Summary:
Adding `xcscheme` files for iOS template back in. They were removed in https://github.com/facebook/react-native/issues/26471 after https://github.com/facebook/react-native/issues/25451. We have reverted the change on `.gitignore`, as this directory is supposed to be in git, but we didn't revert the removed files. Now they are back as well.
You can also see here that the two files are still missing, they got removed on rc.0 and are still missing on rc.3.
https://react-native-community.github.io/upgrade-helper/?from=0.61.5&to=0.62.0-rc.3
## Changelog
[iOS] [Changed] - Add `xcscheme` files for iOS template back in.
Pull Request resolved: https://github.com/facebook/react-native/pull/28198
Test Plan: `init` a nre project, and check if the files are there or not. The paths are `ios/MyApp.xcodeproj/xcshareddata/xcschemes/MyApp.xcscheme` and `ios/MyApp.xcodeproj/xcshareddata/xcschemes/MyApp-tvOS.xcscheme`
Reviewed By: cpojer
Differential Revision: D20179447
Pulled By: TheSavior
fbshipit-source-id: b0f08c0f32d6bb7630179bc2fe46d9ac10f6c1d7
Summary:
This diff implements serialization and deserialization of text attachments as part of the calculation of layout of text components
changelog: [internal]
Reviewed By: JoshuaGross
Differential Revision: D20087251
fbshipit-source-id: dbcbd22f856aadace14343205548154ea80c8464
Summary:
ez diff to remove TODO to support inlinew Views in Fabric + venice
changelog: [internal]
Reviewed By: JoshuaGross
Differential Revision: D20088698
fbshipit-source-id: bd7cbbd76a8c13edf94d4eecac0dd670ccd0d209
Summary:
We are currently unintentionally including both libhermes-executor-release.so
and libhermes-executor-debug.so in all OSS RN builds.
RN tries both in turn, but since they both exist and the release executor is compatible
with the debug build, we always get the release executor without debug functionality.
While we sort this out, switch the load order. Since the debug executor is not compatible
with the release build, so it'll fail to load and try the next one.
ChangeLog: [Android] Fix Hermes debugger being disabled by default
Reviewed By: mhorowitz
Differential Revision: D20163828
fbshipit-source-id: ee4d87f40e42a7c8eedfdb7e1fc17eb3e5966ba5
Summary:
Changelog: [internal]
# Problem
We have node A with N children.
Calling cloning constructor on `YogaLayoutableShadowNode` causes new `yogaNode_` to be created.
However if `fragment.children` is nil in cloning constructor, which basically says children were not changed, then the existing children's `owner` becomes invalid.
# Solution
In the clone constructor, always call `updateYogaChildren`. This updates yoga children with the new `owner`.
Reviewed By: JoshuaGross
Differential Revision: D20139582
fbshipit-source-id: 3932694d4381b601df07dd8a57887ce7c09f1582
Summary:
The sizes of attachments are legit parts of `AttributedString`s identity, so we need to incorporate it into the equality test.
We need that to make measurement cache work correctly with inline views.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: mdvacca
Differential Revision: D20151505
fbshipit-source-id: ae47c24a753eab1e497fcfaa93de557ffb30d874
Summary:
This is a redo of D16969764, with a few extensions.
## Changes
1. Move `RCTDevLoadingView.{h,m}` to `CoreModuels/RCTDevLoadingView.{h,mm}`
2. Extract ObjC API of `RCTDevLodingView` into `RCTDevLoadingViewProtocol` in `ReactInternal`.
3. Create API `RCTDevLoadingViewSetEnabled.h` in `ReactInternal` to enable/disable `RCTDevLoadingView`
Changelog:
[iOS][Added] - Make RCTDevLoadingView TurboModule-compatible
Reviewed By: PeteTheHeat
Differential Revision: D18642554
fbshipit-source-id: 6b62e27e128d98254b7a6d018399ec1c06e274fc
Summary:
Make the Android app bundle size report more legible, and update existing reports if possible.
## Changelog
[Internal] [Changed] - Update existing app bundle size report if possible
[Internal] [Changed] - Generate table for APK sizes report
Pull Request resolved: https://github.com/facebook/react-native/pull/28169
Test Plan:
- App bundle size reports should be updated when possible
- Android apk sizes should be reported using a table
Reviewed By: cpojer
Differential Revision: D20141490
Pulled By: hramos
fbshipit-source-id: 3fc1b808b27e893c0a3e4e34eafc39094448799c
Summary:
`nativeId` type is `NSString`, not `NSNumber`. The `.h` file already has the proper type but `.mm` had a wrong one.
Changelog: [Internal]
Reviewed By: sammy-SC
Differential Revision: D20135465
fbshipit-source-id: e5f9fbf4940d3883334c910f96f0aa850a069d8c
Summary:
This is step 3 in creating a new DevSupportManager with a better async API for loading the JS bundle from Metro.
Right now DevSupportManagerImpl calls a method on its ReactInstanceManagerDevHelper delegate when the JS bundle is done being downloaded from the server. However, I want to move to a callback-based method for bridgeless mode (similar to packager status). In this diff, I'm adding a protected method that uses a callback, and then calling that from `reloadJSFromServer` with a callback that calls the delegate, to preserve existing behavior.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D19871102
fbshipit-source-id: 2dbb6b91c5b927df86c3db42aa11e080d57ea78e
Summary:
This is the second step in creating a new DevSupportManager impl with more async-friendly APIs for bundle loading, which I'll use in bridgeless mode. This diff creates a new interface, AsyncDevSupportManager, and two implementations, mirroring what we currently have.
This new interface doesn't define any methods yet.
Changelog: [Android][Changed] Specify DevSupportManager.setPackagerLocationCustomizer in DevSupportManager impls
Reviewed By: mdvacca
Differential Revision: D19870812
fbshipit-source-id: 042a43e1a3055aba6d7325f948060300b5bf17f3
Summary:
This is the first step for adding my own async APIs for DevSupportManager to use in bridgeless mode.
To avoid having to add these APIs to the DevSupportManager interface + both of its implementations, I'm instead splitting DevSupportManager into a base class + final impl so I can more easily share logic with the new impl I'll create.
Changelog: [Internal]
Reviewed By: RSNara
Differential Revision: D19846619
fbshipit-source-id: c67b0231f4e6581fb484f9a4899ed0b7f08fb684
Summary:
We will soon be enforcing that flow suppressions will only apply when on an error's primary location ([post](https://fb.workplace.com/groups/179614562644215/permalink/559286354677032/)). This diff uses the codemod created in D20008770 to move all suppression comments to their primary locations in the `xplat/js` flow root, and deletes suppression comments that are not on any primary locations.
This diff was generated with:
```
~/fbsource/fbcode/flow/packages/flow-dev-tools/bin/tool suppression-primary-locations --json-file ~/www/errors.json ~/fbsource/xplat/js
hg st -n | xargs grep -l -P '@(partially-)?generated' | xargs hg revert
hg st -n | xargs grep -l 'format' | xargs prettier --write
```
Changelog: [Internal]
bypass-lint
Reviewed By: dsainati1
Differential Revision: D20122544
fbshipit-source-id: d94e409aadb18bb399a1ddbf9f3f2494fe4fb54c
Summary:
Looks like there could be cases when the NativeHirearchyManager may be asking if it should animate on a view that has been removed. Adding a null check
Changelog:
[General] [Fixed] - Check null values in shouldAnimate
Reviewed By: makovkastar
Differential Revision: D20063054
fbshipit-source-id: 5b3b1c27b9aba57a592bd8d4e27a970cf0912b5d
Summary:
Run clang-format and add .clang-tidy with `clang-diagnostic-*` to several more directories in order to catch any problems.
Changelog:
[Internal]
Reviewed By: shergin
Differential Revision: D19860169
fbshipit-source-id: 7785aab010c8e6945cc6b5c9b68cb8ee0cdbb7fa
Summary:
The PlatformConstants native module exposes the ability to query the React
Native version used to build native code. This is managed on iOS and Android by
a version bumping script, which replaces module code based on a template.
It is currently difficult to accurately determine this version for out-of-tree C++
platforms (I.e. React Native Windows). The version of upstream react-native we resolve to is ultimately
dependent on the version of react-native chosen a peer dependency, which is not
neccesarily constant given a build of react-native-windows.
We could try to hack around this, and make our native build try to reason about
the resolved pacakge for react-native using a lockfile, but a much cleaner
solution is to embed version into C++ code, similar to what is done for Android
and iOS. This change does that, adding a header with React Native version and
updating the build stamping script to write to it.
Usage sample:
```c++
constants["reactNativeVersion"] = folly::dynamic::object();
constants["reactNativeVersion"]["major"] = ReactNativeVersion.Major;
constants["reactNativeVersion"]["minor"] = ReactNativeVersion.Minor;
constants["reactNativeVersion"]["patch"] = ReactNativeVersion.Patch;
```
## Changelog
[General] [Added] - Stamp React Native Version Into C++ Code
Pull Request resolved: https://github.com/facebook/react-native/pull/28036
Test Plan: Validated that the bumping script will accurately update the header, can compile under both MSVC and Clang.
Differential Revision: D19865992
Pulled By: hramos
fbshipit-source-id: 9e0b8e9519015bb62c60b9935a234cd367a1926a
Summary:
Changelog: [internal]
# Problem
Implementation of `matricesDiffer` was ignoring first element in the matrix, therefore two matrixes that were different were actually evaluated as same and React didn't pass value to native.
# Solution
Take first element in the matrix into account when comparing matrices.
Reviewed By: cpojer
Differential Revision: D20117210
fbshipit-source-id: 84c3b4e580da44bda4fc8bd8669318282ae9aa32
Summary:
We were already doing this for ViewCommands sent with integer IDs; we should do the same for now with String commands. Otherwise, screens with TextInputs are unusable during development because these exceptions are thrown very often around reload and navigation.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D20127446
fbshipit-source-id: a5d65ff6ae5b520fd0efffa5c325b5cca3bd53a0
Summary:
Cloning subtrees is not something specific to a RootNode, so it makes sense to have it in ShadowNode. Soon we will use that to clone subtrees inside Paragraph component to implement Inline Views.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: mdvacca
Differential Revision: D20090666
fbshipit-source-id: 0a64ef9bda438cd55d5fd21d3ad83b36221fa89e
Summary:
All logic that is performed on the root node only was moved from `layout` to a separate methods `layoutTree`. That makes the code simpler and allows reusing `layoutTree` in InlineViews feature.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: mdvacca
Differential Revision: D20052025
fbshipit-source-id: 3070a1cca4e322c6d077ede751ea80359c96a600
Summary:
Similar to a previous diff but for `setChildren`.
`YogaLayoutableShadowNode::setChildren()` was renamed to `YogaLayoutableShadowNode::updateYogaChildren()`. Now we don't need to pass an argument to this function because the object is already initialized. The new name also disambiguates this method with `getChildren()` from `ShadowNode` (which does something completely different). The rest of the changes is just type adjustments.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: mdvacca
Differential Revision: D20052031
fbshipit-source-id: 6157cad9b55d4cdd97ce04e1278ac1369bfb96bc
Summary:
`YogaLayoutableShadowNode::setProps()` was renamed to `YogaLayoutableShadowNode::updateYogaProps()`. Now we don't need to pass an argument to this function because the object is already initialized. The new name also disambiguates this method with `getProp()` from `ShadowNode` (which does something completely different).
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D20052018
fbshipit-source-id: 4d9651bb4452408708fa1b36bb5c5c8639f7d1bb
Summary:
How we can finally make `ViewProps` inherit `YogaStylableProps` and remove `dynamic_cast`. This overall change is aligned with D19963353.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D20052026
fbshipit-source-id: 1066f9a259b50d4672111c9c03a71c2590d59aca
Summary:
Here we delete several constructors that were needed only because `YogaLayoutableShadowNode::yogaStyle` was private (which is not true anymore).
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC, mdvacca
Differential Revision: D20052033
fbshipit-source-id: c2afcaf2fdebb65cdcde6c214c8b78aa35d45e91
Summary:
Here we refine the `ShadowNode::BaseTraits` static method. Before this change, ConcreteViewShadowNode<> enforces Layoutable and YogaLayoutable traits. This change moves that responsibility to LayoutableShadowNode and YogaLayoutableShadowNode which makes overall logic more coherent.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D20052027
fbshipit-source-id: fd25264204b0232b1dbbff6f9dfefd9fbcb146c4
Summary:
Now we can move custom to `YogaLayoutableShadowNode` code from `ConcreteViewShadowNode<>` template to `YogaLayoutableShadowNode` itself reducing the amount of templated code and reduce overall complexity.
Note, here we have to do `dynamic_cast` for now, we will address that in coming diffs.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: mdvacca
Differential Revision: D20052021
fbshipit-source-id: dac5969a97b75e54c7728a1ca8161922bd2245ca
Summary:
The fact that `LayoutableShadowNode` now inherits `ShadowNode` allows us to de-virtualize `getLayoutableChildNodes` and move that to `LayoutableShadowNode`. Less code, less virtual dispatch, less complexity.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D20052032
fbshipit-source-id: 580e86b5a746028e470788e00027f247bf77126c
Summary:
D19963353 mentioned the infrastructure that re-routes methods calls related to adding and cloning children between YogaLayoutableShadowNode and ShadowNode. `cloneAndReplaceChild` is exactly this. It was implemented as a virtual method that is called from `ConcreteViewShadowNode`. The whole process requires building a list of children of some class and passing that as a list of pointers. Now we don't need it all that because we can call directly and statically. That change will allow us to simplify that infra even more in the future diffs.
With all previous changes, now we can implement `getYogaLayoutableChildren` inside `YogaLayoutableShadowNode` and call that statically. Eventually, that will allow us to remove templated `getChildrenSlice`. Previously the call for that method must be in `ConcreteViewShadowNode`, now it's not true anymore which we will use later to even better goodness.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: mdvacca
Differential Revision: D20052020
fbshipit-source-id: e5c819a4d21b2dbcd08f3439e1783e3a9cba5ef4
Summary:
D19963353 mentioned the infrastructure that re-routes methods calls related to adding and cloning children between YogaLayoutableShadowNode and ShadowNode. `cloneAndReplaceChild` is exactly this. It was implemented as a virtual method that is called from `ConcreteViewShadowNode`. The whole process requires building a list of children of some class and passing that as a list of pointers. Now we don't need it all that because we can call directly and statically. That change will allow us to simplify that infra even more in the future diffs.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D20052022
fbshipit-source-id: ddf341c112edd8a2f79eaf74465a9a360a168541
Summary:
This is another perf-critical place where using traitCast should help with perf.
After this change we practically will not have any `dynamic_cast`s (except those which will be delited (with methods where they are used) in the future diffs).
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D20052029
fbshipit-source-id: 63b589d3c3322f0afd72f18b4701fe31d0e05c2f
Summary:
We use this function in Diffing, and Diffing is one of the hottest pieces of Fabric. Removing `dynamic_cast` here should improve perf. See the previous diff for more details.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D20052023
fbshipit-source-id: d6427ebf076d7f8244f0aeaf16407efa2368957e
Summary:
The Diffing is one of the hottest pieces of Fabric. Removing dynamic_cast here should improve perf. See the previous diff for more details.
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D20052030
fbshipit-source-id: 27fd9f34f2a1f9d22b9da6b1e3c1a2982045c07a
Summary:
`traitCast` is a special form of static_cast that checks additional requirements (similar to `dynamic_cast`) before performing the cast. We will use that in many places in the coming diffs.
Restructuring the class hierarchy in the previous diff finally allows us to do static casts among ShadowNode and LayoutableShadowNode and other classes (previously it wasn't allowed because of lack of common base class).
Why we don't want to use `dynamic_cast`:
* It's expensive (and we need to do the cast on the hottest fragments of the framework). (See: (1) http://www.stroustrup.com/fast_dynamic_casting.pdf and (2) https://www.youtube.com/watch?v=ARYP83yNAWk Herb Sutter & proposal of `down_cast`).
* It's code-size inefficient, whereas `static_cast` has zero runtime and code-size overhead.
* Removing `dynamic_cast` will allow us finally to opt-out `RTTI` (additional code size and perm wins).
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D20052024
fbshipit-source-id: d293c1cf80deb7817d333d5306d6b32bf3abdb27
Summary:
This is a very crucial change, everything else in this stack depends on it.
Here is the set of constraints that we have for ShadowNode-based class hierarchy:
* `ShadowNode` is a base class that defines basic operations on the nodes. It does not have virtual methods by design (virtual dispatch hurts performance; we want to limit and centralize that in `ComponentDescriptor`s).
* `ConcreteShadowNode<>` template *statically* wires particular `ShadowNode` and particular `Props` establishing a type-safe relationship between them ensured on compile time.
* Not all `ShadowNode`s are "layoutable", not all "layoutable" `ShadowNode`s use Yoga to do layout.
* Layotability (and YogaLayotability) feature is implemented as two classes `LayoutableShadowNode` and `YogaLayoutableShadowNode`. These classes essentially need to know something about the ShadowNode nature of the object (get the list of children or replace some children).
Before the change, `LayoutableShadowNode` and `YogaLayoutableShadowNode`classes did not inherit `ShadowNode` because we don't want to use *virtual inheritance* for `ShadowNode`: `ConcreteShadowNode<>` already inherits `ShadowNode`, so if we make `LayoutableShadowNode` inherits `ShadowNode`, we will have to somehow flatten this inheritance hierarchy (and the virtual inheritance is an answer to that). (Yes, C++ supports multiple inheritance and virtual inheritance.)
Before this change, we solved this dilemma this way: We have a subclass-template `ConcreteViewShadowNode<>` that inherits `ConcreteShadowNode<>` and `YogaLayoutableShadowNode`. Then, we had a bunch of methods that implement the rerouting of some functionality from `YogaLayoutableShadowNode` to `ShadowNode` and vise-versa. (See the diagram "Before".)
That worked fine, except the caveats:
* That wiring is nasty, complex, hard to reason about and overall limiting.
* There is no way to statically cast `LayoutableShadowNode` to `ShadowNode` (because there is no common base class). That forces us to use dynamic_cast on some perf critical paths (including layout, diffing and so on).
* Adding features that rely on interop between `LayoutableShadowNode` and `ShadowNode` is a nightmare.
It should be a better way to deal with this dilemma, and this diff implements a different approach: We can have a base class of `ConcreteShadowNode<>` as a template parameter. With this approach, we can make `LayoutableShadowNode` inherit `ShadowNode` and when we need to instantiate a `ConcreteShadowNode<>` that needs to be layoutable, we can just specify `YogaLayoutableShadowNode` as a base class. (See the diagram "After".)
This simple change will allow us to simplify a lot of things. The rest of the stack is about getting rid of unnecessary moving parts. Which will finally allow us to build "Inline Views" feature.
```
╭──────────────────────╮
│ ◎ ○ ○ ░░░░░░░░░░░░░░░│
├──────────────────────┤
│ │
│ │
│ Before │
│ │ ┌────────────────────────────┐ ┌────────────────────────────┐
│ │ │ │ │ │
│ │ │ ShadowNode │ │ LayoutableShadowNode │
└──────────────────────┘ │ │ │ │
└────────────────────────────┘ └────────────────────────────┘
▲ ▲
│ │
╔════════════════════════════╗ ┌────────────────────────────┐
║ ║ │ │
┌────────�║ ConcreteShadowNode<> ║ │ YogaLayoutableShadowNode │
│ ║ ║ │ │
│ ╚════════════════════════════╝ └────────────────────────────┘
│ ▲ ▲
│ │ │
│ │ │
│ │ │
│ │ ╔════════════════════════════╗ │
│ │ ║ ║ │
│ └─────║ ConcreteViewShadowNode<> ║─┘
│ ║ ║
│ ╚════════════════════════════╝
│ ▲
│ │
│ ┌──────────────┴───────────────┐
│ │ │
│ │ │
│ │ │
┌────────────────────────────┐ ┌────────────────────────────┐ ┌────────────────────────────┐
│ │ │ │ │ │
│ TextShadowNode │ │ ViewShadowNode │ │ ParagraphShadowNode │
│ │ │ │ │ │
└────────────────────────────┘ └────────────────────────────┘ └────────────────────────────┘
╭──────────────────────╮
│ ◎ ○ ○ ░░░░░░░░░░░░░░░│
├──────────────────────┤
│ │ ┌────────────────────────────┐
│ │ │ │
│ After │ │ ShadowNode │
│ │ │ │
│ │ └────────────────────────────┘
│ │ ▲
└──────────────────────┘ ┌────────────────────────────────────┤
│ │
│ ╔════════════════════════════╗
┌────────────────────────────┐ ║ ConcreteShadowNode ║
│ │ ║ <Base: ShadowNode> ║
│ LayoutableShadowNode │ ║ ║
│ │ ╚════════════════════════════╝
└────────────────────────────┘ ▲
▲ │
│ ┌────────────────────────────┐
┌────────────────────────────┐ │ │
│ │ │ TextShadowNode │
│ YogaLayoutableShadowNode │ │ │
│ │ └────────────────────────────┘
└────────────────────────────┘
▲
│
╔════════════════════════════╗
║ ConcreteShadowNode ║
║ <Base: ║
║ YogaLayoutableShadowNode> ║
╚════════════════════════════╝
▲
│
╔════════════════════════════╗
║ ║
║ ConcreteViewShadowNode<> ║
║ ║
╚════════════════════════════╝
▲
├─────────────────────────────────────┐
│ │
┌────────────────────────────┐ ┌────────────────────────────┐
│ │ │ │
│ ParagraphShadowNode │ │ ViewShadowNode │
│ │ │ │
└────────────────────────────┘ └────────────────────────────┘
```
Changelog: [Internal] Fabric-specific internal change.
Reviewed By: sammy-SC
Differential Revision: D19963353
fbshipit-source-id: b65c8a5064bdb54ab64f08a8e546aa9e2b5a486b
Summary:
We are moving towards 100%-prettified files. That's another step when we apply Clang Format for `React/Fabric`.
Changelog: [Internal]
Reviewed By: mdvacca
Differential Revision: D20112417
fbshipit-source-id: 020d818e5cb8e2f509c5276738c8f252f6817a56
Summary:
Note: I'm currently testing this, so do not review just yet.
Adding a nightly CI build in preparation for further work on [React Native: Benchmark Suite](https://github.com/react-native-community/discussions-and-proposals/issues/186.)
Currently it will only run the job `nightly_job` which will simply run `$ echo "Nightly build run"`. The plan is to do nightly checks and performance status on the React Native repo, and create a dashboard where we can see performance changes over time. More on this in the proposal: [React Native: Benchmark Suite](https://github.com/react-native-community/discussions-and-proposals/issues/186.)
## Changelog
[General] [Added] - Add nightly CI build
Pull Request resolved: https://github.com/facebook/react-native/pull/28066
Test Plan: Individual jobs will have their own test plan.
Reviewed By: sebmck
Differential Revision: D20008434
Pulled By: hramos
fbshipit-source-id: a5e8aad616c28bfad8f3455b5ebadf7a7d174a55
Summary:
For backwards-compatibility with Paper, we're implementing a feature in Fabric that will allow TextInputs to use the default padding from the theme in Android.
Note that this uses some pretty ugly hacks that probably shouldn't be used inside of components at all: looking directly at rawProps, overriding props/Yoga styles in the component descriptor, etc. I would (personally) really like to kill this feature entirely unless and until we can find a more elegant solution.
Changelog: [Internal]
TextInputs are still not pixel-perfect with Paper, but they're much closer, and the underline visual glitchiness is no longer an issue.
Reviewed By: mdvacca
Differential Revision: D20109605
fbshipit-source-id: 543282843e0a9f03a504d72d7a014431099bd64c
