fbshipit-source-id: 666b99712d35622f87d42f22a4611851df67d905
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43251
Integrates the modern CDP backend with `CatalystInstanceImpl` (the React Native instance implementation) on Android.
This complete the modern CDP integration for Bridge.
Changelog: [Internal]
Reviewed By: motiz88
Differential Revision: D51458010
fbshipit-source-id: 6f73868da9d0d4cc5d086a4569c78444cb1b83ec
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43646
PerformanceEntryReporter maintains a buffer of RawPerformanceEntry objects,
as well as a set of _pointers_ to elements within that buffer, used to find
entries by name.
However, those pointers aren't stable: BoundedConsumableBuffer internally uses
a vector, and there are a few cases where existing references get invalidated:
- When the vector's capacity changes (as new entries get inserted) [1]
- After calling clear-with-predicate, which copies elements into a new vector
This causes nameLookup to contain dangling pointers, and subsequent operations
on it can result in use-after-free.
Fix this by having BoundedConsumableBuffer reserve space for maxSize entries
up front (which ensures that existing pointers remain valid after adding new
elements) and by rebuilding nameLookup after clearing entries by name.
Note that reserve() causes the buffer's memory use to be higher than before in
case where the number of elements is small relative to the max size. Given the
(only) existing usage in PerformanceEntryReporter, as well as the property that
consumed elements remain in the buffer, that cost should be minor.
Changelog: [Internal]
[1] https://en.cppreference.com/w/cpp/container/vector/push_back
Reviewed By: rshest
Differential Revision: D55273403
fbshipit-source-id: c8f33203ae32685e29afa7f8e33edf1284d66e0f
Summary:
Small typo I encountered while trying to build custom C++ type converters :)
## Changelog:
<!-- Help reviewers and the release process by writing your own changelog entry.
Pick one each for the category and type tags:
For more details, see:
https://reactnative.dev/contributing/changelogs-in-pull-requests
-->
[INTERNAL] [FIXED] Fixed a small typo in the "unsupported type" error message
Pull Request resolved: https://github.com/facebook/react-native/pull/43650
Reviewed By: zeyap
Differential Revision: D55364068
Pulled By: cortinico
fbshipit-source-id: 5a1bc9443c82f2473860f379c9ae063cd6e3ceb4
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43655
## Changelog:
[Internal] -
The corresponding prop `ScrollBar.persistentScrollIndicator` was only passed to the Android platform code an `ScrollBar.horizontal` wasn't passed to native at all.
On other platforms we need those props to be available on the C++ side, so this exposes them to the corresponding C++ ScrollViewProps.
Reviewed By: sammy-SC
Differential Revision: D55367445
fbshipit-source-id: e8abca3a2b56a8e7c03593a6c4297f90749ac8fd
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43652
This improves the signature of our existing DOM APIs in 2 ways:
1. It replaces the use of `tuples` in `DOM.{h,cpp}` with safer structs.
2. It removes some unnecessary optionals from the API, returning the default values from the C++ API directly when appropriate.
It still preserves the use of tuples in the native module because objects are not properly supported in the codegen.
Changelog: [internal]
Reviewed By: NickGerleman
Differential Revision: D55316654
fbshipit-source-id: 16ce5ef62ca427cdcd6b9757d77db040e0ccc8b1
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43615
This class is used for a one-time setup of bindings into the jsi::Runtime. The lambda will retain ownership of any java components required, and we can rely on the teardown of the JS runtime to clean up any dependencies.
Changelog: [Internal]
Reviewed By: dmytrorykun
Differential Revision: D55241233
fbshipit-source-id: f7541f28277307be9b3a5f4f780c7eca1a467c57
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43618
Both common and Android implementations of AsyncEventBeat use weak_ptrs
to determine if the object is still valid before invoking the callback.
Move the write to `isBeatCallbackScheduled_` down so it's protected by
that same check.
Changelog: [Internal]
Reviewed By: javache, NickGerleman
Differential Revision: D55226529
fbshipit-source-id: 9e2a34369346d11dcea69d120dfa5935320f9ba1
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43578
Changelog: [internal]
We have a feature to do some validation of the mount operations when committing new trees in Fabric. That's very slow but it was ok before because we were only doing this in debug mode and in the JS thread. We're moving some of this work to the UI thread instead and we're seeing an impact on scroll performance.
This disables this feature by default but leaves it in code to enable it when necessary for debugging.
Reviewed By: NickGerleman, sammy-SC
Differential Revision: D55138795
fbshipit-source-id: 45ca47ae2562cecb070691bf33d95c9108a9eca3
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43580
Changelog: [internal]
This introduces a new feature flag to commit state updates synchronously from the UI thread (generally) instead of dispatching them to the JS thread to be processed there. We can do this now because we introduced a UI consistency mechanism in D55024832 to JS would see a consistent revision during the execution of a specific task.
Reviewed By: sammy-SC
Differential Revision: D55083029
fbshipit-source-id: 8aa84ddaee383f098252fa679cfb07012ba29bf8
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43579
Changelog: [internal]
This updates all the legacy layout APIs in React Native to make use of the UI consistency mechanism introduced in Fabric (if available, otherwise the behavior is the same we have now — we consume the latest version of the tree available).
Reviewed By: sammy-SC
Differential Revision: D55077309
fbshipit-source-id: f6ff2a4f6cd1a2040f0cfeffb1e68f4ec0240f91
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/43577
Changelog: [Internal]
This updates all the DOM APIs in React Native (defined in the `NativeDOM` native module) to make use of the UI consistency mechanism introduced in Fabric (if available, otherwise the behavior is the same we have now — we consume the latest version of the tree available).
As part of that work, this creates a new `DOMMethods` class that implements all methods using purely C++ APIs (not JSI), moves a lot of logic that was specific for this functionality from `UIManager` to `DOMMethods` and modifies `NativeDOM` to use this new class.
This should make it easier to add unit tests for `DOMMethods` in the future.
Reviewed By: NickGerleman
Differential Revision: D55077311
fbshipit-source-id: f96bf5f3a97236fd24dbd2315256de4ce979e151
Ramanpreet Nara
Riccardo Cipolleschi
Christoph Purrer
Samuel Susla
zhongwuzw
Nicola Corti
Soe Lynn
Andrew Datsenko
Vojtech Novak
Riccardo Cipolleschi
Alex Hunt
Eric Rozell
Rubén Norte
Nick Gerleman
TatianaKapos
Tomislav Novak
Kevin Gozali
Marc Rousavy
Ruslan Shestopalyuk
Pieter De Baets