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:
@public
`YGNodeGetInstanceCount` was only ever meant for tests, and caused data races in multi-threaded environments.
It was completely replaced with event-based counting for tests.
Here we remove public API around the counter, and the counter itself.
Reviewed By: SidharthGuglani
Differential Revision: D15174857
fbshipit-source-id: 228e85da565bac9e8485121e956a2e41910b11c8
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: This new target provides dependencies for using folly futures
Reviewed By: willholen
Differential Revision: D15018282
fbshipit-source-id: c38ad4775102b9f0c10b3a52c5a18f00aa398322
Summary: This make iteration work better without needing to clean as much
Reviewed By: willholen
Differential Revision: D15018285
fbshipit-source-id: 034f5529e2e51711aeaa75360ad10bb1f85c7fb8
Summary:
The path to copy log_severity.h could refer to the
destination, which would result in an empty file being copied on some
rebuilds.
Reviewed By: willholen
Differential Revision: D15018283
fbshipit-source-id: 0081526a9686de8c74753738c165753de6dda18d
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
Summary:
This fixes a regression on Android introduced by f3e5cce where JS errors thrown during bundle load were lost (shown only as UnknownCppException). It is especially tough to debug (custom) bundling errors without seeing the javascript error.
Root cause hypothesis: since switching to clang, `JSError`s thrown in ReactCommon's `JSCRuntime::checkException` were never matched as `std::exception` in ReactAndroid's `convertCppExceptionToJavaException` due to these missing flags.
I'm a bit shy on low-level details concerning how C++ rtti works exactly around catching exceptions thrown in other libraries. All I can say that with this change, a `bundle.android.js` that only contains `throw new Error("wtf");` now nicely outputs a message and stack trace in the logcat. Before (and since DanielZlotin's switch to clang) it just outputted:
```
2019-04-29 12:17:59.365 1162-1306/com.rntest E/unknown:ReactNative: Exception in native call
com.facebook.jni.UnknownCppException: Unknown
at com.facebook.react.bridge.queue.NativeRunnable.run(Native Method)
at android.os.Handler.handleCallback(Handler.java:873)
at android.os.Handler.dispatchMessage(Handler.java:99)
at com.facebook.react.bridge.queue.MessageQueueThreadHandler.dispatchMessage(MessageQueueThreadHandler.java:29)
at android.os.Looper.loop(Looper.java:214)
at com.facebook.react.bridge.queue.MessageQueueThreadImpl$4.run(MessageQueueThreadImpl.java:232)
at java.lang.Thread.run(Thread.java:764)
```
[Android] [Fixed] - JS errors during bundle load were reported as UnknownCppException.
Pull Request resolved: https://github.com/facebook/react-native/pull/24648
Differential Revision: D15123525
Pulled By: mdvacca
fbshipit-source-id: 74b5ce9ebae38d172446b6e31739d795c601947b
Summary:
Hello Everyone, this series of commits helps to fix problems related to ART on Android. The main problem in here is that the ART components would disappear if the user turns off the screen and then turn it on again. It's important to note that this behaviour only occurs after Android N (7.1 or higher).
Pull Request resolved: https://github.com/facebook/react-native/pull/22624
Differential Revision: D15122573
Pulled By: cpojer
fbshipit-source-id: e7fb8b9280b4c52562e3d0c1a89759d4d31cd53d
Summary:
On android when borderXColor is null it causes the app to crash with:
```
04-28 17:44:18.021 20114 20213 E unknown:ReactNative: com.facebook.react.bridge.NoSuchKeyException: borderBottomColor
04-28 17:44:18.021 20114 20213 E unknown:ReactNative: at com.facebook.react.bridge.ReadableNativeMap.getValue(ReadableNativeMap.java:111)
04-28 17:44:18.021 20114 20213 E unknown:ReactNative: at com.facebook.react.bridge.ReadableNativeMap.getValue(ReadableNativeMap.java:115)
04-28 17:44:18.021 20114 20213 E unknown:ReactNative: at com.facebook.react.bridge.ReadableNativeMap.getInt(ReadableNativeMap.java:158)
04-28 17:44:18.021 20114 20213 E unknown:ReactNative: at com.facebook.react.uimanager.ViewProps.isLayoutOnly(ViewProps.java:246)
04-28 17:44:18.021 20114 20213 E unknown:ReactNative: at com.facebook.react.uimanager.NativeViewHierarchyOptimizer.isLayoutOnlyAndCollapsable(NativeViewHierarchyOptimizer.java:482)
```
Basically it is missing a `isNull` check on the props map before trying to access the value.
Fixes#22727
[Android] [Fixed] - Fix a crash when borderXColor is null
Pull Request resolved: https://github.com/facebook/react-native/pull/24640
Differential Revision: D15120182
Pulled By: cpojer
fbshipit-source-id: bc41da572f04d8abf733b5a4e94a74a0f40acda1
Summary:
Convert root Gradle script to Kotlin DSL, and cleanup. Currently, there is not much benefit or advantage over Groovy scripts, except IDE support and it'll cache compiled KTS scripts on first run.
[Android] [Changed] - Convert root Gradle script to Kotlin DSL, and cleanup.
Pull Request resolved: https://github.com/facebook/react-native/pull/24631
Differential Revision: D15120190
Pulled By: cpojer
fbshipit-source-id: 86691db5c7746e71bb243ebc263c1a3075ee9a9e
Summary:
This PR bumps Android Gradle Plugin to 3.4.0, which enables R8 shrinker by default and improves build performance significantly.
Disabled R8 for ReactAndroid because it'll strip out AndroidX and other libraries bundled in ReactAndroid.
[Android] [Changed] - bump Android Gradle plugin to 3.4.0
Pull Request resolved: https://github.com/facebook/react-native/pull/24463
Differential Revision: D15107117
Pulled By: hramos
fbshipit-source-id: 35a03dc9955e889c9399faeaf9a862e0fc044fc4
Summary: 9079beda9b wasn't enough to blacklist the strings.xml files from FB infra. So this is a better workaround. According to https://developer.android.com/guide/topics/resources/string-resource the .xml file names doesn't really matter, so this shouldn't break the ability to get the string resource in code, like `R.strings.foo`.
Reviewed By: yungsters
Differential Revision: D15103303
fbshipit-source-id: 6d5174a8dc9598930670d35434e1494f9eaea059
Summary:
Now RN has only ReactActivity which extends AppCompatActivity, subclass of FragmentActivity, therefore no need to check if activity is FragmentActivity or not. This PR changes DialogModule to work only with FragmentActivity.
Also DialogFragment from Android is deprecated in API 28, and recommends to use DialogFragment from Support Library. Excerpt from DialogFragment documentation.
> **This class was deprecated in API level 28.**
> Use the Support Library DialogFragment for consistent behavior across all devices and access to Lifecycle.
**BREAKING CHANGE**: Brown field apps must extend FragmentActivity or its subclasses.
[Android] [Changed] - DialogModule supports only FragmentActivity
Pull Request resolved: https://github.com/facebook/react-native/pull/23365
Differential Revision: D14021986
Pulled By: cpojer
fbshipit-source-id: b0ede60ef19cec48111a12701659a8bc1f66c331
Summary: These strings shouldn't be translated by FB system because each app has its own set of languages and/or translation outputs. We're keeping just values/strings.xml in the repo.
Reviewed By: cpojer
Differential Revision: D15087192
fbshipit-source-id: c4b6112f6dd010d317060ac6640b34e4b725c695
