Summary:
X-link: https://github.com/facebook/react-native/pull/53133
# Changelog:
[Internal] -
As part of the ongoing effort to migrate the React Native codebase to Kotlin, this PR introduces the initial setup required for Kotlin support in Yoga.
- Added initial basic Kotlin configuration to the project.
- Migrated `YogaConstants` as an initial file to try out the first migration steps.
Pull Request resolved: https://github.com/facebook/yoga/pull/1829
Test Plan:
- Tested the migrated class directly against facebook/react-native, see the PR [here](https://github.com/facebook/react-native/pull/52998).
- Run: `./gradlew :yoga:assembleDebug` & `./gradlew :yoga:compileDebugSources`
I am not able to run the Java tests in this repo (even before the initial Kotlin setup) – not sure if I am missing something there but any pointers are welcome – it seems like there is some missing configuration. Currently trying with `./gradlew :yoga:test`
Reviewed By: cortinico
Differential Revision: D79545992
Pulled By: rshest
fbshipit-source-id: 8257ff53e6b6f2436980be98b6c94e1ac526b207
Summary:
X-link: https://github.com/facebook/react-native/pull/52817
Pull Request resolved: https://github.com/facebook/yoga/pull/1828
X-link: https://github.com/facebook/litho/pull/1070
This diff makes the Yoga/Yoga.h header an umbrella header, which means that it includes all of Yoga's public headers. The code changes in each file include adding the IWYU pragma export to each header file, which is a way to tell the compiler to export the header file's symbols to other files that include it. This is necessary for the header file to be used as an umbrella header.
Changelog:
[General][Added] - Code quality fixes
Reviewed By: corporateshark
Differential Revision: D78692457
fbshipit-source-id: 7fcd53d2a6f268fa4377dbd5bd6ba6eebc94b5f8
Summary:
X-link: https://github.com/facebook/react-native/pull/52530
This PR fixes two issues with `display: contents` implementation:
1. When a node with `display: contents` set is a leaf, it won't be cloned after the initial tree is built. The added test case covers this scenario.
2. It was possible for the subtree of `display: contents` nodes not to be cloned during layout. I don't have a minimal reproduction for this one, unfortunately. It was discovered in the Expensify app: https://github.com/Expensify/App/issues/65268, along with a consistent reproduction. In that specific case, it seems to be heavily tied to `react-native-onyx`, which is a state management library.
Changelog: [GENERAL][FIXED] - Fixed nodes with `display: contents` set being cloned with the wrong owner
Pull Request resolved: https://github.com/facebook/yoga/pull/1826
Reviewed By: adityasharat, NickGerleman
Differential Revision: D78084270
Pulled By: j-piasecki
fbshipit-source-id: eb81f6d7dcd1665974d07261ba693e2abea239bb
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1823
X-link: https://github.com/facebook/react-native/pull/52348
Fixes https://github.com/facebook/yoga/issues/1819
Yoga has a fast path when measuring a node, if it thinks it knows its dimensions ahead of time.
This path has some eroneous logic, to set both axis to owner size, if *either* will evaluate to zero, while having an `YGMeasureModeAtMost`/`FitContent` constraint. This means that if a node is given a zero width, and Yoga later measures with with `FitContent`, its height will become the maximum allowable height, even if it shouldn't be that large.
We can fix this, by only allowing if both axis are this fixed case, instead of just one.
This bug has existed for about a decade (going back to at least D3312496).
Changelog:
[General][Fixed] - Fix possible invalid measurements with width or height is zero pixels
Reviewed By: yungsters
Differential Revision: D76851589
fbshipit-source-id: 6f5a0e6beccc51f591726c9e83e9b90f3350ed0f
Summary:
X-link: https://github.com/facebook/react-native/pull/52073
Pull Request resolved: https://github.com/facebook/yoga/pull/1820
Fixes https://github.com/facebook/yoga/issues/1819
Yoga has a fast path when measuring a node, if it thinks it knows its dimensions ahead of time.
This path has some eroneous logic, to set both axis to owner size, if *either* will evaluate to zero, while having an `YGMeasureModeAtMost`/`FitContent` constraint. This means that if a node is given a zero width, and Yoga later measures with with `FitContent`, its height will become the maximum allowable height, even if it shouldn't be that large.
We can fix this, by only allowing if both axis are this fixed case, instead of just one.
This bug has existed for about a decade (going back to at least D3312496).
Changelog:
[General][Fixed] - Fix possible invalid measurements with width or height is zero pixels
Reviewed By: yungsters
Differential Revision: D76793705
fbshipit-source-id: ea4c00e688912a58c08801e4a14ddf1b293a5d86
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1816
This migrates Yoga to publish to Central Portal rather than OSSRC
Reviewed By: NickGerleman
Differential Revision: D76436235
fbshipit-source-id: ddaaa95472054aa2b09399b5cc8d821dae51234f
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1817
This file is not really necessary. As we don't have a Gradle Enterprise instance anymore, let's just remove it and inline it.
Reviewed By: mdvacca
Differential Revision: D76436236
fbshipit-source-id: 27fcc48574905415b147f67dd4a8f57dffbeda3e
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1813
This adds a unit test to Yoga, which emulates the model of "persistent Yoga nodes" and cloning used by React Fabric, including the private (but relied on) Yoga APIs.
It models the over-invalidation exposed in D75287261, which reproduces (due to Yoga incorrectly measuring flex-basis under fit-content, and that constraint changing when sibling changes) but this test for now sets a definite height on A, to show that we only clone what is neccesary, when measure constraints do not have to change.
Having a minimal version of Fabric's model in Yoga unit tests should make some of these interesting interactions a bit easier to debug.
Changelog: [Internal]
Reviewed By: javache
Differential Revision: D75572762
fbshipit-source-id: cda8b3fdd6e538a55dd100494518688c864bd233
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1811
X-link: https://github.com/facebook/react-native/pull/51298
## Resubmit
This was backed out due to being up the stack from another change that was backed out, but should be safe by itself.
## Original
We want to know if an artifact created during measurement can fully be reused after final layout, but the final layout is allowed to be slightly larger due to pixel grid rounding (while still allowing reuse). It's hard to tell after the fact, whether it is larger because of this rounding (though the measure is used), or if it may be a pixel larger for valid reasons.
We can expose the unsnapped dimensions of a node to give us this information, and to correlate measurement artifacts.
This is most of the time the same as the layout's measured dimension, though I don't think it's safe to use this, since anything else measuring the node after could clobber this (I think `YGNodeLayoutGetOverflow` may also be prone to this as a bug).
Changelog: [Internal]
Reviewed By: joevilches
Differential Revision: D74673119
fbshipit-source-id: 06d2eb21e28b76458ec88f4dfcaec809707d0390
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1809
X-link: https://github.com/facebook/react-native/pull/51181
We want to know if an artifact created during measurement can fully be reused after final layout, but the final layout is allowed to be slightly larger due to pixel grid rounding (while still allowing reuse). It's hard to tell after the fact, whether it is larger because of this rounding (though the measure is used), or if it may be a pixel larger for valid reasons.
We can expose the unsnapped dimensions of a node to give us this information, and to correlate measurement artifacts.
This is most of the time the same as the layout's measured dimension, though I don't think it's safe to use this, since anything else measuring the node after could clobber this (I think `YGNodeLayoutGetOverflow` may also be prone to this as a bug).
Changelog: [Internal]
Reviewed By: rshest
Differential Revision: D74292949
fbshipit-source-id: 05011c66a9a9480544313eb1dfe2c46bf7742bac
Summary:
The documentation shall be corrected to specify in which axis the spaces are distributed when flex container is configured 'alignContent' property.
Pull Request resolved: https://github.com/facebook/yoga/pull/1808
Reviewed By: joevilches
Differential Revision: D74347272
Pulled By: philIip
fbshipit-source-id: 1b05840028bca774c9d4a68562bcf537d5a72500
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1791
Disabling RTTI for Yoga is causing std::exception to don't work properly in OSS.
Fixes: https://github.com/facebook/react-native/issues/48027
Not sure why we originally disabled RTTI for Yoga, but we have it enable for the whole
React Native build so it probably makes sense to have it enabled for Yoga as well.
Changelog:
[Internal] [Changed] - Enable RTTI to fix exception pointer issue on React Native
bypass-github-export-checks
Reviewed By: javache, NickGerleman
Differential Revision: D70386744
fbshipit-source-id: 36e3a1ddb38346d31979d5c1b77d6e9796d6a855
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1786
Follow up on comments from D68152410
- Removes minsdk version from manifest
- Update to MIT licence
- Rename targets that don't export existing libs
I've looked into using yoga_java_library and adding a param to pass what libs to exclude, but ultimately provided_deps does what we need and it simpler to maintain; we still need to add the different targets because the aar needs to depend on :jni-server and not :jni.
Reviewed By: NickGerleman
Differential Revision: D69455682
fbshipit-source-id: aa08bc030495d13da9545ec368ab63a7a946dad0
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1782
This adds fb_native.android_aar targets for resources that are needed by Litho at runtime and cannot be exported as jars.
- resource modules referenced by Litho
- Yoga library. Some dependencies required by Yoga neded to be added as "provided_deps" so they don't get exported with the aar
Reviewed By: astreet
Differential Revision: D68152410
fbshipit-source-id: 7dc2ffa0b60a4ca160a011d16d2dc5ab91f608a6
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1781
This new image removed preinstalled libc++, and the ability to install old version of Clang Format we were installing.
This manually installs libc++ when setting up Clang jobs, and fully removes the clang-format validation job, since it wasn't correctly running before, and it's probably more a pain to ask people to run this than to just run Arcanist when importing a change.
Changelog: [Internal]
Reviewed By: joevilches
Differential Revision: D68455129
fbshipit-source-id: b5767be832b2b5d46db7e60e18b66823819ba15a
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1776
# Changelog:
[Internal] -
I was looking into some profiling results and noticed that Yoga has some quirky custom logic when it comes to rounding to the pixel grid, especially for the values that are halfway through.
There are already unit tests in place, but those cases weren't covered, so I am adding these extra corner cases here, to make sure they are covered in case that the actual rounding function may get modified (which it might make sense to, as it's currently can be more expensive than it could have been in some scenarios).
Reviewed By: christophpurrer
Differential Revision: D67712673
fbshipit-source-id: da1b7339a8939ced8d91f15441bc7d1f8af768c8
Summary:
X-link: https://github.com/facebook/litho/pull/1036
Pull Request resolved: https://github.com/facebook/yoga/pull/1775
X-link: https://github.com/facebook/react-native/pull/48404
## Changelog:
[Internal] -
This popped up when profiling some heavy UI performance, calling `fmod` operation in Yoga's `roundLayoutResultsToPixelGrid` in `PixelGrid.cpp` can be expensive, furthermore it turns out that some of the calls were redundant.
This replaces the duplicate calls to fmod with an equivalent single round operation, which for e.g. clang compiler on Windows brings the code in question from ~50 instructions (including 4 call instructions to the fmod function) down to ~30 instructions (without any external calls), and the layout operation being **~1% more efficient** for the particular benchmark I was looking into.
Reviewed By: christophpurrer
Differential Revision: D67689065
fbshipit-source-id: 2a074a1cb81bd7f7a3c414050b9ddda2ba90180f
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1768
Depending on the flex direction text will either be capped to the measured size or to the longest word, so I added that functionality
Reviewed By: mlord93
Differential Revision: D67106199
fbshipit-source-id: 0b4691768809004043a847f3fc5f7b94e92f1575
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1766
While we added support for 16K pages for yoga through React Native,
the standalone version we upload to Maven Central wasn't handled.
This fixes it.
Reviewed By: NickGerleman
Differential Revision: D66975933
fbshipit-source-id: 36a8404e2f2a60a44b9a39e478b23d1829bc542e
Summary:
X-link: https://github.com/facebook/react-native/pull/47973
Gap can be styled using both `points` and `percentages`, but YGNodeStyleGetGap currently returns a float value.
To maintain alignment with the `padding` and `margin` functionalities and allow it to be handled in bridging code, this function has been updated to return YGValue.
Pull Request resolved: https://github.com/facebook/yoga/pull/1753
Reviewed By: joevilches
Differential Revision: D66513236
Pulled By: NickGerleman
fbshipit-source-id: b7110855c037f20780f031f22a945bde4446687d
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1763
X-link: https://github.com/facebook/react-native/pull/48080
Small bug that I noticed while doing intrinsic sizing. We have the ownerHeight as the axis size despite bounding the length of the cross axis. This should therefore be the crossAxisOwnerSize, which might be the width in some cases
Changelog: [Internal]
Reviewed By: NickGerleman
Differential Revision: D66736539
fbshipit-source-id: 528fc438b3327cd6f7890ea0ba408e4ce7b0f02c
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1762
X-link: https://github.com/facebook/react-native/pull/48077
OCD strikes again. Grepped this time to make sure we didn't miss any cases for this specific param name
Changelog: [Internal]
Reviewed By: NickGerleman
Differential Revision: D66715777
fbshipit-source-id: 3e881a15b3b2836a4a55b11d7ec621541b92a05d
Summary:
Pull Request resolved: https://github.com/facebook/yoga/pull/1759
We just have block based tests right now. Intrinsic sizing is commonly used with text so lets add a few there.
Reviewed By: NickGerleman
Differential Revision: D66662940
fbshipit-source-id: f8b91419c89d22d79a91d3bd8c7da70429c827fb
