Summary:
To ensure these Pojos don't get dead code eliminated, this diff adds a DoNotStrip annotation to each Props pojo object.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D26041103
fbshipit-source-id: 02031f39f5fc5ef1bb9f11b729e68437bfe95452
Summary:
The Pojo data structure now keeps a flag `isRoot`, so that we can customize code-generation for the Props pojo vs its helper pojos.
In the following diff (i.e: D26041103), we'll add a DoNotStrip annotation to the props pojo. That doesn't need to exist on the helper pojos, because they'll all be referenced in the root props pojo.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D26041102
fbshipit-source-id: ae2e1ba346d038853b5ff30cd5524a809ef97053
Summary:
## What This Does
- **Phase 1:** Given a component, convert its props (and all its ObjectTypeAnnotations) into this Pojo data structure under some namespace. This is a recursive operation.
- **Phase 2:** Loop over each Pojo data structure, and serialize it to Java class.
So...
- Each Component has its own namespace (i.e: its hasteModuleName) for Java Pojo objects.
- Each Component generates 1 Pojo object, for its props.
- Each Component generates 1 Pojo object for every ObjectTypeAnnotation in its props.
## Decisions
By design, [JNI can read/write to private properties on Java objects](https://stackoverflow.com/a/12208643). So, each Pojo, which represents an ObjectTypeAnnotation, contains only a private member variable for each of its properties, and a getter to retrieve the data from each of its private members.
## Todos
- Improved type-safety:
- ReservedTypeAnnotation (e.g: PointPrimitive). These currently map to ReadableMap.
- String enums, and Int enums don't actually generate Java enums.
- Verify if there are any compilation issues by wiring this up to the Codegen buck infra.
- To actually use the Pojos, we'll need C++/Jni codegen to transform C++ props into these Pojos. Building this out will give a more accurate assessment of the app-size cost of Pojos. However, we can do that if we deem that the app-size increase from adding just the Pojo classes is negligible enough to **not** rule out this entire approach.
- ~~Add some annotations to prevent these Pojo classes from being stripped at compile-time.~~ D26041103
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D26038189
fbshipit-source-id: c137c4ca6b043ee76adb354105aff6e0f270df86
Summary:
Right now, running the codegen parser on regular JavaScript files causes it to throw an error:
https://www.internalfb.com/intern/diffusion/FBS/browsefile/master/xplat/js/react-native-github/packages/react-native-codegen/src/parsers/flow/index.js?commit=1e93f27aac8890f1731650fdcc38b702ae8874e2&lines=59-63%2C106-107
This makes the parser less usable, because you can't simply loop over a list of JavaScript files, and run codegen on them. You need to do some filtering beforehand, or surround each invocation of the parser with a try/catch.
Our codegen schema supports the idea of an empty schema:
```
{
modules: {}
}
```
To improve the parser's ergonomics, this diff migrates the codegen parser over to returning the empty schema when it cannot detect a Component or NativeModule spec inside a JavaScript file.
Changelog: [Internal]
Reviewed By: JoshuaGross
Differential Revision: D26034052
fbshipit-source-id: c2f15aba9c0e052012396eaed2034537f5918e33
Summary:
## Problem
Suppose we have this NativeModule spec, that uses Object literals that contain a property called "id":
```
export interface Spec extends TurboModule {
// Exported methods.
+getConstants: () => {|
id: string,
|};
+getObject: (arg: {id: Object}) => Object;
}
```
For both object literals, we'll generate C++ structs, backed by NSDictionaries. The method in each struct will be named "id_" to avoid a name clash with ObjC's `id` identifier. However, calling that id_ method should still access the "id" property in the corresponding NSDictionary.
## Expected Output
```
inline id<NSObject> JS::NativeSampleTurboModule::SpecGetObjectArg::id_() const
{
id const p = _v[@"id"];
return p;
}
inline JS::NativeSampleTurboModule::Constants::Builder::Builder(const Input i) : _factory(^{
NSMutableDictionary *d = [NSMutableDictionary new];
auto id_ = i.id_.get();
d[@"id"] = id_;
return d;
}) {}
```
## Actual Output
```
inline id<NSObject> JS::NativeSampleTurboModule::SpecGetObjectArg::id_() const
{
id const p = _v[@"id_"]; // <-- HERE!
return p;
}
inline JS::NativeSampleTurboModule::Constants::Builder::Builder(const Input i) : _factory(^{
NSMutableDictionary *d = [NSMutableDictionary new];
auto id_ = i.id_.get();
d[@"id_"] = id_; // <-- HERE!
return d;
}) {}
```
NOTE: This code was generated by running `jf get --version 119805822 && buck build //xplat/js:FBReactNativeSpec_Sample-flow-types-ios --show-output`
This diff fixes this mistake.
Changelog: [Internal]
Reviewed By: fkgozali
Differential Revision: D25907493
fbshipit-source-id: cb37cbf49db4f871b3f4046f7397a7b1b7df0357
Summary:
Changelog: [internal]
Codegen was generating code with return value number instead of boolean.
Reviewed By: RSNara
Differential Revision: D25863062
fbshipit-source-id: 780f88dd2d83e303b03d1ed9cc837ac6733f1702
Summary:
This BUCK target only existed for debugging purposes. Nothing depends on it.
Changelog: [Internal]
Reviewed By: JoshuaGross
Differential Revision: D25844759
fbshipit-source-id: e27fb04809a5c97adbeefd7df60b82992c6e1eda
Summary:
This new type will be valid in Flow strict mode and can be used by native modules and components to replace `Object`, with the same semantics.
This unblocks the migration of the most modules in the React Native package to Flow strict.
Changelog: [Internal] Add UnsafeObject type compatible with Flow strict mode to use in native modules and components
Reviewed By: RSNara
Differential Revision: D25540631
fbshipit-source-id: 60b80bbc84a53aecc747e3a1799cdf551e1859cd
Summary:
## Changes
All `rn_library(codegen_modules = True)` must now also specify native_module_android_package_name, like so:
```
rn_library(
name = "FBAuth"
codegen_modules = True,
native_module_spec_name = "Foo",
native_module_android_package_name = "com.facebook.fbreact.specs",
)
```
This will generate the FBAuth Java spec files under the appropriate directory: "com/facebook/fbreact/specs". It will also make the code-generated specs have the appropriate package name: "com.facebook.fbreact.specs".
Changelog: [Internal]
Reviewed By: shergin
Differential Revision: D25723176
fbshipit-source-id: 6efec1cbee43d70110c0ef23e2422e08609b61d4
Summary:
NativeModule methods are meant to be called from JavaScript. As such, they may not necessarily have call-sites in Java. This means that they're succeptible to being stripped by proguard.
This diff annotates all exported NativeModule methods with DoNotStrip, so that proguard doesn't strip them. We already do this in the legacy codegen.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25723801
fbshipit-source-id: a7c8701e0a5d03a970f5f19cc6ae6b320a2e99a1
Summary:
## Description
Suppose this was the codegen declaration before:
```
rn_library(
name = "FooModule",
native_module_spec_name = "FBReactNativeSpec",
codegen_modules = True,
# ...
)
```
Previously, this would generate the following BUCK targets:
- generated_objcpp_modules-FooModuleApple
- generated_java_modules-FooModuleAndroid
- generated_java_modules-FooModule-jniAndroid
## Changes
We will now generate:
- FBReactNativeSpecApple
- FBReactNativeSpecAndroid
- FBReactNativeSpec-jniAndroid
This matches the naming scheme of the old codegen.
Changelog: [Internal]
Reviewed By: fkgozali
Differential Revision: D25680224
fbshipit-source-id: 617ac18fd915f3277f6bd98072d147f20fb193e5
Summary:
This test was used to compare the old codegen output with the new codegen output. It was written predominately to test the ObjC++ generator. A few reasons why we should delete this:
1. We do not want the old and the new codegen targets co-existing because it will slow down buck query.
2. The new ObjC++ codegen output is different from the old. In the new ObjC++ generator, we aren't generating dead structs. Therefore, this test simply won't work on the most sophisticated generator (i.e: the generator it was designed to run on). The other generators are really simple. We've also been running the ObjC++ and the Java generators in production for over a month for react-native-github specs. These are arguably the most complicated specs we maintain, so generator correctness isn't too much of a concern.
Changelog: [Internal]
Reviewed By: fkgozali
Differential Revision: D25683069
fbshipit-source-id: b2175f34f50a078d80ef738e59024f546628219a
Summary:
Original PR contents:
This pull request updates the Podspecs and associated build scripts, and some source files so they build on OSS. RNTester now compiles with `fabric_enabled` again.
The following changes have been made:
* Various spots that were pointing to the old `ReactCommon/fabric` location have now been updated to `ReactCommon/react/renderer`
* Files that were attempting to use internal FB header `FBRCTFabricComponentsPlugins.h` were changed to use `RCTFabricComponentsPlugins.h`
* `RCTFabricComponentsPlugins` in OSS was updated to include the `Image` fabric component (thanks tsapeta)
* Replaced old `generate-rncore.sh` build script with new `generate-rncore.js` script which does not require `flow-node` and uses the `react-native-codegen` API directly, so there is no longer any need for an interim `schema-rncore.json` file.
* Updated Yoga podspec which wasn't fully synced with changes from the main Yoga repo
* Updated Fabric podspec with additional needed subspecs
Additions to PR by hramos:
* Replaced use of generate-rncore scripts with the original generate-native-modules-specs.sh script, which is now generate-specs.sh and supports both codegen for Native Modules and Components now (TurboModules/Fabric).
* Codegen now runs at build time as part of the Xcode build pipeline instead of as part of `pod install`. The build script is injected by the FBReactNativeSpec pod, as the pod is part of both Fabric and non-Fabric builds.
## Changelog
[General] [Fixed] - RNTester compiles with `fabric_enabled` again
Pull Request resolved: https://github.com/facebook/react-native/pull/29810
Test Plan:
RNTester now compiles and runs in the simulator again when `fabric_enabled` is set to `true`.
```
cd xplat/js/react-native-github/packages/rn-tester
USE_FABRIC=1 pod install
open RNTesterPods.xcworkspace
# Build and run
```
Reviewed By: fkgozali
Differential Revision: D24058507
Pulled By: hramos
fbshipit-source-id: 8b2ea3694e6cb9aa23f83f087e2995fd4320e2bb
Summary:
Generate Fabric C++ files along side TM spec files for RNTester. The combined .so then has both TM and Fabric files.
This commit also removed the checked-in JNI files.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25674313
fbshipit-source-id: 8091d5a00f42849a74cab50e8d24f4010d500e5b
Summary:
For core components, we can start using the codegen output during build time instead of the checked in files in: https://github.com/facebook/react-native/tree/master/ReactAndroid/src/main/java/com/facebook/react/viewmanagers
Note: Some files seemed to be handwritten, so this only removed those that use codegen.
Changelog: [Internal]
Reviewed By: JoshuaGross
Differential Revision: D25453157
fbshipit-source-id: f7eabddfd3fd668bef0c4aef3fddcb38c8b046a0
Summary:
In preparation for flipping the default, marking autoglob as False in places where it isn't explicitly specified.
Changelog: [Internal]
Reviewed By: strulovich
Differential Revision: D25497305
fbshipit-source-id: 142e5caca2d67efcb3c25067a36934f7f6dd4b21
Summary:
Some of the existing files under uimanager/ are self contained and used by the component codegen output. This commit split out those files into a dedicated uimanager/interfaces/ dir/target so that more targets can depend on it without causing dep cycle.
Also, the component codegen output doesn't need LayoutShadowNode at all, so this removed the import.
This will allow us to combine the Java codegen output for TM spec and Fabric components, simplifying build and dependency management (not in this commit yet).
Changelog: [Internal]
Reviewed By: JoshuaGross
Differential Revision: D25451409
fbshipit-source-id: 827545a3d78ebed1815cf9e52da2fa896b012aa1
Summary:
This commit:
* Generate Fabric component Java files along side Java NativeModule specs, when `USE_FABRIC=1` is set
* Adjust the component codegen to place output files in a subdir based on package name
* Adjust existing Buck targets to filter the right nativemodule vs component java files (this avoids duplicated symbols)
* Compiles the Java output during build time on RNTester/ReactAndroid (Gradle)
Not in this commit:
* Fabric C++ files
* Removing checked-in generated component files.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D25416614
fbshipit-source-id: fd670ead2198c9b5a65812c692b7aed9f3d7cd58
Summary:
In the Codegen, we need to answer the following questions:
*Question:* What are all the calls into TurboModuleRegistry?
*Answer:* Find all CallExpressions that represent TurboModuleRegistry.get<Spec>() or TurboModuleRegisty.getEnforcing<Spec>().
*Question:* Is this a component spec?
*Answer:* Does this spec have a CallExpression where the callee is 'codegenNativeComponent'?
*Question:* Is this a module spec?
*Answer:* Does this spec have an interface that extends TurboModule?
All these answers can be implemented using the visitor pattern. Hence, this diff introduces the `visit` utility, and uses it to answer these questions.
**Motivation:** Cleaner code.
Changelog: [Internal]
Reviewed By: hramos
Differential Revision: D25162617
fbshipit-source-id: 66ec95fc07ecb29aa9bf9993cb826204af283d03
Summary:
**Motivation:** After this change, we can show this ParserError using the React Native Module ESLint rule. Previously when we declared more than one NativeModule spec in a file, we incorrectly reported that there were *no* NativeModule specs in the file.
Changelog: [Internal]
Reviewed By: hramos
Differential Revision: D25163299
fbshipit-source-id: 92bc09d09cdbc323e0ac1f317c40a767880f5bc2
Summary:
## What is `guard`?
`guard` accepts some JavaScript function that can throw a ParserError. If a ParserError is thrown by that JavaScript function, it captures and pushes the error to some global array, and returns null. If no ParserError is thrown, it simply returns the return value of the JavaScript function. This utility is used in the NativeModule spec parser to help it continue parsing even after it detects errors. Why do we want to do this? In the NativeModule spec linter, we want to display all these ParserErrors via ESLint.
## Changes
This diff renames `guard` to `tryParse` because `tryParse` more appropriately captures the intent/function of this utility: the work passed to it "tries" to parse some Flow types. A name like "guard" is a bit more ambiguous: What is it guarding against? What is the work doing? ¯\_(ツ)_/¯
Changelog: [Internal]
Reviewed By: hramos
Differential Revision: D25156185
fbshipit-source-id: 516647770579daa8613dbd67535074823f1aa848
Summary:
## Changes
1. In the NativeModule spec parser, the moduleName is now being extracted from the TurboModuleRegistry.get<Spec>(...) call by examining the Flow ast node. Previously, we used regex parsing, which was unsafe because it could be fooled by TurboModuleRegistry.get<Spec>(...) calls in comments.
2. The logic to parse and validate the TurboModuleRegistry.get<Spec>(...) call is now centralized in the NativeModule Spec Parser (it was removed from the react-native-modules ESLint rule). The linter is now only responsible for three things:
1. Detecting if a JavaScript file contains a TurboModuleRegistry.get<Spec> call or a TurboModule interface, and if so
2. Running the NativeModule spec parser on it.
3. It also validates that the Module spec's filename starts with the prefix "Native".
The React Native Modules linter now completely delegates to the NativeModules Spec parser, without doing any error checking of its own. If an error is reported by the React Native Modules linter, and that error doesn't have anything to do with the "Native" prefix, then it *must* be addressed. Otherwise, it will cause the NativeModule Spec Parser to fail on that particular spec.
Changelog: [Internal]
Reviewed By: hramos
Differential Revision: D25153243
fbshipit-source-id: da74dbb66b1d8dca3a2b1952402222c6696b73d6
Summary:
Changes `react-native/babel-plugin-codegen` to generate calls to `NativeComponentRegistry` instead of `registerGeneratedViewConfig`.
The only notable changes in behavior from this will be:
1. In bridgeless mode, all components using `codegenNativeComponent` will no longer access `UIManager`.
2. In bridge mode, all components using `codegenNativeComponent` will no longer verify equivalence in production. Only in `__DEV__`. (This may improve performance slightly.)
This also changes the `ViewConfig` to be lazily allocated and drops support for `__INTERNAL_VIEW_CONFIG`, which we no longer need.
Changelog:
[Internal]
Reviewed By: JoshuaGross
Differential Revision: D25135881
fbshipit-source-id: ca2191872c02622ab2279b808102eeb1f664d207
Summary:
Use pre-built react-native-codegen library from npm in the iOS app template.
Built react-native-codegen from source when used with RNTester.
Published react-native-codegen@0.0.6.
Changelog:
[iOS][Added] - Use react-native-codegen in iOS app template
[Internal] - Bump react-native-codegen: 0.0.6
Reviewed By: fkgozali
Differential Revision: D25128036
fbshipit-source-id: f294c23b9b911aae6f404edc01b62426fb578477
Summary:
When running yarn install from the codegen directory it will reinstall all dependencies for the react-native workspace inside the react-native package. In my case this caused issues with metro because it would now have 2 copies of it (node_modules/metro and node_modules/react-native/node_modules/metro).
To avoid this copy the react-native-codegen source in a temporary directory and yarn install from there, then copy the built files back.
## Changelog
[Internal] - Build rn-codegen in a temporary directory
Pull Request resolved: https://github.com/facebook/react-native/pull/30292
Test Plan: Tested the script in an app with codegen enabled. Fresh install with rn-codegen not built, made sure no extra modules are installed under node_modules/react-native/node_modules.
Reviewed By: yungsters
Differential Revision: D24893216
Pulled By: fkgozali
fbshipit-source-id: 2c372b755632ea6f50ad5d4562248612b349a9a6
Summary:
## Changes
{| ... |} -> { ... }
**Motivation:** In Flow, object literals are exact by default. So, there's no need for the pipes. Also: Now, the syntax for object literals is consistent across react-native-codegen.
Changelog: [Internal]
Reviewed By: hramos
Differential Revision: D24774771
fbshipit-source-id: 24ceb6f5876122aa8ad9e08c7e903215864ad6f5
Summary:
This type annotation was declared inline twice. Just pulling it out into a type alias in this diff.
Changelog: [Internal]
Reviewed By: yungsters
Differential Revision: D24723191
fbshipit-source-id: 9f2061087172979ea838dfdf2533e17b9b559c71
Summary:
Int32EnumTypeAnnotation represents a union of numbers. In the corresponding type annotation, we represent options as `$ReadOnlyArray<{value: number}>`. Since each option is a number, we could instead represent options as `$ReadOnlyArray<number>` - there's no need to use an object with a singular property (i.e: 'value'). The same is could be said of StringEnumTypeAnnotation.
In this diff, we change `Int32EnumTypeAnnotation.options` to `$ReadOnlyArray<number>`, and `StringEnumTypeAnnotation.options` to `$ReadOnlyArray<string>`.
Changelog: [Internal]
Reviewed By: yungsters
Differential Revision: D24723107
fbshipit-source-id: 4734cf72a4a29b6b321d8161bea70cf524ce0963
Summary:
Commands are `FunctionTypeAnnotation`, but they don't have a return type. I this diff, I introduced a `FunctionTypeAnnotation<+P, +R>` utility type, and made the `CommandTypeAnnotation` be an instantiation of it, with the return type fixed to `VoidTypeAnnotation`.
Now, the shape of a FunctionTypeAnnotation is unified across the NativeModule and Component schemas.
Changelog: [Internal]
Reviewed By: yungsters
Differential Revision: D24719965
fbshipit-source-id: 0089c3b23f05b0c534ba28dbe336c7f2db5866b3
Summary:
Everywhere else in the CodegenSchema, type annotation partials are suffixed with "Shape". In the NativeModule schema, we were using the suffix "Schema". In this diff, we standardize on the "Shape" suffix.
Changelog: [Internal]
Reviewed By: yungsters
Differential Revision: D24719395
fbshipit-source-id: 307935f5fe0681c31cd52e9cf4ae579f61c1ae68
Summary:
CodegenSchema exports `NativeModuleMethodParamSchema` and `NativeModuleObjectTypeAnnotationPropertySchema`, which are partials of NativeModule type annotations. This creates unnecessary coupling between the type annotations of CodegenSchema and the files that depend on it.
**Actual Problem:** Suppose that we want to rename one of these partials. Then, all imports in all files would have to be updated, even when the actual shape of the composed type annotation wasn't changed.
This diff removes these partials, which reduces the surface area of the exports of CodegenSchema.js
Changelog: [Internal]
Reviewed By: yungsters
Differential Revision: D24719396
fbshipit-source-id: c822aaa252f156c524f4ef4917ebb61b1a39ff9e
Summary:
The names of events and props flow type annotations are singular. The names of the commands flow types are however plural. This diff renames all "Commands*" flow types to be singular.
**Motivation:** Consistency
Changelog: [Internal]
Reviewed By: yungsters
Differential Revision: D24708276
fbshipit-source-id: 5d5d2123426ca1139953169d0ea764b82b2f3809
Summary:
All throughout the Codegen schema, we re-declare the following shape:
```
{
type: 'ObjectTypeAnnotation',
properties: $ReadOnlyArray<{
name: string,
optional: boolean,
typeAnnotation: ...
}>
}
```
This diff introduces an `ObjectTypeAnnotation<T>` utility type and replaces those re-declarations with instantiations of this type.
**Motivation:** To reduce noise in the CodegenSchema. This should be a pure refactor, and shouldn't actually change any behaviour.
Changelog: [Internal]
Reviewed By: yungsters
Differential Revision: D24707963
fbshipit-source-id: 6b4eb711ddd041f3a041109ade5ad5644fb16924
Summary:
This diff re-organizes CodegenSchema to declare the larger types first, which use smaller undeclared types. The smaller types are declared further down the file, and they themselves use even smaller undeclared types.
**Motivation:** Increase the readability of CodegenSchema.js. Now, if people want to understand the shape of the Codegen Schema, they can just read the file from top to bottom.
Changelog: [Internal]
Reviewed By: yungsters
Differential Revision: D24701424
fbshipit-source-id: 181e87bff5e32d998463221891f459b0df26ef52
Summary:
Reserved type annotations can appear in three different contexts: commands, props, and NativeModules. For now, commands and NativeModules share the same reserved type annotations. In the future, we may want to merge these reserved type annotations with the props reserved type annotations.
**Motivation:** The meaning of FunctionValue in FunctionValueTypeAnnotation isn't clear - in fact, it's downright confusing. Therefore, this diff renames this Flow type to ReservedTypeAnnotation, which I believe sufficiently captures the intent of the type annotation.
Changelog: [Internal]
Reviewed By: yungsters
Differential Revision: D24701322
fbshipit-source-id: bde0273b4a89c9e7175c60ed3468ed870b320044
Summary:
Our CodegenSchema is littered with types that have the following shape
```
{
name: string,
optional: boolean,
typeAnnotation: ...
}
```
In all these types, the only difference is the typeAnnotation. This diff introduces a new utility type called `NamedShape`, that just creates this shape, given a type annotation. This should help reduce the amount of noise in the CodegenSchema, and make it a bit easier to read.
Changelog: [Internal]
Reviewed By: yungsters
Differential Revision: D24701331
fbshipit-source-id: a30d3e22933116e3dabf7929615905febacecba3
Summary:
All ObjectTypeAnnotation *properties* in the codegen have the following shape:
```
{
name: string,
optional: boolean,
typeAnnotation: ...
}
```
EventObjectTypeProperty is a property of some ObjectTypeAnnotation, yet it doesn't follow this pattern. This diff cleans up EventObjectPropertyType. This is a part of a larger effort to clean up the Component Schema and unify the notion of a "type annotation" across the Component and Module schemas.
Reviewed By: yungsters
Differential Revision: D24701027
fbshipit-source-id: edc7dc632a217fb5a82ffd8a62aef990baf398c2
Ramanpreet Nara
Andrei Shikov
Samuel Susla
Rubén Norte
empyrical
Kevin Gozali
Ron Edelstein
Tim Yung
Héctor Ramos
Janic Duplessis