Summary:
Now SurfaceTelemetry records the last 16 full transaction telemetries. We will use it info to report to our trackers.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: mdvacca
Differential Revision: D22887633
fbshipit-source-id: 0d88adff757e4bc5a701b51d4d06d85e1f51f10f
Summary:
This diff removes the inner folder of react/renderer/core, moving all its files into react/renderer/core
This is necessary to simplify the compilation of Fabric in OSS
More details: https://fb.quip.com/amaRA631DX3K
changelog: [internal] Internal
Reviewed By: fkgozali, JoshuaGross
Differential Revision: D22875854
fbshipit-source-id: e2d969c3ec67eab1bbdc9288e5a4285c740fa944
Summary:
If `__turboModuleProxy` is called with a second argument, we'll now forward that `jsi::Value` to TurboModuleManager on iOS and Android, so that the TurboModuleManager can eventually use this `jsi::Value` to read data required to perform method invocation on the TurboModule object.
**Note:** This diff is basically a no-op right now.
Changelog: [Internal]
Reviewed By: PeteTheHeat
Differential Revision: D22828838
fbshipit-source-id: 19db2adcae6a58b4885fcd11bef23f9d5882bfce
Summary:
This diff moves fabric C++ code from ReactCommon/fabric to ReactCommon/react/renderer
As part of this diff I also refactored components, codegen and callsites on CatalystApp, FB4A and venice
Script: P137350694
changelog: [internal] internal refactor
Reviewed By: fkgozali
Differential Revision: D22852139
fbshipit-source-id: f85310ba858b6afd81abfd9cbe6d70b28eca7415
Summary:
The async NativeModule method call event can fail if something goes wrong when we're about do dispatch the method call.
Changelog: [Internal]
Reviewed By: ejanzer
Differential Revision: D22609410
fbshipit-source-id: 699eeb78edfbefa85e417fd82318ae8e4f6f7d90
Summary:
In D17480605 (689233b018), I made all methods with void return types dispatch to the NativeModules thread. This diff makes the same change to methods with promise return types.
**Note:** The changes are disabled for now. I'll add an MC so that we can test this in production in a later diff.
Changelog:
[Android][Fixed] - Make promise NativeModule methods dispatch to NativeModules thread
Reviewed By: PeteTheHeat
Differential Revision: D22489338
fbshipit-source-id: d5b030871f9f7b3f48eb111225516521493cb05e
Summary:
Changelog:[Fabric][iOS] Add full support for AccessibilityState
Since the AccessibilityState checked, busy or expanded only exist in Android and they don't have corresponding AccessibilityTrait to set, so we could just ignore the implementation in RCTViewComponentView.mm. What I did is to update the AccessibilityState values in AccessibilityPrimitives.h and accessibilityPropsConversions.h in order to enable further implementation in Android.
Reviewed By: shergin
Differential Revision: D22807584
fbshipit-source-id: f3ef048055d11314bc833357d8ca061e0fe219a4
Summary:
Changelog: [Internal][Added] Support for toggling all breakpoints in Chrome debuggers
This implements the Debugger.setBreakpointsActive CDP message, allowing
users in Chrome to toggle all exceptions on and off.
As with V8/Chrome, setting a breakpoint will automatically re-activate
all breakpoints.
#utd-hermes-ignore-android
Reviewed By: avp
Differential Revision: D22825209
fbshipit-source-id: bda2cc59aba04443280bca46126c19bb0cdb58d7
Summary: As part of our usage color detection that relies on precise values of colors, it appears that the ration used was 256, but RBGA scale is 0-255
Reviewed By: shergin
Differential Revision: D22785378
fbshipit-source-id: 87a6f9b4611ceaadbbbb75a4566f4cf9b8285b68
Summary:
This diff reverts a revert of D22456266 (0060b5de55) and fixes the reason why it was reverted. Reverted in D22532594 (c0e7e1bd9c).
For reason why it was introduced in the first place, please refer to D22456266 (0060b5de55).
Problem:
`includeTransform` was not taken into account when calculating offset.
I added a unit test covering this specific scenario.
Reviewed By: shergin
Differential Revision: D22763011
fbshipit-source-id: e7d88fc19608ad1a4c7b5e594a9cc48122a2799b
Summary:
The standard says that zIndex should only be defined for non-`static` positioned views. This diff implements it.
For now, it actually enables zIndex for all views in RN because there is no way to specify `position: static` but will will give that ability by changing Flow definitions in future diffs in a couple of weeks (to ensure OTA safety).
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D22098586
fbshipit-source-id: 77dacbee0fe887f667ba81640b8bd759e1df11fd
Summary:
In W3C standard, `zIndex` prop can have `auto` value which is the default.
In classic React Native and in Fabric before this diff, the default value was `0`. The only difference between `auto` and `0` is that nodes with `auto` do not form stacking context whereas nodes with `0` (or any other numeric value) do. This worked fine in pre-Fabric RN where every view always formed staking context but in Fabric we need to finally implement it right.
https://developer.mozilla.org/en-US/docs/Web/CSS/z-indexhttps://stackoverflow.com/a/57892072/496389
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D22403383
fbshipit-source-id: 6532d180a00f22bde763e3ebadd6683b344a0672
Summary:
In the current Fabric model, we compute layout during the commit phase on the caller thread synchronously. Even though, in general, it's by design the correct way to do it, there are cases where it's *not* a requirement. In such cases, it's more optimal to yield early and continue execution of the commit process on a different thread asynchronously.
One of such cases potentially is `completeRoot` call. There we don't need to return anything and can resume JavaScript execution immediately.
The performance implications of that are not clear but there is a hope that it can free up to ~100ms of JavaScript execution time which is currently spent waiting for layout calculation (and other aspects of the commit phase).
This is an implementation in the core. The plan is to test that on iOS first and then, in case if the results of the experiment are positive, to implement it on Android as well.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: JoshuaGross
Differential Revision: D22743723
fbshipit-source-id: 846a13152c5a419de0eaef0e6283ea4277c907dc
Summary: For the JS TurboModule Codegen, we'll have to have `__turboModuleProxy` accept an additional argument: information containing what methods are supported on the TurboModule object. This diff makes calling `__turboModuleProxy` with two arguments valid.
Reviewed By: shergin
Differential Revision: D22743294
fbshipit-source-id: fd0050fc0373b43dc7089695c8e341137cad21f1
Summary:
Unfortunately LayoutAnimations index adjustment is a bit tricky. There was one (hopefully only one!) remaining issue where a specific series of delayed removes, inserts, and removes would trip up index adjustment. Namely in this case: Preexisting delayed removes, and then a later animated commit with immediate inserts followed by higher-index removes.
This case is now resolved and I took time to test and verify other index adjustment paths.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D22709915
fbshipit-source-id: 16ba5bb358608d384f44628dbb6cc691deb8164a
Summary:
Changelog: [Internal] Fabric-specific internal change.
This diff introduces a new value for `YGPositionType`: `YGPositionTypeStatic`.
No part of Yoga, RN, Litho or CK uses this value yet. `relative` and `static` values behave the same way for now. We also do not change any defaults. So, it should be fine.
Reviewed By: SidharthGuglani
Differential Revision: D22386732
fbshipit-source-id: 39cd9e818458ac2a91efb175f24a74c8c303ff08
Summary:
Chrome recently changed schema from 'chrome-devtools' to just 'devtools',
this change updates Hermes inspector cli tool usage string to reflect that change
Changelog: Minor fix in Hermes Inspector cli tool help message
Reviewed By: dulinriley
Differential Revision: D22557806
fbshipit-source-id: 95ec9cbaac445e105e7e92aec2b6c4e5a7d7924f
Summary:
This diff implements TelemetryController, a small tool that can be used to instrument mounting transactions. It abstracts the logic of merging telemetry data of multiple transactions in a thread-safe manner.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D22490580
fbshipit-source-id: 3f3425b88d38fddb555c1390fd8f1ff3ef1c475a
Summary:
SurfaceTelemetry data structure represents telemetry data associated with a particular running Surface. We need it to aggregate data from multiple mount transaction.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D22490579
fbshipit-source-id: e087aa9bc236af4da7010d67941f9ac93ad6da97
Summary:
Changelog:
[Internal][Fixed] - Fix the accessibilityRole crash
I only assigned three values that would be used for embedded links inside text (button, link, image). But the accessibilityRole has a lot of values that may be applied in other places. (https://reactnative.dev/docs/accessibility) To avoid this crash, need to add all possible values to the enum.
Reviewed By: JoshuaGross, shergin
Differential Revision: D22549264
fbshipit-source-id: dbaacf1965624e1d2eabe193b46168c6c8740f46
Summary:
Changelog:
[Internal] - Revert "Change the accessibilityRole from a string to an enum."
Original commit changeset: b46ce0bd156c
Reviewed By: PeteTheHeat
Differential Revision: D22540391
fbshipit-source-id: df4a62c7bb4525782121dea2c70a9101cf0d8e9f
Summary:
Now Yoga.cpp does not use the `YGPositionTypeRelative` value/constant, it uses `YGPositionTypeAbsolute` instead.
Now `YGPositionType` can only be `YGPositionTypeRelative` or `YGPositionTypeAbsolute`, so expressions `x == YGPositionTypeRelative` and `x != YGPositionTypeAbsolute` are equivalent.
The reasoning behind the change is that in all cases we actually check a node to be (or not to be) `absolute`, not `relative`. This will make a difference in the coming diffs in the stack when we will introduce a new value for the type: `static`.
We need to differentiate `static` and `relative` values t implement the `stacking context` feature in the W3C-compliant way (to fix bugs and avoid developer confusion). Read more here:
https://developer.mozilla.org/en-US/docs/Web/CSS/position
Changelog: [Internal] Internal change in Yoga.
Reviewed By: SidharthGuglani
Differential Revision: D22386733
fbshipit-source-id: 8e2c2b88b404660639f845783c8f93f0c62c0fbb
Summary:
During setup, TurboModules may synchronously dispatch to the main queue, if they require main queue setup. This is dangerous because it could cause the app to deadlock during TurboModule require. This is why D21654637 (e206e34175) added a warning aginst this. However, this diff had a mistake. We only want to display the warning if TurboModule eager initialization is enabled, because then, we can eagerly initialize the TurboModules before the bridge starts to avoid the problem. D21654637 (e206e34175) instead showed the warning if TurboModule eager init **wasn't** enabled. This isn't useful, because there's no way to avoid the problem with TurboModu eager initialization off.
Changelog: [Internal]
Reviewed By: sammy-SC
Differential Revision: D22529780
fbshipit-source-id: 15238483758b66b1a6addcad948203c64dca96ad
Summary:
Changelog:
[Internal] - Change the accessibilityRole from a string to an enum.
The change aims to save storage especially when scaling. Also, the enum makes it more difficult to input wrong because it gets checked when compiling.
Reviewed By: sammy-SC
Differential Revision: D22396351
fbshipit-source-id: b46ce0bd156c7705020ef5b061d8ac29e2cf3948
Summary:
Changelog: [Internal]
When calling "measure" on ScrollView's children, origin needs to be adjusted for ScrollView's content offset.
For this scrollView uses `getTransform` to adjust frames of its children.
This is wrong because transform is applied to ScrollView as well.
Example:
ScrollView is scrolled 900 points and its origin is {0, 0}, if you call "measure" on the ScrollView, our current measure infra will report its origin being {0, 900}.
Reviewed By: shergin
Differential Revision: D22456266
fbshipit-source-id: 6fd54661305ad46def8ece93fcf61d66817b3e01
Summary:
Changelog: [Internal]
Add view hierachy drawings to tests to make it easier to picture view hierarchy.
The sketches do not reflect sizing but relationship among the views.
I removed unnecessary reset of transform value to identity matrix.
Reviewed By: JoshuaGross, shergin
Differential Revision: D22456267
fbshipit-source-id: 480d0b938ffd0281fc94148570c412b0fcc22f42
Summary:
TurboModule eager initialization is a bit dangerous if we get it wrong, which we did (twice): T69449176.
This diff gates TurboModule eager init behind a MC, so that we can control (i.e: turn off/on, and do gradually rollout of) TurobModule eager initialization in isolation from the larger TurboModules experiment.
Changelog: [Internal]
Reviewed By: PeteTheHeat
Differential Revision: D22460359
fbshipit-source-id: 3b8dce0529f1739bd68b8b16d6a28aa572d82c2c
Summary:
When you require a TurboModule on thread that isn't the main thread, but the TurboModule requires main queue setup, we are forced to `dispatch_sync` the set up to the main queue. This is hazardous, because it can lead to deadlocks. Therefore, I'm migrating over a warning from the legacy infra that warns against this use-case.
Changelog: [Internal]
Reviewed By: fkgozali
Differential Revision: D21654637
fbshipit-source-id: 99dc45708c533603d2111fe6163d40e807d2a513
Summary:
The changes in this diff are inconsequential. Just thought it was weird that TurboModuleManager was responsible for assigning itself to the bridge. Seems like that should be done by an entity that owns both the bridge and the TurboModuleManager.
Changelog:
[Internal]
Reviewed By: PeteTheHeat
Differential Revision: D22406172
fbshipit-source-id: cb32f097f4d377a3b85a6d165b7b06be8e6a185b
Summary:
## Why?
1. RCTTurboModuleLookupDelegate sounds a bit nebulous.
2. In JS and Java, we use a `TurboModuleRegistry` interface to require TurboModules. So, this diff will make JS, Java, and ObjC consistent.
Changelog:
[Internal]
Reviewed By: PeteTheHeat
Differential Revision: D22405754
fbshipit-source-id: 30c85c246b39d198c5b8c6ca4432a3196ca0ebfd
Summary:
## Context
1. In FBReactModule jsExecutorForBridge, we asynchronously initialize a list of TurboModules on the main queue: https://fburl.com/diffusion/i56wi3px
2. After initializing the bridge, we start executing the JS bundle, here: e23e9328aa/React/CxxBridge/RCTCxxBridge.mm (L414-L417). Since bridge initialization knows nothing about TurboModule eager initialization, this happens concurrently with 1, and starts requiring NativeModules/TurboModules on the JS thread.
## The Race
1. Both the main thread and the JS thread race to create a TurboModule that requires main queue setup.
2. The JS thread wins, and starts creating the TurboModule. Meanwhile, the main thread blocks, waiting on a signal here, in RCTTurboModuleManager: e23e9328aa/ReactCommon/turbomodule/core/platform/ios/RCTTurboModuleManager.mm (L430)
3. The JS thread tries to dispatch_sync to the main queue to setup the TurboModule because the TurboModule requires main queue setup, here: e23e9328aa/ReactCommon/turbomodule/core/platform/ios/RCTTurboModuleManager.mm (L402)
4. We deadlock.
## The fix
Succinctly, NativeModule eager initialization finishes before execute the JS bundle, but TurboModule initialization doesn't. This diff corrects that mistake.
The changes in this diff:
1. The RN application via the TurboModuleManager delegate can now optionally provide the names of all eagerly initialized TurboModules by implementing two methods `getEagerInitModuleNames`, `getEagerInitMainQueueModuleNames`.
2. The TurboModuleManager grabs these two lists from the delegate, and exposes them to its owner via the `RCTTurboModuleRegistry` protocol.
3. The RCTCxxBridge, which already owns a `id<RCTTurboModuleRegistry>` object, uses it to eagerly initialize the TurboModules in these two lists with the correct timing requirements.
This is exactly how we implement eager initialization in Android.
**Note:** Right now, phase one and two of TurboModule eager initialization happen after phase one and two of NativeModule eager initialization. We could make the timing even more correct by initializing the TurboModules at the exact same time we initialize the NativeModules. However, that would require a bit more surgery on the bridge, and the bridge delegate. I think this is good enough for now.
Changelog:
[iOS][Fixed] - Fix TurboModule eager init race
Reviewed By: PeteTheHeat
Differential Revision: D22406171
fbshipit-source-id: 4715be0bceb478a8e4aa206180c0316eaaf287e8
Summary:
Currently the `remove` API returns an owned unique_ptr of the removed ShadowTree but it's not used anywhere, so we can simplify the API.
Because of that change, we can make the API safe even if the SurfaceId has already been removed.
For context, on Android there is a race between RootView.unmountReactApplication and C++ teardown which removes all SurfaceIds. This currently causes a crash, but after this diff the 2nd call to `remove` for a given SurfaceId will noop.
Changelog: [Internal]
Reviewed By: sammy-SC
Differential Revision: D22416471
fbshipit-source-id: dbba44c276aab8e81097b92a89e0becdcb7b28ba
Summary:
We are about to add several additional functions to this file, so we need to split it to keep it manageable. All changes in this diff are cosmetical.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D22384429
fbshipit-source-id: d22592fba17ef04edb44388b4e1ff22112e9c020
Summary:
Changelog:
[Internal] - Add `accessibilityRole` to `RCTAttributedTextUtils`. AccessibilityRole was in `TextAttributes` which is a Fabric's abstraction so cannot be detected when enumerating attributedString. Mapping accessibilityRole from `TextAttributes` to `NSAttributedString` could provide the attributeName when iterating over attributedString and then successfully find the range of the fragment whose `accessibilityRole` has value @"link".
Reviewed By: shergin
Differential Revision: D22286747
fbshipit-source-id: eb039d6a35e77d1860f86ba287391ccb56fbe7b5
Summary:
Changelog:
[iOS][Added] - `getRectWithAttributedString()` aims to get the rect of the fragment with embedded link, which is necessary when building the `accessibilityElement`. In this function, we first enumerate attributedString to find the range of fragments whose `accessibilityRole` is @"link". Then we calculate the rect of the fragment and send to the block and we would define what to do in the block in `RCTParagraphComponentAccessibilityProvider`.
Reviewed By: shergin
Differential Revision: D22286733
fbshipit-source-id: 4d11cb54375a4e9e72869e646fcb484de089b14b
Summary:
ART has been moved out of OSS, but I created the fabric implementation in github repo, this diff migrates the Fabric ART implementation to FB repository again
changelog: [Internal]
Reviewed By: shergin
Differential Revision: D22352930
fbshipit-source-id: 1125d3f27b98be783529134de0d3fbcf1b01a6f4
Summary:
Changelog:
[Internal] - The accessibilityRole will be used to specially distinguish normal fragments and fragments with embedded link, which is helpful when building accessibilityElement.
According to this document: https://reactnative.dev/docs/accessibility, the accessibilityRole is a common accessibility props that could be inherited, and this is why I add accessibilityRole to the TextAttributes.
Reviewed By: sammy-SC
Differential Revision: D22249140
fbshipit-source-id: 408b4415bf44539d8671d3d98f1ec06f8035baf6
Summary:
Changelog: [Internal]
FindNoteAtPoint takes transformations into account. There is however a problem with how transformation in ScrollView. In ScrollViewShadowNode, getTransform is overriden. In the override we apply `Transform::Translate(-contentOffset.x, -contentOffset.y, 0)`, this means that ScrollView has moved by {-x,-y}. However this is only true for its children, not for scroll view itself. This trips off findNodeAtPoint and if you scroll more than than screen's height, point will not be evaluated as inside ScrollView.
Until we can figure out how to deal with this properly, I think it is better to disable it as usage of JS Inspector is more common in ScrollViews than it is with transformed views.
Reviewed By: shergin
Differential Revision: D22332779
fbshipit-source-id: f48c9ae67a595e6954c2b70fb287db7f8c74378b
Summary:
This was an experiment that never shipped. In the meantime we built Fast Refresh which obviates the need to complicate how we load and store bundles on device.
Changelog: [Internal]
Reviewed By: motiz88
Differential Revision: D22330994
fbshipit-source-id: 5a623b2611dd2622f17dd83ed35ef05c3100e40d
Summary:
This introduces a new shadow node trait (`Hidden`) which disables emitting ShadowVides for such nodes and all their descendants.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D22134220
fbshipit-source-id: 8ea1fc3186019005ad687eacaca0ba4a4d8343dc
Summary:
The Hermes inspector always logged the message for `console.assert`.
It instead should check the first argument, and only if that argument
is false should it log.
Also remove the polyfill code that tried to avoid this situation.
Changelog: [Internal]
Reviewed By: mhorowitz
Differential Revision: D22299186
fbshipit-source-id: cdf4f8957d4db3d171d6673a82c7fc32b7152af3
Summary:
It's a Yoga specific feature that does not need to exposed as LayoutableShadowNode.
Removing this we save many virtual calls and add simplicity.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D22304717
fbshipit-source-id: f2dda8309dd86b78cd2775971721d29e5317ffd5
Summary:
`Overflow inset` is a field in LayoutMetrics that describes the external bounding box that covers area formed by all descendants of the view (that might stick out of the frame of the view).
This information can be used for various optimizations across all rendering pipeline. Here is some of them:
Improving hit-testing performance.
Implementing a robust "remove invisible native views" feature.
Improving `clipToBounds` performance.
Improving View Flattening versatility.
Most importantly, we need this for improving <ScrollView> performance (and memory footprint) which we have to do to match parity with Paper's `removeClippedSubviews` feature.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D22297325
fbshipit-source-id: cb238f0505d11ccabbe9db26f36a98b3172c70db
Summary:
After years of trying to design a somehow decomposed and universal layout API, I think we chased a wrong goal. None of the solutions that I tried helped with simplicity, composability, performance, or readability. Besides that, a bunch of execution primitives we use (e.g. `setHasNewLayout`) are heavily influenced by Yoga and are far from being commonly applicable. Finally, we ended up with a situation where the current design does not allow us to implement the features we want (more about that in coming diffs).
The diff changes that. Now we have only one method that implements layout - `layout()`; all possible implementations are free to implement it in any way that makes sense for a particular approach. And, I think, even for the base case - Yoga powered layout - things are much simpler and faster now: it's easy to comprehend how a single method works and now we don't have two loops and an expensive call to `getLayoutableChildNodes`.
But most importantly, after this change will be able to implement the `Inset Overflow` feature quite easily.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: JoshuaGross
Differential Revision: D22297317
fbshipit-source-id: f8662b7e8e3b522bdd6e5b2b0ac6a06efb45be6d
Summary:
Changelog: [Internal]
Introducing InputAccessoryView.
There is one big difference between Fabric's implementation and Paper's implementation.
Fabric searches for text input from InputAccessoryView, unlike Paper where it is the other way around.
Reviewed By: shergin
Differential Revision: D22160445
fbshipit-source-id: 55313fe50afeced7aead5b57137d711dd1cfd3ae
Summary:
View elevation is an Android-only prop, and it's a float, not an int. https://www.codota.com/code/java/methods/android.view.View/setElevation
Changelog: [Internal]
Reviewed By: shergin
Differential Revision: D22298431
fbshipit-source-id: 91edc3aa64e7254a5c4a4cb33f26e64ebae89b8d
Summary:
Here is why:
* It was with us from the very beginning but we never use it.
* The main purpose of this - snap-to-pixel layout - was moved to Yoga, where it should be.
* The current implementation has a bug.
* It's not really correct conceptually because the value becomes incorrect when an immutable subtree is being reused as part of a new tree.
* It over-complicates a new feature I am working on.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D22284645
fbshipit-source-id: c4c2df8d24e8fe924725b465e04e8154d097d226
Summary:
A function in graphics/geometry which adjusts a rectangle by the given edge insets.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D22284644
fbshipit-source-id: ad78abd7889b4457c450b2cc43fb73054aef2c79
Summary:
Same as D19583582 (5b2ea6ec6a).
We need this for implementing some tests.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D22284646
fbshipit-source-id: 9dd6575475c133b9426838021af4d59fb9fc3c65
Summary:
Attempt to fix T68951888 where (1) ComponentHandle is zero, and/or (2) ComponentHandle is missing in the registry. Either will cause a crash and both should be trivial to work around. I was able to repro once accidentally in about 1/200 sessions, so I am not 100% sure if this fixes the root cause.
Changelog: [Internal]
Reviewed By: shergin
Differential Revision: D22216030
fbshipit-source-id: b6986adee6fe739ce58579a2b031a2d252e73e35
Summary:
Previously, we tried to take a Transform matrix, decompose it into parts, and then interpolate between those parts. This will always be risky at best, and in some cases ambiguous or unsolvable. For example, a scale of -1 is identical to a rotation of 180 degrees.
Another issue is that when decomposing a matrix, it is impossible to tell the sign of scaleX, scaleY, scaleZ. This is a problem - flipping a View over an axis via scale then becomes a non-animatable operation.
This caused a number of issues.
To resolve it, we accumulate the "operations" resulting in a particular transform. This allows us to easily interpolate between two Transform matrices without actually decomposing the matrix, since we have the "path" that resulted in each particular matrix.
This will make LayoutAnimations over transforms, including Skew transforms, look and work much better, and more reliably.
Changelog: [Internal]
Reviewed By: shergin
Differential Revision: D22204559
fbshipit-source-id: 0d88ae77e4399a7ea333afbf6062beea977b854a
Summary:
Changelog: [Internal]
When Paper loads images. it calls with with `clipped` being set to NO.
https://github.com/facebook/react-native/blob/master/Libraries/Image/RCTImageView.mm#L346
We should do the same in Fabric as it can result in different images being loaded.
Reviewed By: JoshuaGross, shergin
Differential Revision: D22185286
fbshipit-source-id: 04afafee67fdfac53e9d17f5edeaf9b9317c7a22
Summary:
This diff refactors the theme management for text input in order to avoid extra state updates.
changelog:[Internal]
Reviewed By: JoshuaGross
Differential Revision: D22149754
fbshipit-source-id: 8a6dbe63c8d532986dbf785c7b16323e0a980669
Summary:
Index adjustment is complicated since we defer REMOVE operations. This diff solves some remaining (hopefully final!) bugs found when queueing several conflicting LayoutAnimations repeatedly (this is sort of an "extreme" case since we're continuously re-queueing additional LayoutAnimations instead of allowing any to complete, while reordering and deleting, so it exercises the most extreme LayoutAnimations use-cases).
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D22150085
fbshipit-source-id: db1381bef3ba8bb5ec82dab791f69162d46d23f0
Summary:
This will prevent views from becoming out-of-order as view removals are delayed and there are inserts at the same level.
There is at least one additional issue that crops up if many animations are queued up at the same time, that will be resolved separately.
Changelog: [Internal]
Reviewed By: sammy-SC
Differential Revision: D22130185
fbshipit-source-id: d8041c6afdcd729939dd392c9e2c23fe8ec1b706
Summary:
Changelog: [Internal]
Order of events dispatched to JS matter. If we remove duplicate events, we shouldn't change order of types.
So If we have order of events
A -> B -> C
and inserting a new event A for the same target, we shouldn't delete the first A, as there might be logic that depends on event B being preceded by A.
What we can do though, is delete previous event of the same type and target in queue if we would be adding subsequent event.
So when we have events
A -> B -> C
and we are about to add event C to the queue which goes to the same target, we can delete the first C before inserting the new one.
Reviewed By: shergin
Differential Revision: D22135931
fbshipit-source-id: 1dadd1688516330be07c4251f446f77ca08eaa87
Summary:
When asking for the data of a JSBigFileString, check that the instance satisfies some basic invariants.
This is meant to catch any corruption issues as early as possible.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D22125436
fbshipit-source-id: e0a84752c86151d56b7e7cbed3b95650d8ba1f75
Summary:
Add tooltip to the list of components supported with the legacy view manager, so that it will work in Fabric.
Changelog: [Internal]
Reviewed By: PeteTheHeat
Differential Revision: D21917137
fbshipit-source-id: 1ab6c7cdd0a10a496531d2ea8fbcae7e427ec647
Summary:
Changelog: [Internal]
Add support for dynamic font size.
New class `ThreadStorage` is introduced, which is used to pass LayoutContext to `YogaLayoutableShadowNode::yogaNodeMeasureCallbackConnector`.
## Shortcoming
This implementation doesn't cause re-render, if user changes font size and comes to the app without restarting it, it will show old font size. I believe this is fine for now as most people set their font size before they use the app and keep the same setting for a long time.
Reviewed By: shergin
Differential Revision: D22043728
fbshipit-source-id: 7453d165c280a2f4bcb73f4ee6daf9e64b637ded
Summary:
Changelog: [internal]
Rename `LayoutableShadowNode::measure` to `LayoutableShadowNode::measureContent` and add LayoutContext as parameter.
Pass `LayoutConstraints` by reference rather than value.
Reviewed By: shergin
Differential Revision: D22043727
fbshipit-source-id: b668240c45b658db5b114630b73d7407d35482aa
Summary:
Changelog: [Internal]
A long time ago we experimented with JSC bytecode. We are not experimenting with JSC bytecode any more. This code can be removed.
Reviewed By: mhorowitz
Differential Revision: D22017374
fbshipit-source-id: 6fe3fb7ad7966f92a5cd103605ac5c0bd1f17a8e
Summary:
Fix crashes in LayoutAnimations:
1) It is valid to omit create, update, delete configs
2) When extracting SRT from a matrix, ignoring skew properties
3) Provide valid telemetry from LayoutAnimations transactions
Unrelated to crashes: to help debugging and until onSuccess/onError callbacks are working, log any configuration parsing errors.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D22050736
fbshipit-source-id: e59418ecad0f9bfd20a2b4976557e39020c2d101
Summary:
This diff fixes a race condition that caused a flicker during initial rendering of TextInput in Fabric
The root cause is that the TextInput's State is sometimes initialized with no information from the Theme, this causes text input to be rendered with 0 padding. Later ReactTextInput manager updates TextInput state with theme data and the TextInput is re-rendered with the correct padding information.
changelog: [Internal]
Reviewed By: sammy-SC
Differential Revision: D22034849
fbshipit-source-id: a2fa91f34a8340878f2ec8d231ef6c86cee08f05
Summary:
This pull request updates the Podspecs to reflect recent changes to Fabric, so they build with Fabric enabled again.
In `React-Fabric.podspec`, subspecs for the views `legacyviewmanagerinterop`, `safeareaview`, `textinput`, and `unimplementedview` are added. A minor tweak to the subspec for `rncore` was added blacklisting two codegen-generated files that have issues building, but are not actually needed yet and are empty.
In `React-graphics.podspec`, the header paths for Boost were added.
This is partially a retread of https://github.com/facebook/react-native/issues/26805, which had to be backed out due to issues with the changes to the Yoga podspec.
## Changelog
[Internal] [Fixed] - Fixed building Fabric with the Cocoapods
Pull Request resolved: https://github.com/facebook/react-native/pull/28584
Test Plan: RNTester can now compile.
Reviewed By: shergin
Differential Revision: D20966065
Pulled By: hramos
fbshipit-source-id: 73931be63b72fcb52705643f81c0f787e571eb76
Summary:
Changelog: [Internal]
Round calculated text size to closest larger pixel size.
In Paper we do this as well in https://fburl.com/diffusion/w2pj6ea0
Why do we need this?
For example, we calculate height of the text to be 16.707235, yoga takes this value and assigns text to have height 17 anyway.
I assume Yoga uses this extra 0.3 points of height to move other things around a little bit because Yoga doesn't deal with exact values.
Reviewed By: shergin
Differential Revision: D21999032
fbshipit-source-id: 73923dc3e27fa110adb3f76cfa635e0bfdc672d4
Summary:
There is no good reason why they are `const` and `private`, in the future diffs we will need to mutate them.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D21992197
fbshipit-source-id: f8aee8db9ea1ff8282b38fbe3a966911678ee2c4
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/29090
Moving the logic for calling into JS to handle errors into ErrorUtils, where it can be reused outside of the bridge.
Reviewed By: RSNara
Differential Revision: D21939254
fbshipit-source-id: 0d8f3bd2503720be7619ed8dc8b2389f544049f3
Summary:
While I was investigated the previous issue (T59430348) I found this place where we manipulate managed pointers manually. So, to make it error-prone this diff changes it to use `wrapManagedObject`.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: JoshuaGross
Differential Revision: D21943010
fbshipit-source-id: ecaeffb419ae8d4880187027ca7ec9563e0dfd46
Summary:
ManagedObjectWrapper (`wrapManagedObject`) is used to pass pointers managed by Objective-C runtime object thought C++ parts of the framework. The wrapper consists of a shared pointer to an object with a custom deleter that calls `CFRelease`.
Apparently, there is a caveat here: shared ptr implementation always calls a deleter even if the object points to `nullptr`. It's usually fine because in C++ calling `delete` on a nullptr is a no-op, not an error. But it's an error from the `CFRelease` perspective, which checks the pointer and crashes if it's nullptr.
More info:
https://stackoverflow.com/questions/1135214/why-would-cfreleasenull-crashhttps://stackoverflow.com/questions/50201967/why-unique-ptr-with-custom-deleter-wont-work-for-nullptr-while-shared-ptr-does
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D21943011
fbshipit-source-id: 442ad5e274a146de112e6bd8f3c2d20f0225bf77
Summary:
Changelog: [Internal]
`CFRelease` crashes app if `cf` is `NULL`, `CGColorRelease` doesn't.
From:
https://developer.apple.com/documentation/coregraphics/1586340-cgcolorrelease?language=objc#
Even though this hasn't happened yet, we had a similar crash in D21943011. Also `CGColorCreate` returns nullable so it could happen.
In order to prevent this in the future, let's switch to `CGColorRelease` which doesn't crash if `cf` is `NULL` and also is semantically correct.
Reviewed By: shergin
Differential Revision: D21953404
fbshipit-source-id: 8b969ee345c2507fcba2442fa20e3fdaf2235d8c
Summary:
Changelog: [Internal]
View gets flattened even though it has `display: none` and therefore it and its children do not get hidden.
Reviewed By: shergin, mdvacca
Differential Revision: D21929033
fbshipit-source-id: 994a79fb64fbe66273a70218ebe8056d92cd3cd4
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/29087
D21908523 added an implicit dependency on `jsi.h` to use functions like `asObject`, etc. For some reason this doesn't break the build with BUCK (??) but it does with cocoapods. Adding the dep to the cxxreact podspec and regenerating offline mirrors to unbreak CircleCI. Also adding the BUCK dep and include statement for good measure.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D21924592
fbshipit-source-id: 295c0670c6499e1195ba3c3a3320c6aee13bc025
Summary:
Changelog: [Internal]
The crash is caused by dereferencing invalid pointer.
This can happen because `UIManager` can outlive `RCTScheduler` and `Scheduler`.
Here we make sure when `Scheduler` is being deconstructed, UIManager's pointer is invalidated.
I don't think this is ideal solution but it should fix the crash.
Ideally we want the owner of animation delegate to invalidate the pointer.
Reviewed By: JoshuaGross
Differential Revision: D21922910
fbshipit-source-id: b2a56c1104574cecebaffad1bcbcbff82c1fa0cf
Summary:
Call into our existing JS error handling logic from C++ using JSI in the RuntimeExecutor used by the bridge.
Changelog: [Internal]
Reviewed By: lunaleaps
Differential Revision: D21908523
fbshipit-source-id: ae41196443781b9f2673dcb7bbcb5b5aa8aa2528
Summary:
Changelog: [Internal]
# Problem
`MountingCoordinator` holds a pointer to instance of `MountingOverrideDelegate` which becomes invalid.
# Solution
Use `std::weak_ptr` instead of raw pointer so it is possible to tell whether the pointer is expired.
Reviewed By: JoshuaGross
Differential Revision: D21905351
fbshipit-source-id: c7bf9635742a6ec086a03ba83202e46e1f1f373f
Summary:
This diff adds support elevation prop in Fabric core, additionally it adds this prop as non collapsable on the view flattening algorithm
changelog: [Internal] internal change in fabric to support elevation prop
Reviewed By: JoshuaGross
Differential Revision: D21896465
fbshipit-source-id: e0854acc0b2ac30eaf3f82d615aab1cf378cc530
Summary:
JSError creation can lead to further errors. Make sure these cases
are handled and don't cause weird crashes or other issues.
This solution has a few parts:
* include a ScopedNativeDepthTracker in checkStatus
* If an exception object message or stack property is already a String, don't
call JS String ctor on it
* Verify that a jsi::Value is a String before calling getString on it.
* Add more tests for JSError construction
Changelog: [Internal]
Reviewed By: dulinriley
Differential Revision: D21851645
fbshipit-source-id: 2d10da0e741ad4ede93cd806320f68ad512e5138
Summary:
There are two ways to make a NativeModule method execute synchronously:
- Declare the NativeModule method to be synchronous (i.e: use `RCT_EXPORT_BLOCKING_SYNCHRONOUS_METHOD`).
- Make the NativeModule synchronous (i.e: make its method queue `RCTJSThread`). This way, all its methods are synchronous.
`RCTNativeModule` executes all synchronous methods on the JS thread:
- Executing an async methods on a sync module: https://git.io/JfRPj
- Executing a sync method: https://git.io/JfRXe
However, in TurboModules we block the JS thread, execute the method on the NativeModule's method queue, and then unblock the JS thread. While this approach is thread-safe, and arguably the correct way to dispatch sync methods, it's also much slower than the alternative. Therefore, this diff migrates the legacy behaviour to the TurboModule system.
## Special Case: getConstants()
When an ObjC NativeModule requires main queue setup, and it exports constants, we execute its `constantsToExport` method on the main queue (see: [RCTModuleData gatherConstants](c8d678abcf/React/Base/RCTModuleData.mm (L392-L402))). I replicated this behaviour with TurboModules.
Changelog:
[iOS][Fixed] - Execute ObjC TurboModule async method calls on JS thread for sync modules
Reviewed By: fkgozali
Differential Revision: D21602096
fbshipit-source-id: 42d07b7ad000abeac27091dc3ec440e3836d2eae
Summary:
If a NativeModule requires main queue setup, its `getConstants()` method must be executed on the main thead. The legacy NativeModule infra takes care of this for us. With TurboModules, however, all synchronous methods, including `getConstants()`, execute on the JS thread. Therefore, if a TurboModule requires main queue setup, and exports constants, we must execute its `getConstants()` method body on the main queue explicitly.
**Notes:**
- The changes in this diff should be a noop when TurboModules is off, because `RCTUnsafeExecuteOnMainQueueSync` synchronously execute its block on the current thread if the current thread is the main thread.
- If a NativeModule doens't have the `requiresMainQueueSetup` method, but has the `constantsToExport` method, both NativeModules and TurboModules assume that it requires main queue setup.
## Script
```
const exec = require("../lib/exec");
const abspath = require("../lib/abspath");
const relpath = require("../lib/relpath");
const readFile = (filename) => require("fs").readFileSync(filename, "utf8");
const writeFile = (filename, content) =>
require("fs").writeFileSync(filename, content);
function main() {
const tmFiles = exec("cd ~/fbsource && xbgs -n 10000 -l constantsToExport")
.split("\n")
.filter(Boolean);
const filesWithoutConstantsToExport = [];
const filesWithConstantsToExportButNotGetConstants = [];
const filesExplicitlyNotRequiringMainQueueSetup = [];
tmFiles
.filter((filename) => {
if (filename.includes("microsoft-fork-of-react-native")) {
return false;
}
return /\.mm?$/.test(filename);
})
.map(abspath)
.forEach((filename) => {
const code = readFile(filename);
const relFilename = relpath(filename);
if (!/constantsToExport\s*{/.test(code)) {
filesWithoutConstantsToExport.push(relFilename);
return;
}
if (!/getConstants\s*{/.test(code)) {
filesWithConstantsToExportButNotGetConstants.push(relFilename);
return;
}
if (/requiresMainQueueSetup\s*{/.test(code)) {
const requiresMainQueueSetupRegex = /requiresMainQueueSetup\s*{\s*return\s+(?<requiresMainQueueSetup>YES|NO)/;
const requiresMainQueueSetupRegexMatch = requiresMainQueueSetupRegex.exec(
code
);
if (!requiresMainQueueSetupRegexMatch) {
throw new Error(
"Detected requiresMainQueueSetup method in file " +
relFilename +
" but was unable to parse the method return value"
);
}
const {
requiresMainQueueSetup,
} = requiresMainQueueSetupRegexMatch.groups;
if (requiresMainQueueSetup == "NO") {
filesExplicitlyNotRequiringMainQueueSetup.push(relFilename);
return;
}
}
const getConstantsTypeRegex = () => /-\s*\((?<type>.*)\)getConstants\s*{/;
const getConstantsTypeRegexMatch = getConstantsTypeRegex().exec(code);
if (!getConstantsTypeRegexMatch) {
throw new Error(
`Failed to parse return type of getConstants method in file ${relFilename}`
);
}
const getConstantsType = getConstantsTypeRegexMatch.groups.type;
const getConstantsBody = code
.split(getConstantsTypeRegex())[2]
.split("\n}")[0];
const newGetConstantsBody = `
__block ${getConstantsType} constants;
RCTUnsafeExecuteOnMainQueueSync(^{${getConstantsBody
.replace(/\n/g, "\n ")
.replace(/_bridge/g, "self->_bridge")
.replace(/return /g, "constants = ")}
});
return constants;
`;
writeFile(
filename,
code
.replace(getConstantsBody, newGetConstantsBody)
.replace("#import", "#import <React/RCTUtils.h>\n#import")
);
});
console.log("Files without constantsToExport: ");
filesWithoutConstantsToExport.forEach((file) => console.log(file));
console.log();
console.log("Files with constantsToExport but no getConstants: ");
filesWithConstantsToExportButNotGetConstants.forEach((file) =>
console.log(file)
);
console.log();
console.log("Files with requiresMainQueueSetup = NO: ");
filesExplicitlyNotRequiringMainQueueSetup.forEach((file) =>
console.log(file)
);
}
if (!module.parent) {
main();
}
```
Changelog: [Internal]
Reviewed By: fkgozali
Differential Revision: D21797048
fbshipit-source-id: a822a858fecdbe976e6197f8339e509dc7cd917f
Summary:
At some early stages of Instance initialization, it does not have a `nativeToJsBridge_`. At the same time, `handleMemoryPressure` can be called at any point in time, so we should check the pointer for not being null before calling on it.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: JoshuaGross, mdvacca
Differential Revision: D21798522
fbshipit-source-id: 6384da88784cceb493cf9810408cbb47777d3f4b
Summary:
This diff extends the measurement of Text components in order to support empty strings.
This is required for parity with Paper.
I created a follow up task to analyze support of empty string as part of the Text infrastructure of Fabric in the future.
changelog: [Internal]
Reviewed By: sammy-SC
Differential Revision: D21761171
fbshipit-source-id: d2aa074052b09732af5d35723f19014090fcabbf
Summary:
`getStackTracesTreeNodeForAlloc` only works for memory that was tracked with `newAlloc`.
For now, that is only memory that goes through GC APIs.
Various places were calling this function on native memory which was allocated
with `malloc`, not GC memory. This led to SIGSEGV on nullptr when trying to take heap profiles
of these objects.
`newAlloc` could, in theory, work with native memory as well. But building out support for that
is outside the scope of this fix.
Changelog: [Internal]
Reviewed By: jbower-fb
Differential Revision: D21667842
fbshipit-source-id: 8403a6668e5ec607972ce6819f78fedb89da3f37
Summary:
Index adjustment doesn't work if virtual views are inserted, because those don't actually result in the view hierarchy being mutated, as such.
Add an Android-specific check to solve the crash.
I don't love adding platform-specific checks here, so I'm considering wrapping this logic in a platform-specific class, so that logic lives outside of the core of LayoutAnimations and entirely in platform code.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D21727805
fbshipit-source-id: 5af2cf479beaa4d0e9d94ea16ac989c4268920f8
Summary:
property and type are optional params according to Flow, so we should treat as such.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D21725746
fbshipit-source-id: 3c48a8cef8fa5911c195f582556de6dad871c4f1
Summary:
A key difference in WeakRefs with the Hades GC is that they are mutable,
in the sense that reading the value of a WeakRef while a GC is active might
clear the WeakRef.
For this reason, attempting to lock the WeakObject might mutate the backing
reference.
I preferred to communicate this change in behavior via the API rather than
`const_cast` it away inside the implementation.
Changelog: [Internal] Change jsi::WeakObject to be mutable in lockWeakObject
Reviewed By: mhorowitz
Differential Revision: D21485758
fbshipit-source-id: 3618928be8f8791aed56cb20673896ff5b786ded
Summary:
Simple diff to add ART components into Catalyst app Android
changelog: [Internal] Internal changes to add support of ART for Fabric
Reviewed By: JoshuaGross
Differential Revision: D21621479
fbshipit-source-id: d957c25f447d19d8f349c69aa20f5f19237d867a
Summary:
Here I'm implementing equality methods for ARTGroup, ARTShape and ARTText and I'm using these methods to update the state only when it is necessary.
This will improve perf in rendering of ART
changelog: [Internal]
Reviewed By: JoshuaGross
Differential Revision: D21695127
fbshipit-source-id: b438ddea4c34bd7a0bdf26a6aac4fd62a9f78b49
Summary:
Element, Shape, Group, Text are too generic, we are renaming these classes to ARTElement, ARTBGroup, ARTShape...
changelog: [Internal] internal changes to support ART in android
Reviewed By: JoshuaGross
Differential Revision: D21681878
fbshipit-source-id: f6b35443486a3db210f61dcaf91bd32df47fbc66
Summary:
This diff adds support for Text in ART Fabric Android
changelog: [Internal] internal changes to fully support ART in Fabric
Reviewed By: JoshuaGross
Differential Revision: D21681877
fbshipit-source-id: c92e642cff56b71f8ee8f4eb9af6eea6c490f6c7
Summary:
This diff implements the serialization of Text components to send data from C++ to java
changelog: [Internal] internal changes to support ART in fabric
Reviewed By: JoshuaGross
Differential Revision: D21681875
fbshipit-source-id: eba31f35c95e0a2d3226ec70421832719083d7fa
Summary:
This diff refactors the types of ART Text classes, this is necessary on the next diffs of the stack
closeoncommit
changelog: [internal]
Reviewed By: JoshuaGross
Differential Revision: D21681876
fbshipit-source-id: ea438e89df6d860b3ff8bbdae657ca123b417a1b
Summary:
Changelog: [Internal]
`BatchedEventQueue::enqueueUniqueEvent` goes over event queue and deletes previous event of the same type and same target.
This is useful for ScrollView for example where only the latest event is relevant.
This only affects ScrollView scroll event, other events take the original code path.
Reviewed By: mdvacca
Differential Revision: D21648906
fbshipit-source-id: a80ad652058fd50ebb55e24a87229cdc1764b591
Summary:
Changelog: [Internal]
Separates EventQueues from an array into 4 ivars.
EventQueues are of different type, in the future we will want to call different methods on different kind of EventQueue.
Reviewed By: shergin
Differential Revision: D21648905
fbshipit-source-id: 90ae65edb8a9276eecfea9770f554d8c56804797
Summary:
Changelog: [Internal]
# Problem
Yoga internally uses address of owner to determine whether a node should be cloned or it shouldn't.
During layout, it traverses the tree and looks whether current parent is equal to child's owner. If it isn't it means the yoga node is shared between 2+ trees and need to be cloned before mutated. Parent in yoga is stored as reference to the parent.
We can run into an issue where yoga node is shared between 2+ trees, but its current parent is equal to child's owner. This is because we reallocate new parent at address where its previous parent lived. This happens over few iterations, the old parent is first deallocated.
This is known as ABA problem.
# Solution
When we are cloning node, we loop over all children to see whether any of the is not using address of new `yogaNode_` as its owner. At this point we know this is accidental and set its owner to `0xBADC0FFEE0DDF00D`.
We chose `0xBADC0FFEE0DDF00D` because when someone is debugging this in LLDB and prints this address, it will hint them that it was artificially set and can string search for it in the codebase.
Reviewed By: shergin
Differential Revision: D21641096
fbshipit-source-id: c8b1b4487ea02b367f5831c1cdac055bce79c856
Summary:
This diff serializes ART state into folly::dynamic. this is necessary to send this data to Android
changelog: [Internal]
Reviewed By: shergin
Differential Revision: D21657608
fbshipit-source-id: 6c1b69af7d1dbe7de15e509f83c508a38294d89e
Summary:
This diff replaces the usage of int to represent the type of elements for an Enum
changelog: [Internal] Internal change to support ART in fabric
Reviewed By: shergin
Differential Revision: D21657706
fbshipit-source-id: 7bda0210d50136477f0524695d5406e35074f09c
Summary:
This diff integrates ART state into ARTSurfaceViewShadowNode
changelog: [Internal]
Reviewed By: shergin
Differential Revision: D21657611
fbshipit-source-id: 06bd4d610e2c52e0ef3bca423b93c9ad2318e8df
Summary:
This diff creates the internal state of ART based on its shadow nodes
changelog: [Internal]
Reviewed By: JoshuaGross
Differential Revision: D21657607
fbshipit-source-id: 0a15e90ee7465bf3a2b1001ff9d3198eb22fd708
Summary:
This diff introduces a set of classes that are going to be used to represent the internal State of ART nodes
changeLog: [Internal][Android] Internal change to support ART in Fabric
Reviewed By: JoshuaGross, shergin
Differential Revision: D21657612
fbshipit-source-id: ea6d94b06807ff02d222dfa129a1cae384dceeaa
Summary:
## Motivation
We got this crash T67304907, which shows a `EXC_BAD_ACCESS / KERN_INVALID_ADDRESS` when calling this line:
```
NativeModulePerfLogger::getInstance().asyncMethodCallBatchPreprocessStart();
```
There are no arguments in that call, so I figured the only error could be when we try to invoke `getInstance()` or `asyncMethodCallBatchPreprocessStart()`.
This diff:
1. Removes the `NativeModulePerfLogger::getInstance()` bit. Now NativeModulePerfLogger is used via regular static C functions. So, there's no way that simply invoking one of the logging functions crashes the application: there's no vtable lookup.
2. Inside each logging function, when perf-logging is disabled, the global perflogger should be `nullptr`. This diff makes it so that in that case, we won't execute any code in the control group of the perf-logging experiment.
## Changes
**How do we enable NativeModule perf-logging?**
- Previously:
- `NativeModulePerfLogger::setInstance(std::make_shared<FBReactNativeModulePerfLogger>(...))`
- `TurboModulePerfLogger::setInstance(std::make_shared<FBReactNativeModulePerfLogger>(...))`.
- Now:
- `BridgeNativeModulePerfLogger::enableLogging(std::make_unique<FBReactNativeModulePerfLogger>(...))`
- `TurboModulePerfLogger::enableLogging(std::make_unique<FBReactNativeModulePerfLogger>(...))`
**How do we do NativeModule perf-logging now?**
- Previously:
- `NativeModulePerfLogger::getInstance().command(...args)`
- `TurboModulePerfLogger::getInstance().command(...args)`.
- Now:
- `BridgeNativeModulePerfLogger::command(...args)`
- `TurboModulePerfLogger::command(...args)`.
The benefit of this approach is that each method in `BridgeNativeModulePerfLogger` is guarded with an if check. Example:
```
void moduleCreateConstructStart(const char *moduleName, int32_t id) {
NativeModulePerfLogger *logger = g_perfLogger.get();
if (logger != nullptr) {
logger->moduleCreateConstructStart(moduleName, id);
}
}
```
Therefore, we don't actually execute any code when perf-logging is disabled.
Changelog:
[Internal]
Reviewed By: fkgozali
Differential Revision: D21669888
fbshipit-source-id: 80c73754c430ce787404b563878bad146295e01f
Summary:
1. Split out the prop interpolation function out of the View ComponentDescriptor, into an inline'd function that can be used elsewhere.
2. Call it from View and from Paragraph component descriptors.
This causes animations including Text to look normal on iOS.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D21675804
fbshipit-source-id: c34a317749fd6c108aef072d47f3dcb14ce8aa5c
Summary:
Implement a real Transform interpolation. It uses quaternions/Slerp to interpolate rotations. This allows us to interpolate scale, rotation, and translation simultaneously.
See caveats in code. Because of the way transform matrices work, there isn't much (anything?) we can do about skew, and certain values will look nonsensical. This seems to be true for any variant of this algorithm.
This is a big step up from Classic RN which didn't support this in LayoutAnimations at all.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D21675805
fbshipit-source-id: a33494cc02c73102ca67c1d562efc4b2a7308a4a
Summary:
The LayoutAnimationStatusDelegate exists so that platforms can get a signal when animations are starting or have all completed.
This signal is meant to be used ONLY for driving animations at 60fps, or stopping that process, on the platform side.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D21675807
fbshipit-source-id: e89ba524ea43c266556199c8854e8228869755e3
Summary:
Implement EaseIn, EaseOut, EaseInOut, and Spring with SpringDamping.
Note this does not yet implement Keyboard-type animation for iOS (coming soon), and the spring interpolator is VERY naive. We likely want to replace it with a "real" spring animation ASAP. The spring animation is identical to what Android does today, but would likely be a downgrade for iOS. I will do both in a followup.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D21675810
fbshipit-source-id: ac98554ebc81f0b81fdacefd6d848e3566e424c0
Summary:
This is the V1 implementation of Fabric Core LayoutAnimations.
The intention is to structure this in such a way that it's easy for each platform to customize the "AnimationDriver" class (to do platform-specific optimizations) without changing the KeyFrameManager at all.
In the future, this structure and architecture should allow us to iterate faster on new animation APIs.
Changelog: [Internal] Support for LayoutAnimations in Fabric
Reviewed By: mdvacca
Differential Revision: D21675808
fbshipit-source-id: b3ef44729bb8b6217f90760aec9737276c9601d1
Summary:
1. Split out the prop interpolation function out of the View ComponentDescriptor, into an inline'd function that can be used elsewhere.
2. Call it from View and from Paragraph component descriptors.
This causes animations including Text to look normal on iOS.
Changelog: [Internal]
Reviewed By: shergin
Differential Revision: D21635473
fbshipit-source-id: 470f43fd24a6e80d8696ee2f2a09d9e693b7f280
Summary:
Implement a real Transform interpolation. It uses quaternions/Slerp to interpolate rotations. This allows us to interpolate scale, rotation, and translation simultaneously.
See caveats in code. Because of the way transform matrices work, there isn't much (anything?) we can do about skew, and certain values will look nonsensical. This seems to be true for any variant of this algorithm.
This is a big step up from Classic RN which didn't support this in LayoutAnimations at all.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D21615829
fbshipit-source-id: da6cb931ce857107d4b3d20108fb9bbecbf4f898
Summary:
The LayoutAnimationStatusDelegate exists so that platforms can get a signal when animations are starting or have all completed.
This signal is meant to be used ONLY for driving animations at 60fps, or stopping that process, on the platform side.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D21583109
fbshipit-source-id: 234496841bde226fcd6623c74c1a500e5cd00d99
Summary:
Implement EaseIn, EaseOut, EaseInOut, and Spring with SpringDamping.
Note this does not yet implement Keyboard-type animation for iOS (coming soon), and the spring interpolator is VERY naive. We likely want to replace it with a "real" spring animation ASAP. The spring animation is identical to what Android does today, but would likely be a downgrade for iOS. I will do both in a followup.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D21587648
fbshipit-source-id: 246ab7fd40397a4231bb6b18d2f29602788a1bd2
Summary:
This is the V1 implementation of Fabric Core LayoutAnimations.
The intention is to structure this in such a way that it's easy for each platform to customize the "AnimationDriver" class (to do platform-specific optimizations) without changing the KeyFrameManager at all.
In the future, this structure and architecture should allow us to iterate faster on new animation APIs.
TODOs:
- Use std::chrono for timekeeping
Changelog: [Internal] Support for LayoutAnimations in Fabric
Reviewed By: shergin
Differential Revision: D17486030
fbshipit-source-id: 95c72cf9fc2b4bf3fe652fbd249cf2ad113033c7
Summary:
Changelog: [Internal]
`SafeAreaViewShadowNode.alreadyAppliedPadding` was always {0, 0, 0, 0} because value of previous shadow node was never copied over to new shadow node during clone.
Reviewed By: shergin
Differential Revision: D21617361
fbshipit-source-id: 6d6c91b19ff60271bf7c48145d85faaee0321680
Summary:
Basic implementation of ARTText (shadow node, props and component descriptor)
changelog: [Internal] Internal changes to support art in Fabric
Reviewed By: shergin
Differential Revision: D21621483
fbshipit-source-id: d0886dc149520af13faa1bb936dfcccab1798c37
Summary:
Basic implementation of ARTGroupProps (shadow node, props and component descriptor)
changelog: [Internal] Internal changes to support art in Fabric
Reviewed By: shergin
Differential Revision: D21621480
fbshipit-source-id: 367a479568b8c1a290f3e0f633cc4052a9c95b87
Summary:
Create basic implementation of Shape (shadow node, props and component descriptor)
changelog: [Internal] Internal changes to support art in Fabric
Reviewed By: shergin
Differential Revision: D21621482
fbshipit-source-id: e5b9bb2812ee92bce625301b7521f0578eaca0ff
Summary:
Using of the new API makes clear that we don't need to calculate the newest descendant node.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: mdvacca
Differential Revision: D21480202
fbshipit-source-id: c54998573baffe4a05726e3402da027f052b049a
Summary:
This diff simplifies the implementation of `LayoutableShadowNode::getRelativeLayoutMetrics`.
It fixes a small bug but the most important change is the new interface.
Now the function that does measurements accepts a node and a family instead of two nodes. It prevents misuse and misinterpretation of what the function does. The function needs two things to perform measurement:
* an ancestor node that defines the tree is being measured and the base node of measurement;
* a family of some descendant node being measured relative to the ancestor node.
An API that accepts two nodes is misleading because it implies that the given descendant node will be measured (which is not true).
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D21480200
fbshipit-source-id: 9fddc361417fee47bbf66cc7ac2954eb088a3179
Summary:
One small test was added.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D21480201
fbshipit-source-id: fd6c050143fcdf27d345ee62e74c4368266e6ce0
Summary:
This is an addition to an automatic emergency clean-up algorithm that we have in Scheduler. In addition to committing empty surfaces, we also remove those surfaces from the registry making calling stuff on them impossible. Removing surfaces waits for all commits in flight to be finished, so it theoretically can deadlock (so we gated that).
If we won't face deadlocks in a coming couple of weeks, I would remove gating.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: mdvacca
Differential Revision: D21610683
fbshipit-source-id: 71feeaa0ee4521a0180cdfba6e3a271e7f7d9401
Summary:
## Motivation
This rename will fix the following CircleCI build failures:
- [test_ios_unit_frameworks](https://circleci.com/gh/facebook/react-native/150473?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link)
- [test_ios_detox_frameworks](https://circleci.com/gh/facebook/react-native/150474?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link)
## Investigation
We have 4 podspec targets that map to the same header namespace (i.e: `header_dir`) `ReactCommon`:
- **New:** `React-perflogger`: Directory is `ReactCommon/preflogger`, and contains `NativeModulePerfLogger.{h,cpp}`.
- `React-runtimeexecutor`: Directory is `ReactCommon/runtimeexecutor`, and contains only `RuntimeExecutor.h`
- `React-callinvoker`: Directory is `ReactCommon/callinvoker`, and contains only `CallInvoker.h`
- `ReactCommon/turbomodule/core`: Directory is `ReactCommon/turbomodule`, and contains C++ files, as well has header files.
**The problem:**
We couldn't import headers from `React-perflogger` in `ReactCommon/turbomodule/core` files.
**The cause:**
I'm not entirely sure why, but I was able to discern the following two rules by playing around with the podspecs:
1. If your podspec target has a cpp file, it'll generate a framework when `USE_FRAMEWORKS=1`.
2. Two different frameworks cannot map to the same `module_name` or `header_dir`. (Why? No clue. But something breaks silently when this is the case).
So, this is what happened when I landed `React-perflogger` (D21443610):
1. The TurboModules code generates the `ReactCommon` framework that uses the `ReactCommon` header namespace.
2. `React-runtimeexecutor` and `React-callinvoker` also used the `ReactCommon` header namespace. However, neither generate a framework because of Rule 1.
3. When I comitted `React-perflogger`, I introduced a second framework that competed with the `ReactCommon` framework (i.e: TurboModules code) for the `ReactCommon` header namespace. Rule 2 violation.
## Thoughts on renaming
- `<perflogger/NativeModulePerfLogger.h>` is too generic, and the `perflogger` namepsace is used internally within FB.
- `<react/perflogger/NativeModulePerfLogger.h>` matches our fabric header format, but I'm pretty sure that slashes aren't allowed in `header_dir`: I tested this and it didn't work. IIRC, only alphanumeric and underscore are valid characters for `header_dir` or `module_name`. So, I opted to just use `reactperflogger`.
Changelog: [Internal]
Reviewed By: fkgozali
Differential Revision: D21598852
fbshipit-source-id: 60da5d0f7758eaf13907a080b7d8756688f40723
Summary:
This call is used to continue execution when the app has just been
started in a "wait for debugger" mode. This is the only case
in which it has an effect.
Notably, it should do nothing in the following cases, which a layperson
may be tempted to classify as "WaitingForDebugger":
* The app was running detached and hit a 'debugger;' statement
* The app is paused because of a breakpoint or hitting the Pause button
* The app stopped on an instrumentation breakpoint, and expects
the debugger to collect data and potentially auto-resume.
Changelog: [Internal] Add Hermes support for Debugger.runIfWaitingForDebugger
Reviewed By: mhorowitz
Differential Revision: D21557446
fbshipit-source-id: 790cec7444ddc61908d2ef9d92e4649b535d678f
Summary:
This diff fixes the position of TextInlineViews when nesting multiple Text.
The root is that we were not taking into consideration LayoutOffset of nested TextViews during the calculation of the nested views.
changelog: [Internal] Internal fix in Fabric
Reviewed By: JoshuaGross
Differential Revision: D21586893
fbshipit-source-id: 55e6ad0cf95222588ffe9185f5e22baea1059448
Summary:
## Motivation
This rename will fix the following CircleCI build failures:
- [test_ios_unit_frameworks](https://circleci.com/gh/facebook/react-native/150473?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link)
- [test_ios_detox_frameworks](https://circleci.com/gh/facebook/react-native/150474?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link)
## Investigation
We have 4 podspec targets that map to the same header namespace (i.e: `header_dir`) `ReactCommon`:
- **New:** `React-perflogger`: Directory is `ReactCommon/preflogger`, and contains `NativeModulePerfLogger.{h,cpp}`.
- `React-runtimeexecutor`: Directory is `ReactCommon/runtimeexecutor`, and contains only `RuntimeExecutor.h`
- `React-callinvoker`: Directory is `ReactCommon/callinvoker`, and contains only `CallInvoker.h`
- `ReactCommon/turbomodule/core`: Directory is `ReactCommon/turbomodule`, and contains C++ files, as well has header files.
**The problem:**
We couldn't import headers from `React-perflogger` in `ReactCommon/turbomodule/core` files.
**The cause:**
I'm not entirely sure why, but I was able to discern the following two rules by playing around with the podspecs:
1. If your podspec target has a cpp file, it'll generate a framework when `USE_FRAMEWORKS=1`.
2. Two different frameworks cannot map to the same `module_name` or `header_dir`. (Why? No clue. But something breaks silently when this is the case).
So, this is what happened when I landed `React-perflogger` (D21443610):
1. The TurboModules code generates the `ReactCommon` framework that uses the `ReactCommon` header namespace.
2. `React-runtimeexecutor` and `React-callinvoker` also used the `ReactCommon` header namespace. However, neither generate a framework because of Rule 1.
3. When I comitted `React-perflogger`, I introduced a second framework that competed with the `ReactCommon` framework (i.e: TurboModules code) for the `ReactCommon` header namespace. Rule 2 violation.
## Thoughts on renaming
- `<perflogger/NativeModulePerfLogger.h>` is too generic, and the `perflogger` namepsace is used internally within FB.
- `<react/perflogger/NativeModulePerfLogger.h>` matches our fabric header format, but I'm pretty sure that slashes aren't allowed in `header_dir`: I tested this and it didn't work. IIRC, only alphanumeric and underscore are valid characters for `header_dir` or `module_name`. So, I opted to just use `reactperflogger`.
Changelog: [Internal]
Reviewed By: fkgozali
Differential Revision: D21585006
fbshipit-source-id: e3339273af5dfd65a1454d87213d1221de6a4651
Summary:
For Fabric LayoutAnimations, we need to support interpolating the Transform property (which really ends up just being interpolation of ScaleX, ScaleY, or ScaleXY transforms - not arbitrary matrices).
To support that, we need to be able to convert Transform back to folly::dynamic, and on the Java side we need to support accepting arbitrary matrices instead of transform maps of properties.
Changelog: [Internal] Fabric-only changes
Reviewed By: sammy-SC
Differential Revision: D21564590
fbshipit-source-id: b137f659b27e4b8fae83921a28ccf46035e18651
Summary:
Changelog: [Internal]
Using `empty()` vs `size() == 0` or `size() > 0`.
It is a more semantic way to check whether container is empty or not.
Reviewed By: JoshuaGross
Differential Revision: D21573183
fbshipit-source-id: b83283f687432a037941852114717a0f014e28db
Summary:
Changelog: [Internal]
For consistency, switching west const to east const.
Reviewed By: JoshuaGross, mdvacca
Differential Revision: D21574239
fbshipit-source-id: eb3459c63f731f51b24f40f9f80b574661ffd935
Summary: Changelog: [Internal][Yoga] throw std::logic_error instead of aborting the process and convert to java exception for jni layer
Reviewed By: pasqualeanatriello
Differential Revision: D21301235
fbshipit-source-id: 148b27920e62990a271e1d0df8c85a2cc42f4fd4
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/28893
`JSIExecutor::callSerializableNativeHook` converts the arguments from `JSI::Value` to `folly::dynamic`. Then, `RCTNativeModule` converts the arguments from `folly::dynamic` to ObjC data structures in its `static invokeInner` function.
Therefore, I decided to start the sync markers inside `JSIExecutor::callSerializableNativeHook`, which required me to expose these two methode `ModuleRegistry::getModuleName` and `ModuleRegistry::getModuleSyncMethodName`. This shouldn't modify performance because we eagerly generate a NativeModule's methods when it's first required. So, at worst, this is doing a cache lookup.
Changelog: [Internal]
Reviewed By: PeteTheHeat
Differential Revision: D21443610
fbshipit-source-id: 67cf563b0b06153e56e63ba7e186eea31eafc853
Summary:
NativeModule async method calls are queued up on the JS side, and flushed to C++ on every Native -> JS call. Before we execute the batch of async NativeModule method calls, we convert it (a JS object) from a `jsi::Value` to a `folly::dynamic` object in `JSIExecutor::callNativeModules`. Then, in `JsToNativeBridge::callNativeModules`, we convert this `folly::dynamic` object into an `std::vector<MethodCall>`, before finally looping over these `MethodCall`s and invoking each NativeModule async method call.
The markers I'm adding in this diff measure this `jsi::Value -> folly::dynamic -> std::vector<MethodCall>` pre-processing.
Changelog: [Internal]
Reviewed By: PeteTheHeat
Differential Revision: D21435455
fbshipit-source-id: 4c5a9e2b73c1a2a49d7a8f224a0d30afe3a0c79c
Summary:
This diff instruments two markers:
- JSRequireBeginning: From the start of the JS require to when we start creating the platform NativeModule
- JSRequireEnding: From the end of platform NativeModule create to the end of the JS require
In order to accomplish this, I had modify `ModuleRegistry::ModuleRegistry()` to accept a `std::shared_ptr<NativeModulePerfLogger>`. I also had to implement the public method `ModuleRegistry::getNativeModulePerfLogger()` so that `JSINativeModules` could start logging the JS require beginning and ending.
Changelog: [Internal]
Reviewed By: PeteTheHeat
Differential Revision: D21418803
fbshipit-source-id: 53828817ae41f23f3f04a95b1d3ac0012735da48
Summary:
## Description
This diff introduces `NativeModulePerfLogger`, its BUCK, Cocoapod, android-ndk targets. This diff also wires up those targets into the React Native bridge and TurboModules targets, so that we get signal on if the code compiles.
This diff also introduces `TurboModulePerfLogger`, which is a namespace that holds the `NativeModulePerfLogger` that'll do perf-logging for TurboModules.
## How will perflogging work on iOS?
1. Each application will, during React Native initialization, create a NativeModule perf logger.
2. If TurboModules are enabled, we'll call `TurboModulePerfLogger::setInstance(perfLogger)`. If TurboModules are disabled, we'll call `NativeModulePerfLogger::setInstance(perfLogger)`.
3. TurboModules, when they're created and used, will log events via `TurboModulePerfLogger::getInstance()`. NativeModules (i.e: bridge modules), when they're created and used, will log events via the `NativeModulePerfLogger::getInstance()`.
> **Note:** The NativeModule system will log events for non-TurboModules as well. Maybe we should log events for only NativeModules that conform to the `TurboModule` interface, when TurboModules are disabled. This'll ensure a fair comparison between the two systems.
## How will perflogging work on Android?
Please see the subsequent diff.
allow-large-files
Changelog:
[Both][Added] - Introduce `NativeModulePerfLogger`
Reviewed By: PeteTheHeat
Differential Revision: D21318053
fbshipit-source-id: 6ddf5b5a80bdc4076d2dd6588067e2b0ec8c2c6b
Summary:
The interface of css/TextLayoutManager now matches acual platform-specific implementations, so now CXX tests compiles and usable.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D21518656
fbshipit-source-id: cece3bea14c70410eea22abafb424f7a2cb201c0
Summary:
Each ComponentDescriptor becomes capable of doing its own interpolation over props for animation purposes.
This new custom interpolator is called by default by the ConcreteComponentDescriptor, but `ComponentDescriptor::interpolateProps` is a virtual function and each ComponentDescriptor can provide custom interpolation if necessary.
For now, only View does any actual interpolation, to support LayoutAnimations.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D20965310
fbshipit-source-id: e1c1588107848e94c155efecb0da1cc1619ae544
Summary:
I noticed an issue when the hermes debugger was attached. Console messages would no longer appear in the terminal. This is because the hermes inspector overrides the console object with its own to intercept messages and send them to the debug client.
With this change I made it so that messages are sent to the original console as well instead of completely replacing it.
Changelog:
[General][Fixed] - When Hermes debugger is enabled continue to send log messages to the console
Reviewed By: mhorowitz
Differential Revision: D21509895
fbshipit-source-id: 6d91c4b82682e404679533be14b3e5f12e687e79
Summary:
TextInlineViews in Android was incorrectly converting values to from float to int, this produced to loose precision and to render incomplete texts in some components.
This diff changed the types from int to float, avoiding loose precision.
The impact of this bug is not that high because in the conversion to int we were using Math.ceil(), which was already rounding the result to the next pixel.
changeLog: [Android][Fixed] Fix precision of TextInlineViews in Fabric Android
Reviewed By: JoshuaGross, shergin
Differential Revision: D21541159
fbshipit-source-id: 4741ab96964c35af1c1b7d3e821e505ecef2efce
Summary:
The previous implementation was incorrect causing a memory leak. In the new approach, we use a shared pointer to optional to be able to construct that ahead of time and then fill with actual value.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D21516794
fbshipit-source-id: ddcbf8a1ad2b1f629ae4ff2842edeb7bb2a275a6
Summary:
Changelog: [internal]
Fabric asks server for images of their real size and scale, not images of size that they would take up on the screen once rendered. Also scale of the screen is incorrect. This causes images to be twice as large as they have to be.
Look at the size of the first image size here in Paper, it is `{310.5, 207}`.
{F235394066}
If you compare it with Fabric, there it is `{800, 381}`
{F235394115}
It isn't just the size, but scale of request image as well. Fabric always asks for image of scale 1, unlike Paper which takes screen scale into account.
Reviewed By: shergin
Differential Revision: D21255794
fbshipit-source-id: 9db3ccafec1c09cedc5db5ac0a435a28a4c30c85
Summary:
The approach is quite simple: we check the thread id, and if it matches the caller thread id, we unlock mutexes which lead to the normal uninterrupted execution flow.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: RSNara
Differential Revision: D21328313
fbshipit-source-id: 4290b8a0357dbad3563d7da9464c03ecce5ded7c
Summary:
This diff implements iOS-specific `PlatformRunLoopObserver` (and `MainRunLoopObserver`) that then being glued together with `SynchronousEventBeat` replaces `MainRunLoopEventBeat`, and then same thing glued with `AsynchronousEventBeat` replaces `RuntimeEventBeat`.
So, instead of two platform-specific classes we had on iOS (for that needs), now we have only one (that can be reused for a more broad variety of applications).
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D21341998
fbshipit-source-id: fafde4e678770f7fcf9c1ff87acc02812a37e708
Summary:
We use `EventBeat::OwnerBox` in the `EventBeat` infra to represent weak ownership of classes that contain that, so those classes can ensure own life-time at some critical points.
Before this diff, we stored a pointer to EventDispatcher right after we create it. However, some components that we build require to have a valid pointer from the beginning (even before the EventDispatcher is created). To satisfy this requirement, we create a dummy pointer first and use it as an owner, then we merge (share control block) the pointer to an EventDispatcher into that. It works because the owner pointer never gets dereferenced (it's `void *`), so it does not matter which object it represents while it represents correct ownership.
In the future, this approach will allow us to remove the concept of `OwnerBox` completely and use just `std::weak_ptr<void>` instead.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: mdvacca
Differential Revision: D21441109
fbshipit-source-id: 7457c77bd31cd577f38a26e28e27eb7e33b6ad24
Summary:
The name of the value was incorrect.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: mdvacca
Differential Revision: D21496651
fbshipit-source-id: 464c98436bb8d5f2b6275b7eab1c32d187e2b23c
Summary:
We should check for corner cases.
Another interesting detail is behavior. Now (and previously) we return nullptr if case we could not find the most recent node. But maybe we should return the given node instead?
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: mdvacca
Differential Revision: D21480874
fbshipit-source-id: bb943c4508ec025316a3665d652e4b5e06dd795c
Summary:
We cannot use a raw pointer here because the tree might progress right after we call to `shadowTree.tryCommit` and we will get a dangling pointer as a result.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: mdvacca
Differential Revision: D21464832
fbshipit-source-id: 3998b39d29db672da54d4adcb6be55cd3d44d862
Summary:
In this method we have to maintain a retaining pointer to the node in order to access it as a raw pointer.
See also a comment in the codee.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: JoshuaGross
Differential Revision: D21464834
fbshipit-source-id: 69f6894009509e5feca291fab12c019208933816
Summary:
There is no need to use a raw pointer to a shared pointer as a return value of `UIManager::getNewestCloneOfShadowNode`. If it would be a very hot path, we could you a raw pointer to the node instead but it's not a hot path.
Prooving that using a raw pointer here is safe is quite complex (and I don't know if it's safe or not). So, I changed it to just a shared pointer.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: JoshuaGross
Differential Revision: D21464833
fbshipit-source-id: 813a3c9fca0147afb322db6855a0ce8fd2d47909
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/28851
This diff creates a RuntimeExecutor that uses the bridge and exposes it on CatalystInstanceImpl.
Changelog: [Internal]
Reviewed By: mdvacca, RSNara
Differential Revision: D21051949
fbshipit-source-id: b3977fc14fa19089f33e297d29cedba0d067526d
Summary:
This diff adds support for the "beforeScriptWithSourceMapExecution" instrumentation
breakpoint via "Debugger.setInstrumentationBreakpoint".
CDP describes it as a breakpoint, but we just set a flag in the inspector. A
fake breakpoint ID is synthesized for optional removal later.
Changelog: [Internal] Add Debugger.setInstrumentationBreakpoint to Hermes Inspector
Reviewed By: mhorowitz
Differential Revision: D21418218
fbshipit-source-id: 90fa49c0954980993815322d3a7effee416ed5db
Summary:
This removes the Hermes.setPauseOnLoad. It will be replaced by the more standard
Debugger.setInstrumentationBreakpoint's "beforeScriptExecution" event.
ChangeLog: [Internal] Remove Hermes.setPauseOnLoad message (to be replaced)
Reviewed By: mhorowitz
Differential Revision: D21418219
fbshipit-source-id: 93c53801c23487f9336b322c2bd737663ec21b97
Summary:
This diff exposes the Text.includeFontPadding prop to java, then it uses the prop to calculate the height of Text components correctly.
changelog: [Internal][Fabric] Internal change in Fabric to support Text.includeFontPadding prop in fabric
Reviewed By: shergin
Differential Revision: D21446737
fbshipit-source-id: efe73fb6b0d402c3275ac8c012fa8fa06b743bdd
Summary:
This diff extends the ParagraphAttribute class to store the value of the includeFontPadding prop.
Note that this is an Android only prop, I'm not creating android blocks to improve "cleanliness" of the code.
changelog: [Internal][Fabric] Internal change in Fabric to support Text.includeFontPadding prop in fabric
Reviewed By: shergin
Differential Revision: D21446738
fbshipit-source-id: 0543e86aa18ce10f7a56bbaafe111cce0179ea86
Summary:
We consume Hermes through multiple .so's, which means we have multiple (weak) typeinfo definitions of facebook::jsi::JSError. Previously we were using gnustl, which would strcmp typeinfo to decide whether a certain exception handler applies, which meant this didn't cause any major issues. However since this is deprecated, we recently switched to libc++, which does not have this by behaviour (or it does, but behind a flag I'm not sure how to enable). This causes any JS exceptions to fall through from our exception handlers and fatal the app.
This problem is actually documented in the common Android NDK problems page: https://android.googlesource.com/platform/ndk/+/master/docs/user/common_problems.md#rtti_exceptions-not-working-across-library-boundaries
The suggested solution is to ensure that any exception types have a key function defined (a non-pure, out-of-line virtual function). The simplest one to add is a virtual destructor. This makes the object file that holds the implementation of the destructor export a non-weak typeinfo definition which will at load time override the other weak versions.
I'm not sure why we're the first to hit this. RN's JSIExecutor doesn't explicitly reference JSError which probably helps (https://github.com/facebook/react-native/blob/master/ReactCommon/jsiexecutor/jsireact/JSIExecutor.cpp#L256-L258) and they also don't use unguarded callbacks like we do.
Changelog: [Internal]
Reviewed By: mhorowitz
Differential Revision: D21426524
fbshipit-source-id: 474284ada1ca2810045dc4402c420879447f9308
Summary:
Move and create an empty rule that redirects as well, to handle //arvr rules
Need to do this way, since ovrsource sync rules are in different repo.
allow_many_files
allow-large-files
Steps:
- [X] Move glog from xplat/third-party to /third-party
- [ ] Update references in ovrsource to translate to //third-party instead of //xplat/third-party
- [ ] Get rid of temporary rule
- [ ] Update fbsource/third-party/glog to 0.3.5 (what we have in ovrsource)
Changelog: [Internal] Update reference for glog from xplat/third-party to /third-party.
Reviewed By: yfeldblum
Differential Revision: D21363584
fbshipit-source-id: c1ffe2dd615077170b03d98dcfb77121537793c9
Summary:
`SynchronousEventBeat` and `AsynchronousEventBeat` are a cross-platform re-implementation of run loop related parts of `MainRunLoopEventBeat` and `RuntimeEventBeat` (iOS specific classes for now). In the future, they will replace iOS- and Android-specifc event beat classes.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D21341996
fbshipit-source-id: 8eda9a5df537cd666b7728e32212a8bb5ddb3ab7
Summary:
`RunLoopObserver` is one of the core interfaces that bridge intrinsically platform-specific functionality to cross-platform React Native core. `RunLoopObserver` allows subscribing for notifications about changes in a run loop life cycle. Primarily it supposed to be used for observing UI (aka main) and JavaScript execution thread/run-loop.
Having a `RunLoopObserver` implemented in a platform-specific manner allows building these components in a cross-platform manner:
* Sync and async UI event delivery pipeline;
* Timing for some animation engine;
* Timers (probably additional features are required).
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D21341997
fbshipit-source-id: 7ef61fb51f550dd0f2e89c64af657e0f0de029aa
Summary:
Before this change, in case of incorrect measurements, Fabric's implementation of `measure` and `measureInWindow` incorrectly returned negative height and width. Now it returns zeros (as classic React Native does).
Fabric:
This does not fix `measureLayout` called for virtual nodes. This is not so trivially to fix and it will be done separately.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: JoshuaGross, yungsters, mdvacca
Differential Revision: D21433239
fbshipit-source-id: fbaf5ee35c690506822c634daac4426542c2cdcf
Summary:
We are currently very strict about breakpoint location matching. This
diff allows some fuzz in the column, but not in the line.
Changelog: [Internal] Setting Hermes breakpoints no longer requires exact column match
Reviewed By: avp
Differential Revision: D21343198
fbshipit-source-id: a59786a9d63f9fe1ed576835ed660ba3343affe1
Summary:
The core issue solved in D21389371 was that erased elements of a TinyMap were being iterated over, because TinyMap has somewhat-complicated logic around cleaning out the underlying vector. In some very marginal cases, vectors were not being cleaned and an iterator pointing at erased elements was being returned.
The diff prevents some possible crashes in `begin()` and `find()` while making it much less likely to iterate over erased elements.
We also add a unit test to catch the case fixed in D21389371, in particular.
We also are keeping the code added in D21389371 (for now) since it's a cheap check, and will be a safeguard until we have rigorous testing around TinyMap. To be clear that logic should noop currently, but will prevent crashes in case guarantees around TinyMap change in the future.
Currently there is only one line of code that actually uses the TinyMap iterator, so this should be safe.
Reviewed By: shergin
Differential Revision: D21392762
fbshipit-source-id: 36dc998958c230fad01af93338974f8889cbcf55
Valentin Shergin
David Vacca
Ramanpreet Nara
Jiayan Zhuang
Will Holen
Patrik Tomas Chamelo
Samuel Susla
Joshua Gross
Oleg Bogdanov
Eloy Durán
Christoph Nakazawa
Riley Dulin
Omer Peleg
Vitaly Potlov
Daniel Andersson
Luna Wei
Kacie Bawiec
empyrical
Emily Janzer
Ryan Tremblay
Sidharth Guglani
Paige Sun
Marc Horowitz
Christoph Purrer
Kevin Gozali
generatedunixname89002005287564
Ishan Khot
Jimmy Zhang
Daniel Weaver
Keegan Mendonca
Martin Sherburn
Manvir Singh
Pieter De Baets
Nikita Lutsenko
