Summary:
The hermes profiler doesn't work currently, I tracked down the problem to a couple of things.
- Need to call `registerForProfiling` to enable profiling for a specific runtime. I added the call at the same place where we enable the debugger.
- `runInExecutor` didn't work and call its callback. Not sure exactly why, but using `executor_->add` like we do in a lot of other places to run code on the executor works.
- `GetHeapUsageRequest` seems to cause some deadlocks. JS contexts were not detected reliably, I suspect this is related to deadlocks when trying to run on inspector executor. `GetHeapUsageRequest` doesn't actually need any data from the inspector so there is no need to run it on that queue. To fix it I moved the call to use `runInExecutor` instead.
## Changelog
<!-- Help reviewers and the release process by writing your own changelog entry. For an example, see:
https://github.com/facebook/react-native/wiki/Changelog
-->
[General] [Fixed] - Fix hermes profiler
Pull Request resolved: https://github.com/facebook/react-native/pull/34129
Reviewed By: cortinico
Differential Revision: D37669469
Pulled By: philIip
fbshipit-source-id: 6cf3b2857ac60b0a1518837b9c56b9c093ed222f
Summary:
Current creation of the errorProcessor lambda does a full copy of folly::dynamic object, which for large objects can cause 1000's of memory allocations, and thus increasing app's memory footprint and speed.
Changelog:
[General][Fixed] - Avoid full copy of large folly::dynamic objects in JSIExecutor#defaultTimeoutInvoker
Reviewed By: sammy-SC
Differential Revision: D38368392
fbshipit-source-id: 88579a7069891828cf6dae130c4964db6b494565
Summary:
This Pull Request aims at removing the making of reactnativeutilsjni as it is built from the same sources as reactnativejni. It also replaces references to reactnativeutilsjni with reactnativejni.
This should get rid of `reactnativeutilsjni.so` while reusing `reactnativejni.so` in it's place. This should give us some size improvements in the finally built apk.
## Changelog
<!-- Help reviewers and the release process by writing your own changelog entry. For an example, see:
https://reactnative.dev/contributing/changelogs-in-pull-requests
-->
[Android] [Changed] - Replaced reactnativeutilsjni with reactnativejni in the build process to reduce size
Pull Request resolved: https://github.com/facebook/react-native/pull/34339
Test Plan:
1. Ran the CMakelist.txt file using CMake and I could see that reactnativeutilsjni.dir is no longer generated with my changes.
2. Built the aar from this branch in Android Studio and build happened successfully.
I am not sure if we could run any more tests. Please let me know in case anymore testing is required and I can do accordingly
Reviewed By: cortinico
Differential Revision: D38400481
Pulled By: genkikondo
fbshipit-source-id: 592736e56441328389ae89135667c336ff8018e6
Summary:
Changelog: [iOS][Internal] - Implement offsetX/offsetY properties on the PointerEvent interface
Simple diff implementing the offsetX/offsetY properties on PointerEvent — thankfully the touch events already kept track of these so it was mostly a matter of forwarding that data to the pointer events.
Reviewed By: lunaleaps
Differential Revision: D37830139
fbshipit-source-id: 77f33a99393350d32cbe449e6a009bdeb2a12d08
Summary:
Changelog:
[general][change] - Use RuntimeConfig instead of VM Experiment Flag to set up the micro task queue.
Reviewed By: neildhar
Differential Revision: D37353947
fbshipit-source-id: 5c8f35c0a79d70cb0d234e881e55058cffb44ac8
Summary:
Changelog: [iOS][Internal] - Implement screenX/screenY properties on the PointerEvent interface
This diff implements the screenX/screenY properties which report the position of a pointer in the device's physical screen coordinates.
Reviewed By: lunaleaps
Differential Revision: D37794415
fbshipit-source-id: 6c39c3651812f99e66b93647579a2935598ef6f2
Summary:
The RN Hermes Inspector keeps a mapping of pageId to
Connection. This mapping is created in the call to
enableDebugging. Later, when disableDebugging is invoked,
the inspector will iterate over the pageId to Connection map
to find which page to is being disabled.
The DecoratedRuntime enables debugging on construction, and
disables on destruction. When disabling debugging, the
DecoratedRuntime then passes the "naked" hermesRuntime,
which in turn gets checked against all connections' runtimes.
It turns our that ```HermesExecutorRuntimeAdapter::getRuntime```
returns the runtime **wrapped** by the DecoratedRuntime, and
not hermesRuntime itself.
This means that no connections match the request for
disabling debugging. Which means the Connection never
gets deleted, which in turns keep its underlying adapter (in
this case, the ```HermesRuntimeExecutorAdapter```) alive,
which has a shared_ptr to the runtime wrapped by the DecoratedRuntime.
Other than effectively leaking this connection (and thus the runtime),
this also causes an issue when the app tries to re-load the page -- at
which point the inspector will remove the old entry from the page to
connection mapping, thus initiating the destruction of the (leaked)
Runtime in a thread other than the thread that created the runtime.
Changelog: [Internal]
Reviewed By: javache
Differential Revision: D37809467
fbshipit-source-id: e73caa357a300a0d48e69aa3a1a8b00a97519f24
Summary:
Changelog: [iOS][Internal] - add x/y & pageX/Y implementations to PointerEvent interface
This diff adds the x/y properties, which are defined as aliases of clientX/clientY respectively. In addition this diff adds pageX/pageY which, while not definted as aliases of clientX/clientY, are effectively aliases in React Native since the root view is not scrollable (client and page points only differ on the web when a user has scrolled the document element).
Reviewed By: lunaleaps
Differential Revision: D37766818
fbshipit-source-id: c7ad3750460b1913889c6d1a55b4c1edc6918f5b
Summary:
Changelog: [iOS][internal] - Add twist property to PointerEvent interface
This adds the twist property to the PointerEvent implementation on iOS which similarly to tangentialPressure doesn't really apply so we default to a hard-coded value of 0.
Reviewed By: lunaleaps
Differential Revision: D37760511
fbshipit-source-id: f1fccfd6b5d07024cea83d86925a9bfc2e6cc8cf
Summary:
Changelog: [iOS][internal] - Add tangentialPressure property to PointerEvent interface
This diff adds the tangentialPressure property to the PointerEvent implementation on iOS. This one is pretty simple considering iOS doesn't have the concept of tangential pressure so we just hard code it to 0 as defined by the spec.
Reviewed By: lunaleaps
Differential Revision: D37759634
fbshipit-source-id: 7ca0e47267f5fde76ace2b96d05ea2e154cb4b8f
Summary:
Fix todo and inconsistency across codebase. It's better to have just one way to refer to these types.
Changelog: [Internal]
Reviewed By: philIip
Differential Revision: D37653146
fbshipit-source-id: e82f09caa6cd6eec5512b78f413708d9c04a7a83
Summary:
Changelog: [iOS][Internal] - Add `buttons` implementation to the PointerEvent interface
This diff adds an implementation of the `buttons` property by leveraging `UIEvent`'s `buttonMask` property.
Reviewed By: lunaleaps
Differential Revision: D37430270
fbshipit-source-id: 69fd3aebcb403e665349a24283a04c0eb82ff3e2
Summary: Changelog: [Internal] - If any relevant view events (pointer, touch events, gesture responder, etc.) are declared on view, then the view must form stacking context. We need this change for pointer events specifically to determine whether we've entered/exited a view
Reviewed By: vincentriemer
Differential Revision: D37678352
fbshipit-source-id: 02641549ef608b1c9468ac693c7da629143212cb
Summary: Changelog: [Internal] - We can now remove the '2' suffix as we had an internal implementation that was not truly aligned with W3C pointers but used the same name. We have aligned the internal types to match w3c so we can now remove the suffix that differentiates them.
Reviewed By: vincentriemer
Differential Revision: D37545813
fbshipit-source-id: 6f2336ae9e314066c340161113268c1f28621a71
Summary:
If you don't list `Cxx` as a supported platform, we will use Android builds, even when using `buck run` for local execution.
Changelog: [Internal]
Reviewed By: derolf
Differential Revision: D37600464
fbshipit-source-id: 6ba8566cde4180524351c9d8c647ce1d4ac5279d
Summary:
LA needs to ignore the new RemoveDeleteTree mutation coming from the differ.
Changelog: [Internal]
Reviewed By: jehartzog
Differential Revision: D37531217
fbshipit-source-id: c05b4106b6e955083e5e7e93619a13c4a2858404
Summary:
MSVC doesn't like Clang macros. Oops!
Constraints with this bit of code:
1. I'm trying to enforce constexpr in the most obvious, intuitive way possible. Macros are ugly but a "proper" solution would result in 10x as much code that is, totally subjectively, less readable to me.
2. This is evaluating at compile-time a hash of a string which is usually used in the `case` line of a switch statement.
For now I'm just hoping that MSVC will allow us to shadow `hash` which Xcode doesn't like. We might need to add a special pragma(s) for MSVC if it still doesn't like this.
Changelog: [Internal]
Reviewed By: lyahdav
Differential Revision: D37529949
fbshipit-source-id: 9aa605a9786bf5d194819ef8dade778875ae744e
Summary:
TL;DR: For applications using JS navigation, save 50-95% of CPU during mounting phase in N>2 navigations that replace ~most of screen.
During investigation of performance on the UI thread of React Native applications, I noticed that the /initial/ render of an screen for an application using JS navigation is /mostly/ consumed (on the UI thread) by tearing-down the previous View hierarchy. In one 185ms segment on the UI thread in production, 95% of the CPU time was Remove/Delete instructions and only 5% of CPU time was consumed by actually displaying the new hierarchy (this is specific to Android and also assumes that View Preallocation is being used, so post-commit work consists of Insert and UpdateLayout mutations primarily).
There are /some/ cases where the C++ differ knows that we are deleting an entire subtree and therefore we could communicate this to the mounting layer. All that matters is that these Views are removed from the View hierarchy immediately; and secondarily that their memory is cleaned up ASAP, but that doesn't need to happen immediately.
Some additional constraints and notes:
1) As noted in the comments, we cannot simply stop producing Remove and Delete instructions. We need to produce /both/ the new RemoveDeleteTree instruction, /and/ produce all the Remove/Delete instructions, primarily because LayoutAnimations relies heavily on these Remove/Delete instructions and certain things would break if we removed those instructions entirely. However, we can mark those Remove/Delete instructions as redundant, process them only in LayoutAnimations, and not send them to the Android mounting layer.
2) We want to make sure that View Recycling is not impacted. Since Android cannot take advantage of View Recycling until /after/ the second major render (preallocation of views will happen before any views are recycled), this doesn't impact View Recycling and we'll make sure Views are recycled whenever they are deleted.
Thus, we do two things:
1) Introduce a new RemoveDeleteTree operation that can delete an entire subtree recursively as part of one operation. This allows us to avoid serializing hundreds or thousands of instructions and prevents JNI traffic.
2) Besides removing the topmost View from the View hierarchy, and ensuring it's not drawn, the full teardown and recycling of the tree can happen /after/ the paint.
In some flows with JS navigation this saves us 95% of CPU during the mount phase. In the general case it is probably closer to 25-50% of CPU time that is saved and/or deferred.
Changelog: [Android][Changed] Significant perf optimization to Fabric Remove/Delete operations
Reviewed By: ryancat
Differential Revision: D37257864
fbshipit-source-id: a7d33fc74683939965cfb98be4db7890644110b2
Summary:
`getCurrentPriorityLevel` is a function which should return the priority level on invocation.
Changelog: [Internal]
Reviewed By: ryancat
Differential Revision: D37314727
fbshipit-source-id: fe385af02af49d1ca444beb881a4893f7f0712f0
Summary:
`asHostObject` should throw a `JSINativeException` like the other `as*`
functions. This should be a safe change to make in most cases, since
this function already did not work when JSI was being provided by a
dynamic library. However, I defer to mhorowitz.
Changelog: [Fixed] Throw JSINativeException from asHostObject
Reviewed By: jpporto
Differential Revision: D36873355
fbshipit-source-id: 589ac50ebb4ecec4bacfd923d8b0795c0a6c0e34
Summary:
This improves errors significantly, which is especially helpful as the runtime scheduler only implements a subset of the JS API.
Example error: `Error: Exception in HostObject::get for prop 'unstable_next': undefined property`
Changelog: [Internal]
Reviewed By: rickhanlonii
Differential Revision: D37313924
fbshipit-source-id: b53bc67b9cc36dee34dba86c07fdcf2353338c72
Summary: Add annotations to function parameters required for Flow's Local Type Inference project. This codemod prepares the codebase to match Flow's new typechecking algorithm. The new algorithm will make Flow more reliable and predicatable.
Reviewed By: evanyeung
Differential Revision: D37353648
fbshipit-source-id: e5a0c685ced85a8ff353d578b373f836b376bb28
Summary:
Bumping RTC-Folly version used to address CVE-2022-24440.
## Changelog
<!-- Help reviewers and the release process by writing your own changelog entry. For an example, see:
https://github.com/facebook/react-native/wiki/Changelog
-->
[General][Security] - Bump RTC-Folly to 2021-07-22
Pull Request resolved: https://github.com/facebook/react-native/pull/33841
Reviewed By: Andjeliko, philIip
Differential Revision: D36425598
Pulled By: cortinico
fbshipit-source-id: d38c5f020dbecf794b10f12ed2da30e1825071af
Summary:
`use_frameworks!` is broken again in react-native 0.69 because React-bridging. in the `use_frameworks!` mode, header structures are flattened, so `#include <react/bridging/CallbackWrapper.h>` is not reachable to the header. to somehow workaround the issue without touch React-bridging imports, the pr do these things:
- use `header_mappings_dir` to keep `react/bridging` header structure
- because the header structure is not default framework header structure, explicitly `HEADER_SEARCH_PATHS` is necessary.
- forward declare `CallbackWrapper` and use it internally in ReactCommon. so that we don't need to add `HEADER_SEARCH_PATHS` for React-bridging to every pods depending on `ReactCommon/turbomodule/core`, e.g. React-RCTSettings.podspec.
## Changelog
[iOS] [Fixed] - Fix use_frameworks! for 0.69
Pull Request resolved: https://github.com/facebook/react-native/pull/34011
Test Plan:
```sh
$ npx react-native init RN069 --version next
# add `use_frameworks!` to ios/Podsfile
# comment out use_flipper!() in ios/Podfile
# patch node_modules/react-native with these changes
$ yarn ios
```
Reviewed By: cortinico, cipolleschi
Differential Revision: D37169699
Pulled By: dmitryrykun
fbshipit-source-id: 309c55f1c611a2fc3902a83e8af814daaf2af6a0
Summary:
LogBox was using AppRegistry to render on to the screen. Switch LogBox over to using SurfaceRegistry instead.
Changelog: [Internal]
Reviewed By: sshic
Differential Revision: D37223641
fbshipit-source-id: 59001ad290c1e2c2f14828d38a96f48bd1ab39ca
Distiller
Distiller
Distiller
Kudo Chien
Janic Duplessis
Distiller
Distiller
Distiller
Christoph Purrer
Distiller
Sparsha Saha
Distiller
Vincent Riemer
Michael Anthony Leon
John Porto
Pieter De Baets
Genki Kondo
Kevin Gozali
Luna Wei
Ron Edelstein
Joshua Gross
Neil Dhar
Pieter Vanderwerff
Luis Santana
Ramanpreet Nara