New default merge path (merger.mode="ai"), self-contained in merger-ai.ts and
dispatched from ProjectEngine.onMerge instead of the legacy aiMergeTask pipeline
(kept for merger.mode="deterministic").
Flow: clean-room detached worktree at the target branch tip → AI agent merges
the task branch + squashes (resolving conflicts) → fresh read-only AI reviewer
audits with corrective retries (blocking vs advisory; advisory lands, unfixable
correctness hard-fails via AiMergeBlockedError; fail-safe verdict parsing) →
land via `git merge --ff-only` when the checkout is on the target (else
update-ref CAS) → sync the local checkout (stash → ff → restore; AI reconciles
a conflicting restore and keeps the original edits in a backup stash;
un-stashable dirt advances the ref + warns) → finalize (delete task branch —
never the integration branch — task→done, remove temp worktree).
- Per-task target branch honored (falls back to the default integration
branch); local checkout synced only when on that target.
- Structurally immune to the dirty-clobber and stale-base/non-FF bug classes of
the legacy path (clean room + FF-by-construction).
- Progress surfaced on the task status pill + task log stream.
- Clear error when the target branch has no local ref.
Settings: merger.mode / merger.reviewerModel / merger.maxReviewPasses, surfaced
in Settings → Merge; legacy merge-mechanics settings hidden when AI mode is on.
Tests: merger-ai.test.ts (verdict parser, clean merge, blocking hard-fail,
advisory land, empty no-op, target-branch isolation, missing-target error,
landSquash clean/other-branch/dirty-restore/AI-resolved). Legacy
merge-orchestration tests pinned to deterministic mode.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Implements graceful suppression of transient `task.json` ENOENT errors in the executor, logging a suppression signal and surfacing a banner in the UI, with test coverage for both the executor behavior and notification service. Documentation in `docs/architecture.md` and a changeset for `@runfusion/f
Fusion-Task-Id: FN-5624
Co-authored-by: Fusion (runfusion.ai) <noreply@runfusion.ai>
Fusion-Task-Id: FN-5624
React 19 removed the global JSX namespace; matches the pattern already used in agentHealth.tsx.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Implements source issue delete handling in the task dashboard, including prompt-based user interaction for source-linked issues, new GitHub tracking state logic in `packages/dashboard/src/github-tracking-state.ts`, and corresponding test coverage, with documentation added to `docs/task-management.md
Fusion-Task-Id: FN-5618
Co-authored-by: Fusion (runfusion.ai) <noreply@runfusion.ai>
Fusion-Task-Id: FN-5618
Adds padding expansion to the PR notice block in PrPanel.css with corresponding test assertions to verify the token values, completing the FN-5619 styling pass.
Fusion-Task-Id: FN-5619
Co-authored-by: Fusion (runfusion.ai) <noreply@runfusion.ai>
Fusion-Task-Id: FN-5619
Added OpenAI Responses API as a new custom provider type, wiring `apiType: "responses"` through the core registry, engine routes, and dashboard UI with a dropdown selector; includes test coverage across the registry, routes, and component layers.
Fusion-Task-Id: FN-5601
Co-authored-by: Fusion (runfusion.ai) <noreply@runfusion.ai>
Fusion-Task-Id: FN-5601
This merge implements an OAuth relogin banner feature (FN-5595) that displays in the dashboard when OAuth tokens expire. The feature includes a new `OAuthReloginBanner` component with styling and tests, an OAuth validity logger in the engine for tracking token state, and corresponding API route inte
Fusion-Task-Id: FN-5595
Co-authored-by: Fusion (runfusion.ai) <noreply@runfusion.ai>
Fusion-Task-Id: FN-5595
When the user clicks 'Latest' while loadMoreMessages was in-flight,
the scroll restoration (which maintains position when older messages
are prepended) was undoing the explicit jump-to-bottom.
Fix: clear scrollRestoreSnapshotRef and set isUserScrollingRef=false
before anchorToBottom so the layout effect's restoration is skipped.
Multiple coordinated fixes for the perceived "dashboard takes forever to
load" complaint. Per-page-load HTTP requests drop from ~177 to ~101 and
duplicate per-project InProcessRuntime creation is eliminated.
- engine: shouldUseHybridExecutor no longer auto-enables for local-only
multi-project setups (set FUSION_HYBRID_EXECUTOR=1 to force). The
duplicate-runtime path was running self-healing twice per project and
contending on the same SQLite file. ProjectEngineManager already
handles N local projects with one InProcessRuntime each.
- dashboard cli: parallelized independent store inits, started
CentralCore.init early in background, ran plugin loading concurrently
with extension resolution. Sequenced SQLite store inits to avoid a
TOCTOU race in addColumnIfMissing migrations across TaskStore /
AutomationStore / PluginStore / AgentStore (all open the same
.fusion/fusion.db). Restored try/catch around HybridExecutor.initialize
and engineManager.ensureEngine so a paused or broken cwd project no
longer aborts dashboard startup.
- dashboard client: added in-flight request dedupe wrapped around the
top API offenders. /api/plugins/ui-slots drops from 17x to 1x per load.
dedupe.forceFresh redirects ALL in-flight waiters to receive the fresh
post-mutation response, not just the forcing caller. Generation
counters in useAgents and AgentListModal protect against slow polls
overwriting fresh state.
- dashboard SSE: agent event handler now debounces 250ms with a
trailing-edge guard so multi-agent activity bursts coalesce to at
most 2 refetches per burst.
- dashboard route: PATCH /api/projects/:id with isolationMode change
returns 503 with actionable guidance when HybridExecutor is
unavailable, instead of silently persisting a config the live runtime
won't honor.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds a dedicated Secrets section to the SettingsModal navigation, removes the now-unused secrets footer callback wiring from App and AppModals, updates the related tests, and documents the secrets location in the dashboard guide.
Fusion-Task-Id: FN-5588
Co-authored-by: Fusion (runfusion.ai) <noreply@runfusion.ai>
Fusion-Task-Id: FN-5588
- Remove fetchAllMessagesInChat/fetchAllMessages helpers
- Keep limit:50 for initial load in useChat and useQuickChat
- Add IntersectionObserver sentinel at top of ChatView message list to
trigger loadMoreMessages() when user scrolls to the top
- Keep stale-session guards (activeSessionRef checks) from original PR
- Tests: revert assertions back to { limit: 50 }
Adds planning branch controls to the PlanningModeModal, extending the planning API and routes with supporting tests and documentation. The changeset bumps `@runfusion/fusion` as a minor release.
Fusion-Task-Id: FN-5585
Co-authored-by: Fusion (runfusion.ai) <noreply@runfusion.ai>
Fusion-Task-Id: FN-5585
Implements the allow-resurrection toggle for task deletion, letting users prevent deleted tasks from being automatically restored. Changes span the `ConfirmDialog` component, `TaskDetailModal`, and the `useConfirm` hook, with comprehensive test coverage across the dashboard API and UI layers.
Fusion-Task-Id: FN-5475
Co-authored-by: Fusion (runfusion.ai) <noreply@runfusion.ai>
Fusion-Task-Id: FN-5475
Three coupled fixes to make `pnpm test:full` exit cleanly when the local
`fn` dashboard is running:
1. scripts/check-test-isolation.mjs — replace timing-based "is the
engine writing?" heuristic with a deterministic check: if
`.fusion/engine.lock.lock/` exists (proper-lockfile's held-lock
marker), the dir is engine-active and auto-skipped from violation
reporting. The 2-second mutability probe is retained as a backstop
for dirs with another external writer but no live lock. Also adds
`engine.lock` / `engine.lock.lock/` to RUNTIME_IGNORE_PATTERNS so
a mid-test engine start/stop doesn't trip the signature compare.
2. packages/dashboard/.../__tests__/GitManagerModal.test.tsx — prune
the Status-panel Sync button + Recent-advances-events describe
blocks. Their UI was removed in 5d35b64bd ("remove duplicate
integration-advances UI") but the tests stayed and were timing
out at 1s each. The Remotes-panel Sync describe is kept because
the `remotes-sync-integration-tip-btn` still exists.
3. packages/engine/.../merge-reuse-task-worktree.slow.test.ts —
update the happy-path assertion to reflect 4c31e885b
("merger auto-syncs project-root checkout after ref advance").
Before that change, the merger's `update-ref` advance left the
project root's working tree stale, so `git status --porcelain`
would differ after the merge. With auto-sync, the new file is
tracked + clean at HEAD, so status doesn't change. Verify the
file actually landed via `git ls-files` instead.
After this, `pnpm test:full` exits 0 with the local dashboard running.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two cleanups in Git Manager → Status:
- Removed the duplicate "Sync local tip" button (gm-integration-actions)
and the second "Recent integration advances" list (gm-recent-advances)
that rendered above the highlighted block. Also dropped the dead
mergeAdvanceEvents state, fetcher, and SSE subscription that only
fed the deleted UI.
- Sync working tree is now pure-local. Added skipOriginFetch to
PullGitBranchOptions.integration (and the matching POST /api/git/pull
body field). When set, pullGitBranch skips tryFastForwardFromOrigin
entirely — the sequence is just auto-stash → git reset --hard
refs/heads/<integration> → restore stash. The Sync button passes
skipOriginFetch: true because the "N need action" recovery is for
catching the worktree up to a *local* merger ref-advance; touching
origin could silently pull in unrelated remote commits.
Help disclosure rewritten to reflect the pure-local behavior.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
`notice` is `events.find(...)` which returns `undefined` (not `null`)
when no match. `waitFor(() => expect(...).not.toBeNull())` exited
immediately because `undefined !== null` — the test never actually
waited for the api mock to resolve. Sometimes the followup assertions
happened to land after the events fetched (test passed by luck);
sometimes they ran while notice was still undefined and the assertions
failed.
Switched all five waitFor sites to `.toBeDefined()` so they actually
block on the events-fetch resolution.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Settings inheritance (high):
- Restored the value !== initialProjectValue gate on the non-model
project branch. Previously every effective/inherited project key was
persisted as an explicit override on every save.
Git Manager UI lie in remote-only mode:
- "Local <branch> vs origin" card now renders "no local tracking"
instead of a green "Synced" badge when integrationTipSource ===
"remote-only" (no local branch to compare).
- New dedicated "HEAD vs origin/<branch>" card surfaces a meaningful
distance in remote-only mode.
isIndexStale correctness:
- Walks up to 16 reflog entries so multi-hop misses (A→B→C without
sync) are detected; the prior check only consulted @{1}.
- Gated on isOnIntegrationBranch === true so a feature-branch worktree
whose HEAD happens to descend from <integration>@{1} no longer trips
the FN-INDEX-DESYNC warning.
Enumeration-failed events surfaced:
- collectRecentMergeAdvances pairs events with (taskId, newSha) when
both are present, falls back to taskId-only for early-failure events
(e.g. "enumeration-failed") that have neither path nor newSha. The
diagnostic outcome now surfaces on the matching advance instead of
being silently dropped.
aheadOfIntegration semantics no longer shift:
- Split into three distinct fields: aheadOfIntegration (HEAD vs local),
aheadOfIntegrationRemote (HEAD vs origin/<branch>),
aheadOfOriginIntegration (local vs origin). Consumers no longer have
to read integrationTipSource to know which comparison they got.
currentBranch failure no longer masks wrong-branch state:
- Distinguish "command threw" (transient git error) from "command
succeeded with empty stdout" (legitimate detached HEAD). New
currentBranchDetectionFailed field lets the UI surface "branch
detection unavailable" on a real failure.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Settings (data-loss):
- Non-model project keys now use null-as-delete instead of dropping
undefined via JSON.stringify, so clearing Integration branch (Use
dropdown / auto-detect) actually clears the persisted value.
isIndexStale (false-positive AND false-negative):
- Replaced the empty-worktree heuristic with a reflog-anchored check:
stale iff refs/heads/<integration>@{1} exists, HEAD is descendant of
it, and `git diff-index --cached <prevTip>` is empty.
Auto-sync attribution in collectRecentMergeAdvances:
- Match auto-sync events by (taskId, newSha) instead of taskId-only;
re-merged tasks no longer have older advances mislabeled with the
newest outcome.
- Compare worktreePath after realpathSync on both sides; macOS symlink
paths no longer cause permanent "needs action" false positives.
Extended path no longer 500s:
- Route wraps computeExtendedGitStatus in try/catch and falls back to
basic status on failure. Inner `branch --show-current` wrapped too
so detached HEAD / non-git rootDir doesn't throw.
Integration branch falls back to remote-only ref:
- When refs/heads/<branch> is missing, use refs/remotes/origin/<branch>
as the integration tip. New `integrationTipSource` field
("local"|"remote-only"|"missing") drives a UI badge.
Copy commit hash:
- Short-SHA copy is the default and matches what's displayed; a
separate "full" button copies the 40-char headSha. Previously the
single button silently copied the full SHA when extended was on.
Detached HEAD:
- isOnIntegrationBranch left undefined when currentBranch is empty so
the UI doesn't render "(not on <integration>)" against a
no-branch state.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replaces the plain text input with a <select> that lists the project's
local branches (loaded via fetchGitBranches when the Merge section
becomes visible) plus an (auto-detect) default and a Custom… option for
branches that don't exist locally yet.
Branch list is deduped + sorted with common integration names (main,
master, trunk, develop) pinned to the top so the typical case is one
click. Choosing Custom… swaps in a text input with a "Use dropdown"
link to revert.
A previously-saved value not in the loaded list (branch deleted
locally, or initial render before fetch resolves) falls through to the
custom input automatically so the operator can still see + edit it.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Repository Status panel now answers "what is the actual state of my
project root vs the integration branch?" so operators can be sure of
the picture even when the Merge Advance Notice banner is dismissed.
GET /api/git/status accepts ?extended=1 and returns additional optional
fields:
- integrationBranch + integrationBranchSource (settings|origin-head|fallback)
- integrationTipSha / originIntegrationTipSha
- aheadOfIntegration / behindIntegration (HEAD vs local integration tip)
- aheadOfOriginIntegration / behindOriginIntegration (local tip vs origin)
- dirtyDetails {staged, modified, untracked, conflicted, sample}
- indexStaleVsHead (surfaces the FN-INDEX-DESYNC scenario)
- stashCount
- recentMergeAdvances: up to 5 merge:integration-ref-advance events
joined with merge:auto-sync outcomes; needsAction flag flips when
auto-sync didn't successfully bring this worktree forward
GitManagerModal renders all of it:
- Existing cards get sub-text: branch shows "not on <integration>",
Working Tree shows staged/modified/untracked/conflicted breakdown
- Second row: Integration branch + source, HEAD-vs-integration,
local-vs-origin, stash count
- Yellow warning panel when indexStaleVsHead surfaces the merger's
stale-index situation with a recovery hint
- Recent integration-branch advances list, color-coded by needsAction,
shows the per-advance auto-sync outcome so operators can audit
even after dismissing the banner
All fetchGitStatus calls in GitManagerModal switched to extended:true.
Other callers unaffected — extra fields are optional.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds a text input for the canonical integration branch directly under
the Auto-completion mode select. Visible regardless of direct vs PR
mode since the setting applies to both.
Blank preserves the existing auto-resolution cascade
(integrationBranch → baseBranch → origin/HEAD → main). Setting it to
master/trunk/develop/etc. pins the resolution explicitly.
Field trims whitespace and stores undefined when cleared so the
auto-resolution stays active.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds the missing form control for the auto-sync mode introduced by the
merger hook. Lives next to Direct merge commit routing / Integration
worktree inside the merge-strategy panel and only renders when
mergeStrategy === "direct". Three options with descriptions cribbed from
docs/settings-reference.md:
- Stash + fast-forward (default): preserve local edits across the snap
- Fast-forward only: skip dirty worktrees, surface the banner instead
- Off: legacy behavior; project root stays stale until manual pull
Value is normalized through normalizeMergeAdvanceAutoSyncMode on both the
merged-settings and scoped-settings load paths so a missing/invalid
stored value cleanly falls back to the default.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Implements a pull-based merge workflow by wiring the merger pull helpers from the engine, extending the git pull and stash routes, and aligning the `MergeAdvanceNotice` and `StashConflictModal` components to gate dismissal on stash drop. The `run-audit` module is updated with pull mutation documenta
Fusion-Task-Id: FN-5419
Co-authored-by: Fusion (runfusion.ai) <noreply@runfusion.ai>
Fusion-Task-Id: FN-5419
Adds a push-to-origin workflow to the merge notice system, introducing a new `useMergeAdvanceNotice` hook, a `merge-advance-push-origin` route handler, and corresponding UI affordance in the `MergeAdvanceNotice` banner component. The engine gains TOCTOU and refusal audit assertions, and coverage exp
Fusion-Task-Id: FN-5359
Co-authored-by: Fusion (runfusion.ai) <noreply@runfusion.ai>
Fusion-Task-Id: FN-5359
The visualViewport ICB compensation (intended for Android ICB-stuck-large
and pinch-zoom) was also reacting to iOS soft-keyboard shrinkage, pushing
the mobile bottom nav up above the keyboard. Pass the existing
`keyboardOpen` prop through to a `mobile-nav-bar--keyboard-open` modifier
class that pins `bottom: 0` so the keyboard simply covers the bar.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds a `useOverlayDismiss` hook that suppresses touch-synthesized mouse events on modal/dropdown overlays to prevent unintended close behavior on touch devices, with tests covering TaskCard dismissal and overlay interaction edge cases. Documentation updates in AGENTS.md and docs/architecture.md capt
Fusion-Task-Id: FN-5482
Co-authored-by: Fusion (runfusion.ai) <noreply@runfusion.ai>
Fusion-Task-Id: FN-5482
Test coverage for mobile views: `MobileNavBar.test.tsx` now syncs its mock with the viewport constant, and `SecretsView.mobile.test.tsx` derives secrets media parsing from the hook query.
Fusion-Task-Id: FN-5521
Fusion-Task-Lineage: eb37d760-ab67-49c5-9b57-2c14d648a98a
Co-authored-by: Fusion (runfusion.ai) <noreply@runfusion.ai>
Fusion-Task-Id: FN-5521
- Header.css: mobile padding-top is additive (var(--space-md) + env(safe-area-inset-top)) so the brand row keeps its 12px breathing room below the Android status bar instead of having it replaced by the inset.
- ExecutorStatusBar.css: bottom offset now uses max(env(safe-area-inset-bottom), 12px) to match the floor MobileNavBar already applies, so the footer lands flush on top of the nav instead of inside its padding band.
- useToast.ts: silently drop bare "Failed to fetch" error toasts (from fetch() aborts on tab background/resume); toasts with additional context still pass through.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Android Chrome under viewport-fit=cover intermittently reports
env(safe-area-inset-bottom) as 0 while the address bar is visible or
during URL-bar collapse, causing the bottom nav to render beneath the
system gesture pill. Floor the inset to 12px in MobileNavBar's own
padding and the matching content-padding rules so the bar always clears
the gesture area; devices that report a larger inset are unchanged.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Switch commits Fusion produces (both executor step commits and merger squash
commits) from setting `--author="Fusion <noreply@runfusion.ai>"` to appending
`-m "Co-authored-by: Fusion <noreply@runfusion.ai>"`. The user's configured
git identity now stays as the primary author/committer, and Fusion is recorded
as a co-author (recognized by GitHub for shared attribution). The
`commitAuthorEnabled` toggle and `commitAuthorName`/`commitAuthorEmail`
settings keep their existing keys; the dashboard settings UI relabels them
from "Author" to "Co-author" to match.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two engine merger tests created mkdtempSync workspaces directly in tmpdir()
under the tracked `fusion-test-` prefix; under full-suite concurrent load
the post-run check-test-isolation flagged them as leaks. Route both
(`merger-no-op-fix-finalize.test.ts`, `merger-verification-fix-already-on-main.test.ts`)
through FUSION_TEST_WORKER_ROOT like sibling merger tests so they nest
inside the already-tracked worker root.
Bump engine vitest subprocess guard from 60s to 120s and testTimeout to
30s — plain git commands (branch -d, worktree remove) queued behind
system contention during `pnpm -r --workspace-concurrency=2` runs were
timing out. The guard only fires on hangs, so healthy tests pay nothing.
Also bundles in-progress dashboard mobile-breakpoint regex/CSS test
updates and docs index additions.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
ChatView and QuickChatFAB both had an iOS-specific onTouchStart on the
textarea that called event.preventDefault() and then programmatically
re-focused the input — meant to suppress iOS's visualViewport auto-scroll
on re-focus. On Android, preventDefault on a textarea touchstart blocks
the soft keyboard from opening (programmatic focus() alone does not raise
the Android keyboard — only the default touch action does), so tapping
the main chat or quick chat composer focused the input but the keyboard
never appeared, looking like an instant dismiss.
Gate the touchstart workaround to iOS via isIOS().
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Some Android Chrome builds (multi-window / split-screen / certain WebViews)
leave window.innerWidth/Height stuck larger than the actual rendered canvas.
DOM, body, and visualViewport report the true dimensions, but position:fixed
uses the ICB, pinning fixed-bottom elements offscreen below the visible area.
JS-side meta override (setAttribute and full replacement) does not force
Chrome to recompute the ICB on those builds.
index.html now publishes the ICB→visualViewport delta as CSS variables
(--icb-bottom-offset, --icb-right-offset) on <html>. MobileNavBar.css and
ExecutorStatusBar.css consume them so the bars pin to the visible viewport
edge regardless of ICB drift. Math is visualViewport-relative so it also
handles pinch-zoom in (offsets compensate) and pinch-zoom out (clamp at 0).
Healthy browsers see 0px and behave unchanged.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Broaden mobile media query to include (max-height: 480px) so landscape
phones (which exceed 768 CSS px wide) still render the bottom nav and
mobile board layout instead of desktop horizontally-scrollable columns.
- Guard useMobileKeyboard against pinch-zoom (vv.scale > 1) — Android
Chrome ignores user-scalable=no, and a focused textarea + zoom was
false-positiving keyboard-open and hiding MobileNavBar.
- Read documentElement.clientHeight instead of stale window.innerHeight
when computing keyboard overlap (Android multi-window can leave
innerHeight cached at a wildly different value than the actual layout
viewport — observed 2848 while html was 797).
- Add interactive-widget=resizes-content to the viewport meta so Android
Chrome shrinks the layout viewport with the soft keyboard, matching iOS.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds a 255-line fast-path regression harness (`self-healing-fn-5488-fast-path-regressions.test.ts`) covering unbacked merge fast-paths, retry-exhausted fanout, grace-window boundaries, and overlap preservation invariants. Also includes a small cleanup to `MobileNavBar.test.tsx`.
Fusion-Task-Id: FN-5498
Fusion-Task-Lineage: f39ca8a1-223b-4b43-a844-6cb17d19e099
Fusion-Task-Id: FN-5498
After a long session of guess-and-check without remote debugging, none
of the viewport/portal/keyboardOpen interventions cleanly fixed the
Android Chrome multi-window layout bug. Reverting index.html,
MobileNavBar.tsx, Board.tsx, and the FN-5476 overlap test back to
their pre-session state. The board-squeeze fix in styles.css is kept
(confirmed working).
The remaining workaround: turn off "Request Desktop Site" on the
Android tablet for an acceptable layout. Revisit with chrome://inspect
remote debugging when available.
Reverts (squashed):
- 49839245e portal MobileNavBar
- 4bd9711e8 drop keyboardOpen guard
- d2e31dc10 don't fight pinch zoom
- f881ed22e visualViewport.width override
- 301e0a050 self-healing overlay
- 16b18bf33 publish mNavDebug
- d3dff9442 extend vpdebug
- a9094d79d restore overlap test
- 94bc43e04 restore FN-5476 observer
- 5a9095e54 re-drop viewport flags + add vpdebug
- and the index.html/Board.tsx portions of fbf7e2cb4
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
vpdebug on Android tablet captured the nav bar's getBoundingClientRect
at y=2800 with html height 798 — meaning some ancestor was creating a
containing block for the bar's `position: fixed`, putting it 2000+ px
below the visible window. Any ancestor with `transform`, `filter`,
`will-change`, or `contain: paint` would do that, and it's hard to
audit which of our many ancestors (providers, view wrappers, plugin
hosts) introduces one in a given Android Chrome state.
Sidestep the whole question: render the bar via createPortal into
document.body. Its only ancestor is body, so `position: fixed; bottom: 0`
is now guaranteed to pin to the visual viewport bottom.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
vpdebug overlay on Android tablet captured the smoking gun: with the
corner-cutoff fixed (viewport now full window), the bar still disappears
because useMobileKeyboard's keyboardOpen state can get stuck at true.
Repro: the Planning column textarea autofocuses on mount; Android Chrome
fires a transient visualViewport.resize during focus that registers as
"keyboard open", but the matching dismiss event never fires once the
keyboard never actually appeared. visualViewport stays at full height
yet useMobileKeyboard's state remains true, and MobileNavBar's render
guard hides the bar forever.
The original guard existed to keep the bar off the iOS soft keyboard.
On Android the keyboard pushes content up so the bar isn't covered
anyway; on iOS the worst case (bar overlapping keyboard) is far less
broken than the bar permanently vanishing on every Android tablet
session. Drop the keyboardOpen check from the render guard.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Surface the live mode / modalOpen / keyboardOpen / footerVisible / view
values that MobileNavBar uses for its early-return so the ?vpdebug overlay
can show which one is hiding the bar on Android. Also dumps the
.project-content className so we can correlate with `--with-mobile-nav`.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Add core settings schema/types support for testMode with model-resolution override handling
- Enforce engine session lane overrides in test mode with targeted helper coverage
- Add dashboard settings toggle plus persistent test-mode banner and related component tests
- Update settings documentation and parity/roundtrip tests for the new test mode behavior
- Drop the tablet-tier `.board` grid rule that crammed 6 columns into
≤1024px viewports with no min-width, scrunching column content to
unreadable widths. Tablets now use the default `minmax(300px, 1fr)`
and scroll horizontally like desktop.
- Drop `maximum-scale=1.0, user-scalable=no` from the viewport meta.
Combined with `initial-scale=1.0` those flags trigger Android Chrome
layout bugs in multi-window mode; the Capacitor-feel justification
isn't worth the breakage in a browser-rendered dashboard.
- Broaden the existing iOS scroll-snap stabilization in Board.tsx from
`(max-width: 768px)` to any touch-primary device, and re-run it the
first time tasks populate so Android tablets get the same first-cards-
loaded reflow that mobile already had.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
