Summary:
Default behavior should be unchanged.
If we queue up a bunch of expensive tasks during an interaction, the default
`InteractionManager` behavior would execute them all in one synchronous loop at
the end the JS event loop via one `setImmediate` call, blocking the JS thread
the entire time.
The `setDeadline` addition in this diff enables an option to only execute tasks
until the `eventLoopRunningTime` is hit (added to MessageQueue/BatchedBridge),
allowing the queue execution to be paused if an interaction starts in between
tasks, making the app more responsive.
Additionally, if a task ends up generating a bunch of additional tasks
asynchronously, the previous implementation would execute these new tasks after
already scheduled tasks. This is often fine, but I want it to fully resolve
async tasks and all their dependencies before making progress in the rest of the
queue, so I added support for `type PromiseTask = {gen: () => Promise}` to do
just this. It works by building a stack of queues each time a `PromiseTask` is
started, and pops them off the stack once they are resolved and the queues are
processed.
I also pulled all of the actual queue logic out of `InteractionManager` and into
a new `TaskQueue` class to isolate concerns a bit.
public
Reviewed By: josephsavona
Differential Revision: D2754311
fb-gh-sync-id: bfd6d0c54e6410cb261aa1d2c5024dd91a3959e6
Summary:
public
We were adding all the arguments passed to all the JS functions and callbacks
called over the bridge to marker names, and this args can be huge, meaning a lot
of time spent stringifying arguments and therefore less accurate profile results
Reviewed By: nicklockwood
Differential Revision: D2761809
fb-gh-sync-id: 2d0b5b90cc9e59fe491c108b0360b84ab5fee5b7
Summary:
public
Rename the `BridgeProfiling` JS module to `Systrace`, since it's actually just
an API to Systrace markers.
This should make it clearer as we add more perf tooling.
Reviewed By: jspahrsummers
Differential Revision: D2734001
fb-gh-sync-id: 642848fa7340c545067f2a7cf5cef8af1c8a69a2
Summary:
MessageQueue no longer falls back to require. To do this we need to register all the modules in our internal unit tests. I did this codemod manually.
This is a bit unfortunate boilerplate but there are very few of these modules outside of unit tests. This boilerplate is only a hassle for these test files.
public
Reviewed By: spicyj
Differential Revision: D2736397
fb-gh-sync-id: 59fa4c4e75c538f3577bc9693b93e1b7c4d4d233
Summary:
The JavaScript ecosystem doesn't have the notion of a built-in native module loader. Even Node is decoupled from its module loader. The module loader system is just JS that runs on top of the global `process` object which has all the built-in goodies.
Additionally there is no such thing as a global require. That is something unique to our providesModule system. In other module systems such as node, every require is contextual. Even registered npm names are localized by version.
The only global namespace that is accessible to the host environment is the global object. Normally module systems attaches itself onto the hooks provided by the host environment on the global object.
Currently, we have two forms of dispatch that reaches directly into the module system. executeJSCall which reaches directly into require. Everything now calls through the BatchedBridge module (except one RCTLog edge case that I will fix). I propose that the executors calls directly onto `BatchedBridge` through an instance on the global so that everything is guaranteed to go through it. It becomes the main communication hub.
I also propose that we drop the dynamic requires inside of MessageQueue/BatchBridge and instead have the modules register themselves with the bridge.
executeJSCall was originally modeled after the XHP equivalent. The XHP equivalent was designed that way because the act of doing the call was the thing that defined a dependency on the module from the page. However, that is not how React Native works.
The JS side is driving the dependencies by virtue of requiring new modules and frameworks and the existence of dependencies is driven by the JS side, so this design doesn't make as much sense.
The main driver for this is to be able to introduce a new module system like Prepack's module system. However, it also unlocks the possibility to do dead module elimination even in our current module system. It is currently not possible because we don't know which module might be called from native.
Since the module system now becomes decoupled we could publish all our providesModule modules as npm/CommonJS modules using a rewrite script. That's what React Core does.
That way people could use any CommonJS bundler such as Webpack, Closure Compiler, Rollup or some new innovation to create a JS bundle.
This diff expands the executeJSCalls to the BatchedBridge's three individual pieces to make them first class instead of being dynamic. This removes one layer of abstraction. Hopefully we can also remove more of the things that register themselves with the BatchedBridge (various EventEmitters) and instead have everything go through the public protocol. ReactMethod/RCT_EXPORT_METHOD.
public
Reviewed By: vjeux
Differential Revision: D2717535
fb-gh-sync-id: 70114f05483124f5ac5c4570422bb91a60a727f6
Summary: I have disected lint warnings fixes to several PRs. This one fixes lint warnings under Libraries/Utility path.
Closes https://github.com/facebook/react-native/pull/4444
Reviewed By: svcscm
Differential Revision: D2705303
Pulled By: spicyj
fb-gh-sync-id: c745ac62cbff30d6bb9478a1d2465fe56b305f0c
Summary: public
We were calling constantsToExport twice for every ViewManager, and including two copies of the values in __fbBatchedBridgeConfig. This diff removes the copy from UIManager and then puts it back on the JS side.
Reviewed By: tadeuzagallo
Differential Revision: D2665625
fb-gh-sync-id: 147ec4bfb404835e3875964476ba233d619c28aa
Summary: public
Use arrays instead of dictionaries for encoding module method information.
This further reduces UIExplorer startup JSON from 16104 bytes to 14119 (12% reduction)
Reviewed By: javache
Differential Revision: D2570057
fb-gh-sync-id: 4a53a9ead4365a136e7caeb650375733e1c24c0e
Summary: public
We're sending a lot of module config data when the app first starts, and much of this is redundant.
UIExplorer current sends 19061 bytes of module config JSON. This diff reduces that to 16104 (15% saving) by stripping modules that have no methods or constants, and removing method types unless method is async.
Reviewed By: tadeuzagallo, javache
Differential Revision: D2570010
fb-gh-sync-id: 8c0abbd1cdee3264b37a4f52e852008caaffb9c5
Summary: @public
Take a step back and de-batch the bridge calls so we can have better profiling data and a better starting point to work on future optimisations. Also gave a 10~15% win on first render.
Reviewed By: @javache
Differential Revision: D2493674
fb-gh-sync-id: 05165fdd00645bdf43e844bb0c4300a2f63e7038
Summary:
This will throw an error message with the problematic callback module/method. Previously we would get an invariant in this case when we try to access `callback.apply` later in the method.
Summary:
@public
After refactoring the MessageQueue a guard was missing on around `batchedUpdates`
call.
Test Plan: Introduce an error on `getInitialState` of `AdsManagerTabsModalView.ios.js`
Summary:
@public
The current implementation of `MessageQueue` is huge, over-complicated and spread
across `MethodQueue`, `MethodQueueMixin`, `BatchedBridge` and `BatchedBridgeFactory`
Refactored in a simpler way, were it's just a `MessageQueue` class and `BatchedBridge`
is only an instance of it.
Test Plan:
I had to make some updates to the tests, but no real update to the native side.
There's also tests covering the `remoteAsync` methods, and more integration tests for UIExplorer.
Verified whats being used by Android, and it should be safe, also tests Android tests have been pretty reliable.
Manually testing: Create a big hierarchy, like `<ListView>` example. Use the `TimerMixin` example to generate multiple calls.
Test the failure callback on the `Geolocation` example.
All the calls go through this entry point, so it's hard to miss if it's broken.
Summary:
@public
This removes the last piece of data that was still stored on the DATA section,
`RCT_IMPORT_METHOD`. JS calls now dynamically populate a lookup table simultaneously
on JS and Native, instead of creating a mapping at load time.
Test Plan: Everything still runs, tests are green.
Summary:
@public
`[Bridge] Add support for JS async functions to RCT_EXPORT_METHOD` was imported but broke some internal code, reverting the `MessageQueue` that caused the issues and add a test, since the method is not used yet.
Test Plan: Run the test o/
Summary:
Adds support for JS async methods and helps guide people writing native modules w.r.t. the callbacks. With this diff, on the native side you write:
```objc
RCT_EXPORT_METHOD(getValueAsync:(NSString *)key
resolver:(RCTPromiseResolver)resolve
rejecter:(RCTPromiseRejecter)reject)
{
NSError *error = nil;
id value = [_nativeDataStore valueForKey:key error:&error];
// "resolve" and "reject" are automatically defined blocks that take
// any object (nil is OK) and an NSError, respectively
if (!error) {
resolve(value);
} else {
reject(error);
}
}
```
On the JS side, you can write:
```js
var {DemoDataStore} = require('react-native').NativeModules;
DemoDataStore.getValueAsync('sample-key').then((value) => {
console.log('Got:', value);
}, (error) => {
console.error(error);
// "error" is an Error object whose message is the NSError's description.
// The NSError's code and domain are also set, and the native trace i
Closes https://github.com/facebook/react-native/pull/1232
Github Author: James Ide <ide@jameside.com>
Test Plan: Imported from GitHub, without a `Test Plan:` line.
Summary:
@public
Right now the profiler shows how long the executor took on JS but doesn't show
how long each of the batched calls took, this adds a *very* high level view of JS
execution (still doesn't show properly calls dispatched with setImmediate)
Also added a global property on JS to avoid trips to Native when profiling is
disabled.
Test Plan:
Run the Profiler on any app
{F22491690}
Summary:
Wraps the setImmediate handlers in a `batchUpdates` call before they are synchronously executed at the end of the JS execution loop.
Closes https://github.com/facebook/react-native/pull/1242
Github Author: James Ide <ide@jameside.com>
Test Plan:
Added two `setImmediate` calls to `componentDidMount` in UIExplorerApp. Each handler calls `setState`, and `componentWillUpdate` logs its state. With this diff, we can see the state updates are successfully batched.
```javascript
componentDidMount() {
setImmediate(() => {
console.log('immediate 1');
this.setState({a: 1});
});
setImmediate(() => {
console.log('immediate 2');
this.setState({a: 2});
});
},
componentWillUpdate(nextProps, nextState) {
console.log('componentWillUpdate with next state.a =', nextState.a);
},
```
**Before:**
"immediate 1"
"componentWillUpdate with next state.a =", 1
"immediate 2"
"componentWillUpdate with next state.a =", 2
**After:**
"immediate 1"
"immediate 2"
"componentWillUpdate with next state.a =", 2
Addresses the batching issue in #1232. cc @vjeux @spicyj
Summary:
Adds support for JS async methods and helps guide people writing native modules w.r.t. the callbacks. With this diff, on the native side you write:
```objc
RCT_EXPORT_METHOD(getValueAsync:(NSString *)key
resolver:(RCTPromiseResolver)resolve
rejecter:(RCTPromiseRejecter)reject)
{
NSError *error = nil;
id value = [_nativeDataStore valueForKey:key error:&error];
// "resolve" and "reject" are automatically defined blocks that take
// any object (nil is OK) and an NSError, respectively
if (!error) {
resolve(value);
} else {
reject(error);
}
}
```
On the JS side, you can write:
```js
var {DemoDataStore} = require('react-native').NativeModules;
DemoDataStore.getValueAsync('sample-key').then((value) => {
console.log('Got:', value);
}, (error) => {
console.error(error);
// "error" is an Error object whose message is the NSError's description.
// The NSError's code and domain are also set, and the native trace i
Closes https://github.com/facebook/react-native/pull/1232
Github Author: James Ide <ide@jameside.com>
Test Plan: Imported from GitHub, without a `Test Plan:` line.
Mike Armstrong
Spencer Ahrens
Tadeu Zagallo
Sebastian Markbage
Nick Lockwood
Huang Yu
David Aurelio
Andy Street
Eric Vicenti
Krzysztof Magiera
Olivier Notteghem
James Ide
Chace Liang
Kevin Gozali
Marshall Roch
Christopher Chedeau