Summary:
Hi,
This changelist is for fixing a crash issue on Android devices. For detail please see, https://github.com/facebook/react-native/issues/7080
In short, the crash occurred when a RN app with a Alert dialog shown at front switches back after it is destroyed at background due to lack of phone memory.
My fix is to let those Alert fragment classes accessible from android.support.v4.app.Fragment instantiate() function, so that it won't crash. And since other UI will be reloaded whatever, mListener is set to null to avoid any callback happen.
Closes https://github.com/facebook/react-native/pull/7105
Differential Revision: D3212435
Pulled By: mkonicek
fbshipit-source-id: d900a33a4f0fd49258be94c277def55555ec8d73
Summary:
Got this error when trying to run instrumentation tests locally with `./scripts/run-android-local-integration-test.sh`
```
C:\Users\janic\Developer\react-native\ReactAndroid\src\androidTest\java\com\facebook\react\tests\DatePickerDialogTestCase.java:110: error: cannot access com.facebook.react.modules.core.DefaultHardwareBackBtnHandler
return (DialogFragment) getActivity().getSupportFragmentManager()
^
class file for com.facebook.react.modules.core.DefaultHardwareBackBtnHandler not found
C:\Users\janic\Developer\react-native\ReactAndroid\src\androidTest\java\com\facebook\react\tests\ViewRenderingTestCase.java:82: error: cannot access com.facebook.react.touch.ReactInterceptingViewGroup
assertEquals("Incorrect (or not applied) opacity", expectedOpacity, view.getAlpha());
^
class file for com.facebook.react.touch.ReactInterceptingViewGroup not found
C:\Users\janic\Developer
Closes https://github.com/facebook/react-native/pull/7575
Differential Revision: D3306330
fbshipit-source-id: a6d8afd060b54d56f9e2b97f45b642d7b7f46209
Summary:
We want to give people the ability to log scroll performance (including Fb).
This adds an interface that can be enabled and disabled from the react scroll views.
This is a prerequisite to implementing the actual framerate logger that will log dropped
frames while scrolling in prod.
Reviewed By: astreet
Differential Revision: D3283588
fbshipit-source-id: ed9736cb9ed3f441511647f36b1460092bd91e56
Summary:
More instrumentation tests in OSS means less work for FB engineers to investigate if a PR breaks some internal tests.
+ increased timeouts and retries for OSS tests runner
Reviewed By: andreicoman11
Differential Revision: D3292582
fbshipit-source-id: 3f8aa4d3536450ea3af7acff044b9bb62be0f9db
Summary: Use the more accurate timestamp that we have computed for the touch event rather than the event timestamp that Android provides.
Reviewed By: andreicoman11
Differential Revision: D3292705
fbshipit-source-id: dad082ab74406d391481d16cdac19629751aa1eb
Summary:
I ran into problems when using Clipboard.getString() in Android. The method seems to work fine when there is *something* in the clipboard, but when it's empty the app crashes. I think I've tracked down the bug to ClipboardModule.java@58.
When clipData is null, the promise is resolved with an empty string. However, after that the rest of the block is executed. There should be a return or an 'else if' like in my pr.
Screenshot of the error message when clipboard is empty:
Closes https://github.com/facebook/react-native/pull/7527
Differential Revision: D3292232
fbshipit-source-id: d2191286c49ee31233203fab4648449964b9d950
Summary:
There was a race condition with `SwipeRefreshLayout` that cause the `RefreshControl` to keep refreshing when it shouldn't.
It was caused because we have to use `post` to set the refreshing state otherwise it doesn't work when setting `refreshing=true` on initial mount. What happened is that `post` doesn't guarantee the order the runnables will be called so calling post with `refreshing=true` followed by `refreshing=false` caused the `resfreshing=false` runnable to be called before the `resfreshing=true` one. This made it stay in refreshing state when it should not.
```
D/test ( 6171): setRefreshing true
W/ReactNativeJS( 6171): false
D/test ( 6171): setRefreshing false
D/test ( 6171): setRefreshing post false
D/test ( 6171): setRefreshing post true
```
This change adds an instance variable and uses it in the `post` runnable to make sure the last set value is always used.
**Test plan (required)**
Tested that it fixed the issue in the [original issue app](https://github.com/digisqu
Closes https://github.com/facebook/react-native/pull/7317
Differential Revision: D3290464
Pulled By: andreicoman11
fbshipit-source-id: 15cabcfc6d2f191443be96e8845b924ce66c369f
Summary:
This is a nice feature to have.
I've tested this by copying and renaming the ViewPager java and javascript files from the react-native repo and including them in a project. Whats the best way to test this directly from the repo?
Closes https://github.com/facebook/react-native/pull/5968
Differential Revision: D3240651
Pulled By: mkonicek
fbshipit-source-id: 5f1d157216df4f3314915496188a92aec1b85e91
Summary: This is called on the JS thread so we don't need to post a runnable.
Reviewed By: lexs
Differential Revision: D3281388
fbshipit-source-id: ff292a710442c0b84c6ee31a2644a5c2ea650dc9
Summary:
Older Android devices don't have a large stack by default and with about 16 levels of views the app will crash. Modern Android devices are fine so we can continue to use the system defaults while using a 2MB stack on older Android devices.
Closes https://github.com/facebook/react-native/pull/7416
Differential Revision: D3276805
Pulled By: astreet
fbshipit-source-id: a7b31a1be62e13f333148ca0908eb01a11aa4562
Summary:
Currently the Modal component on Android is rendered below the Status Bar, which changes it's color to grey, and in the UIExplorer example the backdrop is just formatted to look the same color. In some scenarios users may want to preserve the color of their status bar and make it look as though the modal is appearing on top. This PR allows for that.
This GIF shows current behavior and new behavior with the translucentStatusBar prop set to true.
I've updated the UIExplorer app to demonstrate and the docs as shown below
Thanks!
Closes https://github.com/facebook/react-native/pull/7157
Differential Revision: D3264497
Pulled By: dmmiller
fb-gh-sync-id: 61346d99414d331d3420f44a4c5f6341b0973be6
fbshipit-source-id: 61346d99414d331d3420f44a4c5f6341b0973be6
Summary:
We previously were unmounting the react application unconditionally when the ReactRootView#onDetachedFromWindow. This is nice in that it automatically allows us to reclaim memory, but there are many scenarios where a ReactRootView can be embedded in another piece of UI that detaches its children as part of its normal function (e.g. ListView, RecyclerView, ViewPager, etc).
As such, we will now enforce that the hosting Activity/Fragment/??? explicitly calls unmountReactApplication in the same way it calls startReactApplication. For Applications extending ReactActivity/AbstractReactActivity, this will happen automatically in onDestroy.
Reviewed By: foghina
Differential Revision: D3265161
fb-gh-sync-id: 4d49b0c41256213f00874f57e784aa8741dbf394
fbshipit-source-id: 4d49b0c41256213f00874f57e784aa8741dbf394
Summary:
Add the possibility to define a progress view top offset to RefreshControl on android. As i comment in #6740, contentInset does the trick on IOS.
Looking android documentation seems that exists a possible solution:
http://developer.android.com/reference/android/support/v4/widget/SwipeRefreshLayout.html#setProgressViewOffset(boolean, int, int)
This pull request implement that but keeping it simple, only a top offset.
For example, now we could put navigation bar over the scrollview (or listview) and define a progressViewOffset on RefreshView in order to start behind the navigation. At this point we could make some kind of coordinator layout to hide/show navigation on scroll.
To maintain the default behavior, start point is equal to start point minus progress circle diameter in order to create that progress circle before the start point.
Closes https://github.com/facebook/react-native/pull/6759
Differential Revision: D3240664
Pulled By: mkonicek
fb-gh-sync-id: ccf866272e871811c1c6dcc2a34f5c217967feee
fbshipit-source-id: ccf866272e871811c1c6dcc2a34f5c217967feee
Summary: Fixes a race condition where JS module functions could be called in between ##initializeWithInstance(catalystInstance);## and ##CatalystInstance#runJSBundle##, before the BatchedBridge in JS was set up. We now guarantee that all JS module methods that are called after `ReactContext#hasActiveCatalystInstance()` returns true will have the batched bridge created and ready to use.
Reviewed By: lexs
Differential Revision: D3258651
fb-gh-sync-id: 66e66533cc86d185e7c865376d6a5cdc6520d2d4
fbshipit-source-id: 66e66533cc86d185e7c865376d6a5cdc6520d2d4
Summary:
I need to do some refactoring of what happens when a ReactRootView is attached/detached from the window. In doing so, I found ReactRootView is way more complicated than it needs to be and very hard to understand. This simplifies it while maintaining functionality.
Basically:
- When you start a react application, if the root view is measured, we attach it to the instance.
- If it wasn't measured, we do nothing
- In onMeasure, if we have a ReactInstanceManager, but are not attached to it, attach to it.
None of this needs to depend on whether we're attached to the window or whether an attach is scheduled afaict.
Reviewed By: dmmiller
Differential Revision: D3260056
fb-gh-sync-id: 974c19ed75a07f83299472f75346bd6be7414128
fbshipit-source-id: 974c19ed75a07f83299472f75346bd6be7414128
Summary:
This adds support for pagingEnabled to the HorizontalScrollView.
This is an initial implementation.
Because Android doesn't provide great details about what is happening with a scroll view after you are done touching it, we have some post touch handling. This is kicked off either by touch up or a fling call.
Once we are doing that handling, we start a runnable that basically checks if we are still scrolling. If we are, we just schedule that runnable again and check a frame later. If we are done scrolling (no onScrollChanged since we last fired), we could be in one of two states, the fling is done or we are done snapping to the page boundary. If we are in the fling done case, we then check if we need to scroll to a page boundary. If so, we call smoothScrollTo and schedule ourself to check onScroll events again until done with that scroll. If we are done with both (either we only did momentum scroll or we did that and then snapped to page), we can then fire the final event and stop checking. This logic is all in handlePostTouchScrolling.
Because of the decision to only do page scrolling after momentum ends, we do allow you to scroll through with momentum a number of pages and the transition can be a little strange where it stops a sec and then slides to be page aligned. As a follow up, we can probably smooth that up by changing the value we pass to super.fling() that would adjust it to be let momentum carry it to the page boundary.
Reviewed By: weicool
Differential Revision: D3207608
fb-gh-sync-id: 02f62970ed9a5e3a5f9c0d959402756bc4b3699e
fbshipit-source-id: 02f62970ed9a5e3a5f9c0d959402756bc4b3699e
Summary:
This change makes all instrumentation tests to be executed in sequence in independent retriable processes.
With a new test being open sourced recently our CI stability degraded.
This PR should bring back stability because tests won't affect each other and will have shorter lifetime
Closes https://github.com/facebook/react-native/pull/7353
Differential Revision: D3259081
fb-gh-sync-id: 48ccdb5dbd561d416526497ff474378db9ca3c60
fbshipit-source-id: 48ccdb5dbd561d416526497ff474378db9ca3c60
Summary:
We have a crash in prod t11090540, urgent fix same way as in setFontStyle
Closes https://github.com/facebook/react-native/pull/7362
Reviewed By: foghina
Differential Revision: D3254158
Pulled By: dmmiller
fb-gh-sync-id: a497d0d4d0e4df9284e328ebb14fee74ba2f899b
fbshipit-source-id: a497d0d4d0e4df9284e328ebb14fee74ba2f899b
Summary:
transformMatrix only worked on iOS and there is an equivalent API that (mostly)
works cross platform.
decomposedMatrix could technically be passed on Android but it wasn't document and explicitly flagged as not working.
My goal is to deprecate both uses and then the only supported API is the `transform: [{ matrix: ... }]` form.
The only difference is that on Android the matrix gets decomposed.
Currently there is some special cased magic that renames transform -> transformMatrix or decomposedMatrix depending on platform.
https://github.com/facebook/react/blob/master/src/renderers/native/ReactNative/ReactNativeAttributePayload.js#L50
Therefore I'm adding an alias for both native platforms called just "transform".
Next I'll swap over the JS to always target the name "transform". The only difference is how the value is marshalled over the bridge in processTransform.
To do this, I have to clean up a few callers. Mostly that's just swapping to the new API.
For buildInterpolator this is a bit trickier but this fixes it for all our use cases (which is only the Navigator in AdsManager).
Reviewed By: vjeux
Differential Revision: D3239960
fb-gh-sync-id: 838edb6644c6cdd0716834f712042f226ff3136f
fbshipit-source-id: 838edb6644c6cdd0716834f712042f226ff3136f
Summary: This adds support for UIManager.setChildren on Android like D2757388 added for iOS.
Reviewed By: andreicoman11
Differential Revision: D3235369
fb-gh-sync-id: b538556ec4abdb606f9be26d1b74734046bca0cd
fbshipit-source-id: b538556ec4abdb606f9be26d1b74734046bca0cd
Yu Zheng(Sam)
Janic Duplessis
Martin Konicek
Dave Miller
Nathan Spaun
Olivier Notteghem
Konstantin Raev
Michał Gregorczyk
Sakari Tuominen
Kevin Johnson
Alexander Blom
Andy Street
Nikhilesh Sigatapu
Jesse Sessler
UnoDeTantos
Sokovikov
Russell Porter
Jimmy Mayoukou
Srikanth K Hari
Digeff
Charles Dick
Igor Avramovic
Sebastian Markbage
Ronak Patel
