Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix: Android gradle config when bundling for release #28415

Closed
wants to merge 4 commits into from

Conversation

vdmtrv
Copy link
Contributor

@vdmtrv vdmtrv commented Mar 27, 2020

Summary

This fix aims to address the issue when bundling an Android app for release and getting the error exhibited in #28002 which I also encountered myself.

The config was changed sometime in November 2019 (as part of #26940, commit a3b0804) to be very opinionated when it comes to the use of npx which Gradle itself cannot find anyway (I have npx installed globally and it didn't pick it up).

Another issue that the use of npx creates is that Gradle should only ever use the currently installed react-native cli rather than a (possibly) higher version which may not always have backward compatibility.

The proposed change simply throws a more descriptive error rather than defaulting to a tool which may or may not exist on the machine, be it CI or a development environment. I've also modified the RNTester app to reflect the correct config implementation relative to the RNTester app itself.

In real projects, the config inside android/app/build.gradle should look similar to the following snippet:

project.ext.react = [
  cliPath: "$rootDir/../node_modules/react-native/cli.js",
  entryFile: "index.js"
];

Changelog

[Android] [Fixed] - Gradle release config

Test Plan

  • Successfully bundled an Android release build with correct config
  • Works with RNTester app

@facebook-github-bot facebook-github-bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Mar 27, 2020
@react-native-bot react-native-bot added Bug Platform: Android Android applications. labels Mar 27, 2020
@react-native-bot react-native-bot added the Tech: Bundler 📦 This issue is related to the bundler (Metro, Haul, etc) used. label Mar 27, 2020
@analysis-bot
Copy link

Platform Engine Arch Size (bytes) Diff
android hermes arm64-v8a 6,596,588 0
android hermes armeabi-v7a 6,273,741 0
android hermes x86 6,936,744 0
android hermes x86_64 6,866,547 0
android jsc arm64-v8a 8,904,972 0
android jsc armeabi-v7a 8,545,447 0
android jsc x86 8,729,149 0
android jsc x86_64 9,305,192 0

Base commit: d3658bc

@analysis-bot
Copy link

analysis-bot commented Mar 27, 2020

Platform Engine Arch Size (bytes) Diff
ios - universal 966,896 -9,833,328

Base commit: d3658bc

react.gradle Outdated
@@ -110,7 +110,8 @@ afterEvaluate {
execCommand.addAll([*nodeExecutableAndArgs, cliPath])
}
} else {
execCommand.addAll([npx, "react-native"])
throw new Exception("Missing cliPath or nodeExecutableAndArgs from build config. " +
"Please set set project.ext.react.cliPath to the path of the react-native cli.js");
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thank you for taking your time to improve this experience! Do you think we should mention android/app/build.gradle explicitly in the error message so it is clear where the config is located?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You're correct. It's probably worth adding. I was simply following the convention on line 49:

throw new Exception("OS not recognized. Please set project.ext.react.hermesCommand " +
                        "to the path of a working Hermes compiler.");

If we include the path in this exception message then it needs to be added to the one on line 49.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm OK merging this as is, since it follows convention. I'm not as familiar with Gradle, so my main worry is whether these error messages are clear to the average Android developer. Would they know where to go to fix this config? Does the config location vary across workspaces or is it generally expected at the same path?

In other words, if the config is always at android/app/build.gradle then perhaps we can explicitly point to that file.

Copy link
Contributor

@facebook-github-bot facebook-github-bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@hramos has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

@react-native-bot
Copy link
Collaborator

This pull request was successfully merged by @vdmtrv in 0d1fb45.

When will my fix make it into a release? | Upcoming Releases

@react-native-bot react-native-bot added the Merged This PR has been merged. label Mar 28, 2020
@grabbou
Copy link
Contributor

grabbou commented Apr 14, 2020

Thank you @vdmtrv and @hramos for working on this PR!

We have shipped a follow-up PR a few days after that which completely removed the usage of npx to fix this and a few other issues. We rely on pure node resolution mechanism, which works in all environments and is already used by the CLI to resolve additional modules from node_modules.

In this particular case, I'd consider this a downgrade in terms of the DX - we will now need to propagate cliPath to be added back to all React Native projects, including non-standard setups (such as mono-repositories) where this may be confusing. We never required this configuration, it was only meant to enable React Native to work in non-standard project configurations.

By default, the cliPath is defined here. I guess we should use that one instead, when custom config.cliPath is not present, rather than throwing an error.

That said, I am going to submit a follow-up PR which reverts the requirement of cliPath and aligns this with the rest of the CLI work. In other words, it will resolve the CLI path, but without using npx to do so.

@grabbou
Copy link
Contributor

grabbou commented Apr 14, 2020

I am going to send a follow-up PR as well to update code owners to better detect changes around that.

@vdmtrv
Copy link
Contributor Author

vdmtrv commented Apr 14, 2020

Thank you for providing more context, @grabbou!

I'll keep the points you raised in mind should I come across something similar in the future and CC you as requested. I was not aware that this was being worked on as it wasn't mentioned in any of the open issues related to the use of npx.

I did see config.cliPath, but the behaviour I was able to observe was that the default CLI path was not being respected. Because of this, it fell into the else, where it was using npx, which it couldn't find, hence the rationale behind the commit.

We have shipped a follow-up PR a few days after that which completely removed the usage of npx to fix this and a few other issues.

Can we have a link to the commit/PR for reference, please?

Lastly, I think it's worth linking to the PR which is aiming to improve the above for anyone's future reference - #28625

@grabbou
Copy link
Contributor

grabbou commented Apr 14, 2020

No worries @vdmtrv, hard to stay on top of everything - we should update the code owners to better reflect that state. Appreciate time spent on debugging - we wouldn't find a fix if it wasn't for you.

Some additional context: react-native-community/cli#886

alloy added a commit to microsoft/react-native-macos that referenced this pull request Sep 30, 2020
* Upgrade to Metro 0.59

Summary:
Upgrades RN to Metro 0.59.

Changelog: [Internal] Metro Upgrade

Reviewed By: motiz88

Differential Revision: D20533864

fbshipit-source-id: 3c5fb8e37d2363edf0b9a1a8cfbdefba00763415

* Fix mock for TextInput (#28332)

Summary:
This PR adds the `isFocused` method to the mock of the TextInput component. My understanding some of the latest changes on the TextInput to make it use a forwardRef change the way this method is mock giving an error when trying to use in on a mock.
The change suggested here fixes the issue.

## Changelog

[JavaScript] [Fixed] - Fix the mock for TextInput to support the `isFocused` method
Pull Request resolved: https://github.com/facebook/react-native/pull/28332

Reviewed By: cpojer

Differential Revision: D20538044

Pulled By: TheSavior

fbshipit-source-id: be734af105ab62ffdf9ed4017bd70845e207f8cd

* Properly handle LogBox errors during tests

Summary:
This diff fixes an issue where errors in LogBox during tests would cause the tests to crash.

The crash is due to the NativeExceptionsManager module not being mocked (as all native module need to be in tests). The fix is to properly mock the NativeExceptionManger.

This fix exposed an infinite loop issue where failures in LogBox will be logged to the ExceptionManager, which logs to the console, which logs to LogBox, creating a loop. This diff also fixes that look by moving the LogBox internal error check to the top of the monkey patched console methods.

Changelog: [Internal]

Differential Revision: D20428590

fbshipit-source-id: 7289a480c99ba8dee67772178b7629afb40b330a

* Back out "Track animations and flush them"

Summary:
Original commit changeset: b594d0e6e9b6

D20319824 introduced a problem in LayoutAnimations, which makes surfaced as the problem in T63911344. This diff reverts D20319824.

Changelog: [Internal]

Reviewed By: lunaleaps

Differential Revision: D20541918

fbshipit-source-id: ff72b839f57d39051122920a38b2632cbb5ec362

* Consolidate "dispatchMountItems" reentrancy prevention code, and retry code, in one function

Summary:
Simplifying the dispatchMountItems reentrance and retry logic.

Motivation: cleanup so I can work on dispatching ViewCommands before anything else.

Importantly, this gives us the properties that:
1) Only one function is responsible for calling dispatchMountItems
2) Only one function is responsible for deciding if we shouldn't call dispatchMountItems due to reentrance
3) Only one function is responsible for all cleanup
4) Only one function maintains all of the relevant flags (except dispatchPreMountItems... two total now, instead of 4 before)

Changelog: [Internal]

Reviewed By: mdvacca

Differential Revision: D20437035

fbshipit-source-id: 5370366790eb25f653bee6c1950e747458374a61

* Only retry ViewCommand mount items if exception is marked as "Retryable"

Summary:
Instead of just blindly retrying all ViewCommands if they fail - which could be dangerous, since it's arbitrary imperative commands we'd be executing twice, potentially with bad app state - we only retry if the ViewCommand throws a "RetryableMountingLayerException".

Changelog: [Internal] Optimization to ViewCommands

Reviewed By: mdvacca

Differential Revision: D20529985

fbshipit-source-id: 0217b43f4bf92442bcc7ca48c8ae2b9a9e543dc9

* Introduce early dispatch of ViewCommands in FabricUIManager

Summary:
Earlier this week I introduced a change in the old, non-Fabric renderer (D20378633 D20427803) that (gated behind a feature-flag) executes ViewCommands before all other types of commands, as a perf optimization and (I think) a potential fix for a category of race conditions.

I've added more details in comments here.

The Fabric renderer uses the same feature-flag that I introduced for the non-Fabric renderer.

Changelog: [Internal] Fabric

Reviewed By: mdvacca

Differential Revision: D20449186

fbshipit-source-id: bb3649f565f32c417a6247369902333989a043aa

* Change nightly build from hourly to nightly (daily at 00:00) (#28346)

Summary:
We initially added the nightly build test to run every hour, in order to more quickly validate it. Now that it has been validated we can run it every night as it is intended to do.

cc hramos

## Changelog

[General] [Changed] - Change nightly build from hourly to nightly
Pull Request resolved: https://github.com/facebook/react-native/pull/28346

Reviewed By: cpojer

Differential Revision: D20550143

Pulled By: hramos

fbshipit-source-id: 9487c6785684ad6ea7e877290d50a33118090a7f

* Add diffing to app bundle size reports (#28284)

Summary:
Add diffing to app bundle size reports.

## Changelog

[Internal] [Changed] - Add diffing to app bundle size reports

Pull Request resolved: https://github.com/facebook/react-native/pull/28284

Test Plan:
- App bundle size reports should now display a diff where available
  - Right now, the database contains only one entry for the last known good iOS build
- Triggering a new build should not create additional comments

Reviewed By: cpojer

Differential Revision: D20450158

Pulled By: hramos

fbshipit-source-id: 720772275f24d3ff0a49705f4dada2efe2e99bd3

* feat: improve monorepo support by removing redundant PROJECT_ROOT (#28354)

Summary:
Historically, React Native didn't support a lot of custom project structures apart from the standard flat directory with `ios` and `android` folders. The CLI had to be explicitly started from the project root, otherwise Metro didn't work right.

In order to resolve the project root in the most accurate way, React Native assumed that project root is always `../../` from its location in `node_modules` - this is not true when the installation gets hoisted (e.g. in a monorepo).

To address that, janicduplessis brought support for custom [`PROJECT_ROOT`](https://github.com/facebook/react-native/commit/9ccde378b6e6379df61f9d968be6346ca6be7ead) that allowed overriding the `../../` in case it wasn't true.

Today, CLI is able to automatically resolve the project root, no matter where it's started. It will traverse the tree of the directories upwards and stop as soon as it meets `package.json`.

As a result, it doesn't really matter from where we start the CLI anymore as a part of `react-native-xcode.sh`.

By replacing the default value of `$REACT_NATIVE_DIR/../../` with `$PWD, that is default for all Xcode scripts, we can make the setup for monorepo easier - nobody will need to set `$PROJECT_ROOT` in order to override the incorrect defaults.

By default, all scripts defined in Xcode run from `$PWD` directory, which is the location of the iOS project. In the future, we will be able to remove `cd` entirely.

To better understand this PR, let's look a few hypothetical structures as an example:

#### Monorepo:

> tl;dr works out of the box, no need to mess around with paths

```
- package.json
- packages/
  - my-app/
     - index.js
     - package.json
     - ios/
        - MyApp.xcodeproj
```

**Before this PR**, the `react-native-xcode.sh` will start the CLI like this:

```bash
cd $REACT_NATIVE_DIR/../../
node <absolute_path_to_cli.js> bundle --entry-point index.js
```

- Because we change the directory to the root of monorepo, CLI throws an error. All in all, there's no `react-native` dependency at the workspace root.

- Some users turn `no hoist` in an act of troubleshooting the errors, which resolves the problem - `react-native` is moved under `my-app/node_modules` which makes this mechanism resolve properly.

- Some users find out about `PROJECT_ROOT` and set it to overwrite the default value. For example, setting `export PROJECT_ROOT = "$PWD/../` will set the directory to `my-app`, which has a dependency on `react-native` in a `package.json` and makes the CLI happy.

**After this PR**, the `react-native-xcode.sh` will start the CLI like this:

```bash
cd $PWD
node <absolute_path_to_cli.js> bundle --entry-point index.js
```

- The `$PWD` is `packages/my-app/ios/` because that's where the Xcode project is located. CLI will automatically set the root to `../` because that's where it finds `package.json` with `react-native` dependency. It will pass that root to Metro, unless users have set a different one themselves. Thanks to that, all paths to JavaScript files remain working and unaffected.

- No need to set `PROJECT_ROOT` anymore.

- We don't rely on the location of `node_modules`, which is cleaner and future proof.

#### Standard:

> tl;dr no changes

```
- ios/
   - MyApp.xcodeproj
- index.js
- package.json
```

**Before this PR**, the `react-native-xcode.sh` will start the CLI like this:

```bash
cd $REACT_NATIVE_DIR/../../
node <absolute_path_to_cli.js> bundle --entry-point index.js
```

- Everything works fine. Path from `react-native` inside `node_modules` is correct - the project root is set right to `/`

**After this PR**, the `react-native-xcode.sh` will start the CLI like this:

```bash
cd $PWD
node <absolute_path_to_cli.js> bundle --entry-point index.js
```

- The root will be set to where Xcode project is located, which is `/ios`. This is the PWD for all Xcode scripts.

CLI will look for the `package.json` going upwards from `ios` folder. Will stop at `/`, find out it has `react-native` dependency, load it and its commands and proceed further.

## Changelog

[iOS] [Feature] - Better monorepo support when building release apk
Pull Request resolved: https://github.com/facebook/react-native/pull/28354

Test Plan:
- All projects (standard/monorepo) run without issues.
- PROJECT_ROOT is not needed.

CC: Titozzz (who wrote monorepo guide), alloy, bartolkaruza

Reviewed By: cpojer

Differential Revision: D20558005

Pulled By: hramos

fbshipit-source-id: 2551120beadcfd4c2f1393ce8a2c2fa6b93c9290

* Fix `test_android`: Remove references to fbsource cell (#28363)

Summary:
Fixes https://github.com/facebook/react-native/issues/28361.

## Changelog

[Internal] [CI] - Fix test_android
Pull Request resolved: https://github.com/facebook/react-native/pull/28363

Test Plan:
Prior to fix:

```
react-native $ ./scripts/circleci/buck_fetch.sh
Guessing 168a69309928ba16065cdb33b1775a4af9f924a6 as the last one used version.
Using additional configuration options from /Users/hramos/.buckconfig.d/experiments, /etc/buckconfig.d/fb_chef.ini, /etc/buckconfig.d/fb_chef_override.ini
Invalidating internal cached state: Watchman failed to start. This may cause slower builds.
Parsing buck files: finished in 1.5 sec
Buck wasn't able to parse /Users/hramos/git/react-native/ReactAndroid/src/main/java/com/facebook/fbreact/specs/BUCK:
IOError: [Errno 2] No such file or directory: '/Users/hramos/git/react-native/tools/build_defs/platform_defs.bzl'
Call stack:
  File "/Users/hramos/git/react-native/.buckd/resources/168a69309928ba16065cdb33b1775a4af9f924a6/buck_server/buck_parser/profiler.py", line 507, in wrapped
    return func(*args, **kwargs)
  File "/Users/hramos/git/react-native/ReactAndroid/src/main/java/com/facebook/fbreact/specs/BUCK", line 1
    load("//tools/build_defs:platform_defs.bzl", "ANDROID")
  File "/Users/hramos/git/react-native/.buckd/resources/168a69309928ba16065cdb33b1775a4af9f924a6/buck_server/buck_parser/profiler.py", line 507, in wrapped
    return func(*args, **kwargs)

This error happened while trying to get dependency '//ReactAndroid/src/main/java/com/facebook/fbreact/specs:FBReactNativeSpec' of target '//ReactAndroid/src/main/java/com/facebook/react/devsupport:devsupport'
```
After fix:

```
react-native $ ./scripts/circleci/buck_fetch.sh
+ buck fetch ReactAndroid/src/test/java/com/facebook/react/modules
Guessing 168a69309928ba16065cdb33b1775a4af9f924a6 as the last one used version.
Using additional configuration options from /Users/hramos/.buckconfig.d/experiments, /etc/buckconfig.d/fb_chef.ini, /etc/buckconfig.d/fb_chef_override.ini
Invalidating internal cached state: Watchman failed to start. This may cause slower builds.
Parsing buck files: finished in 1.1 sec
Configuration 'ANDROID_SDK' points to an invalid directory '/opt/android_sdk'.
    When creating rule //ReactAndroid/src/main/java/com/facebook/hermes/instrumentation:instrumentation.
```

> Note: I don't have the Android SDK configured in this machine.

Verified on Circle CI. `test_android` is now green: https://circleci.com/gh/facebook/react-native/140682?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link

Reviewed By: cpojer

Differential Revision: D20564934

Pulled By: hramos

fbshipit-source-id: 5d843b8f113c4db5391ee39addc3ff259d962290

* Fix TextInput left/right padding

Summary:
This fixes two things:

1) Currently it only respects Start and End padding, and if there's a Theme default, it will override Left/Right padding. Whoops.
2) Currently it doesn't respect when a TextInput starts with padding, but then is removed.

This resolves both.

It still does not account for RTL support.

Changelog: [Internal] Fix AndroidTextInput padding

Reviewed By: mdvacca

Differential Revision: D20573151

fbshipit-source-id: e89791641b6699e728cde9dbd661a8c21485fbc8

* Validate selection range passed to setTextAndSelection

Summary:
Changelog: [Internal]

# Fabric

1. If `start` and `end` parameters in `setTextAndSelection` are -1, we don't move the cursor. Previously the cursor would be moved to beginning of text input.

2. In view commands, do not validate `eventCount`. It is passed in as undefined from JS because Fabric's text input doesn't use `eventCount`.

# Paper

1. If `start` and `end` parameters in `setTextAndSelection` are -1, we don't move the cursor. Previously the cursor would be moved to beginning of text input.

Reviewed By: shergin

Differential Revision: D20538290

fbshipit-source-id: c7aeddc25f58697254474058ce901df958321f7c

* Remove ReactTypes from fbsource and React sync

Summary:
See https://github.com/facebook/react/pull/18366

This contains a fork of the upstream Flow types. We shouldn't be syncing this since these leads to conflicting types.

As a result, these uses have already been codemodded away. Only the imports remained.

Changelog:

[React Core] - Remove ReactTypes from sync.

Reviewed By: gaearon

Differential Revision: D20583740

fbshipit-source-id: fc86a934cbdca8ff90fe90282b86ecc945a85e5f

* Fix controlled TextInput with child nodes

Summary:
Changelog: [Internal]

# There are three changes in this diff

## _stateRevision is replaced with a BOOL
`_stateRevision` was protecting against setting attributed string that is already visible to the user. Previously this was ok because the change was only coming from native, any changes from JS were ignored.

Imagine following scenario:

1. User taps key.
2. Update state is called on component initiated by native.
3. New state is created with incremented revision by one.
4. `_stateRevision` gets set to new state's revision + 1.
5. Now JS wants to change something because it just learnt that user tapped the key.
6. New state is created again with incremented revision by one.
7. Update state is called on the component, but the change isn't applied to the text view because `_state->getRevision()` will equal `_stateRevision`.

By having a BOOL instead of number, we very explicitly mark the region in which we don't want state changes to be applied to text view.

## Calling [_backedTextInputView setAttributedText] move cursor to the end of text input
This is prevented by storing what the current selection is and applying it after `[_backedTextInputView setAttributedText]` is called.
This was previously invisible because JS wasn't changing contents of `_backedTextInputView`.

## Storing of previously applied JS attributed string in state

This is the mechanism used to detect when value of text input changes come from JavaScript. JavaScript sends text input value changes through props and as children of TextInput.
We compare what previously was set from JavaScript to what is currently being send from JavaScript and if they differ, this change is communicated to the component.
Previously only first attributed string send from JavaScript was send to the component.

# Problem

If children are used to set text input's value, then there is a case in which we can't tell what source of truth should be.

Let's take following example
We have a text field that allows only 4 characters, again this is only a problem if those 4 characters come as children, not as value.
This is a controller text input.

1. User types 1234.
2. User types 5th character.
3. JavaScript updates TextInput, saying that the content should stay 1234.
4. In `TextInputShadowNode` `hasJSUpdatedAttributedString` will be set to false, because previous JS value is the same as current JS value.

Reviewed By: shergin

Differential Revision: D20587681

fbshipit-source-id: 1b8a2efabbfa0fc87cba210570142d162efe61e6

* Daily `arc lint --take BUCKFORMAT`

Reviewed By: zertosh

Differential Revision: D20593906

fbshipit-source-id: b056947c698508119dc9d4d1bba202295b8f0fda

* RN picker - implement background color

Summary:
add support to the android implementation of the Picker component for setting the background color.

Changelog: [Android] [Added] - Support item background color in Dialog Picker

Differential Revision: D20566131

fbshipit-source-id: d693b40803fa1051ec955c5728994c820fecd9e9

* Fabric: Modernizing Yoga Dirty flag test.

Summary:
Now we using TEST_F thing that allows consilidating initialization.

Changelog: [Internal] Fabric-specific internal change.

Reviewed By: mdvacca

Differential Revision: D20578788

fbshipit-source-id: 103bcb8fdeb3dbf297385cfe56415bd646e16791

* Fabric: Changing signature of `ComponentDescriptor::createState`

Summary:
This is pure syntactic change. Often we don't have a shared pointer to ShadowNodeFamily and only have just a reference. At the same time, `ComponentDescriptor::createState` does not have to accept a shared pointer. So, it's better to accept just a reference.

Changelog: [Internal] Fabric-specific internal change.

Reviewed By: sammy-SC

Differential Revision: D20578787

fbshipit-source-id: 905277001e096d41e75007575b59ea2ea15fbf4b

* Fabric: View Test: Changing state should not dirty Yoga tree (in some most cases)

Summary: Changelog: [Internal] Fabric-specific internal change.

Reviewed By: mdvacca

Differential Revision: D20578789

fbshipit-source-id: 4336165217bd39fc8065cfaeb96ef7753433d48a

* Get ReactiveNative compiled with Clang 10 (#28362)

Summary:
Pull Request resolved: https://github.com/facebook/react-native/pull/28362

Fixed a few compilation errors emitted by Clang 10.

Changelog:
[iOS] [Fixed] - Get ready for Clang 10

Differential Revision: D20549970

fbshipit-source-id: dc36a85d90d3e43a05f045feb57c6ab6ded67da7

* Guard against null values in object parameters for bridged methods

Summary:
Handles the case when a value in an object parameter of a turbo module spec is null (even if the type is nullable).

For example, given:
```
export interface Spec extends TurboModule {
  +myFunc: ({|
    foo: ?string,
  |}) => void;
}
```
and calling `NativeModule.myFunc({foo: null})`, we see an error like:
```
JSON value '<null>' of type NSNull cannot be converted to NSString
```
Guarding against this by converting NSNull's to nils

## Changelog:

[iOS] [Fixed] - Fix crash when passing null value in object parameter of bridged method

Reviewed By: fkgozali

Differential Revision: D20591590

fbshipit-source-id: fdb90f34131427a235f2e3c99147bf1e6a9c6732

* Modify pending deletion tags to be cross manageChildren

Summary:
Changelog: [Internal]

Removing historic layout animations index adjustment (D20323928) broke the Dating Secret Crush screen.

Since flushing animations (D20319824) had to be reverted due to issues with Saved + Privacy Shortcuts (https://fburl.com/tasks/eijtmifu) we need to track pending deletions across `manageChildren` operations.

Reviewed By: JoshuaGross

Differential Revision: D20601079

fbshipit-source-id: c6f116683750e97abe7f988cf361d2a6449e90e6

* Enable label-actions on the react-native repository (#28374)

Summary:
Enhance our issue management workflow by having the bot respond automatically whenever a label is applied to the issue.

## Changelog

[Internal] - CI
Pull Request resolved: https://github.com/facebook/react-native/pull/28374

Test Plan: Not tested. If needed, could be applied to a different, test repository.

Reviewed By: cpojer

Differential Revision: D20606887

Pulled By: hramos

fbshipit-source-id: 874d1464527ea76bf51394a7d3e98e4fd8f69345

* Fix Animated Value initialized with undefined in ScrollView (#28349)

Summary:
When passing an object to contentOffset that doesn't have `y` prop set it causes the following error:

```
 Error: AnimatedValue: Attempting to set value to undefined

This error is located at:
    in ScrollView (at src/index.js:638)
...
```

This happens since a runtime check was added to the `AnimatedValue` constructor. (a3aaa471eca58b31597b9a0669f7ade385ccb175)

According to flow types the object passed to contentOffset should always contain both x and y props but since it worked before when y is undefined I think its fine to patch the runtime behaviour defensively, especially since the code change is simple.

## Changelog

[General] [Fixed] - Fix Animated Value initialized with undefined in ScrollView
Pull Request resolved: https://github.com/facebook/react-native/pull/28349

Test Plan: Tested that the crash no longer reproduces when passing an empty object to contentOffset.

Reviewed By: cpojer

Differential Revision: D20601664

Pulled By: hramos

fbshipit-source-id: b098a2dd1e702f995a9a92fa6e4e9a204187dac4

* xplat/js/react-native-github/ReactCommon/fabric/components/textinput/

Summary: Changelog: [Internal]

Reviewed By: scottrice

Differential Revision: D20619227

fbshipit-source-id: 674337e6ce585a4e96d020f9624b874ba86e2d80

* Seed ssh known hosts with github's public key (#28370)

Summary:
The [previous attempt](https://github.com/facebook/react-native/pull/28304) to fix the publish step failed, so now reverting to manually configuring things. This PR adds an entry to SSH’s `known_hosts` file using github.com’s public key that I have verified as per [these instructions](https://serverfault.com/a/807363):

```
~/C/R/react-native [master] » nmap github.com --script ssh-hostkey
Nmap scan report for github.com (140.82.118.4)
rDNS record for 140.82.118.4: lb-140-82-118-4-ams.github.com
PORT     STATE SERVICE
22/tcp   open  ssh
| ssh-hostkey:
|   1024 ad:1c:08:a4:40:e3:6f:9c:f5:66:26:5d:4b:33:5d:8c (DSA)
|_  2048 16:27:ac:a5:76:28:2d:36:63:1b:56:4d:eb:df:a6:48 (RSA)
```

These fingerprints line up with [the ones posted by GitHub](https://help.github.com/en/github/authenticating-to-github/githubs-ssh-key-fingerprints), so my setup should be good and can be trusted to grab the public key from the right host:

```
~/C/R/react-native [master] » ssh-keyscan -t rsa -H github.com
# github.com:22 SSH-2.0-babeld-d48c3acd
|1|If6MU203eXTaaWL678YEfWkVMrw=|kqLeIAyTy8pzpj8x8Ae4Fr8Mtlc= ssh-rsa AAAAB3NzaC1yc2EAAAABIwAAAQEAq2A7hRGmdnm9tUDbO9IDSwBK6TbQa+PXYPCPy6rbTrTtw7PHkccKrpp0yVhp5HdEIcKr6pLlVDBfOLX9QUsyCOV0wzfjIJNlGEYsdlLJizHhbn2mUjvSAHQqZETYP81eFzLQNnPHt4EVVUh7VfDESU84KezmD5QlWpXLmvU31/yMf+Se8xhHTvKSCZIFImWwoG6mbUoWf9nzpIoaSjB+weqqUUmpaaasXVal72J+UX2B+2RPW3RcT0eOzQgqlJL3RKrTJvdsjE3JEAvGq3lGHSZXy28G3skua2SmVi/w4yCE6gbODqnTWlg7+wC604ydGXA8VJiS5ap43JXiUFFAaQ==
```

## Changelog

[Internal] [Fixed] - Make automated publishing of packages from CI work again
Pull Request resolved: https://github.com/facebook/react-native/pull/28370

Test Plan: I used the command being added in this PR in [a failed CI job](https://app.circleci.com/pipelines/github/facebook/react-native/4104/workflows/916127cb-177f-4583-9f90-cae5318041d8/jobs/140810). When I invoked the publish script manually I was not greeted by the blocking prompt and the package was successfully published: https://www.npmjs.com/package/react-native/v/0.0.0-56cf99a96

Reviewed By: cpojer

Differential Revision: D20601527

Pulled By: hramos

fbshipit-source-id: b1a4405228408cfc4a1b3b44ab88c79522af3a66

* Fix app bundle size diff not always being compared against latest commit (#28368)

Summary:
- Timestamp of entries in our Firebase instance sometimes get stored as number. This means that we may not always be diffing against the latest master commit.
- Size report of Android and iOS gets overwritten depending on which build finishes first.

## Changelog

[Internal] [Fixed] - App bundle size diff not always being compared against latest commit
[Internal] [Fixed] - Android and iOS app bundle size diff overwrite each other
Pull Request resolved: https://github.com/facebook/react-native/pull/28368

Test Plan:
- We are now using Firebase's own [firebase.firestore.Timestamp.now](https://firebase.google.com/docs/reference/js/firebase.firestore.Timestamp#now) to ensure that we always get a timestamp in the preferred format. This has been tested locally but can only be verified when merged to master and we start getting new data. In the meantime, I'll manually fix up all the entries in the store.
- There should be one app bundle size comment for Android and one for iOS in this PR.

Reviewed By: cpojer

Differential Revision: D20601620

Pulled By: hramos

fbshipit-source-id: 0c3e4b78a74cbd659f1957a6aa74322b016e0646

* Hopefully fix so loading crashes

Summary:
Changelog:

[Android][Internal] Fix potential initializer interruption threading crashes.

Reviewed By: mdvacca

Differential Revision: D20615755

fbshipit-source-id: 58b706deeb6df1998caff5bf2ae9ec60114313fe

* Fix label-actions configuration

Summary:
Adds back a missing label key, fixes open source issue: https://github.com/facebook/react-native/issues/28378

Changelog:
[Internal] [CI] - Fix label-actions config

Reviewed By: cpojer

Differential Revision: D20625887

fbshipit-source-id: 63c90db249aa9c15369a4b5bcab71cbe75c6d4b8

* Changing Order Of mOverrideColorScheme In Constructor

Summary:
Changelog: [Android] [Updated]

mOverrideColorScheme should be assigned before the first colorSchemeForCurrentConfiguration call, so the initial setting of mColorScheme will reflect the override

Reviewed By: zackargyle

Differential Revision: D20630173

fbshipit-source-id: a2a2d174d3fc40c14f27dce6a7fa8e67203480c9

* hermes | inspector | Don't include posix headers on non-posix systems

Summary:
Changelog: [Internal]

Hermes inspector includes pthreads, arpa and sys headers on all OSes that would break vanilla Windows builds. This diff adds a check for posix-compliance before inclusion

(Note: this ignores all push blocking failures!)

Reviewed By: dulinriley

Differential Revision: D20564449

fbshipit-source-id: 8e264bc3104065dc4315bb291e8560609fe65184

* Upgrade Prettier from 1.17 to 2.0.2.

Summary:
This gets us on the latest Prettier 2.x:
https://prettier.io/blog/2020/03/21/2.0.0.html

Notably, this adds support for TypeScript 3.8,
which introduces new syntax, such as `import type`.

Reviewed By: zertosh

Differential Revision: D20636268

fbshipit-source-id: fca5833d003804333a05ba16325bbbe0e06d6c8a

* Back out "Upgrade Prettier from 1.17 to 2.0.2."

Differential Revision: D20639755

fbshipit-source-id: 5028563f9cf0527a30b4259daac50cdc03934bfd

* Fabric: Additional temporary checks in prop parsing infra

Summary:
While ViewConfig infra isn't perfect we need to check some value for correctness during prop-parsing.

Changelog: [Internal] Fabric-specific internal change.

Reviewed By: JoshuaGross

Differential Revision: D20639055

fbshipit-source-id: 193dcd0769bc7777bc8d60c964ede72ebdaa83e4

* Update React package

Summary:
This just updates the `react` package to the latest stable version. We updated it to experimental internally earlier so this brings the open source version to the latest before the branch cut. This doesn't include any breaking changes.

Changelog:

[General][Changed] - Update to React 16.13.1

Reviewed By: cpojer

Differential Revision: D20642909

fbshipit-source-id: 68a4c74bfe72f1abdb33b0b9071a4f4e8e568318

* Fix sketchy null checks induced by new formatting in Prettier 2.0

Summary:
Update code to prepare for Prettier 2.0, which will
reformat `a || (b || c)` to `a || b || c`.

Changelog: [Internal] prepare for Prettier 2.0

Reviewed By: kassens

Differential Revision: D20639483

fbshipit-source-id: c2932b1495884684172ba9291d56c546f51711b8

* RN picker - fix types in AndroidDialogPickerManagerInterface

Summary:
according to [this crash report](https://our.intern.facebook.com/intern/logview/details/facebook_android_crashes/7ba7056481015482c6166d65cb97e49d/?trace_key=1506fe36a70dd5e50cdc8968f6317f27), `value` was throwing an NPE despite being null-checked. this is because it was an `int` rather than an `Integer`, so the null check wasn't working

Changelog: Fix types in AndroidDialogPickerManagerInterface

Reviewed By: mdvacca

Differential Revision: D20646343

fbshipit-source-id: a27587e0a48f5782bcf5ffddb604018218e65206

* Remove RCTExportModule log spam

Summary:
The bridge complains if modules aren't exported, which isn't really helpful with lazily loaded modules and turbo modules.

I considered only turning this off when TurboModules is enabled, but figured we'd be killing this soon anyways... If anyone feels strongly I can go that approach.

Changelog: [iOS][Internal] Remove RCTExportModule log spam

Reviewed By: shergin

Differential Revision: D20629575

fbshipit-source-id: d32d9fe244c4d06acfee982fca7c7f63da294dc5

* De-jank DevLoadingView

Summary:
## Problems
Repro steps:
1. Disable Fabric (because CMD + R doesn't work with Fabric right now).
2. Open up Marketplace and hit `CMD + OPT + R`
3. **Observe:** The progress bar doesn't show up right away. It also doesn't actually show progress.
https://pxl.cl/140g1

RN Support post: https://fb.workplace.com/groups/rn.support/permalink/3437652016283389/

## Fixes
The first problem is that progress bar doesn't actually show progress.

**Fix:** Bundle load progress is updated in `RCTCxxBridge`, where we first require `RCTDevLoadingView`, and then call its `updateProgress` method. Previously, we wouldn't  lazily load `RCTDevLoadingView`, it already didn't exist. Lazily loading `RCTDevLoadingView` causes the progress view to show up. Here: https://pxl.cl/140gt

If you look at the above video, you'll notice there are two stages to the progress bar: stage 1 displays the actual progress. Stage 2 prompts that we're downloading the JS bundle. As you can see, stage 1 and stage 2 have different background colors, even though both of them are green.

**Fix:** I adjusted the JS to match the Native color. Here: https://pxl.cl/140gT

We're almost there, but the progress bar is dismissed twice?

**Fix:** I dug into the code, and the reason why was because when we hit `CMD + R`, we invalidate the bridge, and immediately re-initialize it. This means that we asynchronously invalidate the old TurboModuleManager, and immediately create a brand new one. Therefore, two `RCTDevLoadingView` modules can (and do) exist at once. So, I moved `RCTDevLoadingView` to be an instance member of `FBReactModule`, to ensure that it doesn't get cleaned up and re-created when TurboModuleManager is deleted and re-created. This finally fixed the progress bar jank:
https://pxl.cl/140hn

Changelog:
[iOS][Fixed] - Remove RCTDevLoadingView jank

Reviewed By: rickhanlonii

Differential Revision: D20607815

fbshipit-source-id: 05825c67adaf3cfda70be0fa2dc92d413dc8921b

* Fix retaining self in block in LogBox impl

Summary:
Logbox has a retain cycle (see linked task for my deeper investigation).

This diff doesn't fix the retain cycle, but it's just good practice to not retain self strongly in blocks.

Changelog: [iOS][Internal] Fix retaining self in block in LogBox implementation

Reviewed By: shergin

Differential Revision: D20630693

fbshipit-source-id: cf399495e9bcd1917932fcc0e9c9d2d2a32bf6f0

* Flow type infoLog

Summary:
Changelog:
[General][Internal] flow type infoLog

Reviewed By: zackargyle

Differential Revision: D20577939

fbshipit-source-id: eed4401b2ae0a6bf845fdcb54c6abe1fe98fe7c1

* Replace fbsource// with // in xplat/js/ files [1]

Summary:
`fbsource//xplat` and `//xplat` are equivalent for FB BUCK targets. Removing extra prefix for consistency.

Changelog: [Internal]

Reviewed By: scottrice

Differential Revision: D20495655

fbshipit-source-id: a57b72f694c533e2e16dffe74eccb8fdec1f55f5

* Deploy Flow 0.121 to Xplat (#901)

Summary:
Deploy Flow 0.121 to Xplat

bypass-lint
allow-large-files

Closes https://github.com/facebook/flipper/pull/901

Changelog: [Internal]

Reviewed By: panagosg7

Differential Revision: D20570316

fbshipit-source-id: a76983d6f46c8b995ce2dd5cd1e014534790698a

* Replace fbsource// with // in xplat/js/ files [3]

Summary:
`fbsource//xplat` and `//xplat` are equivalent for FB BUCK targets. Removing extra prefix for consistency.

Changelog: [Internal]

Reviewed By: JoshuaGross

Differential Revision: D20656211

fbshipit-source-id: deb91b917d349bc500acbb03d734ff621f6e1fc7

* Replace fbsource// with // in xplat/js/ files [4]

Summary:
`fbsource//xplat` and `//xplat` are equivalent for FB BUCK targets. Removing extra prefix for consistency.

Changelog: [Internal]

Reviewed By: JoshuaGross, shergin

Differential Revision: D20656696

fbshipit-source-id: 10f02decb1dc969fd3491ac90d97f09e2bda59e7

* Add Needs: Repro bot action (#28397)

Summary:
Add automated response for Needs: Repro

## Changelog

[Internal] [Added] - Add automated response for Needs: Repro
Pull Request resolved: https://github.com/facebook/react-native/pull/28397

Test Plan: Bot should add a comment with the Needs: Repro label.

Reviewed By: cpojer

Differential Revision: D20665378

Pulled By: hramos

fbshipit-source-id: 1c7d878faacf935a640849f74c81f119e5c7e92d

* Daily `arc lint --take CLANGFORMAT`

Reviewed By: zertosh

Differential Revision: D20666684

fbshipit-source-id: 32255ac7509e0257693969a7b4b044569af30df7

* ✅ Green CI: Fix JavaScript e2e tests, disable failing Android e2e test (#28392)

Summary:
Jobs now have a `run_disabled_tests` argument that allows for the selective execution of disabled tests. When working on re-enabling a failing test, the contributor just needs to set `run_disabled_tests` to `true` in the appropriate workflow in `.circleci/config.yml`.

Tests can be kept green by moving failing tests into the disabled section until a contributor can provide a fix, thus ensuring signal is maintained on master. For example, a failing end-to-end test might be disabled in order to allow the signal from unit tests to be provided, as opposed to flat out failing the entire job.

What was done in this PR:
* The failing `test_js_e2e` job has been fixed, and merged into the `test_js` job. An empty disabled tests section is added for future use.
* The failing `test_ios_e2e` job has been merged into `test_ios`, with all of its steps gated behind the `run_disabled_steps` argument.
* The failing Android end-to-end tests have been added to `test_android`, gated behind the `run_disabled_steps` argument
* The failing Podspecs test has been added back into `test_ios`, gated behind the `run_disabled_steps` argument

## Changelog

[Internal] [CI] - ✅ Green CI, disabled test infrastructure work
Pull Request resolved: https://github.com/facebook/react-native/pull/28392

Test Plan: Verified on Circle CI

Reviewed By: cpojer

Differential Revision: D20665512

Pulled By: hramos

fbshipit-source-id: 831738027f90f4b23313893d8342d7e654f34726

* Upgrade internal packages to support ESLint >= 6 (#28393)

Summary:
Fixes https://github.com/facebook/react-native/issues/28293

I've tested it with https://github.com/react-native-community/react-native-template-typescript and it seems to be working as expected - no warnings, supports typescript 3.8.

(note: I didn't upgrade the package version as I don't know how the releases work for this package)

## Changelog

[CATEGORY] [TYPE] - Message
Pull Request resolved: https://github.com/facebook/react-native/pull/28393

Reviewed By: hramos

Differential Revision: D20647112

Pulled By: cpojer

fbshipit-source-id: ca6b67971f625dc8125a58f9220dfcd86250ba94

* Fabric: Fixing a deadlock in RCTSurfacePresenter

Summary:
This is another attempt to fix an issue very similar to T59424871. The previous attempt was in D19249490. I don't know why we don't see production crashes (stalls) but it happened to me (and not to me) in the debugger. The previous attempt didn't work because we still could have a deadlock because we tried to acquired shared mutex already owned exclusively by the `suspend` method.

Here is another approach: Instead of using one shared mutex, now we use two. One is similar to what we had and another that protects `suspend` and `resume`. Besides that, now we pass a Scheduler to functions that use that explicitly. This way we can be more explicit about acquiring mutexes and the overall flow of execution. The idea is: Now an arbitrary code that can be reentrant does not cover with mutexes, so the deadlock is not possible.

Changelog: [Internal] Fabric-specific internal change.

Reviewed By: sammy-SC

Differential Revision: D20639228

fbshipit-source-id: 98515742f00f2ae94b50b585c9f1f0611e169ebe

* Update React Hooks Plugin

Summary:
Updates `eslint-plugin-react-hooks` to 3.0.0. This introduces a new lint error when you use a Hook inside a class.

Changelog:
[General][Changed] - Updated the React Hooks ESLint Plugin

Reviewed By: cpojer

Differential Revision: D20675528

fbshipit-source-id: d0cbe9748fd15df7a4c6de00bd1462610e0a43d6

* Upgrade React DevTools 4.0.6 -> 4.6.0

Summary:
Upgrading the embedded version of React DevTools, primarily to pull in [this PR](https://github.com/facebook/react/pull/18397) which will reduce the impact of this package on `node_modules` size.

# Update process

Following a similar process as D15973709, I began by searching for [all of the references](https://our.intern.facebook.com/intern/biggrep/?corpus=xplat&filename=.json&case=false&view=default&extre=&s=%22react-devtools&engine=apr_strmatch&context=false&filter[uninteresting]=false&filter[intern]=false&filter[test]=false&grep_regex=) to the `react-devtools-core` package and updated all v4 usage to to point to the new 4.6.0 release:

1: Manually update "react-devtools-core" versions:
```
js/package.json
js/react-native-github/package.json
nuclide/package.json
sonar/desktop/app/package.json
sonar/desktop/plugins/reactdevtools/package.json
vscode/modules/vscode-webview/package.json
```
2: Setup Yarn proxy:
```
yarn config set proxy http://fwdproxy:8080/
yarn config set https-proxy http://fwdproxy:8080
```
3: Run "yarn" in each of the above directories.
4: Run the lockfile shell script:
```
~/xplat/js/scripts/update-oss-yarn-lockfile.sh
```
5: Update the generated `MOBILE_JS_NODE_MODULE_DEPS.bzl` by running
```
js1 build buckfiles
```

## Changelog:

[General] [Changed] - Upgrade embedded React DevTools backend from v4.0.6 to v4.6.0.

Reviewed By: cpojer, gaearon

Differential Revision: D20676091

fbshipit-source-id: 99865bdba9bce45e2a7d582d5fb550cfdbeeca3a

* Make ScrollView use ForwardRef

Summary:
Have ScrollView use forwardRef so that the host component methods like `measure` and `measureLayout` are available without having to call `getNativeScrollRef`. Instead, you can use `<ScrollView ref={myRef} />` and directly call all methods of ScrollView and host components on `myRef`.

Previous usage:
```
const myRef = React.createRef<React.ElementRef<typeof ScrollView>>();
<ScrollView ref={myRef} />

const innerViewRef = myRef.current.getNativeScrollRef();

innerViewRef.measure();
```
New usage:
```
const myRef = React.createRef<React.ElementRef<typeof View>>();
<ScrollView ref={myRef} />

// now, myRef.current can be used directly as the ref
myRef.current.measure();
myRef.current.measureLayout();

// Additionally, myRef still has access to ScrollView methods
myRef.current.scrollTo(...);
```

Changes:

* Added deprecation warnings to ScrollView methods `getNativeScrollRef`, `getScrollableNode`, and `getScrollResponder`
* Added the forwardRef call to create `ForwardedScrollView` - this takes in `ref` and passes it into the class ScrollView as `scrollViewRef`.
* Forwarded the ref to the native scroll view using `setAndForwardRef`.
* Added statics onto `ForwardedScrollView` so that `ScrollView.Context` can still be accessed.
* Added type `ScrollViewImperativeMethods`, which lists the public methods of ScrollView.
* Converted all public methods of ScrollView to arrow functions. This is because they need to be bound to the forwarded ref.
* Bound all public methods of ScrollView to the forwarded ref in the `setAndForwardRef` call.
* Flow typed the final output (ForwardedScrollView) as an abstract component that takes in the props of the `ScrollView` class, and has all methods of both the inner host component (`measure`, `measureLayout`, etc) and the public methods (`scrollTo`, etc).

Changes to mockScrollView:
* Changed mockScrollView to be able to mock the function component instead of a class component
* Updated necessary tests

Changelog:
[General] [Changed] - Make ScrollView use forwardRef

Reviewed By: TheSavior

Differential Revision: D19304480

fbshipit-source-id: 6c359897526d9d5ac6bc6ab6d5f9d82bfc0d8af4

* Fix issue with onEndReached

Summary:
onEndReached can be triggered twice when more items are added to the end of the list. This change makes it so that a second call to onEndReached won't happen until the user scrolls down to the new end of the list.

Changelog:

[General] [Fixed] - Fix double call to onEndReached in VirtualizedList

Reviewed By: sahrens

Differential Revision: D20066740

fbshipit-source-id: 129d7ae6bfd241eeea18fe0bb12b82be67735874

* Remove console warnings from ScrollView methods

Summary:
The newly added console warnings in D19304480 are adding a lot of warning noise due to missed infra callsites. Those callsites need to be updated before these warnings can be added.

Changelog:
[Removed] Remove console warnings from ScrollView methods

Reviewed By: rickhanlonii

Differential Revision: D20700917

fbshipit-source-id: cb618ee3a291d26e1942e4f91bbc02dee41fb78b

* Upgrade react-docgen, jscodeshift and flow-parser

Summary:
In preparation for upgrading babel, I'm updating some of our source transform tools to the latest versions.

Changelog: [Internal]

Reviewed By: GijsWeterings

Differential Revision: D20675201

fbshipit-source-id: fa4fee2652529c6677087e42cdd1325a8080e46f

* Ship State Reconciliation 100% on all platforms

Summary:
State Reconciliation has been running 50/50 for a while and all metrics look stable. This is necessary for providing a good experience so we should ship to everyone unconditionally.

Changelog: [Internal] Fabric diffing reconciliation process improvement

Reviewed By: mdvacca

Differential Revision: D20715694

fbshipit-source-id: 25b2635ecc29b67e2911679c9db66bc84d37dec1

* Core telemetry tests: update so they pass on my machine

Summary:
`std::this_thread::sleep_for` is not really precise and will attempt to sleep for "at least" that much time, but may sleep much longer depending on what CPUs are doing and scheduling policies.

To get this to pass on my machine, I had to substantially increase the thresholds.

Changelog: [Internal]

Reviewed By: shergin

Differential Revision: D20689571

fbshipit-source-id: f159420d24a95da2b5d95d49ed7882e783291e98

* Optimize diff algorithm to produce fewer remove+insert ("move") paired instructions

Summary:
An evolution of D20633188 but more performant.

There are three optimized paths before the slow path.

The first optimized path tries to pair identical nodes from old/new tree, and generate Update mutations, until we hit nodes that are different (indicating either a remove or an insert). This already existed.

The next two optimizations, introduced by Tim in his JS pseudocode, were inspired by ReactJS's diffing algorithm. They work in cases where the rest of the nodes are (1) all removals/deletes or (2) all creates+inserts.

Finally, if those final two optimized paths can't run, it's because there is a mix of delete+remove, create+insert, and "move" operations, mixed at the beginning, middle, and/or end of the list.

This has slightly better average/best-case complexity as the previous implementation.
In particularly pathological cases where all nodes are arbitrarily reordered, or reversed, for instance (ABCDE->EDCBA) the algorithm has the same complexity as the previous algorithm (quadratic).

For now iOS is pinned to the older differ

Changelog: [Internal] Experiment to optimize diffing algorithm in Fabric

Reviewed By: shergin

Differential Revision: D20684094

fbshipit-source-id: d29fba95a0328156c023e1c87804f23770ee1d91

* Unit test for V2 "minimal instruction" diffing algorithm

Summary:
This unit test is to verify that the new diffing algorithm generates a "minimal" instruction set, with regards to removes and inserts ("moves").

These unit tests are here to verify the expected behavior in this new algorithm, but these tests may be modified or deleted in the future if we decide we want to change this behavior.

Changelog: [Internal] fabric unit test

Reviewed By: mdvacca

Differential Revision: D20706592

fbshipit-source-id: 5f9991498e0d788ecbf88d938bfe6d3f0f27af40

* Add ES Lint rules for `DynamicColorIOS()`and `ColorAndroid()` (#28398)

Summary:
The [PlatformColor PR](https://github.com/facebook/react-native/pull/27908) added support for iOS and Android to express platform specific color values.   The primary method for an app to specify such colors is via the `PlatformColor()` method that takes string arguments.   The `PlatformColor` method returns an opaque Flow type enforcing that apps use the PlatformColor method instead of creating Objects from scratch -- doing so would make it harder to write static analysis tools around Color values in the future.   But in addition to `PlatformColor()`, iOS has a `DynamicColorIOS()` method that takes an Object.   The Flow type for this Object cannot be opaque, but we still want to enforce that app code doesn't pass variables instead of Object literals or that values in the Objects are variables.   To ensure `DynamicColorIOS()` can be statically analyzed this change adds an ESLint rule to enforce that `DynamicColorIOS()` takes an Object literal of a specific shape.   A `ColorAndroid()` was also introduced not for practical use but just to test having platform specific methods for more than one platform in the same app.   A second ESLint rule is created for `ColorAndroid` as well.

## Changelog

[General] [Changed] - Add ES Lint rules for `DynamicColorIOS()`and `ColorAndroid()`
Pull Request resolved: https://github.com/facebook/react-native/pull/28398

Test Plan: `yarn lint` passes.

Reviewed By: cpojer

Differential Revision: D20685383

Pulled By: TheSavior

fbshipit-source-id: 9bb37ccc059e74282b119577df0ced63cb9b1f53

* fix: Android gradle config when bundling for release (#28415)

Summary:
This fix aims to address the issue when bundling an Android app for release and getting the error exhibited in https://github.com/facebook/react-native/issues/28002 which I also encountered myself.

The config was changed sometime in November 2019 (as part of https://github.com/facebook/react-native/issues/26940, commit https://github.com/facebook/react-native/commit/a3b08048674e324dbe1f0ca816f35607e9e06a2f) to be very opinionated when it comes to the use of `npx` which Gradle itself cannot find anyway (I have `npx` installed globally and it didn't pick it up).

Another issue that the use of `npx` creates is that Gradle should only ever use the currently installed react-native cli rather than a (possibly) higher version which may not always have backward compatibility.

The proposed change simply throws a more descriptive error rather than defaulting to a tool which may or may not exist on the machine, be it CI or a development environment. I've also modified the RNTester app to reflect the correct config implementation relative to the RNTester app itself.

In real projects, the config inside `android/app/build.gradle` should look similar to the following snippet:

```
project.ext.react = [
  cliPath: "$rootDir/../node_modules/react-native/cli.js",
  entryFile: "index.js"
];
```

## Changelog
[Android] [Fixed] - Gradle release config
Pull Request resolved: https://github.com/facebook/react-native/pull/28415

Test Plan:
- [x] Successfully bundled an Android release build with correct config
- [x] Works with RNTester app

Reviewed By: mdvacca

Differential Revision: D20714372

Pulled By: hramos

fbshipit-source-id: 4d66139249c6f840582a71a48c64e6a6595f7af0

* Reimplement D19965405: Small improvements in Differentiator/TinyMap

Summary:
Two things:
1) I reimplement Valentin's idea in D19965405, so that TinyMaps can be iterated over, with a couple of bugfixes (calling front() or back() on an empty vector will crash).
2) I now use TinyMap instead of better::map in the "optimized" diffing algorithm.
3) `erase` now actually removes elements from the vector, but only when more than half of elements have been erased.
4) If you repeatedly erase elements at the beginning of the vector, they will no longer be iterated over. This is a specific optimization for our heaviest TinyMap use-cases.

These amount to some small but hopefully somewhat meaningful perf improvements.

Changelog: [Internal] Fabric perf

Reviewed By: shergin

Differential Revision: D20718719

fbshipit-source-id: 91f4b2e2e0f6387ae484e43d5b0095103087baa6

* Remove LayoutInspectingPolicy.includeScrollViewContentOffset

Summary:
`LayoutInspectingPolicy` has two flags, `includeTransform` and `includeScrollViewContentOffset`.

`includeScrollViewContentOffset` seems to be redundant for two reasons.

# 1st
From looking at callers, they have always the same value.
I looked at all call sites, and they are either always both set to true or both set to false.

# 2nd
The way we include scroll view content offset, is through transformation, so setting `includeTransform` to true and `includeScrollViewContentOffset` to false will include content offset anyway. In order to make both flags work, we would need to introduce further changes to `getRelativeLayoutMetrics`. But since the flag isn't used anyway, I think it is better to get rid of it for now. If we need it in the future, we could re-introduce it.

Reviewed By: shergin

Differential Revision: D20622256

fbshipit-source-id: fb6156c66b752319ea928239fa723ff90688b0a0

* Add support for translation and rotation to operator * between Rect and Transform

Summary:
Changelog: [Internal]

Until now `Rect operator*(Rect const &rect, Transform const &transform)` supported only scaling. Now it supports translation and rotation as well.

Reviewed By: shergin

Differential Revision: D20622876

fbshipit-source-id: 1b65393bd3fd6fd9a8941903e0f2681a10097e4a

* Include transform property when calling getRelativeLayoutMetrics

Summary:
Changelog: [Internal]

Current implementation of `measure` doesn't take transform into account..

So if you had a view which has width and height 100 and had `Scale(0.5, 0.5, 1)` (this will shrink view by half). Calling `getRelativeLayoutMetrics` would report its size being `{100, 100}`.
This applies if view's parent has transformation as well, because transformation is applied to all subviews of the view as well.

Reviewed By: mdvacca

Differential Revision: D20621590

fbshipit-source-id: 2cf902a0494291c821ecada56f810c5e6620db5a

* feat: migrate appveyor to circleci (#28245)

Summary:
This issue closes https://github.com/facebook/react-native/issues/28241
Migrated Windows test from AppVeyor to CircleCI

## Changelog

[Internal] [Changed] - Migrated Windows test from AppVeyor to CircleCI
Pull Request resolved: https://github.com/facebook/react-native/pull/28245

Test Plan: For CircleCI to Pass

Reviewed By: cpojer

Differential Revision: D20689163

Pulled By: hramos

fbshipit-source-id: 285c762457ef00f7884ee9157b3f336044c0452f

* Remove "Debug with Nuclide" option

Summary: This is no longer needed.

Reviewed By: cpojer

Differential Revision: D20722274

fbshipit-source-id: 5bc3104e90811d724f42aadbf137ab8eff718ca0

* experiment to preload RN bridge after fb4a bookmarks render

Summary:
Changelog:
[Android][Internal] add internal supermodule label

Reviewed By: mdvacca

Differential Revision: D20434200

fbshipit-source-id: fae50309cdd0df4a4523c2f88d1c8e01a7163575

* Fix CursorDrawable Color Tint for Android 10+

Summary:
Accessing this field via reflection is explicitly blacklisted by Google in Android 10 and higher. They have provided a new API to change the color, which I have implemented here. [The old setColorFilter is deprecated](https://developer.android.com/reference/android/graphics/drawable/Drawable#setColorFilter(int,%20android.graphics.PorterDuff.Mode)) so I also updated that method call as well.

Changelog: [General] [Fixed] Use new setTextCursorDrawable API for Android 10

Reviewed By: JoshuaGross

Differential Revision: D20656068

fbshipit-source-id: 58a92b57c0a892c7c87fc5d735e4ceaa4e987ec7

* Early return on tinting CursorDrawable if no color supplied

Summary:
There's (potentially) a lot of expensive reflection calls here that, as best I can tell, end up being ignored if the supplied color is null. Better to early return.

Changelog: [General] [Internal] Preclude reflection when setting cursor color if color is null

Reviewed By: JoshuaGross

Differential Revision: D20670594

fbshipit-source-id: 480a988355bbd79008002c4326d4b35035ec2a95

* Partial React Sync for Inspector

Summary:
Partial sync for React that includes:

- https://github.com/facebook/react/pull/18388
- https://github.com/facebook/react/commit/dd7e5e4f5ac2ffac3171ef61daee2cb1edc69635

Created from this branch: https://github.com/facebook/react/compare/master...rickhanlonii:rh-partial-3-24?expand=1

Changelog: [Internal]

Reviewed By: mdvacca

Differential Revision: D20651395

fbshipit-source-id: 67baf7c407f75d9fd01c17f2203a77a38567100e

* Enable inspector for Fabric

Summary:
## Overview
This diff refactors the Inspector, moving logic to look up view data for a touched view inside the renderer as `getInspectorDataForViewAtPoint`. We then implement that same function for Fabric in order to support the inspector in that renderer.

Requires https://github.com/facebook/react/pull/18388

## Motivation

Reason one for this refactor is that, previously, the inspector held all of the logic to look up view data for a given x,y touch coordinate. To do this, it would take the React tag and coordinates, look up the native view tag, measure it, and then ask React internals for the Fiber information of that tag. All of this is deeply coupled to React internals, yet the logic is outside of React core in the Inspector.

Reason two is that, for Fabric, the logic for getting the view data is different than Paper. In Fabric, we pass the x,y coordinates to native directly, which returns an instance handle. That handle can be used to measure the ShadowNode, or retrieve the Fiber information.

By moving the logic into the renderer in React core, we decouple the implementation details of looking up view data for a tapped point and allow ourselves the ability to add and change renderer-specific code for the actual lookup without impacting outsiders like the Inspector.

Changelog: [Internal]

Reviewed By: TheSavior

Differential Revision: D20291710

fbshipit-source-id: a125223f2e44a6483120c41dc6146ad75a0e3e68

* chore: update url of warning message from deprecated imports (#28452)

Summary:
Some of the repository name of Lean Core(https://github.com/facebook/react-native/issues/23313) libraries has been renamed.
This PR updates the warning message to display the updated url.

## Changelog

[General] [Changed] - Update warning message of deprecated imports
Pull Request resolved: https://github.com/facebook/react-native/pull/28452

Test Plan: updated URL can be accessed.

Reviewed By: cpojer

Differential Revision: D20745184

Pulled By: TheSavior

fbshipit-source-id: 2c3ed6a000b45022ca6c4862305aa567c4d18b2e

* Add `upgrade-support` link on issue creation (#28411)

Summary:
This PR adds a https://github.com/react-native-community/upgrade-support link for the user when creating an issue.
Changelog:
[Internal]

Pull Request resolved: https://github.com/facebook/react-native/pull/28411

Reviewed By: cpojer

Differential Revision: D20714274

Pulled By: hramos

fbshipit-source-id: 4ca42224a50e386b95f21f0fb236a917e1b6b982

* Update PixelRatio 'getFontScale' method description (#28407)

Summary:
Refs facebook/react-native-website#1776.

Despite in-code description `PixelRatio.getFontScale()` is working properly on the iOS (it also reflects the user settings).

This PR updates the in-code description to match current behaviour.

I have decided to skip in the code information about additional setting in `Accessibility` menu and in `Control Centre`, but if you think it is important just let me know, I can update this PR.

## Changelog

[Internal] [Fixed] - Fix PixelRatio getFontScale method description
Pull Request resolved: https://github.com/facebook/react-native/pull/28407

Test Plan: N/A

Differential Revision: D20750260

Pulled By: shergin

fbshipit-source-id: c40ec2fd49cd60e2975351c3a1c453aab0045da4

* Remove allowDisablingImmediateExecutionOfScheduleMountItems feature flag

Summary:
No longer needed.

Changelog: [Internal]

Reviewed By: mdvacca

Differential Revision: D20747684

fbshipit-source-id: a8077519b7670d72e23267b1c1423556ec97be3f

* RuntimeExecutor helpers that modify the way of the callback is being executed.

Summary:
Here we implement a bunch of helper methods that allow customizing the behavior of a RuntimeExecutor "on-demand" on the caller side. We will use it in the next diff(s).

Changelog: [Internal] Fabric-specific internal change.

Reviewed By: PeteTheHeat

Differential Revision: D20551411

fbshipit-source-id: 51d3cd02b69753110c0e1155347c6e52eb882c7d

* Fabric: Using `executeSynchronouslyOnSameThread_CAN_DEADLOCK` in MainRunLoopEventBeat

Summary:
We are replacing inline-ed implementation with practically the same one implemented as the helper method.

Changelog: [Internal] Fabric-specific internal change.

Reviewed By: sammy-SC

Differential Revision: D20551409

fbshipit-source-id: fcc6f497cd240af65fba534051c217fe5746ce82

* Set RootNodeKind trait in ModalHostViewShadowNode

Summary:
Changelog: [internal]

`ModalHostViewShadowNode` didn't have `RootNodeKit` trait, therefore `getRelativeLayoutMetrics`  was including nodes in ancestors that it shouldn't have.

Reviewed By: shergin

Differential Revision: D20735801

fbshipit-source-id: 6b81e3b174c2f82e530abc2bca2da8bebc2270b0

* mention RNTester app in contributor guide (#28042)

Summary:
motivation is following - the RNTester app is imho the best place to try out any changes that a contributor would make, yet it is not directly mentioned in the contributor guide. This fixes it.

## Changelog

[Internal] - Docs
Pull Request resolved: https://github.com/facebook/react-native/pull/28042

Test Plan: not necessary

Reviewed By: TheSavior

Differential Revision: D20401260

Pulled By: hramos

fbshipit-source-id: 01c1b7dff56b59909c94b2feb609650f0baba1a9

* Buck: Use Android SDK 29 during build (#28455)

Summary:
Fixes `test_android` and `test_docker` build failures. Thanks to dulmandakh for identifying the fix.

Changelog:
[Internal] [Android] [Changed] - Use Android SDK 29 to build during CI tests
Pull Request resolved: https://github.com/facebook/react-native/pull/28455

Test Plan: Circle CI shows `test_android` and `test_docker` passing: https://app.circleci.com/jobs/github/facebook/react-native/142273

Reviewed By: sturmen

Differential Revision: D20766589

Pulled By: hramos

fbshipit-source-id: 8ef8a8ce3a6e7353ae47425accb3bd26cf1608c4

* Assign orderIndex_ in ConcreteViewShadowNode constructor instead of ViewShadowNode's constructor

Summary:
Changelog: [Internal]

`orderIndex_` was only being assigned for `ViewShadowNode`, not for other `ShadowNodes` that are later represented on the screen.

Reviewed By: shergin

Differential Revision: D20746477

fbshipit-source-id: c04c2cfea14b9141d22bc3d9e9bb4c0c59925754

* Implement nativePerformanceNow to improve Profiler API results (#27885)

Summary:
When experimenting with React Profiler API (https://reactjs.org/docs/profiler.html), I noticed that durations are integers without a debugger, but they are doubles with higher precision when debugger is attached. After digging into React Profiler code, I found out that it's using `performance.now()` to accumulate execution times of individual units of work. Since this method does not exist in React Native, it falls back to Javascript `Date`, leading to imprecise results.

This PR introduces `global.nativePerformanceNow` function which returns precise native time, and a very basic `performance` polyfill with `now` function.

This will greatly improve React Profiler API results, which is essential for profiling and benchmark tools.

Solves https://github.com/facebook/react-native/issues/27274

## Changelog

[General] [Added] - Implement `nativePerformanceNow` and `performance.now()`
Pull Request resolved: https://github.com/facebook/react-native/pull/27885

Test Plan:
```
const initialTime = global.performance.now();
setTimeout(() => {
  const newTime = global.performance.now();
  console.warn('duration', newTime - initialTime);
}, 1000);
```

### Android + Hermes

![Screenshot_1580198068](https://user-images.githubusercontent.com/13116854/73245757-af0d6c80-41b5-11ea-8130-dde14ebd41a3.png)

### Android + JSC

![Screenshot_1580199089](https://user-images.githubusercontent.com/13116854/73246157-92256900-41b6-11ea-87a6-ac222383200c.png)

### iOS

![Simulator Screen Shot - iPhone 8 - 2020-01-28 at 10 06 49](https://user-images.githubusercontent.com/13116854/73245871-f136ae00-41b5-11ea-9e31-b1eff5717e62.png)

Reviewed By: ejanzer

Differential Revision: D19888289

Pulled By: rickhanlonii

fbshipit-source-id: ab8152382da9aee9b4b3c76f096e45d40f55da6c

* Save/restore IP when leaving the interpreter

Summary:
This diff implements the instruction pointer save/restore trick Tzvetan came up with; allowing us to observe and modify the IP from outside the interpreter loop with negligible overhead.

From Tzvetan's internal post on the subject:

> [Today] the interpreter IP is just a local variable in the interpreter function, so there is no way to get to its value from outside the function. It lives in a register and we don't want to make it a Runtime field since the overhead [of accessing it via memory in the interpeter loop] would kill us.

> However, if you really think about it, it only lives in a register while the interpreter function is running. For the rest of the time, it is spilled by the C++ compiler onto the stack.
So, precisely when we need it, it is actually stored in memory. The only problem is, we don't know where! Admittedly, that is an annoying problem, but it feels like it should be solvable.

> What if, instead of relying on the compiler to spill the IP register, we manually spill it ourselves, to a known location? It works. Example: https://godbolt.org/z/ftSDnp

This diff implements this approach across the whole interpreter loop: whenever we call out of the loop we capture/publish the IP and restore it again immediately after the external call returns. This means we can now see the IP outside the interpret loop and even change it. This is effectively "for free" as the compiler now skips spilling/restoring the IP behind the scenes.

The immediate benefit of this is knowing the current IP allows us to have more accurate stack-traces during execution. In future this may enabled tricks like changing the IP before returning to th…
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Bug CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. Merged This PR has been merged. Platform: Android Android applications. Tech: Bundler 📦 This issue is related to the bundler (Metro, Haul, etc) used.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

7 participants