Summary:
CI is failing to build HermesC on windows due to a version mismatch of the CMake already installed
## Changelog:
[Internal] - Fix Windows CI for HermesC
Pull Request resolved: https://github.com/facebook/react-native/pull/47867
Test Plan: GHA
Reviewed By: robhogan
Differential Revision: D66292617
Pulled By: cipolleschi
fbshipit-source-id: 5e8f4f45e33fbdd9ff163b4e8a09cb98d4366dc7
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47733
In https://github.com/facebook/yoga/pull/1729 I moved the cleanup of `display: contents` nodes to happen before all the early returns, but that change also made it be called **before** `cloneChildrenIfNeeded`. It actually needs to be called after `cloneChildrenIfNeeded` to make sure that children of `display: contents` nodes are properly owned.
It also needs to be called in every short-path, so it's being called in four places in this PR. Please let me know whether it's ok or not.
X-link: https://github.com/facebook/yoga/pull/1743
Reviewed By: NickGerleman
Differential Revision: D65953902
Pulled By: zeyap
fbshipit-source-id: 0b18a5651f19c23564f5b3aa2a50833426e9ca5f
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47732
This pull request addresses the issue where combining align-content and align-items properties resulted in incorrect layout behavior in Yoga version 3.1.0, as reported in [Issue https://github.com/facebook/yoga/issues/1739](https://github.com/facebook/yoga/issues/1739).
# Changes Made:
Alignment Logic Update: Modified the alignment calculations to ensure that the combination of align-content and align-items properties produces the expected layout, consistent with CSS Flexbox standards and previous Yoga versions.
Test Cases Added: Introduced new test cases to cover scenarios involving various combinations of align-content and align-items properties to prevent future regressions.
# Testing:
All existing tests pass successfully.
New test cases confirm that the layout behaves as expected when align-content and align-items are used together.
# Impact:
This fix ensures that layouts using both align-content and align-items properties render correctly, aligning with the behavior observed in Yoga version 1.19.0 and standard web browsers.
X-link: https://github.com/facebook/yoga/pull/1742
Reviewed By: joevilches
Differential Revision: D65953882
Pulled By: zeyap
fbshipit-source-id: 7e12a21b1d442b35c3f3536cad32dc4b82130d15
Summary:
- Current implementation does not follow CSS spec for rectangle boxes with corner angles. It is using non spec compliant algorithm to calculate start and end points. This PR follows the spec compliant algorithm to implement and makes sure Web, iOS and Android gradients are identical with corner angles.
- Also, currently it is using `CAGradientLayer` which does not support spec compliant start and end points i.e. start and end point can be outside of rectangle bounds. This leads to inconsistent gradients on iOS for corner angles compared to web and android. So this PR replaces it with `CGGradient`.
- I have also moved some files to make it easier to add more background image types in future.
## Changelog:
[GENERAL] [FIXED] - Linear gradient start and end point algorithm.
<!-- 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
Pull Request resolved: https://github.com/facebook/react-native/pull/47003
Test Plan:
- Added multiple gradient example which should be identical in all platforms (Web, iOS and Android) and tested thoroughly on all platforms. I think some visual test cases can help here.
- I have referred to [blink's](https://github.com/chromium/chromium/blob/main/third_party/blink/renderer/core/css/css_gradient_value.cc) implementation.
## Aside
Took a while to understand the [spec](https://www.w3.org/TR/css-images-3/#corner-gradient-example), but felt great after getting it. Gradients should be 100% identical on all platforms now. Sorry i missed testing cornered angles + rectangles earlier and I found out it is inconsistent on platforms just this weekend 😅
<img width="1389" alt="Screenshot 2024-10-14 at 12 24 45 AM" src="https://github.com/user-attachments/assets/2f61eb87-502b-4b8c-88f3-d8a3cca9a7a3">
Reviewed By: joevilches
Differential Revision: D64497127
Pulled By: jorge-cab
fbshipit-source-id: 2647176ae2ee74b6c71f9061465b07dccdabcfc1
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47735
We are trying to remove or minimize private header usage in YogaLayoutableShadowNode.cpp and YogaLayoutableShadowNode.h. There are a few "trivial" ones, meaning some private call maps 1:1 to an existing public API, so I just opt to use that instead
Changelog: [Internal]
Reviewed By: NickGerleman
Differential Revision: D66131948
fbshipit-source-id: 7068021513703d1c142d3c298f4804e4e8b81ee2
Summary:
This PR fixes the code for generating EventEmitter C++ code in case nested objects in arrays are used.
```typescript
export interface NativeProps extends ViewProps {
onEvent: DirectEventHandler<
Readonly<{
payloadArray: Readonly<
{
obj: Readonly<{ str: string }>
}[]
>
}>
>;
}
export default codegenNativeComponent<NativeProps>('SomeComponent');
```
In this case the generated `EventEmitters.cpp` code would contain:
```c
obj.setProperty(runtime, "str", payloadArrayValue,obj.str);
```
while
```c
obj.setProperty(runtime, "str", payloadArrayValue.obj.str);
```
is expected.
## Changelog:
<!-- Help reviewers and the release process by writing your own changelog entry.
Pick one each for the category and type tags:
[ANDROID|GENERAL|IOS|INTERNAL] [BREAKING|ADDED|CHANGED|DEPRECATED|REMOVED|FIXED|SECURITY] - Message
For more details, see:
https://reactnative.dev/contributing/changelogs-in-pull-requests
-->
[GENERAL] [FIXED] - Codegen: Support nested objects in arrays
Pull Request resolved: https://github.com/facebook/react-native/pull/47514
Test Plan: Tested with the reproduction case above to verify correct output.
Reviewed By: javache
Differential Revision: D65848670
Pulled By: elicwhite
fbshipit-source-id: 0021e87bc7213fff615465cab8554cc1a2159522
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47737
We may see an NSRangeException when setting new AttributedString content, where setting the AttributedString itself changes selection (before we mutate it later).
It seems like the selection here is not in a good state yet in regards to the AttributedString backing exposed (since we are reading it while modifying it). So let's fold the logic for updating typing attributes into the collection of ignored work from non-user-selection updates, since programatically setting an AttributedString will already trigger updating typing attributes.
I also added a nil check here, which is unrelated to the crash, but it seems like we should have it for safety...
Changelog:
[iOS][Fixed] - Fix possible NSRangeException when updating typing attributes in response to new text content
Reviewed By: cipolleschi
Differential Revision: D66202986
fbshipit-source-id: fded492b5022c5fef5b9563f93a57549d06a7020
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47738
- updating targetSdk to 35 for RN Android and helloworld
- **Note:** `tagetSdk` of RN Android does not have effect on the app's targetSdk. App can set `targetSdk` regardless of what RN targetSdk is. Updating the targetSdk just signals that RN Android has been tested with targetSdk 35 and apps can choose to support lower targetSdk as needed.
Changelog:
[Android][Changed] updating targetSdk to 35 (apps can still choose their own targetSdk regardless of RN version)
Reviewed By: NickGerleman
Differential Revision: D66209381
fbshipit-source-id: 2f26e8f605a383ff662e4b1d65611f186a7d7979
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47734
- With forced edge-to-edge on Android 15 targetSdk 35, `HermesBadge` overlaps with the top status bar.
- Move the location of the HermesBadge to align with bottom of the header so we can avoid UI overlap with minimal changes to the template.
Changelog: [Internal]
Reviewed By: cipolleschi
Differential Revision: D66183918
fbshipit-source-id: 6e9eff337660a133349a2813e9ee033a8d2b749e
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47707
NewAppScreen should be removed from public API test
Changelog: [Internal]
Reviewed By: cipolleschi, blakef
Differential Revision: D66165484
fbshipit-source-id: 2b72a874e443447a2b444921f25cec729bcb89c5
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47761
When breaking the last dependency, we created the `RCTApDependencyProvider` in the Codegen pod.
This is not an issue per sè. The problem comes in with the new template in Swift: ReactCodegen contains some headers with some C++ code that Swift can't really process.
That's why most libraries started failing in jobs like [this one](https://github.com/facebook/react-native/actions/runs/11906196751/job/33177904733): the template app was not able to load the `ReactCodegen` pod in the Swift app delegate.
Given that the app delegate only have to actually load the RCTAppDependencyProvider, I extracted that class in its own pod: ReactAppDependencyProvider.
The name of the pod does not follow the React-RCTXXX structure because that will create issues with the import statements and the `use_frameworks!` use case.
> [!NOTE]
> We need to update the template and change the `import ReactCodegen` to `import ReactAppDependencyProvider`
## Changelog:
[iOS][Added] - Extract RCTAppDependencyProvider in the ReactAppDependencyProvider pod
Reviewed By: blakef
Differential Revision: D66241941
fbshipit-source-id: 6b888109c65d9560fff322ec84a16da78fbcd64b
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47760
Changelog: [internal]
(this is internal because this API isn't enabled in OSS yet)
This fixes 2 bugs in `MutationObserver`:
1. Incorrectly reporting updates for the same node multiple times, if changes are observable by different targets in the same observer.
2. Incorrectly ignoring observations of subsequent targets in a given observer.
These were caught when migrating the unit tests for `MutationObserver` that were mocking Fabric to a Fantom integration test that uses the whole C++ infra.
Reviewed By: sammy-SC
Differential Revision: D66232571
fbshipit-source-id: b6e967ca4deaa1a69d35f14d4f921103fec2bbaf
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47759
Changelog: [internal]
These interfaces should be available in the global scope, so this exposes them in their setup modules.
Reviewed By: javache
Differential Revision: D66232574
fbshipit-source-id: 191c579ffce3fb8b4b454b5c5725661ff160a46f
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47758
Changelog: [internal]
This helps with debugging when adding logs in native code. By default they're logged to stderr and not shown in the Jest output, but we can opt into getting them with this new flag.
Reviewed By: sammy-SC
Differential Revision: D66232572
fbshipit-source-id: d507fa4438ae46988fafc92123d0954ab3d27631
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47673
Update shared monorepo build script:
- Refactor entry point discovery and file iteration
- Add support for "exports" subpath patterns (skip considering as entry point with warning)
This is required for the incoming migration of `react-native-codegen` to this build setup (D51465053).
Changelog: [Internal]
Reviewed By: robhogan
Differential Revision: D66121262
fbshipit-source-id: 0deb2c2b26de442ee348ef9de54a1fa144368b6f
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47685
Currently, we assume any URL with a hostname of `10.0.2.2` or `10.0.3.2` (device-relative) is eligible for rewriting to `localhost` (frontend-relative), because we assume the device is an Android emulator. We rewrite these URLs between device and dev machine so that the rewritten URLs are reachable from the dev machine.
This diff narrows this logic so that we'll only rewrite URLs where the hostname matches the pre-existing list *and* this matches the host the device is actually connected on, according to its headers from the original connection.
The main motivation for this change is to unblock removing assumptions about device-reachable vs server-reachable hosts. Later in the stack we'll drop the hardcoded listing of `10.0.2.2` etc in favour of identifying URLs that target the dev server, from whatever network.
There's also an edge case fix here that `10.0.2.2` etc might actually refer to a remote LAN server, and not be an Android emulator's alias for for an emulator host.
Changelog:
[General][Fixed] RN DevTools: Don't assume 10.0.2.2 is an alias for localhost unless it's used to establish a connection to the server
Reviewed By: huntie
Differential Revision: D66058704
fbshipit-source-id: bad28717b0c9b1ca43e2ea3391cef13f87892e6c
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47719
There is a subset of `ImageProps` that are used by Android but not used by iOS. They were missing in the `ImageProps.h`. It was fine when they were only consumed by the Android mounting layer. But it becomes a problem if we want to write some shared C++ code based on those props.
This diff adds those props to the corresponding C++ type.
This should have no practical effect for both platforms.
Changelog: [Internal]
Reviewed By: javache
Differential Revision: D65426569
fbshipit-source-id: dc1c9fe4a6e0e62e62f84b9b249e1c7d253290f5
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47608
This is a codemod. It was automatically generated and will be landed once it is approved and tests are passing in sandcastle.
You have been added as a reviewer by Sentinel or Butterfly.
Autodiff project: fileops2
Autodiff partition: xplat.js.react-native-github.packages.react-native.ReactCommon.cxxreact
Autodiff bookmark: ad.fileops2.xplat.js.react-native-github.packages.react-native.ReactCommon.cxxreact
This updates `open`, `close`, `read`, `write`, and `pipe` call sites to use
`folly::fileops` qualified name lookup.
This is the 2nd phase in a 3-phase change to remove folly's global definitions
of the posix functions that conflict with windows CRT.
The 1st phase created namespaces for folly's posix functions. The 2nd phase
updates callsites to use the qualified name of folly's `open`, `close`,
`read`, `write`, and `pipe` functions. The 3rd and final phase will remove
folly's globally defined posix functions and have windows CRT define them
again.
**What is the reason for this change?**
Folly's global definitions of posix functions on Windows causes `#include`
order issues if folly is not included first.
For example, when `gtest/gtest.h` is included before folly, gtest includes
`windows.h` and that declares `open`, `read`, and `chdir`, which creates
ambiguous references to folly's `open`, `read`, and `chdir`.
Another example is where posix functions go undeclared when
`folly/portability/windows.h` is included without other portability headers
(e.g., `folly/portability/unistd.h`). `folly/portability/windows.h` includes
`windows.h` in a way that only underscore versions of the posix functions are
available (e.g., `_open`, `_close`).
These issues create friction for windows development.
**Background: What is the purpose of `folly::portability::{fcntl,stdlib,sysstat,unistd}`?**
It is a portability layer to make posix functions available and behave
consistently across platforms. Some posix functions don't exist on windows
(e.g., `sysconf`). Some other posix functions, folly changes to adapt behavior
across platforms. For example, on windows folly defines `open`, `read`,
`write`, and `close` functions to work with sockets. Folly makes these
functions available in the global scope for convenience.
Changelog: [Internal]
Reviewed By: javache
Differential Revision: D65855147
fbshipit-source-id: b06863330ca213b9d1bffe0ee85e0fbf1bc8a845
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47713
`defaultSource` is defined as a single `ImageSource` in docs (https://reactnative.dev/docs/next/image#defaultsource) but its type is a vector of `ImageSource`s in `ImageProps.h`.
This diff changes its C++ type to just `ImageSource`.
Technically, this is a breaking change, however I don't think folks should directly access `ImageProps` outside of `RCTImageComponentView.mm`
Moreover, this prop is actually not implemented in Fabric, so this change should have no practical effect.
Changelog: [Internal]
Facebook
T208171435 - [RN][Fabric][iOS] Implement defaultSource support for Image
Reviewed By: javache
Differential Revision: D65821570
fbshipit-source-id: 38139b0f8d6da495e82c4ef72c19af3db254ba6c
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47637
Changelog: [Internal]
RSNRS does not support off JS thread layout due to current fiber tree corruption when syncing happens on more than one shadow tree at the same time. This diff guarantees that RSNRS will only be enabled on the JS thread, avoiding any state corruption.
Reviewed By: sammy-SC
Differential Revision: D64500893
fbshipit-source-id: aa20f54a0fcfa47534ae099e95307a692bd9fd0f
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47736
This has been 100% enabled in production across all the apps for more than a month, after an experiment a long time ago, and we kinda depend on it now for correct border drawing in some cases. Let's clean it up!
Changelog: [Internal]
Reviewed By: joevilches
Differential Revision: D66208418
fbshipit-source-id: 582a38bd84a2d085ba5c4aac4cd478680dd206cc
Summary:
X-link: https://github.com/facebook/yoga/pull/1746
Chrome made some changes for how overflowed row-reverse containers are laid out which was causing some issues on CI. I updated them here and skipped the new failing tests which we would want to followup on.
For LTR, the differences are seen below
|Before|After|
|--|
|{F1962694149} | {F1962694151}|
The extra space is now extending past the flex start edge vs flex end. RTL is the opposite. NickGerleman had deviated from the spec back in the day to match Chrome and it seems they made the adjustment recently. T208209388 is tracking the followup to align with the spec again. Basically, there is a notion of fallback alignment when certain justification/alignment values cannot actually apply. Right now we are falling back to flex start in all cases but we should fallback to start sometimes.
Changelog: [Internal]
Reviewed By: NickGerleman
Differential Revision: D66138361
fbshipit-source-id: c46d2e9b0cd297069b9cc544e3bded995e4867a6
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47689
We were seeing some segfaults from trying to access `display_` in Yoga's `replaceChild` function. Some memory debugging helped identify this was a use-after-free error and NickGerleman suggested we try swapping these lines. Logic being the shared_ptr of YogaLayoutableShadowNode is replaced right before we go and replace its yoga node in the yoga tree. If this is the last shared_ptr holding this node we will delete this object and thus the yoga node with it. We do not need to do this first, so let's swap the lines.
Changelog: [Internal]
Reviewed By: NickGerleman
Differential Revision: D66142356
fbshipit-source-id: 8fd835346edc91e045ed2ee8945a95af21c47556
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47618
Its not optimal to reconstruct CompositeBackgroundDrawable everytime we add a layer to it.
With this change I'm adding an optimization to modify the underlying `LayerDrawable` in place instead of reconstructing everything. `LayerDrawable` has a pretty constrained API and some weirdish behaviors.
- `addLayer` - This API doesnt set the callback for the recently added layer so after adding the layer we also need to manually set the callback
- Mutating `LayerDrawable` in-place is not straightforward, I had to add a function that will figure out where each layer should be inserted
- `LayerDrawable` doesn't allow deleting an element which means that for this case in particular we do need to re-create `CompositeBackgroundDrawable`
- Newer Android versions allow `null` on `LayerDrawable` layers, but older versions do not, this implementation is mostly done this way to accommodate older versions. But also, even though newer versions can have `null` set on a layer `LayerDrawable` still doesn't handle it well and we get some bugs when removing and inserting a layer which the current implementation handles by not relying on null
This is all feature flagged. since it will only be enabled with the new drawables
Changelog: [Internal]
Reviewed By: joevilches
Differential Revision: D65907786
fbshipit-source-id: c18b898ddfe58faa5b90714945ffcc82d471051d
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/46818
Before we were using lists which meant `CompositeBackgroundDrawable` had a `LayerDrawable` of variable length.
This meant that our layers were not consistent and made it hard to optimize `CompositeBackgroundDrawable` to not create a new instance every time we set a new Layer.
With this change `CompositeBackgroundDrawable` has a consistent amount of layers which will allow us to more easily mutate it and land the optimization present on D63287222
Changelog: [Internal]
Reviewed By: NickGerleman
Differential Revision: D63798842
fbshipit-source-id: bf0b9ee91ff2083b89ca90ace4de00c71bf3153c
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47710
There is Android code that supports remapping `defaultSrc` -> `defaultSource` for `ImageNativeComponent`.
The docs reference this prop as `defaultSource`: https://reactnative.dev/docs/next/image#defaultsource
It is not referenced as `defaultSrc` anywhere. Let's unify it as `defaultSource` across both platforms.
Changelog: [Internal]
Reviewed By: javache
Differential Revision: D65819218
fbshipit-source-id: 0f468e2327ad07285a45e4c9f5e33d74da411c74
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47676
`getModule` is used by all modern ReactPackage base-classes to allow lazy loading of specific module and should be considered stable at this point.
Changelog: [Android][Added] Marked ReactPackage#getModule as stable.
Reviewed By: mdvacca
Differential Revision: D66127068
fbshipit-source-id: de447794435d267619430e3c0cbc65c6e1c614ba
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47669
When investigating [#47476](https://github.com/facebook/react-native/issues/47476), I found that the `secureTextInput` prop was not changing in the Mounting layer when changing it in JS.
I track down the problem to the `UIManager::cloneNode` method.
When we clone the node, we first merge the patch that arrives from React into the props controlled by setNativeProps, ignoring the patch's props that are controlled by React.
But then, we forgot to merge back the React's controlled property into the final props, effectively losing them.
This change adds an extra merging step, merging the props controlled with setNativeProps back into the patch of props controlled by React, and then using this new set of props as source of truth.
## Changelog:
[General][Fixed] - do not discard props in the patch when they are not null while using `useNativeProps`
Reviewed By: sammy-SC
Differential Revision: D65948574
fbshipit-source-id: db4f2b793f4a6348456933c95a151012252b8ebc
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47704
Now that we have a min requirement of Node 18.18 on `main`, drop dependency on this polyfill.
Changelog: [Internal]
Reviewed By: vzaidman
Differential Revision: D66162328
fbshipit-source-id: e8ab6669fe14ed177eccf4b861c01df4fb0d405a
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47718
We have a test in the OSS E2E CI pipeline that is a bit flaky and fails often on Android.
I don't have time to investigate that properly right now, so I'm disabling it to improve the situation on the main branch for the time being.
We are planning to invest more resources in H1 2025 to improve the E2E testing in OSS, so I'll get back to it soon.
## Changelog
[Internal] - Disabling part of the FlatList E2E test in Maestro
Reviewed By: javache
Differential Revision: D66169183
fbshipit-source-id: 5b2b0c45e124a642b626b014b91fa61d17226f9b
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47700
Changelog: [internal]
Add support for the string parameter for `toThrow` to assert for specific error messages.
Reviewed By: sammy-SC
Differential Revision: D66118001
fbshipit-source-id: 8c04cd20d4ad17163ec0c7bf943c429507a97985
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47699
Changelog: [internal]
Implements a `jest.fn()` and a subset of methods in `expect` using them (`.toHaveBeenCalled()` and `.toHaveBeenCalledTimes()`).
Reviewed By: sammy-SC
Differential Revision: D66118002
fbshipit-source-id: 5422307d68967d7d8b4c4d5155a45926f8fc8cae
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/47706
changelog: [internal]
The changelog in D65909191 already has the details. The fix is behind a feature flag. Let's enable it in OSS to stop the bleeding.
Reviewed By: cipolleschi
Differential Revision: D66163663
fbshipit-source-id: 84a3dfb823fdd286e919ebf22b07054b4399662a
