Summary: Adds information which type of delta client is used (if at all) to the `BundleInfo` object.
Reviewed By: fromcelticpark
Differential Revision: D7845139
fbshipit-source-id: e4bf6cda62c71a78aaff97aa69daec263e6d3cdf
Summary:
Adds support for native clients to `ReactAndroid`:
- `.devsupport.BundleDeltaClient` is now abstract with two implementations: the existing Java client, and a native client
- `BundleDeltaClient#processDelta(...)` can now return a native delta client object
- if that client object is non-null, the bridge is started up with that client rather than a script written to disk
Reviewed By: fromcelticpark
Differential Revision: D7845135
fbshipit-source-id: 379a9c6f9319c62eec3c370cda9ffa0969266a29
Summary: Adds glue between `CatalystInstance` and `JSBundleLoader`: Adds the possibility to create a `JSBundleLoader` instance that calls into `CatalystInstanceImpl#loadScriptFromDeltaBundle(...)`.
Reviewed By: fromcelticpark
Differential Revision: D7845134
fbshipit-source-id: 9585b44a6e7c63245c9f634543642be55c928896
Summary:
Adds the possibility to disable native delta clients in `DevInternalSettings`
Depending on the bridge in use, there might not be support for native delta clients, but since the settings are shared app-wide, it can not be enabled individually.
Reviewed By: johnislarry
Differential Revision: D7845137
fbshipit-source-id: ab368e6fed0f4bec49032c4a20466e156d20fdae
Summary:
Adds `loadScriptFromDeltaBundle` / `jniLoadScriptFromDeltaBundle` methods to `CatalystInstanceImpl`.
These methods allow to run JS coming from a native delta client as RAM bundles, to leverage the RAM bundle mechanism for development / reload scenarios.
Reviewed By: fromcelticpark
Differential Revision: D7845140
fbshipit-source-id: b79b340f36c28939a31eb63a3c463d0792a208f7
Summary:
Adds hybrid JNI wrapper class that can hold a native delta client.
This class allows Java code in the `devsupport` package to hold onto a native delta client across reloads.
It also takes care of passing `okhttp` response objects to native code, where we build a `std::string` to hold JSON messages coming from Metro, and parse them with `folly::parseJson`.
In a follow-up, we will switch to a streaming-friendly binary format that will make allocating memory for the whole message unnecessary.
Reviewed By: fromcelticpark
Differential Revision: D7845138
fbshipit-source-id: 7a29b30645331addf843097dd0d05c03b3143991
Summary:
Adds C++ delta client that keeps modules in memory, and can be used as a RAM bundle.
For now, this client expects a `folly::dynamic` object as payload for patches, i.e. the JSON response retrieved from Metro needs to be parsed with `folly::parseJson` by the caller.
In the future, we will replace JSON with a streaming friendly binary format.
Reviewed By: fromcelticpark
Differential Revision: D7845136
fbshipit-source-id: f003f98a2607c8354c427a7e60e01e19e20295b1
Summary:
Adds a menu item to enable handling of JS module deltas in native code.
The check box is dependent on JS Deltas being activated in the first place.
Reviewed By: cpojer
Differential Revision: D7380167
fbshipit-source-id: 3001d12140542b4e52b1ce96be3c5f56e52a15f8
Summary:
Fixing from call response.body.string() twice in DevServerHelper.java.
https://github.com/square/okhttp/issues/1240#issuecomment-68142603
I'm getting error like this
```
05-01 21:16:47.080 22836-23064/com.my.company.bundle E/AndroidRuntime: FATAL EXCEPTION: OkHttp Dispatcher
Process: com.my.company.bundle, PID: 22836
java.lang.IllegalStateException: closed
at okio.RealBufferedSource.rangeEquals(RealBufferedSource.java:398)
at okio.RealBufferedSource.rangeEquals(RealBufferedSource.java:392)
at okhttp3.internal.c.a(Util.java:449)
at okhttp3.v.string(ResponseBody.java:174)
at com.facebook.react.devsupport.f$8.onResponse(DevServerHelper.java:487)
at com.newrelic.agent.android.instrumentation.okhttp3.CallbackExtension.onResponse(CallbackExtension.java:41)
at okhttp3.s$a.c(RealCall.java:153)
at okhttp3.internal.b.run(NamedRunnable.java:32)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1162)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:636)
at java.lang.Thread.run(Thread.java:764)
```
<!--
Required: Write your motivation here.
If this PR fixes an issue, type "Fixes #issueNumber" to automatically close the issue when the PR is merged.
-->
- In my case, My device is using a Proxy, and tried to connect the packager. When it failed from connecting packager through websocket, It crash by this line of code.
<!--
Required: Write your test plan here. If you changed any code, please provide us with
clear instructions on how you verified your changes work. Bonus points for screenshots and videos!
-->
<!--
Does this PR require a documentation change?
Create a PR at https://github.com/facebook/react-native-website and add a link to it here.
-->
<!--
Required.
Help reviewers and the release process by writing your own release notes. See below for an example.
-->
[ANDROID] [BUGFIX] [DevServerHelper] - fixing from calling body.string() twice.
<!--
**INTERNAL and MINOR tagged notes will not be included in the next version's final release notes.**
CATEGORY
[----------] TYPE
[ CLI ] [-------------] LOCATION
[ DOCS ] [ BREAKING ] [-------------]
[ GENERAL ] [ BUGFIX ] [ {Component} ]
[ INTERNAL ] [ ENHANCEMENT ] [ {Filename} ]
[ IOS ] [ FEATURE ] [ {Directory} ] |-----------|
[ ANDROID ] [ MINOR ] [ {Framework} ] - | {Message} |
[----------] [-------------] [-------------] |-----------|
EXAMPLES:
[IOS] [BREAKING] [FlatList] - Change a thing that breaks other things
[ANDROID] [BUGFIX] [TextInput] - Did a thing to TextInput
[CLI] [FEATURE] [local-cli/info/info.js] - CLI easier to do things with
[DOCS] [BUGFIX] [GettingStarted.md] - Accidentally a thing/word
[GENERAL] [ENHANCEMENT] [Yoga] - Added new yoga thing/position
[INTERNAL] [FEATURE] [./scripts] - Added thing to script that nobody will see
-->
Closes https://github.com/facebook/react-native/pull/19088
Differential Revision: D7853822
Pulled By: mdvacca
fbshipit-source-id: c11a73ce2fa6d40b0a7bd8bc0fca7b07c6bc27ed
Summary:
There is a race condition stemming from `ReactRootView.onCreate` that occasionally causes the initial root layout calculation to never occur. In this method currently, `updateRootLayoutSpecs(...)` is called before `enableLayoutCalculation()`, meaning that it's possible for the native modules thread to reach `UIImplementation.updateViewHierarchy` before layout calculation has been enabled (i.e. before the rootViewTag is added to `UIImplementation.mMeasuredRootNodes` on the UI thread).
When this occurs, `UIImplementation.applyUpdatesRecursive` is never called. This causes the app to hang on the splash screen indefinitely, the JS never gets past the first call to `render`, and no layout events are ever sent to the JS.
This PR reverses the order of those two method calls to ensure that the rootViewTag is always added to `mMeasuredRootNodes` before `UIImplementation.updateViewHierarchy` expects it. We inspected all of the code paths of both `enableLayoutCalculation()` and `updateRootLayoutSpecs()` to ensure that in this new order, the rootViewTag will *always* be added to `mMeasuredRootNodes` before the call to `updateViewHierarchy` is dispatched, and that no other side effects would be introduced.
Tested with an app that reliably had this issue (hanging splash screen) on 1 out of every ~10 launches. Logged values to ensure that empty `mMeasuredRootNodes` was the difference between failed and successful launches. After this change, I launched the same app 50+ times without any issues.
[ANDROID][BUGFIX][ReactRootView] - remove race condition in initial layout of ReactRootView
Closes https://github.com/facebook/react-native/pull/19038
Differential Revision: D7834270
Pulled By: hramos
fbshipit-source-id: 6040da900f807dcacbc86ae2c36b4ca004f80178
Summary: Adds the name of the active JS executor when logging the `CREATE_REACT_CONTEXT_START` marker.
Reviewed By: fromcelticpark
Differential Revision: D7800760
fbshipit-source-id: d5090a9f90cf4b7bcfb7218e75621becebd70675
Summary: adds a `getName()` method to all `JavaScriptExecutor` classes that can be used by `ReactInstanceManager` to identify the bridge type when logging markers.
Reviewed By: fromcelticpark
Differential Revision: D7790531
fbshipit-source-id: efe485588738a38936accc4a7956840784b2dd08
Summary:
This PR removes the need for having the `providesModule` tags in all the modules in the repository.
It configures Flow, Jest and Metro to get the module names from the filenames (`Libraries/Animated/src/nodes/AnimatedInterpolation.js` => `AnimatedInterpolation`)
* Checked the Flow configuration by running flow on the project root (no errors):
```
yarn flow
```
* Checked the Jest configuration by running the tests with a clean cache:
```
yarn jest --clearCache && yarn test
```
* Checked the Metro configuration by starting the server with a clean cache and requesting some bundles:
```
yarn run start --reset-cache
curl 'localhost:8081/IntegrationTests/AccessibilityManagerTest.bundle?platform=android'
curl 'localhost:8081/Libraries/Alert/Alert.bundle?platform=ios'
```
[INTERNAL] [FEATURE] [All] - Removed providesModule from all modules and configured tools.
Closes https://github.com/facebook/react-native/pull/18995
Reviewed By: mjesun
Differential Revision: D7729509
Pulled By: rubennorte
fbshipit-source-id: 892f760a05ce1fddb088ff0cd2e97e521fb8e825
Summary:
I found many crash reports of imagepipeline in our android app,
this is caused by fresco webp native memory leak, they have fixed it in v1.9.0
fixed wrong pull request of [18848](https://github.com/facebook/react-native/pull/18848)
Test with local build.
Check CI passes.
[ANDROID] [ENHANCEMENT] [Fresco/OkHttp] - Updates Fresco to 1.9.0, OkHttp to 3.10.0
Closes https://github.com/facebook/react-native/pull/18849
Differential Revision: D7651377
Pulled By: mdvacca
fbshipit-source-id: 0bfe79fd3d232b11a6a9380c961b9401c9e0ced9
Summary:
Switch the equality check to avoid crash on the first item. The check can be on a null object and return the correct result.
Fixes#18709
Just a simple switch on equals, to make sure we're not bombing out by having a null scheme.
No related PRs and does not require a document change.
[ANDROID][BUGFIX][BlobModule] Switch equality check in BlobModule.java
Closes https://github.com/facebook/react-native/pull/18893
Differential Revision: D7658036
Pulled By: hramos
fbshipit-source-id: db61b98dae178dbbb645070f7b0d73ab43d30541
Summary:
Fixes issue #11828 that causes layout animations for removed views to
remove some adjacent views as well. This happens because the animated
views are still present in the ViewGroup, which throws off subsequent
operations that rely on view indices having updated.
This issue was addressed in #11962, which was closed in favour of a more
reliable solution that addresses the issue globally since it’s difficult
to account for animated views everywhere. janicduplessis [recommended][0]
handling the issue through ViewManager.
Since API 11, Android provides `ViewGroup#startViewTransition(View)`
that can be used to keep child views visible even if they have been
removed from the group. ViewGroup keeps an array of these views, which
is only used for drawing. Methods such as `ViewGroup#getChildCount()`
and `ViewGroup#getChildAt(int)` will ignore them.
I believe relying on these framework methods within ViewManager is the
most reliable way to solve this issue because it also works if callers
ignore ViewManager and reach into the native view indices and counts
directly.
[0]: https://github.com/facebook/react-native/pull/11962#pullrequestreview-21862640
I wrote a minimal test app that you can find here:
<https://gist.github.com/lnikkila/87f3825442a5773f17ead433a810d53f>
The expected result is that the red and green squares disappear, a blue
one appears, and the black one stays in place. iOS has this behaviour,
but Android removes the black square as well.
We can see the bug with some breakpoint logging.
Without LayoutAnimation:
```
NativeViewHierarchyOptimizer: Removing node from parent with tag 2 at index 0
NativeViewHierarchyOptimizer: Removing node from parent with tag 4 at index 1
NativeViewHierarchyManager: Removing indices [0] with tags [2]
RootViewManager: Removing child view at index 0 with tag 2
NativeViewHierarchyManager: Removing indices [1] with tags [4]
RootViewManager: Removing child view at index 1 with tag 4
```
With LayoutAnimation tag 3 gets removed when it shouldn’t be:
```
NativeViewHierarchyOptimizer: Removing node from parent with tag 2 at index 0
NativeViewHierarchyOptimizer: Removing node from parent with tag 4 at index 1
NativeViewHierarchyManager: Removing indices [0] with tags [2]
NativeViewHierarchyManager: Removing indices [1] with tags [4]
-> RootViewManager: Removing child view at index 1 with tag 3
RootViewManager: Removing child view at index 2 with tag 4
(Animation listener kicks in here)
RootViewManager: Removing child view at index 1 with tag 2
```
Here are some GIFs to compare, click to expand:
<details>
<summary><b>Current master (iOS vs Android)</b></summary>
<p></p>
<img src="https://user-images.githubusercontent.com/1291143/38695083-fbc29cd4-3e93-11e8-9150-9b8ea75b87aa.gif" height="400" /><img src="https://user-images.githubusercontent.com/1291143/38695108-06eb73a6-3e94-11e8-867a-b95d7f926ccd.gif" height="400" />
</details><p></p>
<details>
<summary><b>With this patch (iOS vs Android, fixed)</b></summary>
<p></p>
<img src="https://user-images.githubusercontent.com/1291143/38695083-fbc29cd4-3e93-11e8-9150-9b8ea75b87aa.gif" height="400" /><img src="https://user-images.githubusercontent.com/1291143/38695137-1090f782-3e94-11e8-94c8-ce33a5d7ebdb.gif" height="400" />
</details><p></p>
Previously addressed in #11962, which wasn’t merged.
Tangentially related to my other LayoutAnimation PR #18651.
No documentation changes needed.
[ANDROID] [BUGFIX] [LayoutAnimation] - Removal LayoutAnimations no longer remove adjacent views as well in certain cases.
Closes https://github.com/facebook/react-native/pull/18830
Reviewed By: achen1
Differential Revision: D7612904
Pulled By: mdvacca
fbshipit-source-id: a04cf47ab80e0e813fa043125b1f907e212b1ad4
Summary:
Before, any calls to ReactContext would either use the default DevSupportManager's exception handler in debug mode OR throw exceptions immediately in non debug mode. In order to intercept these kinds of native exceptions, we should reuse the NativeModuleCallExceptionHandler provided by the ReactInstanceManager. For those who don't specify a NativeModuleCallExceptionHandler, the resulting behavior remains the same. For those who do specify a NativeModuleCallExceptionHandler,
the main difference is that it will now be responsible for handling exceptions for ReactContext.handleException
Reviewed By: mdvacca
Differential Revision: D7641772
fbshipit-source-id: 8f175df687723fcbb8a7620f90d8a08c94798738
