Summary:
Fixes#4740, where views would unnecessarily be retained after performing `navigator.pop()` - this was particularly problematic for big lists and memory-intensive custom views.
This fix causes no functional change: `_previousViews` are only used in the loop starting at line 564 to ensure that the JavaScript and Native navigation stacks are equivalent at all times. As we do in this fix, that loop limits itself to only the views expected to be on the React navigation stack. So overall this change makes the code logically 'more correct'.
Tested by checking that `_previousViews.count` is always equivalent to `previousReactCount` in the loop (which means we could remove the complex `MIN(... MIN(previousReactCount, _previousViews.count)` in the loop too, but I wanted to keep the diff as small as possible for now).
Closes https://github.com/facebook/react-native/pull/10789
Differential Revision: D4140502
Pulled By: ericvicenti
fbshipit-source-id: 4491ad3c16642914c3081295cf95c4cf36be9f94
Summary: Instead of silently ignorning non-leaf nodes with measure functions, we should assert that we don't create those kinds of trees.
Reviewed By: emilsjolander
Differential Revision: D4130770
fbshipit-source-id: a3ef10a2e63bbc12b5aa07977e4b84c8d59e3ffe
Summary:
This adds native support for `Animated.event` on iOS.
**Test plan**
Tested in the native animated UIExplorer example that it works properly like on Android.
Closes https://github.com/facebook/react-native/pull/9598
Differential Revision: D4110331
fbshipit-source-id: 15748d23d0f475f2bcd1040ca3dca33e2620f058
Summary:
UIScrollView's internal logic with scroll indicator dies when bad values (e.g. NaN/Infinity) are set on the position/size.
We already guard UIManager with these checks but the RCTScrollView's underlying scrollview (RCTCustomScrollView) can get these set from other places, and we're seeing crashes in this area.
Reviewed By: javache
Differential Revision: D4088601
fbshipit-source-id: b1185cc7c65ba0266787441169264c94338fc55c
Summary: This method has been deprecated for a while and there are no internal use-cases left (customBubblingEventTypes is still used by RCTViewManager though).
Reviewed By: fkgozali
Differential Revision: D4083327
fbshipit-source-id: 261e0dce3b41714d13b46d146f87fc415eb9e817
Summary:
**Motivation**
I'm working on a project that uses React Native and needs to add direct synchronous bindings to native stuff through the JavaScriptCore C API. This is because it's performance-sensitive and would benefit from the quickest JS->C path. It does this using cross-platform C++ code that works on both iOS and Android. Most of the infrastructure for getting access to the JSC context is already in React Native actually, just had to add a few more things.
(lexs you mentioned to tag you in this pull request)
**Test plan**
Modify the JavaScriptCore context through the `JSContextRef` returned (eg. add an object at global scope) and verify that it exists in JavaScript.
Closes https://github.com/facebook/react-native/pull/10399
Differential Revision: D4080945
Pulled By: lexs
fbshipit-source-id: 6659b7a01e09fd84475adde183c1d3aca2d4cf09
Summary:
Explain the **motivation** for making this change. What existing problem does the pull request solve?
See https://github.com/facebook/react-native/issues/10506. A native `NSError` with `NSUnderlyingErrorKey` set causes a JSON stringify error from the websocket dispatcher if remote debugging is enabled.
**Test plan (required)**
I'm not familiar with the react native testing framework. Happy to add a test for this if someone can point me to where this part of the codebase is exercised :)
I did some spot checks with nil user dictionaries and nil underlying errors here. The case that this solves is testable using https://github.com/superseriouscompany/react-native-error-repro, specifically:
```objective-c
NSError *underlyingError = [NSError errorWithDomain:@"underlyingDomain" code:421 userInfo:nil];
NSError *err = [NSError errorWithDomain:@"domain" code:68 userInfo:@{@"NSUnderlyingError": underlyingError}];
reject(@"foo", @"bar", err);
```
Closes https://github.com/facebook/react-native/pull/10507
Differential Revision: D4080802
Pulled By: lacker
fbshipit-source-id: 93a41d9e9a710e406a6ccac214a5617271b4bede
Summary:
1. Using weak container to hold the currently opened alerts.
2. Using weak reference to alertController in action handler block.
3. BTW, remove the unused vars: _alertCallbacks, _alertButtonKeys.
Test plan (required)
```
- (void)invalidate
{
for (UIAlertController *alertController in _alertControllers) {
[alertController.presentingViewController dismissViewControllerAnimated:YES completion:nil];
}
}
```
Since we use weak container, _alertControllers should only contains the currently opened alerts.
I test this way: Put a breakpoint in invalidate, open the UIExplorer play with the 'Alert' & 'AlertIOS' examples, then fire a reload and see if _alertControllers contains the expected values.
Closes https://github.com/facebook/react-native/pull/10407
Differential Revision: D4078649
Pulled By: lacker
fbshipit-source-id: 8509e7e7142379a81d5b28c9067c085bad8bb5cb
Summary: This is a simple hook to allow native side to fetch any file under the js root folder via packager. Historically, only the `main.jsbundle` is fetched via the packager. This then allows fetching local file like a json file that lives under the same root js folder
Reviewed By: yungsters
Differential Revision: D4037730
fbshipit-source-id: a2d6eb5e30d148fee573d413fc4036d0189f4938
Summary:
- Consolidate common code in iOS and tvOS test scripts
- Start the packager before starting tests, to improve reliability
- Increase timeout value in RCTTestRunner.m
Closes https://github.com/facebook/react-native/pull/10378
Differential Revision: D4028364
Pulled By: bestander
fbshipit-source-id: 24c2124a1c62643a02f0668b60a67b971e08d1a3
Summary:
JS API very similar to web workers and node's child process.
Work has been done by somebody else for the Android implementation over at #7020, so we'd need to have these in sync before anything gets merged.
I've made a prop `messagingEnabled` to be more explicit about creating globals—it might be sufficient to just check for an onMessage handler though.
Closes https://github.com/facebook/react-native/pull/9762
Differential Revision: D4008260
fbshipit-source-id: 84b1afafbc0ab1edc3dfbf1a8fb870218e171a4c
Summary:
This fixes the following problem on iPad (happens when trying to show the devmenu ActionSheet):
> Terminating app due to uncaught exception 'NSGenericException', reason: 'Your application has presented a UIAlertController () of style UIAlertControllerStyleActionSheet. The modalPresentationStyle of a UIAlertController with this style is UIModalPresentationPopover. You must provide location information for this popover through the alert controller's popoverPresentationController. You must provide either a sourceView and sourceRect or a barButtonItem. If this information is not known when you present the alert controller, you may provide it in the UIPopoverPresentationControllerDelegate method -prepareForPopoverPresentation.'
I see 2 ways of fixing the issue: falling back to Alert (always displayed in the center) or adding a fake view somewhere on the screen to display the ActionSheet.
Reviewed By: mmmulani
Differential Revision: D3989672
fbshipit-source-id: 59fabc50ec1aac3ae6ddd7ecf4d8e8e5b9586dba
Summary:
* Motivation *
Second PR for Apple TV support.
* Test plan *
Apple TV tests have been added to scripts/objc-test.sh
Closes https://github.com/facebook/react-native/pull/10227
Differential Revision: D3974064
Pulled By: javache
fbshipit-source-id: 36dffb4517efa489e40fa713a30655d1d76ef646
Summary:
Packager can take a long time to load and the progress is usually displayed in another window (Terminal). I'm adding support for showing a UI inside React Native app for packager's progress when loading a bundle.
This is how it will work:
1. React Native sends request to packager with `Accept: multipart/mixed` header.
2. Packager will detect that header to detect that client supports progress events and will reply with `Content-Type: multipart/mixed`.
3. While building the bundle it will emit chunks with small metadata (like `{progress: 0.3}`). In the end it will send the last chunk with the content of the bundle.
4. RN runtime will be receiving the events, for each progress event it will update the UI. The last chunk will be the actual bundle which will end the download process.
This workflow is totally backwards-compatible -- normally RN doesn't set the `Accept` header.
Reviewed By: mmmulani
Differential Revision: D3845684
fbshipit-source-id: 5b3d2c5a4c6f4718d7e5de060d98f17491e82aba
Summary:
Android (starting from API 23) supports "light status bar", thus it is possible to extend StatusBar and make `barStyle` property work not only for iOS, but also for Android.
This PR introduces one more `barStyle` option `dark-content` in addition to two existing ones (`default` and `light-content`).
Why there are 3 options instead of 2?
Two simple reasons:
1) to make all existing applications fully compatible with these changes;
2) the default status bar on Android is dark with white text and icons, while on iOS it is light with black text and icons on it. Thus the `default` option means something like "I don't really care, just apply the default color for this platform", while two other options (`light-content` and `dark-content`) allow to accurately specify the required result.
Closes https://github.com/facebook/react-native/pull/10185
Differential Revision: D3952346
fbshipit-source-id: 999a67614abff52321fbeb06298ebf1946c3f1d1
Summary:
First commit for Apple TV support: changes to existing Objective-C code so that it will compile correctly for tvOS.
Closes https://github.com/facebook/react-native/pull/9649
Differential Revision: D3916021
Pulled By: javache
fbshipit-source-id: 34acc9daf3efff835ffe38c43ba5d4098a02c830
Summary: Changing the Modal's presentation style ensures that UIKit correctly adjusts its bounds when entering/exiting the in-call status bar.
Reviewed By: fkgozali
Differential Revision: D3916167
fbshipit-source-id: a0bea60751744ac082ba6ec38177fb093a8f2be5
Summary:
This fixes measuring of items in the main axis of a container. Previously items were in a lot of cases measured with UNSPECIFIED instead of AT_MOST. This was to support scrolling containers. The correct way to handle scrolling containers is to instead provide them with their own overflow value to activate this behavior. This is also similar to how the web works.
This is a breaking change. Most of your layouts will continue to function as before however some of them might not. Typically this is due to having a `flex: 1` style where it is currently a no-op due to being measured with an undefined size but after this change it may collapse your component to take zero size due to the implicit `flexBasis: 0` now being correctly treated. Removing the bad `flex: 1` style or changing it to `flexGrow: 1` should solve most if not all layout issues your see after this diff.
Reviewed By: majak
Differential Revision: D3876927
fbshipit-source-id: 81ea1c9d6574dd4564a3333f1b3617cf84b4022f
Summary:
This version of the css-layout project includes support for .NET projects. Up in the air is how many configurations of .NET projects we allow for, such as Portable Class Libraries, Universal Windows Platform, .NET Core, etc. Still needs integration with Buck.
Closes https://github.com/facebook/css-layout/pull/220
Reviewed By: lucasr
Differential Revision: D3909367
Pulled By: emilsjolander
fbshipit-source-id: aaaa9c4ff2d3452649f256c3268cf873cf33a0b9
Summary: In dev mode, this removes all iOS crashes from specifying the incorrect interface orientations. It also includes okay-ish instructions for how to fix the problem.
Reviewed By: fkgozali
Differential Revision: D3904998
fbshipit-source-id: 699599fa77fd87e6615436250e38944e577e75a0
Summary:
If we use RCTLogError, we show a redbox with a native stacktrace before we can actually show the JS stacktrace (which is only done through dispatch_async). Instead use RCTLogWarn so its logged to console but not reported as a redbox.
Also fix call to RCTNSErrorFromJSErrorRef off the main thread RCTJSCExecutor.
Reviewed By: majak, mmmulani
Differential Revision: D3886813
fbshipit-source-id: 119d04858cb0c85c79e12ff8a79806711f8e0b27
Summary: Introduce `overflow:scroll` so that scrolling can be implemented without the current overflow:visible hackiness. Currently we use AT_MOST to measure in the cross axis but not in the main axis. This was done to enable scrolling containers where children are not constraint in the main axis by their parent. This caused problems for non-scrolling containers though as it meant that their children cannot be measured correctly in the main axis. Introducing `overflow:scroll` fixes this.
Reviewed By: astreet
Differential Revision: D3855801
fbshipit-source-id: 3c365f9e6ef612fd9d9caaaa8c650e9702176e77
Summary: Introduce `overflow:scroll` so that scrolling can be implemented without the current overflow:visible hackiness. Currently we use AT_MOST to measure in the cross axis but not in the main axis. This was done to enable scrolling containers where children are not constraint in the main axis by their parent. This caused problems for non-scrolling containers though as it meant that their children cannot be measured correctly in the main axis. Introducing `overflow:scroll` fixes this.
Reviewed By: astreet
Differential Revision: D3855801
fbshipit-source-id: 6077b0bcb68fe5ddd4aa22926acab40ff4d83949
Summary:
Explain the **motivation** for making this change. What existing problem does the pull request solve?
- Being able to call these from JS is really useful in envs where action sheets are not available (eg: App Extensions). I made the functions act as explicit toggles, this way it's easy to set the wanted behavior.
**Test plan (required)**
- I tested the functions manually.
Closes https://github.com/facebook/react-native/pull/9242
Differential Revision: D3749568
fbshipit-source-id: 2cea1d609dc792aec73d4bcf1eae07575725b4b7
Summary: Get rid of the old behaviour of JSON encoding in `nativeRequireModuleConfig` and consistently use the same names for function types "async/promise/sync"
Reviewed By: lexs
Differential Revision: D3819348
fbshipit-source-id: fc798a5abcaf6a3ef9d95bd8654afa7825c83967
Summary:
At the moment, posting RCTReloadNotification in any circumstance causes all RCTBridge instances to reload. This change scopes the notification to the bridge for which it was intended.
Closes https://github.com/facebook/react-native/pull/8762
Differential Revision: D3831914
fbshipit-source-id: ff29574f574ecd1a403057ddd0458dea38f0136e
Summary:
Include CSSLayout headers in the same way as other project headers, ie `#import <CSSLayout/CSSLayout.h>` becomes `#import "CSSLayout.h"`. CSSLayout is not a framework or system dependency, so shouldn't (AFAIK) be included with angle brackets. Doing so breaks framework builds, such as when RN is used as a pod in a swift project.
In combination with https://github.com/facebook/css-layout/pull/217 this fixes https://github.com/facebook/react-native/issues/9014 (specifically swift cocoapods projects). There is then no need for a separate CSSLayout pod subspec.
Tests run on the RN project in isolation (with changes inside `CSSLayout` itself also applied) and against a dummy swift project with RN included as a pod.
NB: This effectively reverts https://github.com/facebook/react-native/pull/9015 and may break non-swift cocoapods projects unless https://github.com/facebook/css-layout/pull/217 is merged and synced first.
Update: As discussed with alloy and emilsjolander, wrap these imports in a preprocess
Closes https://github.com/facebook/react-native/pull/9544
Differential Revision: D3821791
Pulled By: javache
fbshipit-source-id: d27ac8be9ce560d03479b43d3db740cd196c24da
Summary:
First of all,please forgive my poor english. LOL
When I use Promises as callBack,
I followed the guide to make a (RCTPromiseRejectBlock),in this RCTPromiseRejectBlock,
from the guide I should add a (NSString)code a (NSString)message and an (NSError) to a RCTPromiseRejectBlock.
As what did guide said I just make a NSError with
```
NSError *error=[NSError errorWithDomain:@"native stroage error" code:402 userInfo:@{@"data":key}];
```
As you can see, I want to transfer some data in this callback.But I just found the (userInfo) of NSError didn't composed to a RCTPromiseRejectBlock.
So I check out the source code of RCTUtil ,I found this
```
// TODO: Can we just replace RCTMakeError with this function instead?
NSDictionary<NSString *, id> *RCTJSErrorFromCodeMessageAndNSError(NSString *code,
NSString *message,
NSError *__nullable error)
{
NSString *errorMessage;
Closes https://github.com/facebook/react-native/pull/9519
Differential Revision: D3821334
Pulled By: javache
fbshipit-source-id: 0fa40dd93932af70123b8f8a20a5c7dc5afdd6ac
Summary:
Refresh bug #7976
This bug has persisted for 3 versions of React Native. Currently everyone is having to do a silly fix, documented in the bug. I spent an hour trying to find the source of the bug. Failing to find the issue quickly, I've just decided to make this simple bugfix.
According to janicduplessis #7976 we will likely be re-writing this control in iOS 10.
Closes https://github.com/facebook/react-native/pull/9249
Differential Revision: D3817227
fbshipit-source-id: f98282aadeca3e85239836f589eda901d454f2ce
Summary:
When an absolutely positioned node appeared inside an overflow:hidden
parent, we were limiting its height. This is inconsistent with how
layout behaves on the web.
Closes https://github.com/facebook/css-layout/pull/218
Reviewed By: lucasr
Differential Revision: D3797285
Pulled By: emilsjolander
fbshipit-source-id: 98f98e77aa26edce86b9882c1cac284799b69a27
Summary:
The overflow prop needs to be set on the shadow view so that it can make its way into the layout engine. In some situations, the value of the overflow prop affects the calculations of the layout engine.
**Test plan (required)**
Verified in a test app that the `overflow` prop makes its way into the layout engine. Also, my team's app is currently using this change.
Adam Comella
Microsoft Corp.
Closes https://github.com/facebook/react-native/pull/9659
Differential Revision: D3790552
fbshipit-source-id: 61513ece63ae214f48c6cb6f40fb29757a0ac706
Summary:
**Motivation:** We have a project that dynamically changes classes in runtime. This component is initialised within the React Native context. Therefor `RCTBatchedBridge` copies the classes before changes are made to them. While React Native is running, changes are made to the classes dynamically. When the project reloads the `RCTBatchedBridge`, it now has an invalid list of classes containing trash pointers. This causes the project to crash on https://github.com/facebook/react-native/blob/master/React/Base/RCTBatchedBridge.m#L288 with EXC_BAD_ACCESS in the DEBUG-mode.
**Solution:** Copy the class list on each reload to get the current state. Since this is only a DEBUG-feature the overhead of this should not be a major issue.
Closes https://github.com/facebook/react-native/pull/9541
Differential Revision: D3775012
Pulled By: javache
fbshipit-source-id: d55fa0742ca100d8018c73080230cf718aa5a7e9
Summary: added API to in UIManager to find the rootTag/View of any reactTag based on its layout (shadow views) hierarchy (not to be used by JS)
Reviewed By: javache
Differential Revision: D3750410
fbshipit-source-id: 68611e39930d53ece478f25245ddc7f7838daaa6
Summary:
Change & motivation:
Added an environment variable, RCT_NO_LAUNCH_PACKAGER, that when set
Xcode not the launch the packager when building. This switch is needed
for CI builds, where launching the packager, which stays running, causes
the build to hang; port conflicts can be an issue as well. Really the
packager is only appropriate for interactive developer builds. This
is a problem for Visual Studio Team Services where we just added
React Native CI build support with a React Native build extension and
surely a problem for all CI systems. This fix is simple, and now if
the build machine sets RCT_NO_LAUNCH_PACKAGER globally or passes
it as an arg to xcodebuild, the packager won't launch and all is well.
Testing:
I tested by building both with & without the environment variable defined, ensuring
things work probably in both cases. I also added support to our VSTS build ask to use
this and verified it works as expected too, when setting the variable..
Closes https://github.com/facebook/react-native/pull/6180
Differential Revision: D3764648
Pulled By: sahrens
fbshipit-source-id: b41cc06152aec271bbfdb2bd52d4b83a5f9c09a7
Summary:
Thanks for submitting a pull request! Please provide enough information so that others can review your pull request:
> **Unless you are a React Native release maintainer and cherry-picking an *existing* commit into a current release, ensure your pull request is targeting the `master` React Native branch.**
Explain the **motivation** for making this change. What existing problem does the pull request solve?
Prefer **small pull requests**. These are much easier to review and more likely to get merged. Make sure the PR does only one thing, otherwise please split it.
**Test plan (required)**
Demonstrate the code is solid. Example: The exact commands you ran and their output, screenshots / videos if the pull request changes UI.
Make sure tests pass on both Travis and Circle CI.
**Code formatting**
Look around. Match the style of the rest of the codebase. See also the simple [style guide](https://github.com/facebook/react-native/blob/master/CONTRIBUTING.md#style-guide).
For more info, see
Closes https://github.com/facebook/react-native/pull/9560
Differential Revision: D3763027
Pulled By: hramos
fbshipit-source-id: 3d60833fde92a30658f3d9f108708b41ce859a13
Summary:
iOS and css have a different way to draw borders. Decision which rendering to use considers alpha value of given color as well.
If alpha=0 -> border is invisible -> we use ios render.
That doesn't work as expected in a case where only `borderWidth` property is set.
A default black color is used when no `borderColor` is set, but alpha checking logic will test `null`. Which resolves as alpha=0.
Reviewed By: javache
Differential Revision: D3750523
fbshipit-source-id: 453d52a16287c6286e8887a9ab3795d3ded1dec6
Geordie
Andy Street
Dustin Shahidehpour
Ashok Menon
Georgiy Kassabli
Pieter De Baets
Dan Caspi
James Ide
Emil Sjolander
Mehdi Mulani
Kevin Gozali
Nikhilesh Sigatapu
Neil Sarkar
littlesome
Scott Wolchok
Kazuki Sakamoto
dlowder-salesforce
Jacob Parker
Alex Kotliarskyi
Ovidiu Viorel Iepure
Alexander Pantyuhov
Marc Horowitz
Martin Kralik
Antti Moilanen
mattpodwysocki
Yiding Jia
Hugo Dozois
Alexey Lang
Ben Roth
rh389
BaihaoTian
makee
Gant
Indragie Karunaratne
Ben Nham
Adam Comella
Kasper Hirvikoski
Bret Johnson
Matt Johnston
