Summary:
(This diff was decoupled from D4983054 which landing was quite delayed.)
`RCTUnsafeExecuteOnMainQueueOnceSync()` synchronously executes a block once, on main queue.
I found that our old implementation of `RCTScreenScale()` causes deadlock when it is called from main and background thread simultaneously.
After I implemented my own solution I googled this issue and found an article by Ben Alpert with this awesome helper:
https://benalpert.com/2014/04/02/dispatch-once-initialization-on-the-main-thread.html
So, I found it super useful and borrowed it.
Hey spicyj! :)
Reviewed By: fkgozali
Differential Revision: D5632592
fbshipit-source-id: dff43a5780b7404a3cc109c66c131cef4f4df429
Summary:
This shows progress for the download of the JS bundle (different from the packager transform progress that we show already). This is useful especially when loading the JS bundle from a remote source or when developing on device (on simulator + localhost it pretty much just downloads instantly). This will be nice for the expo client since all bundles are loaded over the network and can take several seconds to load.
This depends on https://github.com/facebook/metro-bundler/pull/28 to work but won't crash or anything without it, it just won't show the progress percentage.
![img_05070155d2cc-1](https://user-images.githubusercontent.com/2677334/28293828-2c08d974-6b24-11e7-9334-e106ef3326d9.jpeg)
**Test plan**
Tested that bundle download progress is shown properly in RNTester on both localhost + simulator and on real device with network conditionner to simulate a slow loading bundle.
Tested that it doesn't cause issues if the packager doesn't send the Content-Length header.
Closes https://github.com/facebook/react-native/pull/15066
Differential Revision: D5449073
Pulled By: shergin
fbshipit-source-id: 43a8fb559393bbdc04f77916500e21898695bac5
Summary:
We've simplified a lot of the conditions for eager of the module init so now we can introduce a final switch to allow modules to opt-out (and in the future opt-in if they still require the behaviour).
We now require you to be explicit about the intended behaviour and implement the `+ (BOOL)requiresMainQueueSetup` method on your module. When you return YES from this method, it tells the bridge the module needs to be created on the main thread (and to avoid deadlocks, we do so eagerly during bridge startup). When you return NO, the native module will be initialised when it's first accessed from JS.
The current behaviour is maintained but a warning is emitted until the new API is adopted.
Reviewed By: fkgozali
Differential Revision: D5527788
fbshipit-source-id: 56d38f81e58cf950547b9780e89bfac4667eeaaa
Summary:
**Motivation**
Properly support long presses on the Apple TV remote, and also enable dev menu functionality on a real Apple TV device (shaking an Apple TV doesn't work 😄 )
**Test plan**
New example added to `RNTester`.
Closes https://github.com/facebook/react-native/pull/15221
Differential Revision: D5526463
Pulled By: javache
fbshipit-source-id: a61051e86bc82a9561eefc1704bed6b1f2617e05
Summary: When loading bundle from packager, "application/javascript" and "text/javascript" both refer to JS, so let's allow both for now.
Reviewed By: javache
Differential Revision: D5499446
fbshipit-source-id: f0b42e2fe5dc043a68d2c8df6a9f81e6dd995b57
Summary:
This is the first PR from a series of PRs grabbou and me will make to add blob support to React Native. The next PR will include blob support for XMLHttpRequest.
I'd like to get this merged with minimal changes to preserve the attribution. My next PR can contain bigger changes.
Blobs are used to transfer binary data between server and client. Currently React Native lacks a way to deal with binary data. The only thing that comes close is uploading files through a URI.
Current workarounds to transfer binary data includes encoding and decoding them to base64 and and transferring them as string, which is not ideal, since it increases the payload size and the whole payload needs to be sent via the bridge every time changes are made.
The PR adds a way to deal with blobs via a new native module. The blob is constructed on the native side and the data never needs to pass through the bridge. Currently the only way to create a blob is to receive a blob from the server via websocket.
The PR is largely a direct port of https://github.com/silklabs/silk/tree/master/react-native-blobs by philikon into RN (with changes to integrate with RN), and attributed as such.
> **Note:** This is a breaking change for all people running iOS without CocoaPods. You will have to manually add `RCTBlob.xcodeproj` to your `Libraries` and then, add it to Build Phases. Just follow the process of manual linking. We'll also need to document this process in the release notes.
Related discussion - https://github.com/facebook/react-native/issues/11103
- `Image` can't show image when `URL.createObjectURL` is used with large images on Android
The websocket integration can be tested via a simple server,
```js
const fs = require('fs');
const http = require('http');
const WebSocketServer = require('ws').Server;
const wss = new WebSocketServer({
server: http.createServer().listen(7232),
});
wss.on('connection', (ws) => {
ws.on('message', (d) => {
console.log(d);
});
ws.send(fs.readFileSync('./some-file'));
});
```
Then on the client,
```js
var ws = new WebSocket('ws://localhost:7232');
ws.binaryType = 'blob';
ws.onerror = (error) => {
console.error(error);
};
ws.onmessage = (e) => {
console.log(e.data);
ws.send(e.data);
};
```
cc brentvatne ide
Closes https://github.com/facebook/react-native/pull/11417
Reviewed By: sahrens
Differential Revision: D5188484
Pulled By: javache
fbshipit-source-id: 6afcbc4d19aa7a27b0dc9d52701ba400e7d7e98f
Summary:
When using the packager behind something like an http tunnel it is possible something else than JS is returned, in that case throw an error instead of trying to parse it.
This is useful for Expo since the packager runs behind ngrok. This allows to intercept that error and show a more meaningful error message based on the ngrok response.
**Test plan**
Tested by changing the packager to return text/html content type and validate that the error shows up properly.
Also tested that it works when multipart response is disabled.
<img width="354" alt="screen shot 2017-07-19 at 8 01 58 pm" src="https://user-images.githubusercontent.com/2677334/28394905-39e86d52-6cbe-11e7-9059-13a85816a57e.png">
Closes https://github.com/facebook/react-native/pull/15112
Differential Revision: D5459395
Pulled By: shergin
fbshipit-source-id: aaea7ab2e1311ee8dc10feb579adf9b9701d8d4c
Summary:
Just a little typo fixing and wording clean up around some header docs.
Closes https://github.com/facebook/react-native/pull/14947
Differential Revision: D5398609
Pulled By: javache
fbshipit-source-id: 3eb40ef3308130c1d28b2efc7bb64d493e98825b
Summary:
C++ doesn't provide an implicit cast to an enum value from the enum's backing type. When a `.mm` file calls `RCT_ENUM_CONVERTER`, we end up with the following compiler error:
> Error: cannot initialize return object of type `<TypeName>` with an rvalue of type `NSInteger`
Since `RCT_ENUM_CONVERTER` is a macro, this error will appear whenever we try to expand the macro in a C++ context.
The project compiles and runs as expected when this additional cast is added 😃
Closes https://github.com/facebook/react-native/pull/14408
Reviewed By: javache
Differential Revision: D5215584
Pulled By: jballer
fbshipit-source-id: 7f4464afd7cd9dc9a014f646278bae20731f08ba
Summary:
Our iOS devs frequently turn off wifi and forget to turn it back on. This message should remind them that they need wifi to connect. Often they waste several minutes due to this problem.
I'm not sure if there's a test plan to apply here. Any suggestions?
Closes https://github.com/facebook/react-native/pull/13551
Differential Revision: D5149231
Pulled By: hramos
fbshipit-source-id: 0afc71024f10f802ac1a50435fb57fc10a02c819
Summary:
Thanks for submitting a PR! Please read these instructions carefully:
- [ ] Explain the **motivation** for making this change.
- [ ] Provide a **test plan** demonstrating that the code is solid.
- [ ] Match the **code formatting** of the rest of the codebase.
- [ ] Target the `master` branch, NOT a "stable" branch.
What existing problem does the pull request solve?
XCode [Thread Sanitizer](https://clang.llvm.org/docs/ThreadSanitizer.html) find race condition while read/write `_instance` variable in RCTModuleData class. A bridge can check `hasInstance` method while instance writes.
All tests passed on my device.
These changes remove data race, you can turn it in scheme configuration ![](https://www.shinobicontrols.com/wp-content/uploads/2016/08/Enable_Sanitizer.png)
Closes https://github.com/facebook/react-native/pull/13757
Differential Revision: D4994041
Pulled By: javache
fbshipit-source-id: 631cd59bbcbde193937d8baf8358ff6868717a2e
Summary:
Mitigates the issue in https://github.com/facebook/react-native/issues/10471, by not trying to present
a modal or alert view if the presenting view controller is already being dismissed.
Reviewed By: shergin
Differential Revision: D4962169
fbshipit-source-id: 593e3f21096458651d16677a3e030552f809bf02
Summary:
React Native uses JSON to marshal the data across the bridge.
And because of this we have to avoid using NaN and INF values in events and other pieces of data that suppose to be transfered to/from JS side.
(We also don't want to introduce additional wrapping/escaping semantics for perfomance reasons.)
So, we have to gate all particular cases where there is a possibility of NaN or INF values, and replace these value with something meaningful for each particular case.
We are using `0` as NaN substitution here because:
* NaN in touch event is super rare case;
* Conversion to `0` is fast;
* `0` is okay value for product code in most cases;
* In all cases `0` is decent analog to "undefined position on screen" for touch event;
* Nobody will explicitly handle NaN case in product code, just because it is super rare case and actually indicates that something else went wrong.
Reviewed By: javache
Differential Revision: D4918669
fbshipit-source-id: e95fa29e59dcdc40b57519e307b74c1f293da188
Summary:
In the Docs of `NSURLSession` ,
>IMPORTANT
>The session object keeps a strong reference to the delegate until your app exits or explicitly >invalidates the session. If you do not invalidate the session, your app leaks memory until it exits.
The RCTMultipartDataTask will cause memory leaks, it will make `RCTBatchedBridge` and ` RCTMultipartDataTask` will not release.
So call `[session finishTasksAndInvalidate];` at the end of `startTask` function.
Closes https://github.com/facebook/react-native/pull/12673
Reviewed By: shergin
Differential Revision: D4896497
Pulled By: javache
fbshipit-source-id: eb5f8761f67ad33a7de081a68a9a7e1d4329bfc0
Summary:
Issue Fix: #13485
Yet another race condition that was found by XCode's Thread Sanitizer.
Happens because wasBatchActive is read/write from multiple threads at the same time
- opened UIExplorer and see it works
- npm run test pass
- tested on own project as well
Signed CLA
Closes https://github.com/facebook/react-native/pull/13505
Differential Revision: D4906096
Pulled By: javache
fbshipit-source-id: 5d4329aafcfe9491ce0188fa1e2dd71e09b31031
Summary:
Output the reason for the error when failing to load source code. This was a big help when trying to diagnose https://github.com/facebook/react-native/issues/13299.
~~Unfortunately there still seems to be no way to get the offending line number (because `loadError.userInfo[RCTJSStackTraceKey]` is empty), but this is good enough.~~
Before:
```
[warn][tid:com.facebook.react.JavaScript][RCTBatchedBridge.m:510] Failed to execute source code. Unhandled JS Exception: SyntaxError
```
After:
```
[warn][tid:com.facebook.react.JavaScript][RCTJSCErrorHandling.mm:30] Couldn't get stack trace for http://localhost:8081/index.ios.bundle?platform=ios&dev=true&minify=false:81886
[warn][tid:com.facebook.react.JavaScript][RCTBatchedBridge.m:510] Failed to execute source code. Unhandled JS Exception: SyntaxError Unexpected keyword 'var'
```
Closes https://github.com/facebook/react-native/pull/13561
Differential Revision: D4908501
Pulled By: javache
fbshipit-source-id: a316dc70739b917b3cc690309d0ff37a8bb5d412
Summary:
* It complicates Touch Handling mechanism.
* The same functionality can be (and should be) implemented via overriding standard `hitTest:` method.
* It was marked as deprecated a while ago.
Reviewed By: mmmulani
Differential Revision: D4667776
fbshipit-source-id: 2e047c3308563a2849ea351a242270f0800fead2