[iOS][Fabric] Changing Dynamic Type while the app is running does not re-lay out text (root not dirtied on a layout-context-only change)
I maintainer di solito rispondono entro 2 giorni
Nessuno ha ancora preso questa issue.
Valutazione
- Difficoltà
- 4/5
- Tempo stimato
- 3-5 giorni
- Idoneità per principianti
- 45/100
Direzione di ricerca
Start with RootShadowNode.cpp and inspect RootShadowNode::clone, then follow layoutContext into YogaLayoutableShadowNode::configureYogaTree. The issue identifies the three layout-context inputs that need to trigger layout dirtiness and explains why comparing the full context would be too broad. Reproduce the Dynamic Type change on iOS with the New Architecture, then verify existing layout tests cover the context changes and that text relays out after the change.
Scritto dal modello di indicizzazione a partire dal testo della issue.
Descrizione
Description
On iOS with the New Architecture, changing the system text size (Dynamic Type) while the app is running has no visible effect on text already on screen. React Native receives the new fontSizeMultiplier, but no layout pass runs, so text keeps its launch-time size until the app is restarted. Text mounted after the change can pick up the new size, which produces mixed layouts.
Steps to reproduce
- Build a Release app with React Native 0.88.1 (New Architecture) and any
<Text>that hasallowFontScalingleft at its default. - Launch it on an iOS simulator or device.
- Without closing the app, change the text size (Settings › Display & Brightness › Text Size, or on a simulator
xcrun simctl ui <udid> content_size accessibility-extra-large). - Return to the app.
Expected: text re-lays out at the new size.
Actual: nothing changes, in either direction. Relaunching renders the new size correctly.
Root cause
-
RCTFabricSurfacehandlesUIContentSizeCategoryDidChangeNotificationby calling_updateLayoutContextwith the newfontSizeMultiplier, which reachesSurfaceHandler::constraintLayout. That commitsRootShadowNode::clone(propsParserContext, layoutConstraints, layoutContext). -
RootShadowNode::cloneonly callsdirtyLayout()when the layout constraints change:if (layoutConstraints != getConcreteProps().layoutConstraints) { newRootShadowNode->dirtyLayout(); } -
For a layout-context-only change, the root stays clean.
RootShadowNode::layoutIfNeededthen returns early (getIsLayoutClean()), solayoutTree→configureYogaTreenever runs. -
That means the existing font-scale handling in
YogaLayoutableShadowNode::configureYogaTreenever gets a chance to run, even thoughenableFontScaleChangesUpdatingLayoutdefaults totrue. That code marks measurable nodes dirty when theirfontSizeMultiplierdiffers from the new one. -
pointScaleFactorandswapLeftAndRightInRTLare consumed byconfigureYogaTreein the same way, so a change to either alone has the same problem.
The same code is on main as of 2026-11-10 (RootShadowNode.cpp last changed in 178c30d7ac). #59285 makes ParagraphShadowNode drop content built for another fontSizeMultiplier, but that only helps once a layout pass actually runs.
Proposed fix
Dirty the root when any layout-context input that configureYogaTree consumes changes:
- if (layoutConstraints != getConcreteProps().layoutConstraints) {
+ const auto& previousContext = getConcreteProps().layoutContext;
+ if (layoutConstraints != getConcreteProps().layoutConstraints ||
+ layoutContext.fontSizeMultiplier != previousContext.fontSizeMultiplier ||
+ layoutContext.pointScaleFactor != previousContext.pointScaleFactor ||
+ layoutContext.swapLeftAndRightInRTL !=
+ previousContext.swapLeftAndRightInRTL) {
newRootShadowNode->dirtyLayout();
}
The full LayoutContext == is deliberately not used: it also compares affectedNodes (a per-pass pointer) and the viewport fields, which would dirty the root on unrelated commits.
I traced this in source and reproduced the symptom in a Release build. I have not run the patched build myself, because our app uses the prebuilt React Native core.
Versions
React Native 0.88.1 (also present on 0.88-stable and main), Expo SDK 58, iOS 26.5 simulator, New Architecture, Hermes.
- Lingua principale
- C++
- Stelle
- 127k
- Fork
- 25.5k
- Merge medio
- 3g 18h
- PR unite (30g)
- 188
Preparare l'ambiente
- Nessun Dockerfile né file Docker Compose
- Ha un modello di pull request
- Leggi la guida per i contributori
Come iniziare
- Leggi tutta la issue e poi la guida ai contributi del progetto.
- Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
- Fai un fork del repository e lavora su un branch.
- Apri una pull request che faccia riferimento al numero della issue.
Altre issue di react/react-native
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 76/100
react/react-native#59945 ·
I maintainer di solito rispondono entro 2 giorni
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 82/100
react/react-native#59944 ·
I maintainer di solito rispondono entro 2 giorni
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 78/100
react/react-native#59942 ·
I maintainer di solito rispondono entro 2 giorni
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 76/100
react/react-native#59939 ·
I maintainer di solito rispondono entro 2 giorni
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 78/100
react/react-native#59929 · 1 commento ·
I maintainer di solito rispondono entro 2 giorni
Tutte le issue di react/react-native
Issue simili
-
test(tools): two testkit cache tests fail unless GOCACHE and GOMODCACHE are in the environmentApertabacklog
Difficoltà 2/5 1-3 ore Idoneità per principianti 78/100
EchoTools/nevr-runtime#2385 ·
I maintainer di solito rispondono entro 1 giorno
-
type:bug
Difficoltà 2/5 1-3 ore Idoneità per principianti 65/100
tensorflow/tflite-micro#3995 ·
I maintainer di solito rispondono entro 1 giorno
-
CPU update 1.31xApertaspeed-report
Difficoltà 1/5 Meno di un'ora Idoneità per principianti 75/100
mw00/project-maya#961 · 1 commento ·
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 78/100
-
not for AI lovers
Difficoltà 2/5 1-3 ore Idoneità per principianti 76/100
OSGeo/gdal#15515 · 1 commento ·
I maintainer di solito rispondono entro 1 giorno