Summary:
We're working on a custom EditText that supports some rich text editing, and one of the places where our logic has to be different from the textinput provided by RN is the text setting logic:
https://github.com/facebook/react-native/blob/7abfd23b90db08b426c3c91b0cb6d01d161a9b9e/ReactAndroid/src/main/java/com/facebook/react/views/textinput/ReactEditText.java#L377
However, some of the important members are private and our subclass cannot access them (we work around this now with reflection). It would be nice if we could work with them directly, either using getters and setters or by changing the access. Let me know what you think about this. Thanks.
## Changelog
[Android] [Added] - allow custom maybeSetText logic for ReactEditText subclasses
Pull Request resolved: https://github.com/facebook/react-native/pull/24927
Differential Revision: D15431682
Pulled By: cpojer
fbshipit-source-id: 91860cadac0798a101ff7df6f6b868f3980ba9b1
Summary:
This pull request enhances the Keyboard API event emitter for Android upon `keyboardDidHide` by returning a `KeyboardEvent` with a meaningful `endCoordinates` property (instead of emitting a null as of current implementation). This change standardizes the `keyboardDidHide` keyboard event emission across both iOS and Android, which makes it easier for developers to use the API.
In particular, the semantics of `endCoordinates` emitted during the `keyboardDidHide` event on Android will match nicely with semantics of the same event emitted on iOS:
- `screenY` will be height of the screen, as that the keyboard has collapsed to the bottom of the screen
- `screenX` will be 0, as the keyboard will always be flush to the sides of the screen
- `height` will be 0, as the keyboard has fully collapsed
- `width` will be the width of the screen, as the keyboard will always extend to the width of the screen
Also, the flowtypes for `KeyboardEvent` have been further improved and are more explicit to better highlight the different object shapes (see `IOSKeyboardEvent` and `AndroidKeyboardEvent`) depending on the platform.
## Changelog
[Android] [Added] - Return endCoordinates for keyboardDidHide keyboard event
Pull Request resolved: https://github.com/facebook/react-native/pull/24947
Differential Revision: D15413441
Pulled By: cpojer
fbshipit-source-id: aa3998542b7068e9852467038f57310355018379
Summary:
Easy diff to refactor the sComponentNames map out of the FabricUIManager class.
This is a necessary clean-up to perform a slightly major refactor of the Fabric classes
Reviewed By: JoshuaGross
Differential Revision: D15421769
fbshipit-source-id: 3be73a6e20b338c8cea23ef0c88db417df7e3aa9
Summary:
Quick diff to refactor RootTag for surfaceId in Binding.cpp class
This is the first diff to start moving away from rootTag naming in Fabric
Reviewed By: JoshuaGross
Differential Revision: D15421770
fbshipit-source-id: 7bca7782f96be3d7148ee93f5d5a3a54e0d768dd
Summary:
This is a reconstitution of #24190. It extends accessibility actions to include both a name and user facing label. These extensions support both standard and custom actions.
We've also added actions support on Android, and added examples to RNTester showing how both standard and custom accessibility actions are used.
## Changelog
[general] [changed] - Enhanced accessibility actions support
Pull Request resolved: https://github.com/facebook/react-native/pull/24695
Differential Revision: D15391408
Pulled By: cpojer
fbshipit-source-id: 5ed48004d46d9887da53baea7fdcd0e7e15c5739
Summary:
This is already handled cleanly on the JS side of things in AnimatedInterpolation.js: https://github.com/facebook/react-native/blob/0ee5f68929610106ee6864baa04ea90be0fc5160/Libraries/Animated/src/nodes/AnimatedInterpolation.js#L133-L142
However, the native driver interpolation will try to divide by 0, produce NaN and then crash. This change just copies the logic from the JS version of the interpolation logic and adds it to the Java version.
Note that this bug only reproduces on Android Q. It seems that RenderNode::setCameraDistance now crashes when receiving NaN on Android Q.
Reviewed By: sahrens
Differential Revision: D15380844
fbshipit-source-id: cfa82d8f58574e1040a851aaa5b5af1e23c9daa8
Summary: `android.util.ArrayMap` ins't available on API < 19. Let's use the one that Android Support Library (aka AndroidX) provides.
Reviewed By: mdvacca
Differential Revision: D15372704
fbshipit-source-id: 5c2ea3c7ea7368bb75ff22c54af0b258558556b5
Summary:
The `measure` API receives LocalData and Props, it should also receive State.
This will also be used in future diffs.
Reviewed By: mdvacca
Differential Revision: D15325182
fbshipit-source-id: 6cb46dd603ce7d46673def16f0ddb517e2cf0c4f
Summary:
Fixes redbox/yellowbox symbolication when the Java delta client is enabled. Previously the modules would get concatenated in a nondeterministic order (owing to Metro's parallelism) which differed from their order in the source map, where they're explicitly sorted by module ID.
This diff changes the data structure holding modules in memory from a `LinkedHashMap` (which iterates in insertion order) to a `TreeMap` (which iterates in key order).
NOTE: Similar to this change in the Chrome debugger's delta client: https://github.com/react-native-community/cli/pull/279
Reviewed By: dcaspi
Differential Revision: D15301927
fbshipit-source-id: 27bdecfb3d6963aa358e4d542c8b7663fd9eb437
Summary:
When running Android app for the first time, the packager is requesting delta bundles from metro instead of a bundle (in dev settings delta bundles are disabled by default and marked as experimental). UI of dev settings is not consistent with the current state, to turn off delta bundles you have to enable them and then disable.
[Android] [Fixed] - Disable delta bundles on the first app run
Pull Request resolved: https://github.com/facebook/react-native/pull/24848
Differential Revision: D15334059
Pulled By: cpojer
fbshipit-source-id: 384a8abba64c54db3656a4d5d0e24acc825870c8
Summary:
We are getting errors where views are being dropped twice.
This is following from logic that viewManagers are only removed from `mTagsToViewManagers` from `dropView`.
This log will hopefully identify if we are getting improper operations because we shouldn't be re-using tags
Reviewed By: mdvacca
Differential Revision: D15152869
fbshipit-source-id: 914ee9c1772fa066adefde0753075ecba6377a0c
Summary:
This diff ensures the method scheduler.constraintSurfaceLayout is executed before the JS run application start.
This is necessary to properly set the pointScaleFactor for the Root before running JS.
This is a workaround to fix a bug when the pointScaleFactor changes over time for the rootShadowNode. The bug is easily reproducible when rendering the "fabric" indicator on Fabric screens. During the first render of a Fabric screen this method was called before "JS run application" starts, and the Fabric indicator was render correctly.
Beacuse of timing of measure APIS, the second time a Fabric screen is rendered the method is called after the "JS run application process started", as a consecuence the Fabric indicator is not rendered correctlly (the pointScaleFactor is incorrectly assigned into the layout metrics of the Fabric indicator text).
We still need to analyze why the pointScaleFactor is not correctly assigned when it is set after the "JS run application process started", but this will be part of another diff.
Reviewed By: shergin
Differential Revision: D15303554
fbshipit-source-id: 7d985cefee20fd40dbe04166c1a1358b3f3ddc85
Summary: `YogaStylableProps.yogaStyle` is designed to be consumed by Yoga only. Making it `protected` allows us to avoid confusion and misuse of this props.
Reviewed By: JoshuaGross
Differential Revision: D15296474
fbshipit-source-id: cf9e416afee99fb426d72765557b34d303a63dbe
Summary:
Fixes#13351
Two root causes:
1. Android Spinner will reset selection to undefined after setAdapter()
which will trigger onValueChange().
The behavior is not expected for RN.
And the solution is to setSelection() explicitly
2. In original implementation, it setups `items` immediately,
but delays the `selected` setup after update transaction.
There will be some race condition and incosistency
if update `items` only.
The fix will do the setup all after update transaction.
[Android] [Fixed] - Fix#13351 PickerAndroid will reset selected value during items update.
Pull Request resolved: https://github.com/facebook/react-native/pull/24793
Differential Revision: D15293516
Pulled By: cpojer
fbshipit-source-id: 5a99a60015c7e1b2968252cdc0b2661d52a15b9d
Summary:
Fixes this issue:
https://github.com/facebook/react-native/issues/24468
It was incorrectly closed by a fix on react-native-community CameraRoll implementation. CameraRoll in react-native still crashes when finding a file with # sign
[Android] [Fix] - Fix Android Camera Roll crash on mime type guessing
Pull Request resolved: https://github.com/facebook/react-native/pull/24780
Reviewed By: mdvacca
Differential Revision: D15281062
Pulled By: lunaleaps
fbshipit-source-id: ca3364c8478d9bfc9a0a6657b531ffb384145d8c
Summary: In the future we're planning to decouple ThemedReactContext from the bridge (CatalystInstance). For now, we just need to be able to create a ThemedReactContext with a ReactContext that has no Catalyst instance.
Reviewed By: mdvacca
Differential Revision: D15246442
fbshipit-source-id: 99ebda6521f4df72969011ea0e6ea41b046875c8
Summary: Refactoring ReactContext to move message queue initialization into its own function that can be called independently of initializeWithInstance. This allows you to create a ReactContext with message queue threads without a CatalystInstance.
Reviewed By: mdvacca
Differential Revision: D15246287
fbshipit-source-id: 4b8c53e68112af7eded47d8c31311500cc296dfe
Summary: Right now TurboModuleManager gets the JSCallInvokerHolder from the bridge in its constructor; this diff changes the constructor to make the JSCallInvokerHolder a required argument so that TurboModuleManager doesn't directly depend on the bridge.
Reviewed By: axe-fb, RSNara
Differential Revision: D15227184
fbshipit-source-id: b16e6abaa727587986a132d0e124163acdf55408
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/24764
The `test_android` CI build was failing:
```
./scripts/circleci/buck_fetch.sh
+ buck fetch ReactAndroid/src/test/java/com/facebook/react/modules
Not using buckd because watchman isn't installed.
Picked up _JAVA_OPTIONS: -XX:+UnlockExperimentalVMOptions -XX:+UseCGroupMemoryLimitForHeap
PARSING BUCK FILES: FINISHED IN 1.1s
No build file at ReactAndroid/src/main/libraries/fbjni/BUCK when resolving target //ReactAndroid/src/main/libraries/fbjni:java.
This error happened while trying to get dependency '//ReactAndroid/src/main/libraries/fbjni:java' of target '//ReactAndroid/src/main/java/com/facebook/react/turbomodule/core:jscallinvokerholder'
Exited with code 1
```
The problem was that I was using `react_native_dep("libraries/fbjni:java")` to access JNI classes like `HybridData`. In open source this target translates to the path `//ReactAndroid/src/main/libraries/fbjni`, which doesn't exist. Instead, the actual classes are available at the path `//ReactAndroid/src/main/java/com/facebook/jni`. Therefore, I changed the target to `react_native_dep("java/com/facebook/jni:jni")`. This is exactly how the bridge (i.e: `//ReactAndroid/src/main/java/com/facebook/react/bridge:bridge`) accesses JNI clases.
Reviewed By: hramos
Differential Revision: D15261218
fbshipit-source-id: 659a5627389bbca3603db7de347618cd400d4ffc
Summary:
AccessibilityInfo.announceForAccessibility is currently only available on iOS. I've added the Android specific implementation, updated RNTester, and the documentation.
[Android] [Added] - Added AccessibilityInfo.announceForAccessibility for Android
[General] [Added] - RNTester example for AccessibilityInfo.announceForAccessibility
Pull Request resolved: https://github.com/facebook/react-native/pull/24746
Differential Revision: D15258054
Pulled By: cpojer
fbshipit-source-id: 3e057a5c32b28e30ea2ee74a18854b012cd2dbfd
Summary:
In https://github.com/facebook/react-native/pull/23865, RN introduced support for custom fonts the Android Way. But it introduced performance regression because it'll lookup for a font using getIdentifier() every time fontFamily changed. This PR fixes regression by requiring custom fonts to be listed in **fonts** array, and populating **mTypeCache** at first use using the list.
[Android] [Changed] - Require custom fonts to list in **fonts** array. Fixes performance regression.
Pull Request resolved: https://github.com/facebook/react-native/pull/24595
Reviewed By: mdvacca
Differential Revision: D15184590
Pulled By: fkgozali
fbshipit-source-id: e3feb2396609583ebc95101130186a1f5af931da
Summary:
Resolve#24690
This is very simple unicode detecting. Should I improve this solution creating StringsUtils for detecting unicodes in whole react-native project ?
[Android][Fixed] onKeyPress method is calling, when user type emoji
Pull Request resolved: https://github.com/facebook/react-native/pull/24717
Differential Revision: D15238388
Pulled By: cpojer
fbshipit-source-id: 038b1040e1c44fd6f9401a3988a782f5778e1209
Summary:
In #24095, we removed the code that changes the underlying Android view's enabled state to false when "disabled" is included in the accessibility states, which seems correct. The Android view's state shouldn't mirror the accessibility state, it should be the other way round-- the accessibility state should represent the state of the view.
It seems that the existing test is brokenly setting the disabled state in the Javascript object, and then querying the Android view's enabled state to confirm the change. If the Button implementation is actually the culprit, then IMHO, the correct fix would be to have the Button implementation manipulate the Android View's enabled state, not the accessibilityStates code.
[android] [fix] - Fix internal test case around disabled state of buttons
Pull Request resolved: https://github.com/facebook/react-native/pull/24678
Differential Revision: D15237590
Pulled By: cpojer
fbshipit-source-id: d7ebefbcba9d4d9425da4285175302e4b6435df7
Summary: I'm not sure if this is a good idea. Right now FabricUIManager creates a ThemedReactContext in addRootView() using the RAC you pass in. If you pass in an RAC without a Catalyst instance, this will throw; this diff makes it so it'll throw the next time you try to actually try to access the CatalystInstance, instead. I don't know if we're really relying on this right now, but we need to be able to create a ThemedReactContext without a CatalystInstance for Venice (for now, until we actually go through and get rid of TRC's dependency on the CatalystInstance entirely - but that'll be a lot more work)
Reviewed By: mdvacca
Differential Revision: D15194220
fbshipit-source-id: 64689cbe79c84ae33fe16e3dc396e3c69ec8e20f
Summary: Refactoring ReactContext to move message queue initialization into its own function that can be called independently of initializeWithInstance. This allows you to create a ReactContext with message queue threads without a CatalystInstance.
Reviewed By: mdvacca
Differential Revision: D14817741
fbshipit-source-id: f314a526c6534792714e5ba55dd873f1728c6b9f
Summary:
[General] [Fix] - Reorder operations of native view hierarchy
When we update the native view hierarchy in `manageChildren` we:
1. iterate through all views we plan to remove and remove them from the native hierarchy
2. iterate through all views we plan to add and add them to the native hierarchy
3. iterate through all views we plan to delete and delete them (remove them from memory)
This covers these cases:
a. A view is moved from one place to another -- handled by steps 1 & 2
b. A view is added -- handled by step 2
c. A view is deleted -- handled by step 1 & 3
> Note the difference between remove and delete
Everything above sounds fine! But...
The important bit:
**A view that is going to be deleted asynchronously (by a layout animation) is NOT removed in step 1. It is removed and deleted all in step 3.** See: https://fburl.com/ryxp626i
If the reader may recall we solved a similar problem in D14529038 where we introduced the `pendingIndicesToDelete` data structure to keep track of views that were marked for deletion but had not yet been deleted. An example of an order of operations that we would've solved with D14529038 is:
* we "delete" the view asynchronously (view A) in one operation
* we add a view B that shares a parent with view A in subsequent operation
* view A finally calls its callback after the animation is complete and removes itself from the native view hierarchy
A case that D14529038 would not fix:
1. we add a view B in one operation
2. we delete a view A in the same operation asynchronously because it's in a layout animation
3. ... etc.
What we must remember is that the index we use to add view B in step 1 is based on the indices assuming that all deletions are synchronous as this [comment notes](https://fburl.com/j9uillje): removals (both deletions and moveFroms) use the indices of the current order of the views and are assumed independent of each other. Whereas additions are indexed on the updated order (after deletions!)
This diff re-arranges the order in how we act upon the operations to update the native view hierarchy -- similar to how UIImplementation does its operations on the shadow tree here: https://fburl.com/j9uillje
By doing the removals and deletions first, we know that the addAt indices will be correct because either the view is removed from the native view hierarchy or `pendingIndicesToDelete` should be tracking an async delete already so the addAt index will be normalized
Reviewed By: mdvacca
Differential Revision: D15112664
fbshipit-source-id: 85d4b21211ac802183ca2f0fd28fc4437d312100
Summary:
This diff forces the eager initialization of some additional classes into FabricJSIModuleProvider.loadClasses().
This is a "hack" that will be removed in the near future
Reviewed By: JoshuaGross
Differential Revision: D15208977
fbshipit-source-id: 2e2c7856839b6c6888452800ef6da7f269e46735
Summary: When passing StateWrapper objects across the JNI, we were not ensuring that the Java objects would own the C++ state. This was initially done because I assumed that in Java, State would either be used immediately or discarded, so this wouldn't be unsafe. As it turns out, it makes sense in some cases to store the StateWrapper in Java and use it later, potentially even in other threads, so we need to make sure we maintain ownership of the C++ object from the Java object.
Reviewed By: shergin
Differential Revision: D15206194
fbshipit-source-id: a437d921ba00b194cf08bad80666bd99baf11d52
Summary: This diff implements encapsulating all time metrics in a single class for better extensibility and readability.
Reviewed By: JoshuaGross
Differential Revision: D15179835
fbshipit-source-id: 62bdf94435a0d37a87ad9bad613cc8e38043a235
Summary:
`CatalystInstanceImpl.cpp` now depends on `JSCallInvoker` and `JavaJSCallInvokerHolder`. Therefore, we need to correctly adjust the OSS builds to include these dependencies into the `libreactnativejni.so` file.
I made `ReactCommon/turbomodule/jscallinvoker` a static library `libjscallinvoker.a`. I then made `ReactAndroid/src/main/jni/react/jni`'s `libreactnativejni.so` depend on that static library. Also, because the Android NDK build system doesn't support header namespaces, I had to use the filesystem to simulate them. This is why all the `.cpp` and `.h` files for `JSCallInvoker` were moved to a `jsireact` folder.
Reviewed By: mdvacca
Differential Revision: D15194821
fbshipit-source-id: 0cee682e41db53d0619f56ad017d5882f6d554aa
Summary:
`TurboModuleManagerDelegate` is an abstract base class with the following API:
```
public TurboModule getModule(String name, ReactApplicationContext reactApplicationContext);
public CxxModuleWrapper getLegacyCxxModule(String name, ReactApplicationContext reactApplicationContext);
```
```
std::shared_ptr<TurboModule> getTurboModule(std::string name, jni::global_ref<JTurboModule> turboModule, std::shared_ptr<JSCallInvoker> jsInvoker) override;
std::shared_ptr<TurboModule> getTurboModule(std::string name, std::shared_ptr<JSCallInvoker> jsInvoker) override;
```
On the C++ side, when asked to provide a TurboModule with name `name`:
1. First, is this a CxxModule? If so:
1. Create the CxxModule and return
2. Otherwise:
1. Somehow get a Java instance of the TurboModule
2. If this Java object represents a CxxModule, like `FbReactLibSodiumModule`:
1. Grab the C++ part of this object, and wrap it in a `TurboCxxModule.cpp` and return.
3. Otherwise:
1. Wrap the Java object in a C++ HostObject and return.
This pseudocode demonstrates how we'd use `TurboModuleManagerDelegate` to implement `__turboModuleProxy`.
```
__turboModuleProxy(name, jsCallInvoker):
let cxxModule = TurboModuleManagerDelegate::getTurboModule(name, jsCallInvoker)
if (!cxxModule) {
return cxxModule;
}
// JNI Call that forwards to TurboModuleManagerDelegate.getLegacyCxxModule
let javaCxxModule : CxxModuleWrapper = TurboModuleManager.getLegacyCxxModule(name)
if (!javaCxxModule) {
return std::shared_ptr<react::TurboCxxModule>(javaCxxModule.getModule())
}
// JNI Call that forwards to TurboModuleManagerDelegate.getModule
let javaModule : TurboModule = TurboModuleManager.getModule(name)
if (!javaCxxModule) {
return TurboModuleManagerDelegate::getTurboModule(name, javaModule, jsCallInvoker)
}
return null
```
Reviewed By: mdvacca
Differential Revision: D15111335
fbshipit-source-id: c7b0aeda0e4565e3a2729e7f9604775782b6f893
Summary:
JSCallInvoker requires a `std::weak_ptr<Instance>` to create. In our C++, `CatalystInstance` is responsible for creating this `Instance` object. This `CatalystInstance` C++ initialization is separate from the `TurboModuleManager` C++ initialization. Therefore, in this diff, I made `CatalystInstance` responsible for creating the `JSCallInvoker`. It then exposes the `JSCallInvoker` using a hybrid class called `JSCallInvokerHolder`, which contains a `std::shared_ptr<JSCallInvoker>` member variable. Using `CatalystInstance.getJSCallInvokerHolder()` in TurboModuleManager.java, we get a handle to this hybrid container. Then, we pass it this hybrid object to `TurboModuleManager::initHybrid`, which retrieves the `std::shared_ptr<JSCallInvoker>` from the `JavaJSCallInvokerHandler`.
There were a few cyclic dependencies, so I had to break down the buck targets:
- `CatalystInstanceImpl.java` depends on `JSCallInvokerHolderImpl.java`, and `TurboModuleManager.java` depends on classes that are packaged with `CatalystInstanceImpl.java`. So, I had to put `JSCallInvokerHolderImpl.java` in its own buck target.
- `CatalystInstance.cpp` depends on `JavaJSCallInvokerHolder.cpp`, and `TurboModuleManager.cpp` depends on classes that are build with `CatalystInstance.cpp`. So, I had to put `JavaJSCallInvokerHolder.cpp` in its own buck target. To make things simpler, I also moved `JSCallInvoker.{cpp,h}` files into the same buck target as `JavaJSCallInvokerHolder.{cpp,h}`.
I think these steps should be enough to create the TurboModuleManager without needing a bridge:
1. Make `JSCallInvoker` an abstract base class.
2. On Android, create another derived class of `JSCallInvoker` that doesn't depend on Instance.
3. Create `JavaJSCallInvokerHolder` using an instance of this new class somewhere in C++.
4. Pass this instance of `JavaJSCallInvokerHolder` to Java and use it to create/instatiate `TurboModuleManager`.
Regarding steps 1 and 2, we can also make JSCallInvoker accept a lambda.
Reviewed By: mdvacca
Differential Revision: D15055511
fbshipit-source-id: 0ad72a86599819ec35d421dbee7e140959a26ab6
Summary: Convert FabricUIManager.measure params to floats. Currently we convert parameters to ints across the JNI boundary, and then back to floats several times in Java. This is unnecessary and actually makes measurements trickier. The new implementation uses floats across the JNI boundary and uses Float.POSITIVE_INFINITY to represent unconstrained values, which is consistent with Fabric C++ as well.
Reviewed By: shergin, mdvacca
Differential Revision: D15176108
fbshipit-source-id: cf849b3773007637f059279460163872f300a4aa
Summary:
When acquiring the `PARTIAL_WAKE_LOCK`, Android requires a tag to identify the source, normally the class name. This tag will show on dumpsys call and Google Play developer console.
`getSimpleName` will work fine as long as not enable ProGuard, in my case, it transformed the class name to just `"c"`, and I take my half day to find where the `c` comes from.
`getCanonicalName` will add the package path, which is more friendly for developers.
Later we can even let the developer choose the tag name, but this will require API break changes.
[Android] [Changed] - Use class canonical name for PARTIAL_WAKE_LOCK tag
Pull Request resolved: https://github.com/facebook/react-native/pull/24673
Differential Revision: D15164306
Pulled By: cpojer
fbshipit-source-id: fd65f9e5250c180b0053940b17877fe36af5d48b
Summary: The map of sComponentNames ONLY contains the names of components that are different between JS and Android. This diff adds a method to unify the way we use this map.
Reviewed By: shergin
Differential Revision: D15076549
fbshipit-source-id: 9df750dca305e55cb44037bc63f3ebb6476c8b81
Summary: Trivial diff that adds extra logging information on Exceptions that are thrown by the FabricViewTest
Reviewed By: shergin
Differential Revision: D14817899
fbshipit-source-id: 32e1d1fcd1292715dfcf2750d3f14c668927c8b8
Summary:
This diff fixes a bug that is reproducible when a view is reordered in a different level of hierarchy in the react tree.
Even if this is not supported by react, this can still happen because of viewFlattening.
Reviewed By: shergin
Differential Revision: D14817452
fbshipit-source-id: 13425b0e6a280affe681e80b4a6daa17ee56251a
Summary: This diff refactors the way we synchronize in ReactChoreographer using a lock object
Reviewed By: ejanzer
Differential Revision: D14913056
fbshipit-source-id: e86c4395d5d3c3fd5b7330b72c14920b536f74ce
Vojtech Novak
nossbigg
David Vacca
Eric Lewis
Uilque Messias
Sharon Gong
luancurti
Oleg Lokhvitsky
Alexey Lang
Joshua Gross
Moti Zilberman
wojteg1337
Luna Wei
Valentin Shergin
Kudo Chien
Maciej Srokowski
Emily Janzer
Ramanpreet Nara
Alan Kenyon
Dulmandakh
Piotr Trocki
Marc Mulcahy
Eileen Zhang
Tim Wang
Michał Pierzchała
