chore(deps): update anthropics/claude-code-action digest to e58dfa5#1498
Merged
Merged
Conversation
|
renovate
Bot
force-pushed
the
renovate/anthropics-claude-code-action-digest
branch
from
ce63811 to
d0a32a6
Compare
Contributor
Alpha Preview (Stackflow SPA)
|
Contributor
Alpha Preview (Storybook)
|
Contributor
Alpha Preview (Docs)
|
renovate
Bot
changed the title
renovate
Bot
force-pushed
the
renovate/anthropics-claude-code-action-digest
branch
from
d0a32a6 to
9ddec87
Compare
renovate
Bot
changed the title
renovate
Bot
force-pushed
the
renovate/anthropics-claude-code-action-digest
branch
from
9ddec87 to
9d041bc
Compare
renovate
Bot
force-pushed
the
renovate/anthropics-claude-code-action-digest
branch
from
9d041bc to
7a8953e
Compare
renovate
Bot
changed the title
Contributor
|
/renovate-review |
|
Claude finished @junghyeonsu's task in 2m 26s —— View job 🤖 Renovate 의존성 분석 완료 📋 분석 개요이 PR은 GitHub Actions용 🛡️ 보안 분석GitHub Action 특성
사전 검사 결과📦 업데이트 상세anthropics/claude-code-action
🔍 주요 변경사항 (13개 커밋)✨ 중요 버그 수정1. ripgrep 바이너리 경로 수정 (#1235)
2. 브랜치명 검증 완화 (#1248)
3. Node.js 24 지원 (#1238)
🔄 정기 업데이트
💡 우리 저장소 사용처이 액션은 2개 워크플로우에서 사용 중: 1.
|
junghyeonsu
added a commit
that referenced
this pull request
May 7, 2026
* feat(top-navigation): apply font scaling with component-level clamp bounds (#1477) * feat(top-navigation): apply font scaling with component-level clamp bounds - Wrap title/subtitle in clamp() with per-layout min/max font-size and line-height scale bounds (nested with global limits; component range narrows global) - Introduce main/left/right slot tokens for inter-slot spacing; tokenize android 16px left padding - Rename root.minHeight to root.height - Switch recipe to typed tokens.$fontSize.tNStatic / tokens.$lineHeight.tNStatic references Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * chore(top-navigation): drop spacing slot tokens, keep only font scaling - Revert main/left/right slot additions from rootage; title/subtitle scale tokens kept - Restore cupertino main padding to raw var(--centered-title-padding-x, 0) - Restore android left paddingRight hardcoded 16px Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs: changeset --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(top-navigation): raise min font/line-height scale to 1 Prevent downscaling of title and subtitle under user font size settings by raising minFontSizeScale and minLineHeightScale from 0.8 to 1. Only the upper clamp (1.2x) remains active. * docs(list): update Align Property description (#1494) Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * chore(deps): update dependency bun to v1.3.12 (#1468) Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com> Co-authored-by: 정현수 <hyeonsu.github@gmail.com> * chore(deps): update anthropics/claude-code-action digest to 905d4eb (#1467) Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com> Co-authored-by: 정현수 <hyeonsu.github@gmail.com> * docs * feat(docs): add deprecation notices and replacement links to llms.txt (#1474) * feat(docs): add deprecation notices and replacement links to llms.txt Deprecated components now surface their deprecation message and a link to the replacement component in llms.txt output. Adds `replacement` frontmatter field to all deprecated MDX files and fills missing `deprecated` fields on 8 React component docs. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(docs): use options object for getLLMTextForFullCompilation to fix type error The positional optional params conflicted with Array.map's (value, index, array) signature in breeze/lynx/ai-integration routes. Using an options object avoids this. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(docs): simplify getLLMTextForFullCompilation to remove Section dependency Keep replacement links only in individual page output (getLLMText). Full compilation output now shows only the deprecated notice without replacement links, eliminating the need for Section/allPages params and fixing the .map() signature conflict with other routes. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(docs): remove unnecessary lambda wrapper in llms-full routes pages.map(getLLMTextForFullCompilation) works directly since the function takes a single parameter. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(docs): revert formatting-only changes in llms-full routes Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(docs): address PR review feedback for llms.txt deprecation - Fix replacement slugs: hstack→h-stack, vstack→v-stack to match actual page slugs - Sort llms.txt index by displayTitle instead of full string with (Deprecated) suffix - Unify deprecation notice logic into buildDeprecationNotice with optional params - Remove unnecessary titleToSlug since replacement values are already slugs Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(docs): add missing description field to deprecated MDX files Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * Revert "fix(docs): add missing description field to deprecated MDX files" This reverts commit 5fdd4d6. * refactor(docs): remove replacement field, embed markdown links in deprecated text Instead of a separate `replacement` slug field with matching logic, embed markdown links directly in the `deprecated` frontmatter text. This simplifies the codebase by removing the replacement field from schema, types, and the slug-matching logic in get-llm-text.ts. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * feat(docs): render markdown links in deprecated frontmatter text Fumadocs does not render markdown in frontmatter fields by default. Add renderInlineMarkdown helper to convert [text](url) patterns in deprecated messages into clickable Next.js Link components. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * refactor(docs): move deprecation message to MDX body as Callout Unify UI and llms.txt deprecation rendering through a single source. Change `deprecated` frontmatter from message string to boolean flag, and move the actual guidance (with markdown link to the replacement) into each MDX body as a Callout. Fumadocs renders the Callout natively for the UI, and the processed body flows into llms.txt with the markdown link intact, so LLMs can follow it to the replacement's llms.txt. The boolean flag now drives all structural branches (title suffix, sidebar badge, index label, component-grid filtering), while the body Callout carries the human-readable message. This removes the manual renderInlineMarkdown helper in page.tsx and the buildDeprecationNotice helper in get-llm-text.ts. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * docs(deprecated): align Callout guidance with rootage deprecation field Make each deprecated component's Callout match its rootage `deprecated` message so the UI/llms.txt guidance agrees with the canonical source. - action-chip: point to Chip.Button with variant="solid" - control-chip: point to Chip.Toggle or Chip.Button - link-content: point to Action Button with variant="ghost" - fab: point to contextual-floating-button (was floating-action-button) Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * docs(deprecated): add "Deprecated" title to deprecation Callouts Add `title="Deprecated"` to the first Callout of every deprecated component MDX so the notice is visually heavier — the title renders as a bold label above the message, making it easier to spot the warning at a glance. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(docs): use gray-matter for frontmatter parsing Replace naive regex-based frontmatter parser with gray-matter, which correctly types YAML boolean values. This also fixes the CLI `docs` command not surfacing deprecation hints — the old parser stored all values as strings, so `frontmatter.deprecated === "true"` happened to work with the previous string-message schema but broke after the schema changed to a proper boolean. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(docs): drop gray-matter, extend naive parser to strip quotes Reverts the gray-matter dependency added in df01f9f. The naive regex parser already handles this script's needs once the frontmatter.deprecated comparison is kept as the raw string "true" — which matches the boolean YAML value the parser captures as text. Adds quote-stripping so YAML values that must be quoted (e.g. descriptions starting with `@`) parse cleanly. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * refactor(docs): replace manual frontmatter parsing with gray-matter (#1508) Use gray-matter in generate-docs-index and import-components-to-sanity scripts instead of custom regex-based parsers. Boolean frontmatter values (e.g., deprecated) are now parsed as real booleans, removing the need for string comparisons. Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com> * fix(ci): unblock claude auto-fix in sync-figma-entities workflow (#1512) The build step's `bunx tsc --noEmit | tee` line died on the pipe under `set -e -o pipefail` before reaching the `tsc_failed=true` output write, which silently disabled the claude-fix step (its `if` guards on that output). Wrap the pipe in `if !` so the failure is captured explicitly without tripping `set -e`, allowing the auto-fix flow to actually run. * fix(ci): build slack payload with jq to escape claude summary safely The Slack notification step interpolated claude-summary's free-form korean text directly into a YAML mrkdwn string, which broke the payload parser whenever the summary contained quotes, colons, or newlines (the exact failure on run #24819145934). Build the JSON payload in a prior bash+jq step where --arg auto-escapes the summary, then hand it to slack-github-action via payload-file-path. Also fixes a latent bug: the inline `format('\\n...')` produced a literal `\n` (two chars), so summaries never actually line-broke in mrkdwn. printf '%s\\n%s' now produces a real newline that jq encodes into a proper JSON `\\n`. * feat(notification-badge): add min width/height/aspect-ratio for size=large (#1500) * feat(notification-badge): add min width/height/aspect-ratio for size=large Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(notification-badge): add Positioner usage story Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs(notification-badge): simplify Positioner story condition map --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * refactor(top-navigation): separate button rootage declarations (#1516) * chore(cli): Add PostHog docs env wiring and CLI outcome telemetry (#1513) * feat: add PostHog telemetry for docs and CLI outcomes * chore(cli): align telemetry log messaging * chore(deps): update dependency bun to v1.3.13 (#1496) Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com> * feat(docs): transform ProgressBoardTable in llms.txt output (#1519) * feat(docs): transform ProgressBoardTable in llms.txt output Adds a normalize rule that replaces <ProgressBoardTable /> with markdown summary and per-component status tables sourced from Sanity, mirroring the existing platformStatusRule pattern. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * test(docs): cover progressBoardRule with fixtures and guard empty status - Skip link rendering when status is missing to avoid misleading "Not Ready" link - Add fixture-based happy-path and 0/0 not-planned edge-case tests Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com> * feat(snackbar): change default duration from 5s to 4s (#1517) Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * docs: update changeset * feat(snackbar): add immediate/queued strategy for multi-snackbar behavior (#1413) * test(snackbar): add useSnackbar tests for immediate/queued strategy Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * feat(snackbar): add immediate/queued strategy for multi-snackbar behavior Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * docs(snackbar): add strategy example and stackflow-spa queued demo Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * test(snackbar): add richer mixed strategy test cases Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(snackbar): use dismissing state for immediate replacement to preserve animations Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * test(snackbar): add dismiss+setTimeout workaround demo to stackflow-spa Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * docs(snackbar): simplify strategy example with positive/critical variants Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(snackbar): use functional setState to prevent stale closure race Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor(snackbar): use switch statement for state machine entry events Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(snackbar): wrap strategy example with SnackbarProvider Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor(snackbar): migrate state machine to useReducer for atomic transitions Fixes two real bugs that arose from mixing side-effects into setState updaters and splitting onClose invocation across event handler and effect: - StrictMode double-invocation of setState updaters caused pop() (which itself mutated queue via setQueue) to run twice, corrupting the queue on rapid synchronous pushes. - dismiss() invoked onClose directly while the dismissing effect also invoked it after removeDelay, resulting in duplicate calls. Consolidates state, queue, and currentSnackbar into a single reducer so all transitions are pure. Timers and onClose now live only in the effect. events memo drops its state dependency as a side benefit. * docs(snackbar): annotate dismiss+setTimeout workaround example Adds a brief comment explaining the pattern and makes the setTimeout delay explicit (0) so readers don't have to infer intent from an omitted argument. * test(snackbar): lock in onClose semantics for dropped queue items Pins the intended contract of onClose: it fires only for snackbars that were actually presented (became currentSnackbar). Queued items cleared by an immediate-strategy push never entered the display lifecycle, so their onClose must not be invoked. --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * docs * feat(notification-badge): use static typo for size=large label (#1521) Use $font-size.t1-static and $line-height.t1-static so the label is unaffected by font scaling, and remove the now-unnecessary minAspectRatio spec. Refs DES-1700 * chore(deps): update anthropics/claude-code-action digest to e58dfa5 (#1498) Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com> * chore(deps): update actions/github-script action to v9 (#1470) Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com> Co-authored-by: 정현수 <hyeonsu.github@gmail.com> * release: version packages (#1483) * Version Packages * Update public rootage version --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> * feat(docs): expand ComponentGrid in llms.txt output (#1509) * feat(docs): expand ComponentGrid in llms.txt output Add a rule that transforms <ComponentGrid /> into a categorized markdown list of components so /llms/docs/components.txt exposes the full component index instead of a bare MDX tag. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(docs): address ComponentGrid rule review feedback - Treat deprecated: false / "no" / empty as not deprecated so valid components are no longer filtered out by the string-coerced parser. - Switch tests to fixture-based equality via readFixture / normalizeForAssert to match the project's _llms test convention, and add unit coverage for isDeprecatedValue. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(docs): use getLLMMarkdownUrl helper in ComponentGrid rule Replace hardcoded /llms/docs/components/{slug}.txt template with the existing getLLMMarkdownUrl helper so URL construction stays centralized. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(docs): use gray-matter in ComponentGrid rule Replace the ad-hoc regex parser and string-based isDeprecatedValue helper with gray-matter, consistent with generate-docs-index. Boolean frontmatter values are now parsed as real booleans. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(docs): warn when ComponentGrid components directory is missing Silently returning null masked the original symptom this PR fixes — the <ComponentGrid /> MDX tag leaking into llms.txt output. Log the attempted candidate paths so CI/builds can surface the failure. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(docs): warn when ComponentGrid transform falls back Mirror the resolveComponentsDir warning: surfacing the fallback so a failing getEntries/buildMarkdown can be caught in CI instead of silently leaving the <ComponentGrid /> tag in llms.txt output. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com> * fix(deps): update dependency uuid to v14 [security] (#1511) * fix(deps): update dependency uuid to v14 [security] * chore(mcp): add uuid security update changeset --------- Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com> Co-authored-by: 정현수 <hyeonsu.github@gmail.com> * chore(mcp): use caret range for zod (#1529) Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * chore(docs): upgrade fumadocs to 16.8.2 (#1530) * chore: bump fumadocs * refactor(docs): streamline fumadocs source configuration - Replace getTransformed*PageTree async walker with a sync LoaderPlugin that reads frontmatter via this.storage.read(filePath); removes the rootage coupling from the sidebar-badge path - Drop per-source *Options wrappers in layout.config.tsx; each layout now composes baseOptions + tree inline - Align zod to ^4.3.6 to match fumadocs-mdx; loader() now infers the extended pageData schema correctly, removing the FumadocsSource<TSrc> narrowing workaround Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * chore(deps): remove 11 unused root devDependencies (#1531) Removed from root package.json (verified with isolated per-workspace typecheck + build comparison; results identical across 67 workspaces): - 7 @changesets/* subpackages (assemble-release-plan, config, get-dependents-graph, get-github-info, pre, read, types) - transitively provided by @changesets/cli - @manypkg/get-packages - transitively provided by @changesets/cli - @seed-design/qvism-cli - already declared by packages/css - semver - already declared by packages/cli and packages/codemod - @types/semver - redundant since semver v7.7+ ships built-in types Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * feat(top-navigation-text-button): add maxWidth to preserve main slot space (#1536) Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * chore: pin npmjs.org registry and bump minimumReleaseAge to 3 days (#1541) - Add .npmrc with registry=https://registry.npmjs.org/ so the internal proxy URL does not leak into bun.lock, and external contributors can run `bun i` without extra setup. - bunfig.toml [install].registry has no effect because bun gives user ~/.npmrc precedence over the project bunfig.toml; project .npmrc wins. - Bump minimumReleaseAge from 1 day to 3 days to match the registry proxy policy (only packages older than 3 days are installable). * feat(top-navigation): rename theme to ios/android and add gradient/layer support (#1542) * refactor(top-navigation-text-button): scope maxWidth to theme=ios Move `maxWidth: 96px` from base to the `theme=ios` definition only, and introduce an empty `theme=android` definition so Android can opt out of the iOS-style ellipsis cap. * feat(top-navigation): rename theme=cupertino to theme=ios and add gradient/layer support - Rename `theme=cupertino` to `theme=ios` to match platform terminology - Add `gradient` and `bleedBottom` properties to the root slot - Add `tone=layer` variant so consumers can swap to `$color.bg.layer-basement` - Add `gradient` variant (only effective when `tone=transparent`); split `tone=transparent` into `gradient=false` (existing transparent fill) and `gradient=true` (new linear-gradient with bottom bleed) - Update qvism stackflow `app-bar`/`app-screen` consumers for the renamed vars * chore(skills): add snapshot-release skill Add project skill that orchestrates the `/snapshot` snapshot-release flow on PR comments end-to-end: - Discovers the PR for the current branch, any existing `/snapshot` comment, and the local git state in one Bash call. - If no `/snapshot` exists, asks the user before posting; warns first when there are uncommitted changes or unpushed commits, since the CI builds from the PR's remote HEAD and would silently omit local-only work. - Finds the matching Continuous Releases workflow run, waits for it to finish, and reports the resulting `📦 Snapshot Release` tarball URLs. - Bundles every `gh` invocation into one of two Bash calls (discover + orchestrate); re-entrant when the orchestrator dies mid-flight. * feat(skills): offer "push then post" option in snapshot-release When the working tree is clean but the local branch has unpushed commits, add a `Push, then post /snapshot` option to the trigger prompt so the user can land their work and trigger the snapshot in one step. The orchestrator honors a new `PUSH_FIRST` flag and runs `git push` before `gh pr comment`. The push option is intentionally **not** offered when there are uncommitted changes — `git push` cannot carry those, and offering it would just shift the source of confusion ("I pushed, why aren't my changes in the snapshot?"). * fix(skills): respect wait-only intent in snapshot-release When the user's request is clearly wait-only ("snapshot 기다려줘", "did the snapshot finish?", etc.) and no `/snapshot` comment exists, don't pivot into the trigger prompt. Tell them no comment was found and stop, so the skill matches the user's stated intent instead of silently shifting to a different workflow. Trigger or ambiguous phrasings still fall through to the AskUserQuestion. * chore: upgrade TypeScript to 6.0 * chore(figma): sync entities from Figma (#1493) Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> * ci(figma-sync): skip claude rerun on matching output and stage all src/ changes * fix(figma): make componentPropertyDefinitions optional in ComponentMetadata * Revert "fix(figma): make componentPropertyDefinitions optional in ComponentMetadata" This reverts commit af722da. * fix(figma): always emit componentPropertyDefinitions field in extracted entities * docs: add changeset * chore(rootage): add divider=false * perf(stackflow): replace CSS animation with WAAPI for AppScreen transitions (#1444) * perf(stackflow): replace CSS animation system with WAAPI for all AppScreen transitions Migrate all AppScreen transition animations (push, pop, swipe back) from CSS recipe-based animations to Web Animations API (WAAPI). This eliminates the CSS-to-JS handoff that caused flicker, double-animation, and timing bugs during swipe-back gestures. Key changes: - Remove all transition animation selectors from app-screen.ts and app-bar.ts recipes - Add transition-animation.ts with WAAPI implementations for iOS/Android/fadeIn styles - Rewrite useGlobalInteraction to detect transitionState changes and trigger WAAPI - Add rAF lock pattern and passive-ready touchmove in useSwipeBack - Manage element positions via explicit inline styles instead of CSS variables Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor(stackflow): simplify transition-animation.ts structure Extract shared helpers to reduce duplication without changing behavior: - Unify iosAnimatePush/Pop into parameterized iosAnimate - Unify animateSwipeComplete/Cancel into parameterized animateSwipe - Extract collectAnimations helper for repeated filter+waitAll pattern - Fix file header comment to match actual fill:"forwards" strategy Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(stackflow): make idle/exit positions transition-style-aware setIdlePositions and setPostExitPositions now accept TransitionStyle parameter. iOS-specific behind layer offset (-30%) and title/icon hiding only apply to slideFromRightIOS. Android and fadeIn transitions no longer incorrectly offset the behind activity. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor(stackflow): split transition-animation into dom.ts + animation.ts - Add data-part attributes to AppBarMain, AppBarIconButton, AppBarSlot to replace fragile className-based selectors - Split transition-animation.ts into: - dom.ts: DOM discovery (findTransitionTargets) and inline style management - animation.ts: WAAPI animation functions - Unify SwipeEndpoints with IOS_ONSCREEN/IOS_OFFSCREEN constants - Remove transition-animation.ts Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(stackflow): add push flash prevention for Android and fadeIn transitions Set start positions as inline styles before animation begins for fadeFromBottomAndroid (opacity:0 + translateY:8vh) and fadeIn (opacity:0), matching the existing pattern in iOS push. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(stackflow): read transitionStyle from DOM instead of stale React state useTopActivity().transitionStyle updates one render cycle late because it uses useState + useLayoutEffect. During push, the captured value was the previous activity's style, so fadeFromBottomAndroid/fadeIn always fell back to slideFromRightIOS. Add readTransitionStyle() that reads data-transition-style directly from the top activity DOM element at animation time, bypassing the stale state. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(stackflow): hide behind appBar after Android/fadeIn push For fadeFromBottomAndroid and fadeIn transitions, the behind activity stays in place (no parallax offset). After push completes, its appBar was visible behind the top activity's appBar. Hide it with opacity:0. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor(stackflow): clearAllStyles as foundation for position management Root cause: clearAllStyles didn't clear transform/opacity on appBar roots, leaving stale inline styles from pre-animation setup (e.g. Android push sets transform:translate3d(0,8vh,0) which was never cleaned up). Pattern change: setIdlePositions and setPostExitPositions now call clearAllStyles first (clean slate), then set only non-default positions. No more selective clearing that misses elements. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(stackflow): restore behind appBar before Android/fadeIn pop starts Behind appBar is hidden (opacity:0) during idle to prevent bleed-through. Pop must restore it before animation starts so it doesn't flash in at the end. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * docs: add changeset for AppScreen WAAPI transition migration Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(stackflow): cancel pending push rAF on swipe start If a push rAF is scheduled but hasn't fired yet when the user starts swiping, the rAF would execute after swipe begins and overlay push animations on top of swipe styles. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor(stackflow): extract GlobalInteraction constants to constants.ts * refactor(stackflow): introduce anatomy modules for data-part * refactor(stackflow): use anatomy constants for WAAPI DOM queries * chore(qvism-preset): remove unused stackflow animation/pseudo files * chore: regenerate css artifacts after dev merge * fix(stackflow): scrub app-bar background via WAAPI during swipe back The top AppBar background (::before) was not moving with the swipe gesture because nothing updated the pseudo-element's transform while the finger dragged. Add a WAAPI scrub animation that replaces itself on every touchmove and keeps the top app-bar background aligned with the displacement, uncovering the behind AppBar as intended. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(stackflow): scope WAAPI swipe tracking to slideFromRightIOS fadeFromBottomAndroid and fadeIn activities should not follow the finger during swipe back. Cache the top activity's transition style at gesture start and only apply displacement/complete animations for slideFromRightIOS; for other styles, rely on stackflow's normal exit-active lifecycle so the activity's own exit animation runs. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * refactor(stackflow): move AppBar background from ::before to real DOM slot The previous WAAPI implementation animated the top AppBar background via `el.animate(frames, { pseudoElement: "::before" })`. That option is only supported on Chrome 82+ / Safari 16.4+ / Firefox 97+, so on older WebViews (e.g. Chrome 77) the pseudo-element animation became a no-op and the front AppBar background stayed put during swipe back, hiding the back AppBar. Relocate the background into a dedicated `appBarBackground` slot rendered as a regular `<div>`. Query targets and the `scrubAppBarBackground` swipe helper now operate on the real element. - Add `background` slot to `appBar` recipe and anatomy - Render `<div data-part="appBarBackground" aria-hidden>` inside AppBarRoot - Move `::before` layout/background/divider styles to the slot - Expose `topAppBarBackground` / `behindAppBarBackground` on TransitionTargets - Regenerate css artifacts Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(stackflow): harden WAAPI push/pop lifecycle Three fixes bundled together since they touch the same transition coordination code path. 1. Pre-pin inline styles before push (`pinIosInlineStyles` helper). Browsers could paint the freshly mounted top activity at its final onscreen position for one frame before the compositor picked up `el.animate()`, so push animations appeared to start mid-way on heavy activities. Pin the iOS offscreen coordinates (and the Android / fadeIn equivalents) inline within the same task before invoking animate, guaranteeing the first paint lands at the start of the animation curve. 2. Polyfill `Animation.finished` for Chrome <84. Older WebViews ship WAAPI without the `finished` promise. The old `waitAll` simply called `a.finished.catch(...)` which threw and left the transition's cleanup handler unreachable — subsequent pops ran over stale inline styles and rendered a blank screen. `waitOne` now falls back to `onfinish` / `oncancel` listeners and races every animation against a `duration + 100ms` timeout so cleanup always runs even if the browser never dispatches finish. 3. Defer push animation one rAF for DOM readiness. Running push sync inside `useLayoutEffect` raced with stackflow's own subscription update — `[data-activity-is-top]` was not yet observable on the new top, so `findTransitionTargets` returned nulls and no enter animation played. Wrap the push path in `requestAnimationFrame` (guarded by `pendingPushRAFRef`) to let stackflow finish committing before querying targets. Pop stays sync since its top activity already lives in the DOM. Also keep the scrub animation `appBarBgScrubAnimRef` in sync with the new real background element renamed from the legacy pseudo naming. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * chore(changeset): document WAAPI stabilization fixes Extend the existing AppScreen WAAPI changeset with the three follow-up bullets for the jank, older-WebView, and pseudo-element regressions addressed in this branch. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * refactor(stackflow): clean up WAAPI hooks and tighten types (#1540) * refactor(stackflow): clean up WAAPI hooks and tighten types - Move prop callback handling out of getSwipeBackEvents factory into useSwipeBack to fix a rules-of-hooks violation (useCallback / useMemo / useCallbackRef were called inside another useCallback) - Add runningAnimsRef race guard to all finished.then handlers so a stale handler can't clobber a newer transition's pinned inline styles (the comment promised this but the guard was missing) - Add useEffect cleanup to cancel pending RAF / running animations on unmount so late-firing finished handlers can't run against a torn-down stack - Extract stopRunningAnims, stopAppBarBgScrub, cancelPendingPushRAF helpers to remove repeated cancel + null-out boilerplate - Introduce frozen IDLE_CONTEXT constant for swipe-back context resets - Tighten IosPositions title fields from Keyframe to a strict TitleKeyframe { opacity: string; transform: string } to drop casts - Drop redundant null guards in pinIosInlineStyles - Replace findTransitionTargets backwards loop with findIndex - Mutate swipeBackContextRef.current in place during 60fps moveSwipeBack to avoid per-frame object allocation - Drop unused setSwipeBackState, swipeBackContextRef, swipeBackStateRef from useGlobalInteraction return * fix(stackflow): always cleanup inline styles after WAAPI transition The race-condition guard added in d4c7547 was too strict: when stackflow fires a follow-up transitionState change near the end of a pop (e.g. an enter-active for the activity that becomes top), the new transition cancels the running animation, the cancel drops fill:forwards, and the previous guard returned without running setPostExitPositions / setIdlePositions — leaving the stale idle inline styles (behind layer at -30%) visible on screen. Reorder so cleanup runs unconditionally on finish: - Always run setPostExitPositions / setIdlePositions / pin + cancelAll. - Only the ref / per-gesture cache reset is guarded so a newer transition's state isn't clobbered. Repro: push → pop quickly. Pop ends with behind layer stuck at translate3d(-30%, 0, 0). * fix(stackflow): pin swipe-end duration to 350ms to match CSS-era timing Restore the pre-WAAPI swipe transition feel. The dynamic calculateSwipeDuration (150-500ms based on remaining distance / velocity) shortened cancel/complete transitions enough — especially for short gestures or fast flicks — to feel abrupt. Pin to IOS_DURATION (350ms) exactly as the CSS-era createPresence configured it. - Drop calculateSwipeDuration, MIN/MAX_SWIPE_DURATION constants. - Drop unused velocity parameter from animateSwipeComplete/Cancel. * fix(stackflow): keep swipe-time inline styles to avoid 1-frame snap endSwipeBack used to call clearAllStyles(targets) right before starting the WAAPI cancel/complete animation. On heavy main-thread sessions where a reflow/paint can interleave between the inline clear and the animation's first paint, the screen briefly snapped to default (transform: 0) before the keyframe[0] reapplied the swipe displacement — felt as a "stutter" mid-gesture. The clear is redundant: WAAPI keyframe[0] already matches the current swipe displacement, and onFinish's pin() resets inline state at the end. Drop the call so the inline values cover the 1-frame gap before the animation's first paint. --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Version Packages (#1526) Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> * docs(agents): strengthen component guidance (#1520) * docs(agents): strengthen component guidance from accordion review * docs(agents): clarify headless hook boundaries * docs(agents): refresh generic component guidance * docs(agents): clarify affix slot guidance * docs(agents): address guidance review comments * docs(create-component): clarify component review guidance * docs(stackflow): document modal-on-modal stacking for Bottom Sheet (#1547) Add a "Stacking Another Modal Overlay on Top" section under Keeping Bottom Sheet Mounted in the Bottom Sheet Stackflow guide. It explains when to use `modal` (always true) instead of `modal={isActive}` so the backdrop layering does not break when another modal Activity is pushed on top, with a tradeoff callout for the general AppScreen case. Add example activities under examples/stackflow-spa for live reference: - ActivityNestedBottomSheet: embedded into the new docs section via StackflowExample at /nested-bottom-sheet - ActivityBottomSheetWithAlertDialogStep: kept as an internal Step-pattern reference (not embedded in docs) to support 1:1 follow-ups when the Step-with-Step nested overlay case is asked Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(stackflow): tighten peerDependencies `@seed-design/css` to WAAPI compatible range (#1551) * fix(stackflow): tighten peerDependencies['@seed-design/css'] to WAAPI-compatible range * Update many-mails-sleep.md * Version Packages (#1553) Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> * chore(deps): update anthropics/claude-code-action digest to 9db782c (#1534) Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com> * chore(deps): update dependency chromatic to v16 (#1523) Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com> * chore: regenerate artifacts after dev merge - bun install: refresh lockfile - bun generate:all: regenerate css min bundles + lynx-css vars (top-navigation icon/text button vars added from #1542) * fix(changeset): drop private @seed-design/qvism-preset from lynx-checkbox-radio-group The changeset bundled a private package (@seed-design/qvism-preset, marked "private": true in its package.json) with two public packages (@seed-design/lynx-react, @seed-design/lynx-css). changesets refuses such mixed changesets and the Version Packages CI on lynx kept failing with: Found mixed changeset lynx-checkbox-radio-group Found ignored packages: @seed-design/qvism-preset Found not ignored packages: @seed-design/lynx-react @seed-design/lynx-css Mixed changesets that contain both ignored and not ignored packages are not allowed Removing the private package's bump line fixes the error. qvism-preset is private so it never gets a release tag/version anyway; the recipe-add context stays documented in the body text. * fix(docs): add leading './' to seed-design path mapping TypeScript 6.0 (merged from dev via c5012d1) tightens 'paths' validation and refuses non-relative entries when 'baseUrl' is not set, breaking the docs production build with: Type error: Non-relative paths are not allowed when 'baseUrl' is not set. Did you forget a leading './'? The lynx branch had 'registry/react/*' (no leading dot-slash), which used to work on the older TS version. Match dev's existing convention of prefixing the path with './'. --------- Co-authored-by: Joo Chanhwi <56245920+te6-in@users.noreply.github.com> Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com> Co-authored-by: 유세은 <101736358+SeieunYoo@users.noreply.github.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.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 join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
This PR contains the following updates:
905d4eb→e58dfa5Warning
Some dependencies could not be looked up. Check the Dependency Dashboard for more information.
Configuration
📅 Schedule: (UTC)
🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.
♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.
🔕 Ignore: Close this PR and you won't be reminded about this update again.
This PR was generated by Mend Renovate. View the repository job log.