Commit Graph
17586 Commits
Author SHA1 Message Date
OrcaWinandOrcaWin 651f4f3928 test: remove shallow find suite and duplicate parameter rows (#43209)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-16 10:54:51 -08:00
Brennan Benson 4c5052d4f6 fix(editor): reopen unsaved drafts of workspaces that disappeared while Orca was closed (#42943)
* test(editor): keep a draft whose workspace disappeared while closed

* fix(editor): keep unsaved drafts of workspaces gone at startup and offer Reopen

* fix(editor): settle recovered stranded drafts by identity; keep them over a clean returning tab

* fix(editor): reopen stranded drafts as dirty untitled tabs in the active workspace

* fix(editor): stamp the recovered tab's empty baseline and retry stranded drafts on host changes

* refactor(session): move the editor session builder to its own file

* fix(editor): retry stranded drafts after any dropped trigger, including paired-server SSH changes

* Revert "refactor(session): move the editor session builder to its own file"

This reverts commit 47d3a79eb8.

* refactor(session): move the terminal session builders to their own file

* refactor(session): keep terminal builders in place; move only the tab sanitizer

* fix(editor): don't flag a recovered draft's own placeholder create as a change on disk

* fix(editor): retry stranded drafts when the active workspace's host changes
2026-11-16 10:54:46 -08:00
Brennan BensonandClaudio 41005c9603 Claude sessions restart on the new account right away when you switch accounts (#37943)
* fix(accounts): restart busy Claude sessions at once when the account is switched

A restart accepted after a Claude account switch waited for each busy
session's turn to end, so long turns, subagents and background shells
kept spending the old account until it hit its limit. Accepted restarts
now interrupt the session and resume it on the new account right away,
the same as Codex.

* fix(accounts): say in the restart setting that busy sessions stop their work

* fix(accounts): keep each tab's queued Claude restart selection local to that tab

* fix(accounts): run a queued account restart once its pane's terminal attaches

* fix(accounts): wake queued restarts from the leaf bind path too

* test: mark the restart mount-race test's stub casts as safe

* fix(accounts): run the queued-restart wake without a new React hook

* Wake queued account restarts when mounted ownership ends

* Move or report every running Claude terminal on an account switch

A switch only saw panes whose hook status row said Claude, so Agent Teams panes (an `Orca` foreground), panes with hooks off, and panes whose row was dismissed were skipped without a word. The scan now also finds Claude from a saved session record, the pane's own process read, or a one-pane tab's launch; resumes from the saved record when the row is gone; and, when nothing on the outgoing account moves, warns about running sessions Orca can't resume that may still be on it, with Review.

Tracked in STA-15061.

Co-authored-by: Claudio <claudiofernandopinto@gmail.com>

* Keep Claude account restart discovery aligned with saved sessions

* Restart WSL Claude panes only with live evidence

* Keep the original conversation pane during account restart

* Keep retrying account restarts in their conversation pane

* Place the restart fixture cast rationale on its assertion

---------

Co-authored-by: Claudio <claudiofernandopinto@gmail.com>
2026-11-16 10:51:19 -08:00
Brennan Benson f912850762 fix(native-chat): send other agents' mail as a pointer and let the person approve a denied action (#43093)
* fix(native-chat): send other agents' mail as a pointer and let the person approve a denied action

Another agent's orchestration message reached a structured chat's agent as
ordinary user input, so Codex's automatic reviewer counted a coordinator's
words as the user's authorization. The chat card still shows the whole
message, labeled with its sender; the agent is now sent Orca's existing
pointer at `check`, so the content arrives as tool output. Mail stays unread
until `check` takes it, and a new coordinator is pointed at mail the old one
never read.

Codex automatic-review denials now show as a chat row with the reviewer's
reason and an "Approve once" button, which only this computer's own Orca
window may press; it uses Codex's thread/approveGuardianDeniedAction and
records the approval in the chat.

Tracked in STA-15406.

* fix(orchestration): wake any new mailbox owner after re-pointing stale mail

* fix(native-chat): settle check:changed for pointer mail and denial approval

Move the Codex approval and mobile review copy into their own modules, reuse the shared record guard, and update the tests that encoded whole-mail consumption to pointer delivery with mail read at check.

* style: format review notice copy

* fix(native-chat): hold accepted mail while custody is unreadable and show the whole denied action

* fix(orchestration): hold in-doubt mail while custody is unreadable and wake the new owner after a send-time handoff

* test(orchestration): keep crew reads off a chat host another test file left installed
2026-11-16 10:50:24 -08:00
OrcaWinandOrcaWin 9230d5041a test: remove synthetic availability and shallow helper examples (#43207)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-16 10:28:30 -08:00
Brennan Benson 27ad794b48 AI-vault title sync stops rebuilding the conversation list every 20 s (#43142)
* perf: reuse AI-vault title requests across a sync pass until an input changes

The title pass rebuilt every tab's conversation 2 + 2 x batches times per
20 s pass. Keep one request snapshot per sync and rebuild it only when a
conversation, tab, layout or owner input changed; request timing and
titles are unchanged.

* perf: look up saved titles once per pass and keep only request inputs

Index the request workspaces' saved titles once per pass instead of
searching a workspace's tabs per request, and keep only the store fields
the request list reads so an old snapshot doesn't pin unrelated data.

* fix: read a request's saved title from its own workspace

Two workspaces can hold the same tab id; key the per-pass lookup by
workspace and tab, keeping the first match like the per-workspace search.

* chore: build title request inputs from same-name copies

The layout-writer ratchet reads field-name string lists as computed-key
writes; spell the picked inputs out instead. Also apply formatting.

* chore: move title request inputs to their own file and read workspace records via selectors
2026-11-16 10:26:20 -08:00
Brennan BensonandClaude Opus 5.5 4ecea0a693 fix(native-chat): resend a subscriber's snapshot when its capture was abandoned (#43020)
* fix(agent-session): resend a chat's first snapshot when its capture is abandoned

A subscriber opened without a cursor gets a snapshot first. When live
authority moved during that read (a hand-off finishing, a start going
live), the capture was abandoned and the following delivery sent batches
from sequence 0 instead, so the client never received the conversation
and stayed empty. The subscriber now owes that snapshot until one is
emitted, and the next delivery sends it.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* test(native-chat): pin the owed snapshot after an abandoned capture

The opening snapshot's authority changes mid-read, so its capture is
abandoned; the next delivery must send that snapshot before any other
frame. Without the owed-snapshot check the client's first frame is a reset.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-11-16 10:24:19 -08:00
Brennan BensonandClaude Opus 5.5 9e685dc8a6 fix(native-chat): an upgrade no longer reopens closed chats or copies unchanged ones (#43007)
* test(native-chat): a chat store from before revisions were kept keeps its tabs on upgrade

* fix(native-chat): an upgrade no longer reopens closed chats or copies unchanged ones

A chat store copied by a build from before revisions were kept reopened every
closed chat and copied every converted one on its first launch after #42783,
with nothing written by an older Orca.

- The original's receipt is compared by meaning: a receipt saved before WAL
  identity was kept matches on the fields it has, so the newer shape alone is
  no write.
- A chat with no revision here takes what this file shows as its baseline. A
  converted chat is kept as a copy only when the original holds an item it
  never saw; one not converted yet converts from the current original. Only a
  chat this file doesn't know joins, and only when the original changed.
- So only a chat an older Orca started or positively continued gets a tab;
  every other chat keeps its open or closed state.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* fix(native-chat): tell a continued chat by its restored history, not raw rows

An older Orca saves a chat's turns as outcome records, and a rewind can retire
an alias, so a converted chat's raw journal rows are not what it shows. With no
revision to compare, a converted chat is now kept as a copy only when its
restored history, as conversion reads it, has an item the chat here never
held; nothing depends on save timestamps. An original whose identity can't be
certified records its baselines instead of rereading every launch.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* fix(native-chat): an original as copied takes every chat as shown

With no write since the copy, a converted chat this store moved on can't be
mistaken for one an older Orca continued, and the pass restores nothing. The
test of a store from before revisions were kept now checks that its copy saw
the original's WAL, as in the reported profile.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* fix(native-chat): keep deleted empty chats deleted and finish a partial pass

- On a store copied before revisions were kept, every chat with history was
  cataloged at the copy, so only one with history is new when this store
  doesn't know it; a deleted chat with no history stays deleted.
- A converted chat counts as continued also when the older Orca revised an
  item past the revision it holds.
- A pass where one chat failed still records the baselines it read, so only
  that chat is retried next launch.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* fix(native-chat): decide unknown chats by the copy's catalog alone

A chat with history always has a catalog row from the copy and one without
history has no revision, so the record lookup added nothing. The downgrade
test's original records its tab index, as every older Orca that saves tabs
does.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* style: format

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-11-16 10:11:30 -08:00
Brennan Benson ef36e6c1ca Remote sender updates rejoin only their host and notify only the cards they move (#43112)
* Remote sender updates rejoin only their host and notify only the cards they move

With paired hosts, every remote agent-sender scope update rebuilt the window's whole sender index, including every unchanged local row, and every mounted sender label re-ran its store selectors on every store write; the sender-locations channel told every card to re-resolve.

The index is now partitioned: each publication is joined to this window's panes once (per alias change), each host's part is rebuilt only when that host's scopes change, and a sender's answer is resolved on demand from the host parts, with the cross-host principal/attachment joins and the pane-owner collision check read through reverse pane maps. A publication update notifies only the subscribed sender channels whose answer it moved; tab, owner, handle and publishing-host changes still notify every card. The label reads through the same leased selector as the avatar and link.

Resolution is unchanged: an equivalence test checks every sender against the old whole-window rebuild over random local and remote scope updates, cross-host joins, ambiguity and aliased chats.

* Build the sender label's selector outside render

* Type the equivalence test's hosts without the moved execution-host module
2026-11-16 10:10:14 -08:00
OrcaWinandOrcaWin 52b72b2f56 test: remove source-export mapping tests and empty browser TODOs (#43204)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-16 10:09:19 -08:00
Vitalii KulishandNeil a5b7b499fa fix(agent-availability): ignore unexpanded ~ PATH entries when checking chat programs (#43185)
Some installers write a literal `~/...` entry to PATH (e.g. the macOS .NET
installer's /etc/paths.d/dotnet-cli-tools). observeChatProgram treated it as
workspace-relative and reported every uninstalled chat program as `unknown`,
so the new-tab menu showed "Couldn't check chat availability" instead of
"Not installed" or the setup chat. execvp/libuv never expand `~`, so bare `~`
and `~/`/`~\` entries no longer force workspace resolution; other relative
entries still yield `unknown`.

Fixes #43184

Co-authored-by: Neil <neil@stably.ai>
2026-11-16 10:02:40 -08:00
OrcaWinandOrcaWin b091d23dff test: remove low-value mobile projection and permission preset tests (#43199)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-16 09:45:44 -08:00
OrcaWinandOrcaWin 32e9c4a59c test: remove skill target, port routing and branch helper suites (#43196)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-16 09:35:32 -08:00
OrcaWinandOrcaWin b0426b563d test: remove low-value suites and repeated cases (#43187)
* test: remove narrow renderer and chat visual examples

* test: retire resource and PR target projection suites

* test: retire isolated renderer adapters and repeated mod lifecycle cases

* test: retire route adapter and skill runtime projection suites

* test: retire Jira board preference helper examples

* test: await queued edit acquisition before Activity switch

---------

Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-16 09:18:41 -08:00
Kelvin Amoaba df23e549d3 fix(native-chat): refresh the /mcp tray when the chat starts or /mcp is run again (#43188)
The tray read the chat's MCP servers once when it opened. Opened while the agent was at rest, it said "Connect when you send your next message" and kept saying it after the agent started, and running /mcp again on the open tray did nothing. The tray now re-reads when the host reports the agent starting or stopping and each time /mcp runs, on desktop and mobile. A read asked for while one is in flight now runs after it instead of being dropped.
2026-11-16 16:55:10 +00:00
Neil 3ae47e0464 refactor: organize workspace composer modules by owner (#43182) 2026-11-16 08:16:39 -08:00
tiantiantianblueandNeil 906d697ba6 fix(mobile): canonicalize runtime host IDs from workspace routes (#43165)
Expo Router decodes query params twice, so a WSL runtime host id such as
runtime:wsl%3Aubuntu-22.04 reached screens as runtime:wsl:ubuntu-22.04 and
never matched the desktop host catalog, leaving the workspace on the
offline screen. Read the route executionHost through a shared
canonicalExecutionHostId that re-encodes runtime and SSH targets, and use
it in the workspace route, notification viewing state, and web page
location readers.

Fixes #43160

Co-authored-by: Neil <neil@stably.ai>
2026-11-16 08:02:16 -08:00
OrcaWinandOrcaWin 8ae7f0292d test: remove sidebar disclosure and narrow divider UI suites (#43138)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-16 07:57:13 -08:00
NeilandClaude d35be59b4c fix(sidebar): use one chevron convention across the agent tree (#43177)
Child-agent and idle-group disclosures pointed up when expanded while the
agents group header (and other sidebar trees) point down when expanded and
right when collapsed. Rotate the child disclosures the same way.

Fixes #43169

Co-authored-by: Claude <noreply@anthropic.com>
2026-11-16 07:51:58 -08:00
Kelvin AmoabaandMikel Vallejo c7f83f4e19 fix(computer): stop macOS permission status probes from leaking suspended helpers (#43172)
Spawn the status helper directly instead of through LaunchServices, and have it re-exec itself as its own responsible process so TCC still checks the helper's grant. A timeout now stops the probe by PID, so a late LaunchServices launch can no longer leave a suspended helper behind. Startup kills suspended probes left by older builds.

Co-authored-by: Mikel Vallejo <58292517+VallejoM@users.noreply.github.com>
2026-11-16 15:26:25 +00:00
Kelvin Amoaba 1a23722a74 fix(browser): check the Browser Use skill on the host the settings pane shows (#43170) 2026-11-16 15:21:37 +00:00
Neil 9f96bd57e3 refactor: organize tab and workspace modules by owner (#43167) 2026-11-16 07:17:09 -08:00
Brennan Benson f727d4e5b0 Reuse unchanged jump palette search entries (#43148)
* Reuse unchanged jump palette tab entries

* Preserve palette ordering across skipped status writes

* Apply palette cache lint requirements

* Keep palette cache derivation compatible with static checks

* Restore palette worktree type and memo factory call
2026-11-16 07:01:44 -08:00
Neil 4ae7ff27b7 refactor: organize filesystem IPC and local file I/O (#43120)
* refactor: organize filesystem IPC and local file I/O

* test: retire unreachable filesystem rename alarm
2026-11-16 05:32:37 -08:00
Tauri-EPOandNeil 2bc8fc1ad5 fix(codex-accounts): broadcast Codex account selection changes to windows (#17265)
Selecting, removing, or adding a managed Codex account from the CLI or a
paired phone updated settings without notifying listeners, so the desktop
status bar kept naming the previous account over the new account's usage.
Route every Codex selection write through CodexAccountSelection.saveSettings,
which passes notifyListeners: true, matching the Claude side.

Fixes #43107

Co-authored-by: Neil <neil@stably.ai>
2026-11-16 04:56:13 -08:00
OrcaWinandOrcaWin 93a3aa2f1a test: remove low-value tests (#43099)
* test: remove low-value notification and display tests

* test: remove update preference and host mute predicate suites

* test: remove redundant daemon and checkpoint work-count tests

* test: remove narrow desktop pane-title forwarding tests

* test: remove title-card and title-input work-count tests

* test: remove duplicated tool-pairing adapter tests

* test: remove narrow appearance preference tests

* test: remove turn-status display timing tests

* test: remove visual Activity read-count tests

* test: remove Linear view preference resume tests

* test: remove synthetic pending-question work-count suite

---------

Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-16 04:40:57 -08:00
Brennan Benson 173beb41a2 Give new agents plain UUID IDs (#42940)
* Give new agents plain UUID IDs

New Orca session IDs (chats, host-launched chats, terminal agent IDs) are
plain random UUIDs minted by createOrcaSessionId; the agent stays on the
record. Remove the two readers of the agent prefix: the agent.launch
sessionId refinement and the automation Open run guard. Hosts advertise
agent.launch.plain-session-id.v1; the phone reserves a plain ID only for
hosts that advertise it and keeps the legacy spelling otherwise. A replay
an older replacement host refuses as invalid params is now uncertain, not
a definite failure. Old IDs stay valid; nothing is migrated.

* Repair plain agent ID recordings and compatibility checks
2026-11-16 04:20:56 -08:00
Neil 6018fca4de refactor: group shared settings, profiles and execution hosts (#43047)
* refactor: place shared settings profiles and host contracts with owners

* fix: retire stale cleanup validation contacts

* refactor: consolidate imports from concrete owners

* refactor: bind new card fixture to profile defaults owner

* style: format rebased concrete host import

* refactor: consolidate project header color owner imports

* fix: load released phone readers after folder moves

* fix: preserve released chat readers and current owner imports
2026-11-16 04:03:33 -08:00
81835f3582 fix(agent-session): end a subscriber's stream when a queued delivery fails (#43043)
A failed queued delivery (catch-up read, frame capture, work projection)
dropped the subscriber without ending its stream, so the client never
resubscribed and the structured chat froze. Route both delivery-failure
paths through one guarded drop-and-end that skips subscribers already
replaced by a same-id successor.

Fixes #43030

Co-authored-by: Neil <neil@stably.ai>
Co-authored-by: Hayden Flinner <14845292+haydenflinner@users.noreply.github.com>
2026-11-16 03:30:51 -08:00
Brennan Benson ece3008ad5 Pause hidden chats to speed up switching (#42710)
* refactor: give chat tabs ownership of composer editors

* Retain pasted-text file paths with composer undo history

* Keep composer selection ownership with its editor

* perf: suspend hidden structured chat effects with Activity

* fix: keep pending chat actions independent of view replay

* Preserve hidden chat handoff and pending host actions

* Keep autocomplete ref writes in their owning hook

* Preserve disclosure geometry and pending images across chat reveals

* Use the native Blob implementation for image lifetime coverage

* Retain sent messages while hidden chats await history

* Wait for the held send request before hiding its test pane

* Preserve pending host reconnect state across chat switches

* Reset retained outage state after a hidden host recovery

* Observe retained chat outage transitions through shared host subscriptions

* Start outage observation for initially hidden chats

* Keep chat feedback timers alive across hidden views

* Settle the retained rail while its view ref is disconnected

* Track the rail node lit by the latest wake

* Preserve the lit rail when hidden input has no view node
2026-11-16 02:35:23 -08:00
Jinjing 7dad69cfc2 Preserve unopened chat statuses during capacity eviction (#43061) 2026-11-16 02:21:03 -08:00
fd8198f001 feat(jira): reorder the Tasks board lanes by dragging their headers (#30597)
* feat(jira): reorder the Tasks board lanes by dragging their headers

Jira's board configuration decided the lane order and there was no way to change it in Orca. Column headers are now drag handles, with ArrowLeft/ArrowRight as the keyboard equivalent, and the resulting order is saved per board on the device. Jira's Agile API exposes board column order read-only, so this is an Orca view preference and never edits the Jira board.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

* fix(jira): make the saved lane order survive Jira column inserts and filtered lanes

Review follow-up. Board column keys were derived from the column index, so inserting a column in Jira orphaned every saved lane position; they now come from the column's status ids, which Jira keeps stable across config edits. Saving a reorder also replaced the stored order with only the lanes currently on screen, dropping the saved position of any lane a filter had hidden; the new order is merged into the saved one instead. A keyboard move now restores focus to the lane it moved, and the rendered probe asserts two arrow presses in a row land.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

* fix(jira): tell screen readers a lane's issue count and its arrow-key move

Review follow-up. The lane header's aria-label replaced the visible header text, so assistive technology no longer heard how many issues the lane held, and nothing said the header moves with the Left and Right arrow keys. Both now ride on aria-description, translated in every catalog including the newly added Portuguese one, which also gains the existing reorder label.

Also registers the lane-order storage module in the WSL-move localStorage inventory, as the sibling view-mode change did, and follows the jira-types move on main.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(jira): keep the saved lane order across a runtime reconnect

Review follow-up. The lane order was stored under the status-order cache key, which ends in the provider runtime session generation. That counter restarts at zero on every launch and moves on every runtime focus change and remote session start, so a reconnect or a relaunch read the order back under a different key and silently fell back to Jira's order.

The storage key is now built from the task source, the owner's identity without the generation, and the Jira project. The generation-free identity is split out of getEnvironmentRuntimeContextKey rather than duplicated, and the status-order cache keeps its generation.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-authored-by: Neil <neil@stably.ai>
2026-11-16 02:12:19 -08:00
Brennan BensonandoutworkAnton cbd75eb008 fix(acp): keep saved chats retryable after temporary restore failures (#42981)
* test(acp): cover temporary restore recovery and cancellation

Co-authored-by: outworkAnton <anton.mazurov@reksoft.com>

* fix(acp): preserve saved conversations through temporary restore failures

Share temporary provider-answer classification and bounded abort-aware retries. Isolate each load attempt so partial failed replay cannot enter successful history catch-up.

Co-authored-by: outworkAnton <anton.mazurov@reksoft.com>

* fix(acp): end automatic startup after the restore retry budget

* test(acp): use the advertised resume preference type

* fix(acp): retry temporary moved-session lookup failures within the same flight

---------

Co-authored-by: outworkAnton <anton.mazurov@reksoft.com>
2026-11-16 02:05:16 -08:00
Nikolai ShcheglovandNeil ba1f8f9eb4 fix(browser): open SSH links in Orca when the host browses from this device (#41589)
With "open links in Orca" on and an SSH host's browsing set to "This
device", links from SSH and managed-SSH workspaces (terminal, native chat,
markdown, dashboard preview, app links) skipped Orca's browser and went to
the system browser, even though New Browser Tab already opened a desktop
tab in that workspace.

canOpenRemoteLinkInWorkspaceBrowser now owns the remote-link rule in
lib/browser and passes allowSshDeviceBrowser (only while links open in
Orca) to the SSH route predicates, and executeHttpLink forwards the same
flag to openWorkspaceBrowserTab. Ports, links-in-Orca off, modifier
escapes, paired runtimes, WSL and local links are unchanged. The managed
SSH egress indicator now shows "Browsing from this device".

Fixes #41586

Co-authored-by: Neil <neil@stably.ai>
2026-11-16 02:04:51 -08:00
Neil 3a21fb4106 fix(plugins): re-fetch the catalog before an update preview (#43063)
* fix(plugins): re-fetch the catalog before an update preview

Check for update on a plugin previewed against the cached catalog, so a
release published since the last fetch showed as 'X -> X, already
installed'. The update preview now refreshes the plugin's own catalog
first (one fetch shared by a multi-plugin review), falling back to the
cache when offline.

* fix(plugins): refresh from the lock entry's catalog, so a lost listing still refreshes

* ci: retrigger PR checks
2026-11-16 02:04:18 -08:00
Jinjing 8e5fcac855 Keep completion checks visible for unread agents (#43065) 2026-11-16 01:49:24 -08:00
Neil 571bdb69db fix(sidebar): omit done and active counts from child workspaces toggle (#43064) 2026-11-16 01:48:03 -08:00
026fccf821 feat(notifications): mute notifications per worktree (#26779)
* feat(notifications): mute notifications per worktree

Add a per-worktree notification mute, toggled from the worktree context
menu. Muting silences the interruptions for that worktree (OS banner,
notification sound, mobile push, tray attention dot) while keeping the
passive unread indicators, and shows a small bell-off icon on the card.

The flag is stored in PersistedUIState as notificationsMutedByWorktree
(opt-in entries only), riding the same rails as showDotfilesByWorktree:
debounced writer baseline, hydration sanitizer, paired-client UiUpdate
schema, worktree delete cleanup, folder-rename re-keying, repo removal
pruning, and profile transfer.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs(notifications): document the touched per-worktree flag helpers

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(notifications): adapt the mute gate to the attention-sink and WSL-move ledgers

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Co-authored-by: Neil <neil@stably.ai>
2026-11-16 01:47:43 -08:00
0799a39889 fix(orchestration): keep a refused Dispatch-less escalation out of the recipient's inbox (#43016)
A sender with no active Dispatch (for example a coordinator) that sends
--type escalation or decision_gate is refused with sender_not_assignee,
but the rewritten "Rejected ..." row was still delivered to the recipient
as unread mail and woke it, including on idempotent retries via the
replayed receipt nudge. Nobody can act on such a refusal, so keep the row
for audit but mark it read, record the receipt without a replay nudge,
and skip the wake. Refusals tied to a real Dispatch still reach the
coordinator. The refusal text now points the sender at --type status or
gate-create.

Fixes #42731

Co-authored-by: Neil <neil@stably.ai>
Co-authored-by: Shaukat Ibrahim <shaukati@amazon.com>
2026-11-16 01:44:18 -08:00
Brennan Benson 636d990f34 Keep chat composer state when its view reconnects (#42695)
* refactor: give chat tabs ownership of composer editors

* Retain pasted-text file paths with composer undo history

* Keep composer selection ownership with its editor
2026-11-16 01:36:50 -08:00
Brennan Benson 0f3d81f791 perf(mac): reuse verified terminal names across process-table captures (#43004)
* perf(mac): reuse verified terminal names across process-table captures

Each full macOS process-table capture ran one lstat per terminal on the
machine. On busy Macs that is millions of libuv pool jobs a day in the
terminal daemon, which is the load that wedges macOS's pool and makes
createOrAttach time out (#31926, #42958). Verified ttys names now last 60s;
misses and /dev scan aliases are still read live.

* perf(mac): keep verified terminal names for five minutes

Closed ptys stop reporting their device, so the TTL is only a safety
margin; a longer one removes the largest remaining pool-job source.
2026-11-16 01:30:45 -08:00
OrcaWinandOrcaWin 1c37b4a2e4 test: remove narrow terminal rendering and query-order suites (#43056)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-16 01:30:01 -08:00
Brennan Benson b26087ffcc fix(chat): keep the reply before a mid-turn message visible in finished turns (#42335)
* fix(chat): keep replies before a mid-turn message visible in folded turns

A person's message is a boundary that keeps the reply before it, and
back-to-back replies stay visible, matching the common pattern for
completed turns. Tracked in #41393.

* fix(chat): fold a reply followed by its own tool work, judged within its turn

Tool-only rows merge into the reply before them, so a reply row that
draws tools is followed by work, not by the next reply. Adjacency is
judged within each turn, and orphaned tool results are not a person's
boundary.

* fix(chat): keep generated-image operations out of back-to-back replies

A generated-image row stripped of its preview still carries its
operation label; it is work, not a spoken reply.

* fix(chat): fold Codex image views and image status lines as work in finished turns

Codex's "Viewed image" rows and image-status lines ("Image generation failed", "preview unavailable") were stored as plain assistant text, so a finished turn counted them as replies and kept them, plus the sentence before them, visible. The Codex producer now marks them as image operations when it writes them, the same way it already marks generated images, and the fold treats them as work.

Tracked in STA-14968.

* fix(chat): keep response actions on the reply, not an image operation; tighten interleaved-turn fold test
2026-11-16 01:18:53 -08:00
Brennan Benson 8f07f2749a perf(codex): skip singleton trust table classification (#43024) 2026-11-16 01:18:21 -08:00
Brennan Benson 1b87d1be25 fix(browser): free hidden pages based on the current page only (#42830)
* fix(browser): free hidden pages based on the current page only

Earlier back/forward entries no longer keep a hidden browser page alive. Edits,
child frames, audio, picture-in-picture, PDF, capture and file/document URLs on
the current page still protect it; freed pages restore their history as before.

* fix(browser): keep a hidden page whose current document is a POST result

The freed record keeps only URLs, so reopening a POST result would load a
different page by GET.

* fix(browser): restore scroll and form contents when reopening a freed page

Keep Chromium pageState in the in-memory freed-page record so reopen and
Back restore scroll position and form contents. It is never persisted,
logged or sent over IPC, and is dropped when the page closes or restores.

* fix(browser): drop a closed page's freed-page record immediately

* fix(browser): reset background-activity protection when a new document commits

* fix(browser): keep the favicon baseline across commits

* fix(browser): keep title and favicon baselines across commits
2026-11-16 01:17:58 -08:00
OrcaWinandOrcaWin 84da54ae3d test: remove two low-value copy and language retry suites (#43051)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-16 01:16:18 -08:00
Jinwoo Hong 083a9a8f4d fix(orchestration): count an approved terminal prompt as answered while its tool runs (#42708)
* fix(orchestration): count an approved terminal prompt as answered while its tool runs

* fix(orchestration): a deny counts as answered only once Claude records it

* test(orchestration): stop ticking once the slow allow is answered
2026-11-16 04:14:46 -05:00
Brennan Benson 6db9aabbe1 perf(orchestration): stop orchestration.db growth by pruning settled chat output copies (#42839)
* perf(orchestration): prune settled chat output copies and reclaim orchestration.db space

Background maintenance rewrites chat workers' saved output copies of Runs settled over a day ago
to the reference later releases store, every six hours at a tenth of the thread, and returns
free pages in bounded steps once the file uses incremental auto_vacuum. New databases start
that way; Settings > Agent orchestration can restart Orca to rewrite an existing one.

* fix(orchestration): prune only copies whose chat is readable here, in bounded pages

Pages now examine a fixed window of archive rows and check each Run once, so a large Run
cannot make one slice quadratic. A copy whose chat this host cannot read stays. The restart
rewrite keeps its temporary copy on disk and records a failure for Settings to show.

* fix(orchestration): check chat readability without re-registering released workers

The readability probe builds the identity from the durable row instead of rehydrating it into
the live registry. Free-space checks use SQLite's own temp folder; a page's Run verdicts are
shared across a pass; a failed failure record no longer hides the VACUUM error.

* chore(orchestration): annotate prune casts, follow Button spacing, sync the runtime catalog

* chore(i18n): resync the runtime English catalog after rebase

* fix(orchestration): share Run verdicts across a prune pass; check the temp drive SQLite uses on Windows

* perf(orchestration): compact legacy storage during update shutdown

* test(orchestration): mock filesystem space checks through ESM factory

* Align update maintenance checks with shutdown-only work

* Drop the one-time update VACUUM; keep prune and incremental reclaim

Existing non-incremental orchestration databases stop growing because freed
pages are reused; nothing shrinks them automatically.
2026-11-16 01:01:58 -08:00
Brennan Benson 980352d897 Reduce native chat memory by replacing per-chat Electron supervisors (#42760)
* Launch provider CLIs without a per-chat runtime supervisor

* Initialize process-owner fixtures with an isolated app environment

* Retire provider output transports and failed ownership drafts

* Finish provider stops when the session group empties

* Complete native cleanup from verified watcher exit and cancel held launches

* Let native watcher finish one-shot provider exits

* Retire late ownership records after canceled provider launches settle

* Reap each provider subtree in an isolated native parent

* Preserve provider outcomes and compatibility with native parents

* Publish actual provider identities and retire native parent observers

* Retire numeric writer PIDs when their group is proven absent

* Retire provider transports and accept proven group absence

* Fence retired process identities across asynchronous cleanup

* Use native timer types and readonly process fixtures

* Reuse bounded timeout for compatibility output drain

* Release launch scopes held by late transport error handlers

* Keep processless launch failures out of watcher waits

* Align lifecycle validation with final ownership contracts

* Remove duplicate helper preparation from the dev launcher

* Keep legacy fallback in place and simplify native ownership handoff

* Defer background provider maintenance until the first window

* Release startup hook maintenance for terminal launches

* Use relocated artifact module in native prebuild scripts

* Preserve agent failure outcomes and clean tools before ownership is lost

* Budget ownership capture for its bounded identity reads

* Keep lifecycle regressions at their behavior-owning seam

* Explain the fake-clock process-read fixture

* Capture tool ownership before restart recovery removes the CLI

* Skip descendant inventories for absent provider roots

* Make final output drain regression test deterministic

* Keep output drain fixture assertions type-safe with IPC

* Use existing provider fixtures across Node and Bun

* Prepare native provider helper for Codex contract CI

* Verify provider ownership using its recorded start precision

* Preserve exact provider ownership through reader failures

* Retire exhausted recovery records without blocking watcher cleanup

* Distinguish exited descendants from unreaped zombies

* Preserve group retirement without trusting an exited root

* Keep CLI input owned through native parent loss

* Keep watcher cleanup proof through lost-parent retirement

* Guard retirement against late watcher cleanup proof

* Contain provider launches that fail before stdio setup

* fix(claude): preserve startup diagnostics after input pipe failures

* fix(provider): capture lost outcomes before native completion cleanup

* fix(provider): retire speculative captures after a valid outcome

* fix(provider): share spawn absence classification through its existing boundary
2026-11-16 01:01:24 -08:00
OrcaWinandOrcaWin 348bc2a7b8 test: remove two low-value settings and update disclosure suites (#43046)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-16 00:59:30 -08:00