Summary:
We have a custom STUB_VIEW_ASSERT that helps debugging stub view issues on platforms (like iOS) where flushing logs around assert-time isn't 100% guaranteed.
Move that logic into react_native_assert since it's generally useful.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D26567218
fbshipit-source-id: 79f5ae66fc65a0af48dbcf4c7204ac8245911cb0
Summary:
See react_native_assert.{h,cpp}. Because of the BUCK+Android issue where NDEBUG is always defined, we use react_native_assert instead of assert to enable xplat asserts in debug/dev mode.
This migrates most of the codebase, but probably not 100%. The goal is to increase assertion coverage on Android, not to get to 100% (yet).
Changelog: [Internal]
Reviewed By: RSNara
Differential Revision: D26562866
fbshipit-source-id: a7bf2055b973e1d3650ed8d68a6d02d556604af9
Summary:
This will allow these asserts to crash on Android debug builds.
We will migrate more sites as we confirm this is stable through testing.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D26409354
fbshipit-source-id: fb35cd8de29890f7c2b761435eaa02de377bdd1e
Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/30988
We have a bunch of flags scattered throughout the codebase with poor hygiene and commenting. Consolidate.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D26392518
fbshipit-source-id: 2823de123a5009d6b8c358e8a3f451b9fa0e05b7
Summary:
Building ReactAndroid from source on Windows has recently hit the limitation of maximum path lengths.
At build time, during the `:ReactAndroid:buildReactNdkLib` task, the linker tries to access several of the intermediate binaries located deep in the tmp folder hierarchy, eg.
```
D:\r\ReactAndroid\build\tmp\buildReactNdkLib/local/armeabi-v7a/objs/react_render_components_progressbar/D_/r/ReactAndroid/__/ReactCommon/react/renderer/components/progressbar/android/react/renderer/components/progressbar/AndroidProgressBarMeasurementsManager.o
```
**Suggested fix:** for modules such as `react_render_components_progressbar` and `react_render_components_picker`, rename them to `rrc_progressbar` etc.
**NOTE**: this assumes that the fix from https://github.com/facebook/react-native/issues/30535 is in place. This regression happened while https://github.com/facebook/react-native/issues/30535 has been pending checkin.
**Other mitigations I've tried:**
- setting [`LOCAL_SHORT_COMMANDS`](https://developer.android.com/ndk/guides/android_mk#local_short_commands) for the problematic modules or `APP_SHORT_COMMANDS` for the root project. Turns out those commands don't work on the NDK version RN requires, but even after manually applying a [patch ](https://android-review.googlesource.com/c/platform/ndk/+/1126440) to my local copy of the NDK, these flags had no effect.
- moving the repo directory higher in the file system tree, and using short directory names `D:\r\...` was not enough
- creating virtual drive letters for specific long paths with the [`sust`](https://docs.microsoft.com/en-us/windows-server/administration/windows-commands/subst#examples) command is not workable, since they depend on the source folder structure, and get partly generated by the build system, which I can't plug into
- just enabling long path support on Windows is not enough, since the compiler toolchain doesn't support them.
## 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
-->
[Android] [Fixed] - Fix source build on Windows machines vol. 2
Pull Request resolved: https://github.com/facebook/react-native/pull/30776
Test Plan:
Run `.\gradlew installArchives`
Before:
Now:
Differential Revision: D26194286
Pulled By: mdvacca
fbshipit-source-id: 778b5a0515148e2ace19e7902f74301831ebed94
Summary:
iOS and Android platform code already explicitly check this invariant: nodes cannot have parents when they're inserted into the View hierarchy.
Check this in the core so we get these checks in unit tests, and earlier in the core before platform code runs.
Changelog: [Internal]
Reviewed By: shergin, mdvacca
Differential Revision: D26331842
fbshipit-source-id: c12bc9066d280cb85ccc9e754c9fa475927e6080
Summary:
During earlier testing I didn't fully realize that Android disables a core bit of View Flattening: Views can be concrete or non-concrete; and their children can be flattened or not. None of these properties are mutually exclusive with each other.
Except on Android - that functionality is currently disabled. A View can be either flattened and non-concrete, or non-flat and concrete. So there are some flattening edge-cases hit on iOS but not Android, due to the larger state-space on iOS.
To test, I forced Android to align with iOS and tested; and then tested on iOS; and ensured no mounting errors, assertions, or crashes were hit during some specific tests.
Changelog: [Internal]
Differential Revision: D26298872
fbshipit-source-id: 2f0f78127a7bf057c7cf109005f1dae74f0ff6ba
Summary:
On iOS, log lines are not entirely flushed when an assert is hit right after (or during) this mutate function. Make sure to flush log lines regularly during this function so that debugging is easier.
Changelog: [Internal]
Reviewed By: sammy-SC
Differential Revision: D26291789
fbshipit-source-id: 47af109cdc3dcfc6bf08cbb41db06e9260bfaa08
Summary:
We still have usages of "fbsource//tools/build_defs/apple:flag_defs.bzl" in react-native-github. But this should get us closer towards not using the fbsource cell. Hopefully, this is enough to unbreak the test_docker CircleCI build.
Changelog: [Internal]
Reviewed By: fkgozali
Differential Revision: D26289304
fbshipit-source-id: 1c6464bb84df4f82f8a797321a73a1ed324e319a
Summary:
Changelog: [internal]
Name `shouldCancel` is misleading. It implies the commit is cancelled and doesn't happen.
`shouldYield` expresses the intent better, because the commit's changes do eventually reach mounting layer but the current commit is rather yielding to the next one.
Reviewed By: mdvacca
Differential Revision: D26126121
fbshipit-source-id: 18988f217cbc651f0010f6e2381682bdbaed8bd4
Summary:
Changelog: [internal]
`updateMountedFlag` needs to be called only when new revision will be mounted. In case a commit is throttled, this wasn't the case. Therefore, moving cancellation of commit so it happens before `updateMountedFlag`. We already cancel commits on that place so it should be safe.
Reviewed By: shergin
Differential Revision: D26049262
fbshipit-source-id: e0ecdd2d8f0cdb09d0c0a07ad3931ce77bcf03cf
Summary:
Before this change, `mountingOverrideDelegate` was proxied via `Scheduler::startSurface` and `ShadowTree::ShadowTree` constructor down to `MountingCordinator`. Now we set it on the `MountingCoordinator` directly.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D26049076
fbshipit-source-id: 7f1ecf2c8b6f264a7e59d19881464fe529c53d30
Summary:
This diff simplifies `ShadowTree` constructor by removing `rootComponentDescriptor` argument. It was previously stored and supplied by `Scheduler`; now `ShadowTree` class allocates one instance of it for all instances of `ShadowTree`. The `RootComponentDescriptor` instance of it is only needed to clone a `RootShadowNode` instance; it cannot issue events, state updates, or anything like that because it does not have React counterpart.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: JoshuaGross, sammy-SC
Differential Revision: D26048466
fbshipit-source-id: ec02b1b4bcc917efe17cef58112fa870b341c85f
Summary:
See title. Not used in this diff; see next diff in stack.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D26001014
fbshipit-source-id: 475eb265f1c27bd9aff978a49652764a50287e47
Summary:
Original commit changeset: 3ed8e78e31b0
Backing-out D25938851 (https://github.com/facebook/react-native/commit/69b3016171bb2f994dd4a62c34c2c4645b5a7d56) and D25935785 (https://github.com/facebook/react-native/commit/bdea479a1faa0f1f7d7c9d9162212cce94bc9720). Based on analysis documented in T83141606, I believe this issue should be fixed in JS.
Additionally, this crash actually has nothing to do with (un)flattening or the differ; it is a side-effect of stale ShadowNodes being cloned, which I believe is either UB or a contract violation. Either way, it should probably be fixed either in JS, or in node cloning. So this isn't the right solution for this issue and should be reverted.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25949569
fbshipit-source-id: 8cf1094a767da98fff4430da60d223412e029545
Summary:
Android has some optimizations around view allocation and pre-allocation that, in the case of View Unflattening, can cause "Create" mutations to be skipped.
To make sure that doesn't happen, we add a flag to ShadowViewMutation (in the core) that any platform can consume, that indicates if the mutation is a "recreation" mutation.
It is still a bit unclear why this is needed, in the sense that I would expect props revision to increment if a view is unflattened. However, there is at least one documented reproduction where that is *not* the case. So for now, we'll have a hack pending further investigation.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25935785
fbshipit-source-id: 6fb4f0a6dedba0fe46ba3cd558ac1daa70f671f5
Summary:
CommitMode allows customizing the side-effects of commit operations. In `Suspended` mode, the results of commit operations will not be passed down to a MountingCoordinator which will result in skipping the mounting phase completely.
This is one of the core elements of the Pre-rendering infrastructure.
Changelog: [Internal] Fabric-specific internal change.
Differential Revision: D24290773
fbshipit-source-id: c10ec20d13f3131fc632352ef22f4465c9dfb3c2
Summary:
While pulling in additional react-native code into react-native-windows to support fabric, I hit some bugs with the MSVC compiler. These minor changes work around those bugs and shouldn't have any functional effect on the code
## Changelog
[Internal] [Fixed] - Minor code tweaks to work around a compiler bug in MSVC.
Pull Request resolved: https://github.com/facebook/react-native/pull/30696
Test Plan: Code compiles in MSVC and clang.
Reviewed By: shergin
Differential Revision: D25830710
Pulled By: appden
fbshipit-source-id: 796bd8191ea62fe62e09dd16cc56b8edc092297f
Summary:
This diff adds another complexity level for Stacking Context and zIndex tests: Now we test not only "noting-to-something" cases but also cases where we apply changes to an already non-empty tree, validating our assumption of each step.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D25623176
fbshipit-source-id: 29a447d93522a2a4e7912351f907c46ac7b218bb
Summary:
With the change, a new delegate method allows a receiver to alter a new (proposed) shadow tree with another tree by returning the altered tree.
We will it use in future diffs to implement Commit Hooks.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D25221313
fbshipit-source-id: 9f83577d862b713fff71fa365ce660cc1de87c84
Summary:
After fixing `calculateShadowViewMutationsForNewTree` I realized that it will be even better to test Stacking Context and mutation instructions infra using both functions: `calculateShadowViewMutationsForNewTree` (used for testing) and the Differentiator itself. This diff implements it.
Now we have two similarly working functions with different implementations that we can use for testing Differentiator and other parts of the infra.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D25576922
fbshipit-source-id: 7922e9ebfb9d6ef1792566554ba0c4a14f835ae2
Summary:
Working on zIndex tests I found that calculateShadowViewMutationsForNewTree (that we use for testing) does not take zIndex into account. This fixes it.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D25576923
fbshipit-source-id: a71b3a4630430488c783cd5010c0fbb7273bdaa5
Summary:
The test covers most props that must generate views.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D25576921
fbshipit-source-id: df5bedb8f6d409b5142e472ca2edcb1953bee4e1
Summary:
The `index` parameter for UpdateMutation is optional, and is normally just -1. It's not useful, so remove it. `parentShadowView` is also not relevant and is not used; in some existing use-cases the actual parent view of the updated view is available, and in some contexts the parent view is not set.
The function now will always set the index to -1 for UpdateMutations, and `{}` for ParentShadowView.
This should have no impact on iOS or Android, as this parameter is not used. It could theoretically have an impact on lifetimes of objects retained (now not retained) by not passing parentShadowView into the mutation. For example, any shared props or state associated with the parent will not be retained in the Update mutation now.
Changelog: [Internal]
Reviewed By: shergin
Differential Revision: D25342943
fbshipit-source-id: 0ddbef76a6e2eefc2629c9729f721d8674d7737e
Summary:
The "reparenting differ" has been the default differ for several months; ship it by removing config and the old differ.
Some functions can't be deleted yet because unit testing relies on it heavily; this can be refactored in the future if we care a lot.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25257205
fbshipit-source-id: 6f1dcc490bb1efe3d12506addf5f0843ca48c5c6
Summary:
Changelog: [internal]
The assert is still firing, let's disable it until we can investigate why layout animations creates two delete mount instructions.
Reviewed By: majak
Differential Revision: D25216794
fbshipit-source-id: 6328a2afb5eaf7fceebdc05bc75804f2eb44ddd2
Summary:
The current implementation of LayoutAnimations assumed that the "previous/old" ShadowView passed into the diff mutation didn't matter except for purposes of diffing.
As it turns out, iOS components could possibly use the "old" version of props, state, etc - so we should try to keep track of the current value in the tree as much as possible.
This diff accomplishes that by keeping track of the "previous" view, which the AnimationDriver will update over time. This also allows us to simplify logic around conflicting animations.
I'm also adding a few additional asserts to assist in debugging.
This doesn't totally eliminate all asserts hit on iOS, yet, but it does reduce the number of times the asserts are hit in StubViewTree.
Changelog: [Internal]
Reviewed By: sammy-SC
Differential Revision: D25048644
fbshipit-source-id: d00aeece5af04624d8193063be453c7ce4a6e565
Summary:
This diff adds more enforcement for consistency of `ShadowNodeMutation`s, including:
* `Props` object for newly created or updated view must not be nullptr;
* `oldShadowView` must describe the previous state of the view for `Update` instruction;
* `ignoreDuplicateCreates` option was removed.
I suspect some of the crashes we see in Fabric are caused by a violation of one of these constraints. If one of these fails in debug builds, we will get an early signal.
Changelog: [Internal]
Reviewed By: JoshuaGross
Differential Revision: D24880821
fbshipit-source-id: 8c8a3d8e205ce34f6e0335e8a2b0cf676930c284
Summary:
Changelog: [internal]
Background executor performs unnecessary operations when second `completeRoot` message from React arrives before first `completeRoot` was finished. This produces unnecessary `ShadowViewMutations`.
Mechanism:
Everytime `completeRoot` is received, before the call is dispatched on the background queue, `completeRootEventCounter_DO_NOT_USE_` is incremented.
Inside `ShadowTree::tryCommit` we check if the value has been incremented to determine if another `completeRoot` is queued.
Reviewed By: JoshuaGross
Differential Revision: D24419160
fbshipit-source-id: 11e19026feca01db6c8981b093a691a6b58a006f
Summary:
Changelog: [internal]
There are two exceptions inside `TelemetryController::pullTransaction`:
- Empty Optional cannot be unwrapped
- mutex lock failed: Invalid argument
By marking this method `noexcept`, stack trace is lost and it makes it more difficult to track down the issue.
What does compiler do if a method is marked `noexcept`?
```
void f() noexcept {
try {
// do work
}
catch (...) {
std::terminate(); // This is the std::terminate() we are seeing in stack traces.
}
}
```
Removing noexcept specifier might give us more information about the exception.
Reviewed By: JoshuaGross
Differential Revision: D24477861
fbshipit-source-id: 80f26e9ab160a5330c2848b89a01d60bfc0a4611
Summary:
changelog: [internal]
Prevents 2 type converions:
1. int <-> size_t
2. int <-> int32_t
# Why is using size_t better when working with indexes.
## 1. Type conversion isn't for free.
Take this example
```
size_t calculate(int number) {
return number + 1;
}
```
It generates following assembly (generated with armv8-a clang 10.0.0):
```
calculate(int): // calculate(int)
sub sp, sp, #16 // =16
str w0, [sp, #12]
ldr w8, [sp, #12]
add w9, w8, #1 // =1
mov w8, w9
sxtw x0, w8
add sp, sp, #16 // =16
ret
```
That's 9 instructions.
If we get rid of type conversion:
```
size_t calculate(size_t number) {
return number + 1;
}
```
Assembly (generated with armv8-a clang 10.0.0):
```
calculate(unsigned long): // calculate(unsigned long)
sub sp, sp, #16 // =16
str x0, [sp, #8]
ldr x8, [sp, #8]
add x0, x8, #1 // =1
add sp, sp, #16 // =16
ret
```
Compiler now produces only 7 instructions.
## Semantics
When using int for indexing, the type doesn't say much. By using `size_t`, just by looking at the type, it gives the reader more information about where it is coming from.
Reviewed By: JoshuaGross
Differential Revision: D24332248
fbshipit-source-id: 87ef982829ec14906ed9e002ea2e875fda4a0cd8
Summary:
Index adjustment is tricky. Seems more reliable to adjust each immediate mutation, and then immediately adjust delayed mutations based on it, rinse and repeat.
Previously it was possible to construct examples where the UI would get into a weird state because index adjustment caused items to be inserted in the wrong location.
Changelog: [Internal]
Reviewed By: kacieb
Differential Revision: D24232926
fbshipit-source-id: f8c445213528c2d2aedacf3e0c73c5bbeb62bc3d
Summary:
In the new Flattening differ, I experimentally verified that these two code paths are not hit (or redundant) and deleted them.
One of the branches did nothing and the other produced duplicate DELETE mutations for the same tag, that is handled elsewhere.
Changelog: [Internal]
Reviewed By: fkgozali
Differential Revision: D23806161
fbshipit-source-id: 9ad2929e2d719a7b9b34640ed35f7a696103604b
Summary:
See additional assertion. Tests still pass, so no other change was necessary.
Changelog: [Internal]
Differential Revision: D23775553
fbshipit-source-id: 57d3191f25dd55ab4da189207f6d201a31b175e0
Summary:
Currently, MountingCoordinator's RN_SHADOW_TREE_INTROSPECTION code will crash often because it assumes there is always a "new" tree to compare the old tree to. In the LayoutAnimations context this is not always the case - in fact, the majority of the time, LayoutAnimations is producing mutations for animation without a "new" tree.
Just check that the tree exists before trying to print it.
Changelog: [Internal]
Reviewed By: shergin
Differential Revision: D23747289
fbshipit-source-id: a1ba22aeae32ed8915a53bc33cdc199e8ce5128a
Summary:
In MountingCoordinator override mode (used in LayoutAnimations) we must set the start and end `diff` time when no real diff happens, otherwise we will hit an assert in telemetry later.
I also ensure that the TransactionNumber is incremented in that case.
Changelog: [Internal]
Reviewed By: shergin
Differential Revision: D23746684
fbshipit-source-id: b1fe3864e453fdba89d43cc827bd37434abf7a4d
Summary:
We don't need a shared_ptr here and without it the code will be faster and simpler.
This change is aligned with any clone-line callbacks we have in the Core which accepts a `const &` and return `shared_ptr<>`.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: JoshuaGross, sammy-SC
Differential Revision: D23725687
fbshipit-source-id: 1cd959f4273913175d342302e2f12752f0114768
