Summary: Adds a pool to recycle CSSNodes within UIManager. A follow-up diff will hook this up to a memory pressure listener to drop the pool on memory pressure.
Reviewed By: emilsjolander
Differential Revision: D4189532
fbshipit-source-id: 46583546f71a8c59853e1dd124de31657b3c617b
Summary:
Moves from CSSNodeDEPRECATED to CSSNode. This has shown to be a huge performance win for layout time within FB.
This is BREAKING because CSSNode contains bug fixes that were not migrated to CSSNodeDEPRECATED which may change the way your layout appears. The most common of these by far involves `flex: 1`.
Previously, developers had to put `flex: 1` in many places it didn't belong in order to work around a bug in css-layout. Now `flex: 1` is treated properly and, unfortunately, this means that your layout may no longer look correct. Specifically, you may see that your layout looks collapsed, or children don't render. The fix is to simply remove `flex: 1` from those containers.
Reviewed By: emilsjolander
Differential Revision: D3992787
fbshipit-source-id: 7a3a2a34a8941c0524e6ba3c5379e434d3e03247
Summary: The java enum was recently changed from NOWRAP -> NO_WRAP so the translation from js failed. This fixes that.
Reviewed By: limichaelc
Differential Revision: D4186869
fbshipit-source-id: fe35211a6632d80356d35a01a079279ef4bd7006
Summary:
This diff makes it so ReactShadowNode holds a CSSNode instead of extending one. This will enable us to pool and re-use CSSNodes and will allow us to keep from breaking the CSSNode api assumption that nodes that have measure functions don't have children (right now, text nodes have measure functions, but they also have raw text children).
BREAKING
This diff makes ReactShadowNode no longer extend CSSNodeDEPRECATED. If you have code that depended on that, e.g. via instanceof checks, that will no longer work as expected. Subclasses that override getChildAt/addChildAt/etc will need to update your method signatures. There should be no runtime behavior changes.
Reviewed By: emilsjolander
Differential Revision: D4153818
fbshipit-source-id: 2836434dd925d8e4651b9bb94b602c235e1e7665
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 way `UIImplementation` can hold on to it and use it outside of calls from the `UIManagerModule`.
Reviewed By: lexs
Differential Revision: D3899774
fbshipit-source-id: 01e4956c4540bcdf30774a3f40a625e934714ee9
Summary: Provide a base `HeadlessJsTaskService` class that can be extended to run JS in headless mode in response to some event. Added `HeadlessJsTaskEventListener` for modules that are interested in background lifecycle events, and `HeadlessJsTaskContext` that basically extends `ReactContext` without touching it. The react instance is shared with the rest of the app (e.g. activities) through the `ReactNativeHost`.
Reviewed By: astreet
Differential Revision: D3225753
fbshipit-source-id: 2c5e7679636f31e0e7842d8a67aeb95baf47c563
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 adds support for `Animated.event` driven natively. This is WIP and would like feedback on how this is implemented.
At the moment, it works by providing a mapping between a view tag, an event name, an event path and an animated value when a view has a prop with a `AnimatedEvent` object. Then we can hook into `EventDispatcher`, check for events that target our view + event name and update the animated value using the event path.
For now it works with the onScroll event but it should be generic enough to work with anything.
Closes https://github.com/facebook/react-native/pull/9253
Differential Revision: D3759844
Pulled By: foghina
fbshipit-source-id: 86989c705847955bd65e6cf5a7d572ec7ccd3eb4
Summary:
The original method getNativeProps in ViewManagerPropertyUpdater.java create more HashMaps and putAll method need to re-hash the key again to avoid conflicts. This pull request pass the map as params to avoid the problem and update ReactPropertyProcessor.java to adapt the change.
Closes https://github.com/facebook/react-native/pull/9916
Differential Revision: D3873152
fbshipit-source-id: 089840e5272265662cdbf58d88580f9203153b69
Summary:
This adds support for sticky headers on Android. The implementation if based primarily on the iOS one (https://github.com/facebook/react-native/blob/master/React/Views/RCTScrollView.m#L272) and adds some stuff that was missing to be able to handle z-index, view clipping, view hierarchy optimization and touch handling properly.
Some notable changes:
- Add `ChildDrawingOrderDelegate` interface to allow changing the `ViewGroup` drawing order using `ViewGroup#getChildDrawingOrder`. This is used to change the content view drawing order to make sure headers are drawn over the other cells. Right now I'm only reversing the drawing order as drawing only the header views last added a lot of complexity especially because of view clipping and I don't think it should cause issues.
- Add `collapsableChildren` prop that works like `collapsable` but applies to every child of the view. This is needed to be able to reference sticky headers by their indices otherwise some subviews can get optimized out and break indexes.
Closes https://github.com/facebook/react-native/pull/9456
Differential Revision: D3827366
Pulled By: fred2028
fbshipit-source-id: d346068734c5b987518794ab23e13914ed13b5c4
Summary:
This adds support for sticky headers on Android. The implementation if based primarily on the iOS one (https://github.com/facebook/react-native/blob/master/React/Views/RCTScrollView.m#L272) and adds some stuff that was missing to be able to handle z-index, view clipping, view hierarchy optimization and touch handling properly.
Some notable changes:
- Add `ChildDrawingOrderDelegate` interface to allow changing the `ViewGroup` drawing order using `ViewGroup#getChildDrawingOrder`. This is used to change the content view drawing order to make sure headers are drawn over the other cells. Right now I'm only reversing the drawing order as drawing only the header views last added a lot of complexity especially because of view clipping and I don't think it should cause issues.
- Add `collapsableChildren` prop that works like `collapsable` but applies to every child of the view. This is needed to be able to reference sticky headers by their indices otherwise some subviews can get optimized out and break indexes.
Closes https://github.com/facebook/react-native/pull/9456
Differential Revision: D3827366
fbshipit-source-id: cab044cfdbe2ccb98e1ecd3e02ed3ceaa253eb78
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: This is to be able to depend on ReactClippingViewGroup from BaseViewManager. Devs using ReactClippingViewGroup may need to update their imports when updating past this commit.
Reviewed By: lexs
Differential Revision: D3835328
fbshipit-source-id: 290c08b130d837e553b68a90377bd9a30b7ec6dc
Summary:
This automatically changes the size of the modal by listening to dialog size changes and propagating
those changes through UIManager.
In detail: I've looked into three ways of doing this:
1. Send `onSizeChanged` events/info from the View to the CSSNode directly. This is kinda hacky because you would need to hold a reference to the CSSNode somewhere, either in the View or in the ViewManager. But then you'll have to take care of the lifecycle of the CSSNode, so that you don't update it after it has been dismissed. Not great.
2. The version we went for, is to just update the size of the corresponding CSSNode in the same way we do it for root nodes: we inform the UIManager that the size of the root node has changed, and it will propagate that change, triggering a `dispatchViewUpdates` if none is underway, so that the layout is updated.
3. The other solution we thought of is to treat the Modal as a root view. This would mean rendering an application with the tag of the Modal as the root of the application. That tag would be received by calling some method into UIManager and ReactModalHostManager to create a new RootView, create a Dialog and plop the root view in it. The idea was to maintain the JS API that we now have, but make the implementation more correct (ie. since both RootView and the Modal must deal with touch handling), and could have other benefits (ie. no hacks necessary for making the inspector work on top of modals). However, the change is not trivial and I don't know just how much code would have to be changed to make this work correctly. We might revisit this at a later stage, after we've done more work on having several root views at the same time in the app.
Reviewed By: foghina
Differential Revision: D3841379
fbshipit-source-id: f5e363e27041b785cf44eb59da04bc789306ddb9
Summary:
Here's a little background. Resizing is inferior to scaling. See http://frescolib.org/docs/resizing-rotating.html#_
Currently, React Native has a heuristic to use resize when the image is likely to be from the device's camera. However, there may be other cases where a developer wants to use resize. For example, when the developer knows they'll be downloading a large image from a service but the image will be rendered at a small size on the device.
This change adds a `resizeMethod` prop to the `Image` component so developers can choose how Fresco resizes the image. The options are 'auto', 'resize', or 'scale'. When 'auto' is specified, a heuristic is used to choose between 'resize' and 'scale'. The default value is 'auto'.
**Test plan (required)**
In a small test app, verified that the `resizeMethod` prop properly influences the mechanism that is used to resize the image (e.g. resize or scale).
Adam Comella
Microsoft Corp.
Closes https://github.com/facebook/react-native/pull/9652
Differential Revision: D3841322
Pulled By: foghina
fbshipit-source-id: 6c78b5c75ea73053aa10386afd4cbff45f5b8ffe
Summary: It's called `timestamp` on iOS, making it consistent.
Reviewed By: foghina
Differential Revision: D3820937
fbshipit-source-id: 2805f1fc10d6445d8b31676e0e3dca348510ffe7
Summary:
This adds the `didUpdateDimensions` event that already exists on iOS, and
updates the internal native dimensions data that needs repopulation after
orientation change.
Reviewed By: foghina
Differential Revision: D3819364
fbshipit-source-id: fc2b3b4d1991d101e5de4439ccef2189bc65fd58
Summary:
Landing D3528215 again, now that D3593884 has landed and makes that easier.
Copy-paste summary from previous diff:
So PanReponder.onPanResponderRelease/onPanResponderTerminate receive a
gestureState object containing a onPanResponderTerminate.vx/vy property. On
Android and iOS, they appear to be orders of magnitude different, which appear
to be due to the different scale of timestamps that are used when generating
touch events.
This pull request fixes the timestamps to be milliseconds on both platforms
(since I assume iOS is the more authoritative one, and is the one that
react-native-viewpager's vx thresholds written written to compare against.)
As far as I can tell, the RN code doesn't use the vx/vy properties, so they
should be okay. And looks like the RN code only cares about relative values of
startTimestamp/currentTimestamp/previousTimestamp though, so should be fine too.
it's quite possible there will be downstream android breakage with this change,
particularly for those who are already compensating for the RN discrepancy.
Reviewed By: foghina
Differential Revision: D3819761
fbshipit-source-id: fd2d85748ae6a9cde6af715aabb620f340c2220c
Summary:
There are a couple of buggy behaviors in the current implementation of the `ellipsizeMode` prop on Android:
- Setting the `numberOfLines` prop stomps on whatever value you provided for `ellipsizeMode` earlier.
- The value you've provided for `ellipsizeMode` is used even if you've configured your view to have an unlimited size (i.e. `numberOfLines` is 0 or unspecified).
This change fixes these issues which makes Android's `ellipsizeMode` prop more consistent with iOS's. Additionally, it renames LineBreakMode to EllipsizeMode in a couple of places.
**Test plan (required)**
Verified that the `numberOfLines` and `ellipsizeMode` props work correctly in an Android test app.
Adam Comella
Microsoft Corp.
Closes https://github.com/facebook/react-native/pull/9594
Differential Revision: D3810166
Pulled By: foghina
fbshipit-source-id: 229c9bfc3ef10670a1090311ea9d095cb2c1121a
Summary:
`mRCTEventEmitter` is used by 2 different threads. It's assigned on the UI thread and it's accessed on the JavaScript thread. Currently, it can be the case that the UI thread assigns `mRCTEventEmitter` and later the JS thread accesses it but still sees null.
This change fixes the issue by marking the `mRCTEventEmitter` variable as `volatile` to ensure that both threads see the same value for `mRCTEventEmitter`.
**Test plan (required)**
This change is currently used in my team's app. We're no longer seeing a crash in `EventDispatcher`.
Adam Comella
Microsoft Corp.
Closes https://github.com/facebook/react-native/pull/9655
Differential Revision: D3790888
Pulled By: andreicoman11
fbshipit-source-id: 68cdbc74faffb36dc2bca8ad3d4a78929badbe9c
Summary: Added support for WebViews to take in an onContentSizeChange prop, which will return a native event that contains the width and height of the html content in the WebView. Also moved the ContentSizeChangeEvent from the recyclerview dir to the uimanager/events dir
Reviewed By: andreicoman11
Differential Revision: D3775399
fbshipit-source-id: 19a0579f8345e5853cc7311b80f1f1393c77ab58
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:
`ThemedReactContext` wraps the actual `ReactContext` but doesn't actually receive any lifecycle events, which would set `mCurrentActivity`, so that's always stuck as `null`. To fix, we override `has/getCurrentActivity` and forward the call to the wrapped context, which actually has the correct lifecycle status.
Fixes issue #9310 on github.
Reviewed By: mkonicek
Differential Revision: D3703005
fbshipit-source-id: 363e87ac91d50516899b413e823d5312cbb807f4
Summary:
This adds support for the `transform` animated node. This brings feature parity with the iOS implementation and allows running the NativeAnimated UIExplorer example that was created with the iOS implementation on Android. This is based on some work by kmagiera in the exponent RN fork.
This also adds support for mixing static values with animated ones in the same transform as well which is not supported on iOS at the moment. It is also implemented in a way that rebuilds the transform matrix the same way as we build it in JS so it will be easy to remove some of the current limitations like forcing the transforms order and only supporting one of each type.
**Test plan (required)**
Tested with the NativeAnimated example on Android and iOS. Also tested mixing in static values in a transform (`[{ rotate: '45deg' }, { translateX: animatedValue }]`).
Closes https://github.com/facebook/react-native/pull/8839
Differential Revision: D3682143
fbshipit-source-id: 5e6fd4b0b8be6a76053f24a36d1785771690a6f8
Summary:
In #7916 I moved transform matrix decomposition logic from JS to java. The next step is to accept list of transforms instead oftransform matrix as a transform ReactProp. This way there is no extra processing required on JS side for the transform param (at least for android now) and this on the other hand allow us to execute transform updates (through offloaded animation) solely on the UI thread.
After this change there is a whole bunch of stuff from `Libraries/Utilities/MatrixMath.js` that can be deleted (methods like: determinant, inverse, transpose). Although astreet mentioned under one of my previous commits that the code is still being referenced internally at fb, so I decided not to delete it here.
**Test plan (required)**
Run UIExplorer Transform example before and after - compare the results
Run android unit test: com.facebook.react.uimanager.MatrixMathHelperTest
Closes https://github.com/facebook/react-native/pull/8892
Differential Revision: D3676017
Pulled By: astreet
fbshipit-source-id: 5275e30805a85c12c89bea44e8b3a2b2ec7b33fa
Summary:
The make current RCTShadowView support RTL layout.
1 Change all left/right to start/end for margin, padding, boarder and position
2 Calculate position in the same way as margin, padding and boarder
Reviewed By: fkgozali
Differential Revision: D3619292
fbshipit-source-id: eaaa6faeee93c964d59bb6f498d89effc09ed567
Summary: Make UIOperation public so that custom implementations can expose instances of it.
Reviewed By: ahmedre
Differential Revision: D3618197
fbshipit-source-id: 89f764818cc2e8e94cf9678814e56e7074eb9067
Andy Street
Emil Sjolander
Howard Yeh
Aaron Chiu
Jacob Parker
Andrew Y. Chen
Felix Oghina
Janic Duplessis
ransj
Andrei Coman
Adam Comella
Andrei Coman
Scott Buckfelder
Kevin Gozali
Krzysztof Magiera
Mengjue Wang
Seth Kirby
