Conversation
…ws (#49518) <!-- ccr-slack-attribution --> _Requested by **Keith Kurak** · [Slack thread](https://exponent-internal.slack.com/archives/D0BSRLHKD7V/p1787979054839969)_ # Why While building an agentic workflow triggered by PR events, Claude got hung up on whether someone could run workflows on my EAS account without my permissions by forking my public repo and opening PR's. However, internally, it's been noted that this is not possible: expo/universe#28147 # How Added a note indicating that forks can't trigger PR events with EAS Workflows # Test Plan Documentation-only change. The diff adds two lines to `docs/pages/eas/workflows/syntax.mdx` and touches no code or package sources, so there is nothing to test at runtime. - The `vale` and `docs-pr` checks pass on the head commit (`7e7594c`). - A reviewer can check the rendered page by running the docs site locally — `pnpm install && pnpm dev` in `docs/` — and opening http://localhost:3002/eas/workflows/syntax/, where the new note appears under the `pull_request` trigger section. # Checklist - [ ] I added a `changelog.md` entry and rebuilt the package sources according to [this short guide](https://github.com/expo/expo/blob/main/CONTRIBUTING.md#-before-submitting) - [ ] This diff will work correctly for `npx expo prebuild` & EAS Build (eg: updated a module plugin). - [x] Conforms with the [Documentation Writing Style Guide](https://github.com/expo/expo/blob/main/guides/Expo%20Documentation%20Writing%20Style%20Guide.md) Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: Keith Kurak <keith@expo.dev>
…grouping (#49386) # Why We had two bugs in the `et github-metrics` command that caused the CI health report to be wrong. 1. Success rate counted unresolved runs as failures. 2. Workflow breakdown grouped by name - breaks when workflow sets dynamic name. # How 1. Fixed by moving the denominator: `computeRunConclusionCounts` now returns `resolvedRuns = successfulRuns + failedRuns + cancelledRuns` and divides by that. 2. Runs are grouped by `workflow_id` and not name anymore # Test Plan ✅ Unit tests (`tools/src/commands/GitHubMetricsCommand.test.ts` covers both fixes) # Checklist - [x] Builds, type-checks, lints, and tests pass (the three checks in `tools/AGENTS.md`, which is what `.github/workflows/expotools.yml` runs). - [x] Conforms to the documentation style guide. Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
# Why `Grid` and `GridItem` in `ui/components/Authentication/` have no importers anywhere in the docs; the only page using this folder imports `ASSETS` and `Box`. Found during the border-radius migration (ENG-26237). # How - Delete `Grid.tsx` and `GridItem.tsx` and drop their exports from the folder index. `Box`, `Icon`, and `CreateAppButton` stay in use on /guides/authentication/. # Test Plan <!-- Please describe how you tested this change and how a reviewer could reproduce your test, especially if this PR does not include automated tests! If possible, please also provide terminal output and/or screenshots demonstrating your test/reproduction. --> N/A # Checklist <!-- Please check the appropriate items below if they apply to your diff. --> - [ ] I added a `changelog.md` entry and rebuilt the package sources according to [this short guide](https://github.com/expo/expo/blob/main/CONTRIBUTING.md#-before-submitting) - [ ] This diff will work correctly for `npx expo prebuild` & EAS Build (eg: updated a module plugin). - [ ] Conforms with the [Documentation Writing Style Guide](https://github.com/expo/expo/blob/main/guides/Expo%20Documentation%20Writing%20Style%20Guide.md)
…49439) # Why Fix ENG-26229 ENG-26230 ENG-26231 ENG-26236 <!-- Please describe the motivation for this PR, and link to relevant GitHub issues, forums posts, or feature requests. --> # How <!-- How did you build this feature or fix this bug and why? --> - Add `Vary: Accept` to negotiated markdown responses in `public/_worker.js` so caches key on the `Accept` header (ENG-26229) - Return a markdown 404 body linking `llms.txt`, `sitemap.xml`, and the docs home instead of a plain `Not found` (ENG-26230) - Pass 3xx responses through to markdown requests so moved pages redirect instead of returning 404 (ENG-26231) - Replace the `text/markdown` substring match with an `acceptsMarkdown` q-value parser: markdown is served only when explicitly listed, above zero, and at or above HTML's quality (ENG-26236) - Cover all four behaviors in `scripts/test-worker.ts` # Test Plan <!-- Please describe how you tested this change and how a reviewer could reproduce your test, especially if this PR does not include automated tests! If possible, please also provide terminal output and/or screenshots demonstrating your test/reproduction. --> `pnpm test:worker` should pass. # Checklist <!-- Please check the appropriate items below if they apply to your diff. --> - [ ] I added a `changelog.md` entry and rebuilt the package sources according to [this short guide](https://github.com/expo/expo/blob/main/CONTRIBUTING.md#-before-submitting) - [ ] This diff will work correctly for `npx expo prebuild` & EAS Build (eg: updated a module plugin). - [ ] Conforms with the [Documentation Writing Style Guide](https://github.com/expo/expo/blob/main/guides/Expo%20Documentation%20Writing%20Style%20Guide.md)
# Why Fixes - #47883, #48059 Fixes the layout loops that happens in some cases when `Host` and `RNHostView` has `matchContents` and child of `RNHostView` depends upon parent on sizing. Explained it in more detail here - #48059 (comment) <!-- Please describe the motivation for this PR, and link to relevant GitHub issues, forums posts, or feature requests. --> # How Set `RNHostView` node as a leaf and measurable node using shadow node traits. This removes it from the yoga layout hierarchy (shadow node hierarchy is still preserved). Override `measureContent` and `layout` shadow node functions for `RNHostView`. The `measureContent` is called by Yoga during layout and `layout` is called during commit. A better fix would be if Yoga supports `max-content` so we could just do `<RNHostView style={{ width: "max-content", height: "max-content" }} />` if someone passes `matchContents`. Made a PR in Yoga for that react/yoga#2012 <!-- How did you build this feature or fix this bug and why? --> # Test Plan Added regression testcases for RNHostView which leads to loops. <!-- Please describe how you tested this change and how a reviewer could reproduce your test, especially if this PR does not include automated tests! If possible, please also provide terminal output and/or screenshots demonstrating your test/reproduction. --> # Checklist <!-- Please check the appropriate items below if they apply to your diff. --> - [ ] I added a `changelog.md` entry and rebuilt the package sources according to [this short guide](https://github.com/expo/expo/blob/main/CONTRIBUTING.md#-before-submitting) - [ ] This diff will work correctly for `npx expo prebuild` & EAS Build (eg: updated a module plugin). - [ ] Conforms with the [Documentation Writing Style Guide](https://github.com/expo/expo/blob/main/guides/Expo%20Documentation%20Writing%20Style%20Guide.md)
…pProcedure (#49374) # Why `start` ran BuildData.ensureBuildDataIsConsistent synchronously on the caller's thread, which is the main thread when apps call UpdatesController.initialize from `onCreate`. Room opens the database lazily on first query, so this also ran the database open and every migration on the main thread during cold start. It is the only production code path that queries the updates database on the main thread. # How Moved the call into the top of StartupProcedure.run(), keeping the hasUpdatesOverride guard. The state machine's serial executor queue runs procedures on Dispatchers.IO and guarantees the check completes before any other procedure touches the database. This matches iOS, where ensureBuildDataIsConsistentAsync already dispatches to the database queue. # Test Plan New test `StartupProcedureTest`
…se (#49375) # Why The database was built with allowMainThreadQueries(), which hides regressions. With the build data check moved off the main thread in the previous PR, no production code path queries on main anymore, so the escape hatch can go. # How Removed .allowMainThreadQueries() from the builder in UpdatesDatabase.getInstance. Any future main-thread query now throws immediately instead of janking startup. Tests are unaffected because they build their own in-memory databases. # Test Plan CI and local runs
…ectly (#49376) # Why DatabaseHolder claimed to serialize high-level database operations, but its database property acquired and released the mutex before returning, so no caller ever held the lock during a query. It also ran runBlocking on every access, which could stall the main thread if a withDatabase block ever held the mutex. The real serialization mechanism is the state machine's serial procedure queue. # How Deleted the class. Controllers, procedures, and LoaderTask now take UpdatesDatabase directly, and the two withDatabase launch-count updates became direct DAO calls inside the existing IO coroutines. Updated Expo Go's NativeModuleDepsProvider and ExpoUpdatesAppLoader to inject UpdatesDatabase instead. Updated the UpdatesDatabase doc comment to describe the actual concurrency model. # Test Plan CI E2E, instrumentation, and unit tests all pass
…e` to 0.2.0 (#49241) # Why `oxlint-config-universe` 0.2.0 is out and requires `oxlint` 1.79, so this updates both. # How oxlint 1.79 turns on a set of React Compiler-derived rules by default and the repo has a few hundred pre-existing findings (mostly in expo-router), so the base config keeps them off to make this update behavior-neutral. We can re-enable them rule by rule as the findings get triaged; the stacked #49242 starts with the five cheapest ones. A few more adjustments in the base config: `no-inner-declarations` started flagging exported functions inside TS namespaces, which the new `namespaces: 'allow'` option fixes without losing the rule. The `no-did-mount-set-state` / `no-did-update-set-state` pair is off because oxlint (unlike ESLint's default mode) also flags `setState` inside async callbacks, hitting the common fetch-on-mount pattern. The TS overrides that 0.2.0 now applies itself (or no longer enables) are dropped. Source changes are small and mechanical: two useless `export {}` statements flagged by the newly enabled `typescript/no-useless-empty-export` are removed (an expo-observe test and an `@expo/cli` e2e helper), the LogBox DOM polyfill's inline disables now also cover `import/namespace` (oxlint 1.79 newly reports assignments to namespace-import members), and expo-router's navigation tree reducer gets inline disables where the stricter `rules-of-hooks` rejects effect-event calls made outside of effects it recognizes (event handlers and a custom effect wrapper). # Test Plan Linted all 130 oxlint packages plus the three oxlint apps: no errors anywhere, and the only warnings are in `@expo/metro-config` test fixtures and exist on main as well. `et check-packages` passes for the touched packages, including the expo-observe and expo-router jest suites. <!-- disable:changelog-checks -->
# Why Fixes #49534. `TextView` is one of the Expo UI views that applies common modifiers internally so that concatenated children remain a native `SwiftUI.Text`. The widget renderer currently wraps it in `UIBaseView`, which applies the same modifiers a second time. A modifier such as `padding` or `strokeBorder` therefore appears twice in iOS widgets. `WidgetsDynamicView` is also used by Live Activities, so they share the same behavior. The public minimal reproduction is available at https://github.com/gee1k/expo-widgets-text-modifier-repro. # How Added an opt-out to `WidgetsDynamicView.render` for views that already handle common modifiers, and use it only for `TextView`. All other widget views continue to render through `UIBaseView` unchanged. Also added the fix to the unpublished `expo-widgets` changelog. # Test Plan The reproduction renders the same `padding` and `strokeBorder` values in two ways: on an `HStack` (green) and directly on `Text` (red). | Before | After | | --- | --- | |  |  | Local verification on an iPhone 17 Pro Max Simulator running iOS 26.5: - `npx expo-doctor@latest`: 21/21 checks passed. - `./node_modules/.bin/tsc --noEmit`: passed in the reproduction. - Built and installed the unmodified `expo-widgets@57.0.15` reproduction with `npx expo run:ios`; observed two nested red borders. - Rebuilt the same project with this `DynamicView.swift`; build completed with 0 errors and 0 warnings, and the red `Text` rendered one border with the expected padding. - `git diff --check`: passed. There is currently no native `WidgetsDynamicView` test target; the package's existing automated tests cover the JavaScript bundle and configuration layer, so the regression is demonstrated with the minimal native reproduction and before/after Simulator evidence. # Checklist - [x] I added a `changelog.md` entry and rebuilt the package sources according to [this short guide](https://github.com/expo/expo/blob/main/CONTRIBUTING.md#-before-submitting) - [x] This diff will work correctly for `npx expo prebuild` & EAS Build (eg: updated a module plugin). - [ ] Conforms with the [Documentation Writing Style Guide](https://github.com/expo/expo/blob/main/guides/Expo%20Documentation%20Writing%20Style%20Guide.md) (not applicable; no documentation changes) --------- Co-authored-by: Codex <codex@openai.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
6 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
See Commits and Changes for more details.
Created by
pull[bot] (v2.0.0-alpha.4)
Can you help keep this open source service alive? 💖 Please sponsor : )