Add targeted recovery for rejected PTY source frames instead of
terminating the relay channel. Classify rejection reasons (malformed,
generation mismatch, range invalid) and attempt recovery based on the
rejection type. Implement admission control at publication time to ensure
frames aren't delivered after ownership changes. Bound recovery attempts
and retry with backoff to prevent exhaustion. Diagnose and log rejection
reasons to aid debugging.
An owner-capable `pty.openClient` had two failure modes that presented as something else.
If the relay still held an owner record but the request carried no matching resume proof, admission fell through to a SUBSCRIBER grant — a success-shaped response the client cannot use, which it then rejected as "did not grant an authenticated PTY session owner". And if the relay had forgotten the record the client named, admission threw a stale-recovery error, which the client answered by deleting its own recovery row — `clientInstanceId` included — and reopening. Two round trips, and the identity that lets it resume that target at all went with the deletion.
Now every owner grant carries a required `resumed` flag, a forgotten record mints a fresh claim in one round trip, a held claim returns one of three coded refusals, duplicate opens on one connection are rejected even when identical, and an attached-holder refusal becomes a typed error routed through the terminal-relay-error callback instead of feeding redeploy backoff a link that is working fine.
Independent review caught two regressions in the first attempt, both now fixed and both with tests that fail without them:
**A backpressure teardown could take a live owner's session.** The safety argument was that a record only becomes `disconnected` from an observed peer close — but two of the six paths there are capacity paths, where the relay destroys the client's socket itself because its lane queue filled. That is the signature of a client that is ALIVE but not draining fast enough. Demonstrated: the real owner is torn down for backpressure, a rival is granted ownership 270ms into a nominal 30s grace, and the owner's later reconnect with a valid resume proof is refused permanently, backoff cleared, no retry. Closes now carry a cause (`peer-closed` | `local`, defaulting to `local`, which only ever widens a grace), and the floor applies only to closes the transport actually observed on the peer's side. Capacity teardowns, decode faults and sink failures keep the default.
**A client's own zombie connection blocked it permanently.** Only `SshRelaySession` ever requests owner, and every endpoint-credential client shares one principal — so in a normal single-app deployment an `active` incumbent refusing you is almost always your own half-open connection the relay never saw close. That was refused as terminal, where main recovered on bounded backoff once keepalive noticed. The refusal already held both client identities; a match is now a distinct transient refusal that falls through to relay-lost backoff, restoring that recovery. A genuinely different client is still blocked.
Also: each retry deadline now starts when its own phase begins, instead of both being computed at entry where a slow first phase could leave the second with zero attempts.
Fixes STA-3365.
Per-target SSH teardown awaited `removeAllForwards` BEFORE anything marked the lease detached, so a slow forward close let the final store flush snapshot while leases still said `attached` — and the later durable write was rejected because persistence had already finalized. On the next launch those leases described a state that never existed.
`beginSshShutdown()` now performs every in-memory transition synchronously before returning, and the quit path calls it immediately before `store.flushAsync()` with no await between. The whole drain shares one deadline that REPORTS unfinished `{targetId, phase}` rather than concluding anything about it, and `waitForSystemSshForwardStop` gained a post-SIGKILL bound.
Nothing here destroys a session. `detached` means this app let go of the lease, not that the shell died — the pre-pass exists precisely so still-running PTYs are recorded as detached-but-alive instead of being lost to an `attached` snapshot. Review confirmed every reader honors that: reattach enumeration and persistence restore filter only `terminated`/`expired`, lease normalization has no age-based expiry, and attempt exhaustion leaves a lease alone. The drain deadline's only consumers are a warning log and a join that discards the value — nothing reads it as "gone".
Review also caught a defect the refactor introduced, now fixed: making the pre-pass synchronous meant a throw from `beginShutdownDetach` — via `webContents.send` on a renderer that quit had already destroyed — escaped the non-async `will-quit` listener and skipped `killAllPty()`, the watchers, `store.flushAsync()`, the teardown barrier and `app.quit()`. That would have lost the exact snapshot this PR exists to make correct. Each call is now wrapped per session, collecting errors and continuing. Proven: the test throws from the first of two sessions and fails without the fix with "Object has been destroyed".
A second test could only fail via timeout rather than assertion; the ordering is corrected so removing the post-SIGKILL bound now fails in 5ms with a clean assertion instead of a 5s timeout.
Rebased onto main and verified independent of #12673 (zero references to its owner-admission changes), which is being reworked separately. Fixes STA-3366.
* Open TaskPage GitHub items directly in workspace composer
Remove background creation indirection. The composer prefills with
issue/pull-request metadata from the GitHub work item.
* Generalize workspace composer tests to support multiple sources
Rename test from task-page-github-composer-boundary to task-page-workspace-composer-boundary. Add test verifying Linear items open directly in the workspace composer, expanding beyond GitHub-specific routing.
* Support issue command automation in workspace composer
- Pass GitHub work items through composer for issue context
- Quick create resolves commands at workspace creation time
- Extract command building and trust logic to dedicated module
* Derive composer hooks by host context for React Doctor
Key loaded hooks and issue commands by execution-host context so
repo/host switches no longer reset derived state in effects (the
static-analysis gate). Also wire SSH-aware workspace targets and
cancel-safe submit settlement for workspace creation.
* Support duplicate repo IDs across hosts in workspace creation
When the same repo exists on multiple hosts (local and SSH), resolve the
workspace creation target to the ready setup on the preferred host instead of
failing closed. Also add resilience to hook checks by clearing the cache on
transient IPC failures.
With a desktop client paired to a remote Orca runtime, terminal panes could report connected, writable, and `terminal.send` returning accepted — yet keystrokes never reached the agent. No error, no banner, no recovery; input silently vanished.
The ticket was really two bugs. The attach half was already fixed by #12589 (subscriber-driven daemon attach), confirmed by reproducing against current main. This fixes the remaining half: a write the host refuses had no way to tell anyone.
A capability-negotiated `WriteUnavailable` opcode carries that refusal back to the client, where it feeds the pane's pre-existing recovery hook. Capability gating matters because decoders reject unknown opcodes on desktop — and, worse, silently drop them on mobile — so the signal is negotiated in the subscribe handshake. Verified per direction: an old host strips the unknown Subscribe key, an old client omits it so the host never emits, and capability cannot be inherited across resubscribe.
Independent review then found the signal was being delivered and discarded: recovery demanded an authoritative liveness answer, and `pty:hasPty` had no `remote:` guard, so a paired pane's id fell through to the LOCAL provider, which returned false, and recovery bailed before remounting. Every test stopped at the transport boundary, so all of them passed while the pane stayed just as stuck. `pty:kill` already had exactly that guard.
The fix makes main answer LESS rather than claim more: `pty:hasPty` now returns unknown for a `remote:` id instead of a fabricated false, because main cannot speak for another host's PTY. The remount is then authorized by positive evidence — the process that owns the PTY stating it refused this specific write over a live negotiated connection — not by inference from silence. Local and app-SSH ids keep the probe, where a false genuinely means the shell died. Nothing is destroyed on this path; the remount rebuilds the renderer over the session it already had.
An end-to-end test now carries a rejected write from the host through to an actual remount, which no prior test did. A surviving mutant was also killed: the legacy-binary capability gate could previously be deleted with nothing turning red.
The reliability gate stays experimental — live paired journeys and mixed installed-release evidence remain uncollected. Fixes STA-2830.
Mixed versions are the normal state of the remote-server feature: users update clients and servers independently. Until now nothing tested that. Every cross-version claim was made by code reading plus unit tests with hand-written old/new shapes — enough to catch design problems, not enough to catch a real skew regression.
This runs the REAL protocol implementations from two builds against each other in one process: the actual host methods and RPC dispatcher on one side, the actual renderer multiplexer on the other, with a transport that reproduces the production asymmetry — each side decodes with its OWN codec and drops frames whose opcode it does not know. A frame survives only if the RECEIVING build understands it, which is what makes this level sufficient without launching two apps. The old side is a genuine checkout extracted from the release tag; the extracted client was confirmed to lack a symbol that exists only on main.
Journey: subscribe, first snapshot, input reaching the process, live output, hide/reveal snapshot, transport drop, resubscribe, input landing again — across old->new, new->old, and a current/current control. Every step ends on an observed-state barrier; no sleeps. The oracle asserts the recorded step list, the exact 16-frame named sequence, negotiated capabilities, the exact input the host wrote to the PTY, rendered content, and zero decoder-rejected frames. A host method the stub lacks is recorded by name and asserted empty, so a harness gap cannot masquerade as a wire break.
Detection is proven per violation shape, and it attributes each to the correct side: an unnegotiated opcode goes red only where a decoder would reject it, a removed published field goes red only where an old client consumes it, and a legal additive field stays green in all three pairings so the harness will not cry wolf on safe changes.
It also documents the three compatibility rules in docs/reference/remote-wire-compatibility.md, linked from AGENTS.md, since they previously existed only as folklore — notably that "decoders reject unknown opcodes" is true for the desktop decoder but NOT for mobile, which silently drops them.
Deliberately scoped: terminal stream only. The session-tab sync channel is not covered, nor agent-session publications, file/Git RPCs, mobile E2EE framing, or the relay transport. Two version points, so a regression introduced and reverted between them is invisible.
CI selection was verified rather than assumed — `vitest list` confirms 0 matches under the shard's exclude and 4 under the dedicated job — because a lane silently running zero tests is precisely how a host-side defect escaped CI earlier in this series. Closes STA-3469.
`activateMobileSessionTab` gated only on `publicTab.status !== 'ready'`. A deliberately slept pane publishes as `pending-handle` indefinitely — indistinguishable at that call site from a pane awaiting reconnect — so the reconnect probe added by #11542 respawned it with a re-resolved agent launch, waking something the user had deliberately put to sleep.
The first attempt refused activation for any pane with a `worktree-sleep` record, applied to every path. Independent review found that broke the documented wake gesture: opening the tab IS how those panes are meant to cold-restore (`wake-sleeping-agents-in-background.ts`: "Those panes cold-restore --resume when their own tab is opened"). A mobile tap sends the byte-identical call the reproduction test used, and in three of four topologies no wake clears the record first — so the tap became a permanent no-op with no feedback.
This carries intent explicitly instead of inferring it. A new shared `TabActivationIntent` ('user' | 'automatic') rides the existing ActivateTab schema as an optional additive field; `isAutomaticTabActivation` returns true only for an explicit 'automatic', so an absent value is permissive BY CONSTRUCTION in one place — an older client that does not send it keeps today's behavior rather than silently losing its wake gesture. The field is required on the mobile helper's params, so no call site can be added without declaring who asked.
Every user path (mobile tab switches, paired tab clicks, shortcuts, palette, the pane's own open) is labelled 'user'. The only automatic sender in the codebase is `waitForResubscribeHostSessionHandle`, the #11542 reconnect probe.
Verified per topology: user activation materializes a parked pane under headless serve, a paired runtime client, a completed agent with restoreOnTabOpenOnly, and a running agent whose wake cleared the record. The automatic probe is refused without retiring the surface, and #11542's reconnect tests stay green.
Also fixes a test fixture that made a real bug untestable: the store stub ignored the host id, so mutating the partition lookup to 'local' left the suite green. Correcting it exposed three existing SSH reattach tests that had been relying on that looseness — their workspace session sat in the local partition while their repo was SSH-hosted, a store production would never read. Production was always right; the tests described an impossible world.
Fixes STA-3465.
Opening any webpage in the remote browser dropped the paired runtime connection, and the client then retried forever without recovering.
Causal chain: the screencast travels host->client, a direction that admits up to 8 MiB. The host's encrypted channel rejects anything larger with close code 1013 "Outbound reply buffer overflow" — killing every subscription on that connection. The producer treats a false return as backpressure and retries the identical frame, which for an over-limit frame can never succeed. A permanent condition was being treated as transient.
Two changes:
1. A paired-runtime admission wrapper: an over-limit frame is dropped rather than handed to the transport, and reported as handled so the producer advances instead of retrying something doomed. The generic Chromium producer is untouched, so local browser behavior is unchanged.
2. The actual source of over-limit frames. Live frames are hard-bounded by maxWidth/maxHeight, but the navigation snapshot path ignored those bounds entirely, feeding capturePage device pixels straight into the encoder — capturePage's rect is CSS pixels while the bitmap is device pixels, so at deviceScaleFactor 2 a snapshot could be 4x the pixel area the live path is allowed to send. That path fires on page load, which is literally the reported trigger. Applying the caller's own clamp there makes the drop a backstop rather than the mitigation.
Dropping a frame is safe here because frames are complete standalone images, not deltas — each replaces the client image wholesale, so the next frame fully repaints. Disclosed in the PR: mobile web-view mode sends no viewport and takes the unclipped screenshot branch, where the drop guard remains the only protection; still strictly better than a 1013 that kills every subscription.
Verified by reverting in place: neutralizing the admission guard fails 3 oracles, with the integration test emitting the real [1013, "Outbound reply buffer overflow"] from an actual E2EEChannel — the production symptom, not a mock. Neutralizing the snapshot clamp fails its own oracle, re-proven after the test was relocated.
The second half of the report — never recovering without an app restart — is only partly addressed here and is now tracked as STA-3483: the browser stream restart arms a single 500ms retry and never reschedules, so any connection loss can strand the pane. Fixes STA-2970.
Since #12667, a present error code short-circuits classification: a code that is genuinely transient but missing from `RECOVERABLE_CODES` classifies as FATAL. That is the shape that dead-ends terminal panes — #12650 fixed exactly that for a different error, where a transient failure misclassified as fatal unmounted the Reconnect banner and left recreating the session as the only escape.
Today the code and fragment sets agree. Nothing prevented a future code from being added without a matching entry, and the failure would have been silent.
This pins that agreement: for every reachable transport error, a code that classifies fatal must not carry a message that would have classified recoverable. 57 coded pairs plus 8 code-less ones, derived by invoking the producers where possible so a reworded message updates the corpus instead of leaving a stale copy silently passing. The failure message names the offending fragment and says what to do about it.
Enumeration turned up producers beyond the obvious ones — notably the Tailscale-hinted variants, where `runtime-environment-transport-routing.ts` mutates the message on an already-coded error before it crosses IPC, making those distinct corpus members.
Also documented (not asserted, because it is unreachable today): `runtime_rpc_queue_overloaded` is absent from both host passthrough allowlists, so if it ever crossed `mapRuntimeError` it would flatten to `runtime_error` while keeping its "queue is full" message — precisely the dangerous shape. The queue pool is never instantiated on the server dispatcher, so it cannot happen now.
The known exception is pinned rather than silently exempted: a dedicated test records WHY the guard cannot see `remote_runtime_busy` (fatal by code, matching no fragment, so the two sides have nothing to disagree about). If someone rewords a busy message into connection wording, that test fails and points at STA-3479.
Proven non-vacuous by four separate injections. The only production change is two `const` to `export const`.
After a laptop sleep against a remote machine, the sidebar agent count came back lower than the number of open terminal tabs — rows vanished for panes whose tab and host process were both still alive.
The client mirror deletes a mirrored pane's agent status whenever the host snapshot carries none for it, unless the client's own byte-derived entry is still fresh. But for a remote pane the client is the ONLY writer of that status, and a laptop closed past the 30-minute staleness boundary makes every such entry stale by definition — so the first snapshot after wake erased the sidebar row of every pane the client owned.
Freshness was the wrong gate. It exists to arbitrate between two competing writers, but on the delete branch the host published nothing, so there was nothing to arbitrate and "my status is old" quietly became "delete this pane". The branch now gates on ownership: a pane this renderer claimed and wrote keeps its entry and decays to idle through the normal staleness boundary, exactly like a local pane. Teardown releases the claim, which is how the host takes the pane back.
That reverses a contract #12641 pinned, so its test was updated in place with the reasoning inline rather than deleted — the old premise was that going stale hands the pane back to the host, which does not hold when the host has no value to hand back. The assertion is now stricter: the entry must be retained AND read as stale so consumers render it idle.
This is the sidebar-count half of STA-3107. The blank-terminal half was fixed by #11542 and is proven so: reverting that fix in a six-pane harness makes exactly one of six panes fail to resubscribe while its siblings recover, matching the report.
A remaining gap is documented in the PR: a pane the client never wrote status for stays host-authoritative and can still lose its row. Separating "the host has no opinion" from "the host proved there is no agent" needs the origin marker tracked as STA-3455.
`retireMobileSessionSurfacesForPty` called `getWorkspaceSession()` / `setWorkspaceSession()` with no host id, so every retirement wrote to the LOCAL partition — while its sibling `retirePersistedStablePaneOwner` correctly scopes to the SSH execution host.
For an SSH pane exiting cleanly this is not a harmless misdirected write. Measured on main: the write went to the local partition instead of `ssh:conn-1`; the SSH partition still held the dead PTY binding; the local partition gained a bogus topology revision for an SSH repo; and the published tab list contained a RESURRECTED leaf hydrated back from the stale SSH partition. The wrong-partition write was accepted — a tombstone recorded and the revision advanced for a surface that could not be found.
Found during independent review of #11542; pre-existing, not caused by it. Fixes STA-3463.
When a remote terminal tab is hidden, the host stops sending its output and discards what it queued, so on reveal the only way to recover the missed output is to ask the host to serialize its buffer. That reply was ambiguous — one empty answer covered several unrelated situations — so the client inferred "output is lost" from elapsed time, using budgets sized for local IPC. Over a network that guess was routinely wrong: users saw "[Orca skipped hidden terminal output because main recovery was unavailable.]" on a healthy pane and got a permanent scrollback gap, worst exactly when an agent was streaming heavily and there was the most to lose.
The key insight is that there is no provable-absence case at all. A pane with genuinely no retained output returns a SUCCESSFUL snapshot with empty data, because the host serialized fine and found nothing. The real defect was the host sending an untagged empty reply when no serializer answered — reporting an unprovable failure as proven emptiness.
The host now states why a snapshot is unavailable and the client acts on that reason: an empty snapshot is success; retry-worthy retries and then gives up honestly; permanently-unavailable banners immediately with no waiting; and a host too old to say latches that pane to the pre-existing timer heuristic. Local panes are unchanged. The self-heal repaint no longer yanks the viewport of a user scrolled back reading — it waits for the terminal to return to following output.
Retries are bounded by COUNTING REPORTED OUTCOMES, never elapsed time. Independent review found that the single budget also charged attempts for causes returned locally, where the host was never asked — meaning a re-arming resync could exhaust it and banner on a perfectly healthy host, a residual instance of this very bug. Host answers and local gates now have separate budgets; local gates send zero frames, so retrying them cannot pressure the host.
Review also found a duplicate-banner path where a repaint timer armed before a permanent answer survived the abandon; the clear is scoped to the branch that banners, since the retry loop deliberately arms that timer.
Wire change is additive: an optional field on an existing frame, dropped on the success path, so old clients see an unchanged frame. STA-3476 tracks replacing the legacy-host detection (currently inferred from an absent field) with a positive capability signal. Closes STA-3457.
Errors thrown across Electron's `ipcMain.handle` lose their structured error code — only the message survives. So the renderer classified transport failures by matching substrings of English message text. That is how a queue-overload rejection escaped classification during a remote outage and surfaced as a raw error wall: the code was stripped in transit and its message fragment was not in the recoverable list.
This converts `RemoteRuntimeClientError` and `RuntimeRpcCallQueueOverloadError` rejections from `runtimeEnvironments:call` into the existing structured `{ok:false, error:{code,message}}` response, which the preload already passes through unchanged and `unwrapRuntimeRpcResult` already reconstructs with the code intact. Classification now treats a present code as authoritative and consults message fragments only when there is no code.
The fragment list is deliberately RETAINED as a backstop, not deleted: untyped main-handler rejections, subscription-start failures, and older code-less paths still rely on it.
Proven real rather than cosmetic: a test-only patch applied to unmodified main fails (4 failed / 66 passed) because the code does not survive the boundary today, and passes on this branch.
Independent review specifically chased the risk that a present-but-unrecognized code would now short-circuit to fatal where a message fragment previously rescued it — the shape that dead-ends a pane. It enumerated all 34 reachable code/message pairs and confirmed no pair flips recoverable to fatal, that the newly-serialized code set is closed and client-local, and that host-forwarded codes preserve recoverable classification by design. A differential harness over that corpus was verified non-vacuous by injecting the bad shape.
Nothing crosses the paired-runtime wire: desktop main -> IPC -> preload -> renderer only, reusing an existing response shape, no new fields or opcodes.
The connection-level offline state with a single reconnect affordance remains as STA-3456 follow-up work.
STA-1716 reported that a packaged `orca serve` could become the single-instance owner after the desktop app exits, leaving Dock/Finder unable to restore a window — and that forcing a reopen made the headless process hydrate a renderer that interrupted and DUPLICATED live agent sessions.
Verification against main found every criterion already fixed (#8646 for desktop promotion and the fail-closed CLI, #12212 for duplicate serve activation, #12574 + #9729 for the resume/ownership guards). The genuine gap was criterion 6: the ticket's own automated regression never existed. An existing reliability gate asserted PTY identity survives promotion, but nothing asserted what the incident was actually about — how many agents the promoted renderer resumes.
This adds that coverage: a unit/service-level journey that drives the real single-instance lock, activation gate, settle and focus paths, then runs the real resume logic against a store seeded as a renderer freshly mounted inside the serve process, asserting zero duplicate resumes.
`settleServeDesktopActivation` moved from `index.ts` into its own module with identical semantics, so the test drives the real decision rather than re-implementing it — the earlier repro had to mirror that logic locally, which is the "test passes without running the scenario" failure mode.
Proven to be a real oracle: breaking each guard individually turns it red, and reverting the pre-#12574 pane form reproduces the incident exactly (two duplicate `codex resume` tabs).
* fix(ssh): recover instead of wedging when the relay channel dies mid-connect
Three coupled defects made a dropped SSH relay look like a permanent bug:
1. SshRelaySession.establish()/reconnect() ran their last liveness gate before
configureRelayGraceTime(), whose mux.notify() can dispose the mux
synchronously (writer control-lane admission cap, or a throwing transport).
The session then latched _state='ready' + _onReady (status bar "connected")
while watchMuxForRelayLoss() silently no-op'd on the dead mux, so the bounded
relay backoff in ipc/ssh.ts never ran and the fs/pty/git providers stayed
registered against a dead multiplexer. Both sites now re-check
mux.isDisposed() after the notify and take the existing failure path.
2. SshChannelMultiplexer.request()/notifyWithSettlement() always reported the
permanent-shutdown string 'Multiplexer disposed' with no code, even when the
recorded dispose reason was connection_lost. The reason is now recorded and a
shared disposedError() factory serves dispose(), request(),
notifyWithSettlement(), so a transient drop reports
'SSH connection lost, reconnecting...' / CONNECTION_LOST. onDispose() on an
already-disposed mux now fires the handler synchronously with that reason
instead of returning a silent no-op (without retaining it).
3. TerminalErrorToast no longer renders a transient relay drop in the red
"please file an issue" style. The marker is matched with includes() because
the message reaches the toast IPC-wrapped.
ssh-git-response-stream-reader registers its onDispose subscriber after the
abort wiring, since an already-dead mux now fails synchronously there and the
cleanup must be able to drop the caller's abort listener.
Closes#11953
* fix(ssh): treat a mux killed during PTY reattach as relay loss
reconnect()'s post-reattach gate bare-returned when ownsAttempt() went false,
and reattachKnownPtys swallows every per-PTY error, so a control-lane failure
during a large reattach burst disposed the mux without ever reaching the catch:
providers stayed bound to the dead mux, no relay-loss watcher was installed, and
the session wedged in 'reconnecting' until restart. Take the failure path when
our own mux is the one that died so ssh.ts's bounded backoff retries.
Co-authored-by: Orca <help@stably.ai>
* fix(ssh): recover when relay dies during setup instead of wedging
Introduce verifyRelayAttempt() to detect mux disposal at each setup phase
(consumer session, home resolution, provider registration, PTY reattach).
Routes mid-setup connection loss into relay-loss recovery instead of
hanging in reconnecting state.
* Extract SSH disposal error factory
Multiple sites were duplicating the disposal error creation logic with
specific message and code values. The renderer uses these to distinguish
temporary disconnects (show reconnection overlay) from permanent shutdown
(show error toast), so all producers must use the same factory to avoid
silent UI degradation.
---------
Co-authored-by: Orca <help@stably.ai>
Co-authored-by: Jinjing <6427696+AmethystLiang@users.noreply.github.com>
* Add host and project filtering to the worktree jump palette
Filter the search results by execution host (local, SSH, runtime) and project/repo, with a drill-down options menu, chips for active selections, and overflow hints for large result sets. Filters reset on open to prevent silently hiding results, and stale selections auto-prune. Caps rendered rows per section to prevent DOM bloat from single-character queries. Host badges appear when filtering to clarify which rows survived the cut.
* Add E2E test for worktree jump-palette host filtering
Tests filter interaction via keyboard, filter/project intersection, empty state when filters exclude all results, and ephemeral filter reset on modal close.
* Fix worktree jump-palette filter persistence and search UX
- Persist filter state when filter model changes to prevent dropped IDs from silently re-activating
- Fix result count to distinguish between query matches (all items) vs. empty list (capped sections)
- Improve filter field options: use state for scroller to handle unmount/remount, clamp highlight index to valid range, only reset on query/field change, not re-ranks
- Context-aware space key: allow toggle in listbox only, not in search input (preserves for typing)
- Replace generic "Clear field" translation with field-specific strings to preserve capitalization in non-English languages
* fix(cmd-j): reset filter highlight without prop-change effect
Derive the active option index from field/query identity instead of
resetting it in a useEffect so React Doctor and first-paint stay correct.
* fix static analysis issue
* fix(e2e): rename project entity in jump-palette filter seed
Filter options use project.displayName when a Project exists, so only
renaming the repo left the local option labeled with the path basename.
A remote-paired terminal tab flickered several times per second between the agent-generated title with a running status, and the plain title "Terminal" with "Done - Claude" in the sidebar.
Two writers owned the same state. For remote panes the client parses agent status out of the terminal byte stream, while every host tab snapshot rebuilt the mirrored tab WITHOUT the client's generated title and re-decided status by comparing timestamps taken on two different machines. The host also treated a neutral live title ("Terminal") as proof the agent had finished, and re-stamped that conclusion with the pane's last-output time — so it advanced with every output byte and always looked newer. Neither writer could ever win.
This removes the second writer rather than trying to arbitrate two clocks: the client is authoritative for panes whose status it parses (only while attached, released on teardown), the host no longer invents a finished state from a neutral title, and the generated title is carried through snapshot rebuilds. The client was chosen as the authority because the host snapshot format carries no generated title at all — making the host authoritative would permanently lose generated titles on paired clients.
Purely local and plain SSH panes are structurally unaffected: they have only one writer.
Verified with a reproduction that is red on main (frames show done -> working -> done with the label flipping on every publication) and green with the fix. Independent review additionally found and fixed a defect where a superseded pane's late cleanup could permanently strip a live pane's authority, reinstating the very flap being fixed.
Deferred follow-up STA-3455: host `blocked`/interactive-prompt states can still pierce the fence and fall back to cross-machine timestamps; fixing that properly needs an origin marker on the status entry.
When a remote runtime went unreachable (laptop sleep, Tailscale drop), the UI filled with dozens of repeated timeout errors until it was nearly unusable, and the affected terminal then accepted no input after connectivity returned — leaving "close the session and resume it in a new one" as the only escape.
Four causes, three of which were still live:
- Errors accumulated into one ever-growing surface with no de-duplication or cap.
- Queue-overload rejections lose their structured error code crossing the IPC boundary, so they were never classified as recoverable and surfaced raw.
- A transient failure misclassified as fatal called `recovery.cancel()`, setting the pane to an idle phase — which unmounts the Reconnect banner and makes manual retry, online and resume triggers all no-ops. A true dead end, and the reason recreating the session was the only way out.
- Dismissing an error cleared the surface but not the dedup memory, so an identical fatal error recurring in the same outage was suppressed forever while the pane looked healthy; dedup also compared single lines, so multi-line errors never matched and stacked without bound.
The ordinary reconnect loop was already fixed in v1.4.150/160 — bounded backoff, a Reconnect banner and auto-recovery already ship. This fixes what remained.
Note the fix routes fatal resubscribe failures back through the shared terminal error handler: bypassing it had silently dropped stale-handle re-resolution, terminal-gone retirement, SSH-expired recovery and oversized-snapshot suppression — a stuck-pane regression inside the stuck-pane fix, caught in review and covered by 6 dedicated tests.
Verified: reproductions red on main before the fix; after rebasing onto #11542, reverting the dead-end fix still turns its test red. Follow-up STA-3456 tracks preserving typed error codes across the IPC boundary so classification stops matching message text.
Every SSH reattach failure abandons the remote terminal without shutting it down, and abandoned terminals then became structurally unreachable — excluded from reattach enumeration and, critically, filtered out of the user-facing "Terminate sessions" action, so a user could not kill them even manually.
The core problem is a naming trap: `expired` never meant the remote shell died. It means the app gave up reattaching. It is written on reattach failure, on spawn-time expiry, and in bulk by a relay reset inside a `finally` that runs even when the force-stop threw. So the leases most likely to name a still-live orphan were exactly the ones the terminate path excluded.
This change is reachability only. Expired leases are now reachable by an explicit user-initiated terminate, with the relay's response used as evidence: a shutdown that reports the PTY gone tombstones the lease, and leases already proven terminated are left alone.
**No automatic kills were added.** Every abandon path was enumerated and none of them proves abandonment: attempts-exhausted knows nothing (the relay never answered), identity mismatch means a *live* PTY belongs to a different pane so killing it would destroy someone else's terminal, and not-found is the one branch with real proof of death — where the process is already gone and needs no shutdown. Per the rule that unprovable liveness never authorizes destroying a session, the abandon paths deliberately leave the process running.
Relay-side automatic collection of unattached PTYs is deliberately NOT implemented: the relay cannot distinguish an abandoned terminal from a deliberately detached one, and the unlimited default grace exists precisely so long-running work survives disconnects and host sleep. Any bounded reaper would be killing on absence of evidence.
Verified: 3 tests red on main. Negative tests assert each abandon path leaves the shell running and the lease terminable, proven real by mutation — adding a shutdown to the exhausted branch or expiring on identity mismatch each turns them red.
Fixes STA-3376.
AI commit-message and PR-field generation over SSH failed deterministically at exactly 30 seconds whenever the remote agent CLI took longer, reporting "Claude could not be reached on the remote PATH. Try again after the SSH connection recovers."
That message was wrong twice over: the SSH connection was healthy (terminals and git kept working on it) and the agent binary existed — it was simply still running.
Cause: the SSH channel multiplexer applies a 30s default deadline when a request omits its own, and the generation call passed none, even though the operation itself carries a 60s budget. The shorter transport deadline always won, and the resulting rejection was then mapped onto the generic connection/PATH error.
Fix: derive the transport deadline from the operation's own budget plus a margin at that call site (the global default is deliberately unchanged, since other callers depend on it), and classify a transport timeout as a timeout — reporting that the agent exceeded its budget and may still be running — while genuine connection and PATH failures keep their existing guidance.
Verified red on main first: a 45s response rejected by the 30s default, and a typed timeout mapped to the PATH message. Both green after. Caller audit covered commit messages, PR fields, branch naming and model discovery.
Fixes STA-3073.
Reconnect could never recover a terminal pane whose host-side process was gone (host restarted, or the workspace was never opened there): recovery only polled the tab inventory, which can never create the surface it is waiting for, so Reconnect spun for ~60s and gave up permanently.
Verified with a deterministic reproduction: on main the recovery path issues 51 inventory polls and zero activations across both an automatic online trigger and a manual Reconnect click; with this change the pane re-materializes, rebinds and accepts input.
Review found and fixed three further defects beyond the original change:
- an activation answered with a stale ready handle left the loop polling forever instead of re-activating;
- a non-missing activation failure (e.g. an older host without the method) never fell back to inventory;
- host-side, activating a parked surface permanently deleted the host tab, because an already-absent persisted binding was read as a competing owner *after* the destructive retirement had already run.
Independent review confirmed by mutation testing that every production change is covered by a test that fails when it is reverted, that only an authoritative inventory can retire a pane, that the loop is bounded under every failure mode, and that the unknown-liveness guard (proven death required before retirement) is intact.
Fixes STA-3002.
* feat(linear): add 'Has Workspace' mode to show issues linked to local wo
Enable users to view and open existing workspaces attached to Linear issues
instead of accidentally starting duplicates. Includes shared worktree attachment
labeling for consistent UX across GitHub and Linear surfaces.
* fix(linear): apply search filter in 'in-orca' mode to prevent drops
- Apply search filter in 'in-orca' mode even without active context label to prevent
team filters from silently hiding linked tickets (no "Fetch more" recovery path)
- Add aria-label to workspace-open button for accessibility
- Update tooltip from "local worktree" to "Orca workspace"
- Reorganize i18n: move workspace.open from lib.linear to components.issue
- Expand test coverage for workspace start and activation scenarios
* fix(linear): avoid mutating in-orca linked refs during render
React Doctor fails static analysis when refs are written during render.
Keep the latest linked refs in an effect so the in-orca loader can still
read them without re-running on identity-only worktree churn.
* test(diff): repro for STA-3420 combined-diff invalidation freeze
Co-authored-by: Orca <help@stably.ai>
* Fix diff-view freeze when large diff invalidated by rebase writes
Staged-diff sections now reload in-place on external file changes instead of remounting every visible Monaco editor and bumping the virtualizer generation, which wedged the renderer during rebase bursts.
* test(diff): calibrate STA-3420 burst assertions against an idle baseline
The burst window's peak lag is dominated by a one-off stall from opening 8x15k-line
Monaco editors, which reproduces identically with invalidation disabled. Measure an
equal-length idle window first and assert p95, sample coverage, and lag relative to
that floor. Adds unit coverage for isUnchangedDiffSectionReload.
Co-authored-by: Orca <help@stably.ai>
* fix(diff): keep renderedIndicesRef pure during render
React Doctor blocks ref mutation during render; sync the on-screen
section set in a layout effect instead so static analysis can pass.
* Fix unchanged diff-section reload detection for truncated diffs
When a diff exceeds render limits, content is pruned to '' for memory.
The old check compared content equality, so limited reloads always
appeared changed, triggering unnecessary revalidation that froze the UI.
Compare render-limit metadata instead — it's the sole change signal
and full description of what the fallback banner displays.
Also calibrate STA-3420 e2e assertions relative to idle baseline for
machine independence instead of absolute thresholds.
* fix(diff): defer invalidation reloads for in-flight stale-token loads
When a diff section is invalidated while a large-diff load is in-flight:
- Don't delete the in-flight load from loadingIndicesRef, since a newer load may own it
- Bump the reload token but defer the reload if there's still an in-flight load
- Let the in-flight load settle first, then reschedule the reload at settle-time
- Prevents the freeze by avoiding race conditions that leave sections stuck loading
This fixes STA-3420 where rebase-driven invalidations could hang the diff view.
* test(diff): relax STA-3420 burst assertions to inclusive comparisons
Switch from strict inequality checks (toBeLessThan, toBeGreaterThan) to
inclusive variants (toBeLessThanOrEqual, toBeGreaterThanOrEqual) to allow
measurements landing exactly on the threshold boundaries.
---------
Co-authored-by: Orca <help@stably.ai>
* Add linked issue guidance and ELI5 sections to PR generation prompts
Include linked GitHub issues in PR descriptions with Fixes/Refs guidance, and require ELI5 Problem and Solution sections before implementation details. Tests verify linked issue substitution and prompt structure enforcement.
* Include linked issue details in PR description generation
- Fetch the linked GitHub/GitLab issue title and body so generated PRs reference real issue context instead of just a number
- Use provider-specific reference syntax (Fixes/Refs, Closes/Related to, AB#) and label the issue by the active provider
- Feed issue title and description into the generation prompt while treating them as untrusted context, never as instructions
- Fall back to a cached work-item title when the provider lookup fails, and skip cross-provider issue attachment
* fix(mobile): keep healthy relays green through focus and network nudges (F1+F2)
Focus/app-resume nudges probe the active relay instead of suspending it;
network-change nudges replace it make-before-break, suspending only after a
failed dial. Mount, Retry, and host-swap windows read 'connecting' instead of
'disconnected'; the host list keeps last-known worktrees for every
not-connected state and spins instead of rendering nothing.
* feat(mobile): surface the pairing relay path in the pairing log (F3)
The relay candidate was silent during pairing: dialing, E2EE handshake,
director recovery, and the winning path now emit redacted phase lines through
the same connectOptions.onLog the direct path already used.
* docs(mobile): relay UX investigation findings and F0-F10 fix plan
* feat(mobile): name and narrate relay dials while they happen (F5)
migrateTo forwards the dialing session's connecting/handshaking/reconnecting
phases whenever the client is suspended or disconnected — never downgrading a
live session — and exposes getPendingPath so the host card can say
'· Orca Relay' during the dial instead of only after it.
* feat(mobile): race a relay dial when the direct dial stalls (F6)
A 2.5s grace timer starts relay recovery while an unauthenticated direct dial
is still inside its 12s connect window; the race gets one attempt through the
existing mutex/cooldown machinery, cancels when direct authenticates, and
never arms for hosts without a relay endpoint.
* fix(mobile): overlay the protocol gate instead of unmounting the host stack (F9)
A pending status.get used to swap the mounted HostStack for a spinner at the
moment the socket connected, destroying in-flight nested navigation. Once
children have rendered for a host they stay mounted under an opaque
touch-blocking overlay; first visits and blocked verdicts keep the old
behavior.
* fix(mobile): keep loaded data through transient connection blips (F10)
Git history no longer blanks on reconnect (and commit files refetch instead
of caching an offline empty answer), the repo picker keeps its last-good list
when an in-flight repo.list rejects, the diff review's ready-state
preservation actually runs, and proven host capabilities survive a drop
flagged unverified instead of being wiped.
* feat(mobile): coordinate every home deep push and bounce dead resume targets (F4+F7+F8)
Notification taps, the Accounts card, and host-edit now use the shared
mount-then-replace transition (with a focused-route walker so root-layout
scope works); the Resume card renders from the snapshot in a disabled state
so its late arrival can't shift Tasks under the thumb; resume targets are
validated against proven catalog data, and a session route whose worktree
the host proves missing bounces to the host index with a notice banner
instead of stranding on a dead screen.
* test(mobile): cover the resume-target and notice policies (F7)
Key notice dismissal by code so closing one banner cannot swallow a later,
different one, and move the visibility rule into host-route-notice.ts where it
is testable without a screen.
Adds the missing units for F7's decision points: isResumeTargetConfirmedMissing
(unproven catalog is silence, synthetic routes exempt), the validating
last-visited reader, and the notice visibility rule.
* fix(mobile): review-pass hardening for the gate overlay and diff preservation
Adversarial review findings: the reader's hunk position now survives a
connection blip (reset only on item change), the covered stack is hidden from
TalkBack while the gate overlay is up, and the overlay's hit-test comment is
scoped honestly to in-tree views (native-Modal drawers present above it —
follow-up).
* fix(mobile): CI + CodeRabbit review fixes for #12609
Move the findings doc under docs/ (root directory guard), drop two unused
eslint-disable directives, and address review findings: an unproven snapshot
seed can no longer downgrade a proven worktree catalog; a locally-aborted
relay dial skips the director fallback; post-migration bookkeeping failures
log instead of masquerading as dial failures (which could suspend the healthy
session); the auth wait arms its timeout before subscribing; forwarded dial
phases stop at close(); the legacy selector_not_found fallback requires
runtime_error; the diff-loading effect depends on the fields it reads; and
host-edit auto-cancellation is now pinned by a test.
* fix(mobile): second review round — queued replacements, race fence, confirmed bounces
A network-change replacement now survives the recovery mutex and cooldowns as
a queued intent instead of being dropped or suspending a healthy session —
only a failed dial or a dead probe tears one down. The happy-eyeballs
migration withdraws when direct authenticated during the relay dial
(first-authenticated-wins). A worktree bounce requires two consecutive
host-proven misses, since a transient desktop repo-scan rejection answers
selector_not_found for a live worktree. Background network flaps no longer
wake a billed relay splice, the lifecycle foreground flag stays in sync, a
screen unmount cancels only its own pending host-stack transition, and diff
review keeps the loaded review when its reconnect refresh rejects.
Extracted mobile-endpoint-nudge-router.ts and the establisher's dialEligible
pass, and split the supervisor nudge tests, to stay under max-lines.
* fix(mobile): satisfy the React Doctor changed-code gate
Render-phase ref writes move into effects: the protocol gate's resolved/mounted
latches now record committed outcomes only (a discarded children render can no
longer count as mounted), and the bounce hook syncs its callback ref in an
effect. Array<T> annotations become T[] in the extracted modules.
* fix(mobile): keep the loaded diff when the reconnect refetch rejects (F10)
The diff-loading hook's catch was the one path still erasing a ready diff —
the same keepLoadedDiff guard its disconnect and loading branches already use,
now pinned by a reject-after-ready test.
* fix(mobile): process foreground revival nudges
---------
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
* fix(mobile): label tool rows with a clean summary, expand full input (STA-3333)
Mobile tool rows showed the raw input JSON (`{"file_path":…}`) as the row
label, and the expanded detail just repeated that same truncated string.
- `describeToolInput` labels a row with the target file path, else the
primary argument (command/cmd/query/pattern/url/description), else the
bounded JSON preview.
- Codex delivers tool arguments as a JSON string; normalize those into the
object shape the helpers already understand, so labels, file links,
run summaries and the expanded detail all work for Codex calls too.
- The expanded detail now renders the fully formatted input, capped at
MAX_TOOL_RESULT_CHARS like desktop's tool detail (and like the result
body), and a structured input makes the row expandable.
* fix(mobile): name search rows by their term and keep the filename in path labels (STA-3333)
Review follow-ups to the tool-row summary, all in the shared helper:
- A Grep/Glob row labelled itself with the directory it scanned and dropped
the pattern entirely, because `toolFilePath` treats `path` as a file target.
That path is a scan root, so it also rendered a tap-to-open link that asked
the app to open a folder. `toolFilePath` now ignores the generic `path` key
for search-shaped input, which lets the pattern win the label and drops the
bogus link; an explicit `file_path` still wins.
- An overlong path was truncated from the head, cutting off the basename —
the one part that tells two rows apart. Trim from the front instead, so
the label reads `…/session/MobileNativeChatMessage.tsx`.
- The primary-argument chain used `??`, so a present-but-blank key selected
itself and swallowed the keys ranked after it, dropping the label all the
way back to raw JSON. Take the first key that actually yields a label.
Refs STA-3333.
* fix(mobile): don't offer an expander whose detail repeats the row (STA-3333)
An empty tool input formats back to the row label verbatim, so `{}` and `[]`
advertised an expander and then re-showed the label — the same repeat-the-JSON
problem this change set out to remove. Gate `isStructuredToolInput` on the
collection actually having contents; the lazy detail path is untouched.
Also pins the overlong-path test to the path itself: asserting only length<=80
plus a `…` passed just as well with path labelling deleted.
* fix(mobile): gate the tool detail panel on having detail (STA-3333)
The Tools toggle opens every row at once, bypassing the row's tap guard,
so a row with nothing to expand rendered its own label again underneath
itself — and the tap that would dismiss it is a no-op. Matches desktop.
* fix(mobile): keep a blank tool argument out of the run header (STA-3333)
Skipping a present-but-blank primary key let `briefToolArg` fall through
to the raw JSON preview, so a run header read `Bash {"command":""}` where
it used to read `Bash`. Also state the search-path trade-off honestly:
suppressing the link costs a file-scoped search its tap target.
* fix(mobile): only treat a blank primary key as a missing argument (STA-3333)
The previous guard tested key presence, so a populated but non-string
argument — a mixed argv like ['kill','-9',pid], or a structured query —
dropped out of the run header instead of falling back to the preview.
* test(mobile): pin the tool-row chevron to the detail panel (STA-3333)
The panel gate was covered but the chevron beside it was not: swapping
`showDetail` back to `expanded` on the icon alone left all 909 mobile
tests green, so the affordance lie this branch fixes could return
unnoticed — a down-chevron over no panel, on a row whose tap is guarded
off.
Asserts both icon counts on the fixture that test already renders. The
two halves now die for distinct reasons: the panel gate on the duplicate
label text, the chevron on the icon count.
* test(shared): pin the blank-search-key guard in the tool label (STA-3333)
Dropping `.trim()` from summarizePrimaryToolArg left all 32 tests green,
yet it leaks through isSearchToolInput: a whitespace-only `query` starts
counting as a search term, which suppresses `path`. One character takes
the row's label, its tap-to-open link and its run-header argument at
once, and puts the raw JSON label back — the bug this branch removes.
Asserts all three outputs on that shape. Kills only that mutant; the
isSearchToolInput mutant still dies on the existing search test.
* fix(native-chat): share tool input display semantics (STA-3333)
Build the tool row label, file target, detail eligibility and bounded detail from one normalized input model. Mobile no longer reparses JSON-string input across independent helpers or repeats an already-complete plain label, and desktop now uses the same clean row summary instead of retaining raw JSON.\n\nKeep full detail formatting lazy for collapsed rows and share the 4000-character detail cap across both renderers. Tests pin desktop adoption, mobile disclosure parity, one-pass JSON parsing and the shared bound.
* fix(mobile): keep native chat ask dismissals tab-scoped and gated
Dismissal state lived in the chat view subtree, which unmounts on a
chat<->terminal toggle, so an answered ask card came back on return. It
also had no tab scope and no waiting/blocked gate.
- move dismissal into the controller, keyed per session tab
- gate ask cards on waiting/blocked like the permission path already is,
and retire a dismissal off the ungated detected prompt so a working/done
status can't be mistaken for the prompt clearing
- ignore a dismissal that settles after its prompt cleared or was replaced
Refs STA-3333.
* fix(mobile): keep an ask dismissal through the transcript re-subscribe
A view toggle or tab switch re-subscribes the native-chat transcript, and
useMobileNativeChatSession withholds `messages` until that read settles. A
transcript-derived ask therefore reads as null while the chat surface is
already visible, so the reset effect took it as "the agent moved on" and
retired a live dismissal — the answered card came back, which is the bug
the off-chat guard was meant to close.
Treat an unobserved null as unobserved: `observing` now also requires the
read to have settled. A prompt that is already detected stays observable on
its own, so a status-derived ask still registers on first paint and an
answer taken during that first load is still accepted.
* fix(mobile): keep the transcript-derived ask outside the paused gate
A hook row idle past AGENT_STATUS_STALE_AFTER_MS (30m) projects to `done`
with no interactivePrompt, so the transcript fallback is the only source
left for a still-pending question. Gating it behind waiting/blocked made
that question unanswerable from mobile. Only the sticky status payload
needs the gate; `extractPendingAsk` clears itself on the tool result.
Also pins the load-window clause in the ask-observability guard, which
was behaviourally load-bearing but killed no test.
* fix(mobile): treat a never-read transcript as unobserved, not as "no ask"
The ask-observability guard only excused `transcriptLoading`, which is true
for an in-flight read alone. useMobileNativeChatSession also withholds
`messages` when the client is gone ('idle') or the tab has not reported a
provider session yet ('waiting-session') — both leave the flag false over an
empty list that was never read. The derived prompt then read as null, the
reset effect took that as "the agent moved on", and a live dismissal was
retired; when the read landed with the question still pending the answered
card came back — the resurfacing bug this guard exists to close.
Gate on the read having actually settled instead. 'error' still counts: it
keeps the last successful read in `messages`, so a prompt that clears under
it is real evidence, unlike a list that was never populated.
Also locks three guards that killed no test: the sticky-status suppression
of the transcript fallback (which is what makes the new paused gate hold in
the post-answer window), the reset effect's identity bail-out, and showAsk's
empty-prompt case. The transcript stand-in now derives `transcriptLoading`
from `status` the way the real hook couples them, so these tests can only
express states the session hook can reach.
Refs STA-3333.
* test(mobile): pin the ask dismissal's tab scope and ungated retirement input
Both wirings were unpinned: swapping `scopeKey` to a constant or feeding the
gated `ask` in as `detectedAsk` left the whole mobile suite green.
* fix(mobile): require a landed read before an errored transcript retires a dismissal
`status === 'error'` was treated as settled on the claim that an error keeps
the last successful read in `messages`. That only holds for an error that lands
on top of an earlier read. The host forwards an initial-drain failure as an
error frame carrying an EMPTY list (transcript-watch-error.test.ts), the mobile
frame applier checks `frame.error` before the messages array so those rows are
discarded, and the session hook's error path never calls `setMessages` — so a
first-read error leaves `messages` at the `[]` the identity-change effect wrote.
That frame is also not terminal: the watcher keeps `initialDrain` true and a
real snapshot follows once the read recovers. So a re-subscribe whose first
read errors made the never-populated list read as "no ask", retired the live
dismissal, and the recovered snapshot brought the answered card back over the
composer — the exact resurfacing this guard exists to close, and most likely on
remote/SSH transcript reads.
Require rows for the error case. Rows can only be present once a read landed,
so the predicate is never wrong in the resurfacing direction; it only declines
to retire a dismissal when the transcript was never observed.
Also drop the dismiss hook's `detectedAsk = ask` default and make both prompts
required. That default silently fed the gated prompt in as the detected one,
which is the pre-fix behavior: a paused-out card would read as "prompt gone"
and retire the dismissal. tsc now enforces the ungated payload at every call
site instead of leaving a trap for the next caller.
* fix(mobile): scope the ask dismissal to the provider session, not the tab
A restart, /clear, or resume swaps the provider session inside one tab. The
next session's first question is often byte-identical, so a tab-keyed dismissal
hid the live card and left the turn blocked with nothing to act on.
* chore: restore upstream formatting
* feat(mobile): native-chat model/session-option picker + shared slash catalog (STA-3332)
Piece A — shared slash catalog + send classification:
- Mobile composer now serves getVerifiedNativeChatCommands from the shared
catalog (agent-aware, with description rows) instead of a hardcoded
provider-agnostic list that advertised commands Claude does not have.
- classifyNativeChatSend moves to src/shared/native-chat-slash-commands.ts
(renderer re-exports keep desktop import paths stable); mobile's send seam
now gates optimistic echoes on it, so slash sends no longer create a
'Queued' bubble that no transcript echo can ever retire, and the
ack-lost hold only arms for chat sends.
Piece B — mobile model/session-option pickers:
- New per-tab session-option tracking (state/commands/labels modules) ported
from the desktop live flow, reading the shared agent-session-option
catalog for Claude AND Codex.
- Composer pill row (model + options) opening an inline choice card in the
proven Ask-card pattern; applies use catalog modelApply semantics
(/model <value> via the existing send path), Codex-style agent-picker
entries dispatch the picker command and flip the tab to the terminal view.
- Current model seeds from the hook-reported provider model when derivable;
typed /model-style commands update tracked state (recordOutgoingCommand
parity); dispatched values render as sent-not-confirmed.
* fix(mobile): keep session option sends scoped
* fix(mobile): synchronize native chat refs after commit
* refactor: share native chat session option logic
* fix(mobile): keep the live tab's session-option record from eviction
`getScopedRecord` returned an existing record without re-inserting it, so the
per-tab record map evicted by insertion order rather than recency. A long-lived
active tab is the oldest key, so crossing the 32-scope cap silently dropped its
tracked model and reset the pill to "Model". Desktop's scope cache does
delete-then-set for exactly this reason.
Also moves the shared session-option tests to src/shared so the root suite runs
them (they only exercised src/shared logic the Electron renderer consumes, but
sat under mobile/ where only mobile's vitest project sees them), and restores
two "why" comments dropped while extracting the shared modules.
* fix(mobile): stop a stale session-start report reverting a model pick
Re-entering a chat tab re-delivers the same `agentStatus.model`, and the
reported-model effect re-applied it unconditionally — so picking a model, moving
to another tab, and coming back reverted the pill to the model the agent reported
at session start, which cannot have observed the `/model` sent after it. The
status stream reconnecting had the same effect.
A report is now only treated as evidence when the matched catalog id CHANGES for
that scope; a genuinely new report still supersedes a local pick. Mobile has no
screen read to confirm a switch against, so the repeat is all we can key off.
* fix(mobile): close four session-option picker defects found in review
D1 — a picker apply could interleave with a composer send. The composer already
blocks a text send while an apply is dispatching, but not the reverse: the host
spaces a send's body and its Enter ~500ms apart, so an apply tapped inside that
window was submitted as part of the user's prompt, and the pill then claimed a
model change that never ran as a command. The pickers render inside the composer,
so they now take its in-flight state directly — the same guard, mirrored.
D2 — an option was filed under the wrong model. `setTrackedSessionOption` resolves
the owning model when it commits, not when the command was built, and the report
effect mutates the same record off-queue. A report landing mid-dispatch therefore
recorded `/effort low` against the model it switched TO. Ports desktop's
supersession guard, which skips the commit when the baseline moved.
D3 — a command template's prefix also matches prose that starts with it, so
"/model is a weird word" tracked that prose as the current model, rendered it as
the pill label, and matched no catalog model, dropping every per-model option.
Parsed values are now canonicalized against the catalog; a typed value containing
whitespace is treated as a prompt rather than a command.
Perf — `/` on a Codex tab returned all 45 commands into a non-virtualized
ScrollView showing ~5, re-reconciled on every streaming tick above the transcript.
Capped at 12.
Also splits the row primitives out of MobileNativeChatSessionOptionPickers.tsx,
which the D1 guard pushed to 402 effective lines against a 400 cap.
* refactor: share the session-option display ordering
CATEGORY_ORDER and the non-model sort were byte-identical in
NativeChatSessionOptionPickers.tsx and mobile's labels module — pure logic with
no i18n in it, so there was no reason for two copies that can drift. Both now
call sortNativeChatSessionOptions from the shared snapshot module.
* refactor(mobile): align model picker layout
* style(mobile): round native chat composer
* fix(mobile): inset rounded chat composer
Breaks two independent React #185 (Maximum update depth exceeded) crash loops.
- Activity portal publication is idempotent by descriptor value, so a semantic
no-op no longer bounces synchronously through Terminal and back into Activity.
- The Activity readiness burst budget survives slot, target, pane, and tab
retargeting, and a quiet loading pane is rechecked when the window expires.
- Terminal cold-parking pins verdict bursts to the safe mounted side before
React reaches its nested-update limit, with an expiry so tabs can park again.
Supersedes #12492 and #12485.
Portal descriptor equality is keyed off keyof ActivityTerminalPortalTarget so a
new field fails the build instead of silently suppressing a publish. Park-verdict
damping and breadcrumbs gate on pin liveness, and churn crumbs coalesce by
trigger so a burst cannot collapse into a slow-churn slot.
* perf(tabs): take split-divider drag off the store (STA-3328)
Every pointermove committed a global store write (60-120 publications/s
against every subscriber) plus a forced reflow from per-move
getBoundingClientRect. The drag now writes the two panes' flex styles
directly (identical visuals) and commits setTabGroupSplitRatio once on
release/unmount; the action bails without minting state when the ratio is
unchanged.
* fix(tabs): keep deferred divider commits coherent
* fix(tabs): preserve divider pointer ownership
* perf(terminal): stop the IME candidate anchor forcing layout per compositionupdate
* fix(terminal): refresh deferred IME anchor after refit
* fix(terminal): preserve deferred IME anchor ordering
files.resolveTerminalPath began returning a foreign worktree id + relativePath
for absolute paths owned by a sibling workspace, with no protocol or capability
gate. Mobile 0.0.36 in the field ignores resolved.worktree and reuses its own
worktree id for the follow-up files.open, so a tap on a sibling-worktree path
opened the WRONG worktree's copy of that file (on 1.4.168 the tap was a safe
no-op).
Gate the sibling-workspace lookup behind a new optional crossWorkspace request
field: clients that honor resolved.worktree opt in; everything else keeps the
pre-sibling-resolution contract. Old servers strip the unknown field (zod), so
every version pairing degrades to the safe legacy behavior. Optional-field
addition, so no RUNTIME_PROTOCOL_VERSION bump per protocol-version.ts rules.
The terminal-path RPC tests move to files-terminal-path-resolution.test.ts
because files.test.ts sits at the max-lines cap.
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
* perf(runtime): gate terminal.list visual layouts and stop the false writable claim
visualLayouts is ~31% of a large terminal.list payload (44,208 B of 137,412 B on a live 134-terminal remote runtime) and has exactly one consumer: the human-readable CLI formatter. Gate it behind an includeVisualLayouts request param that defaults to included, so pre-flag clients are unaffected, and have every --json/internal caller opt out.
Also drop the record-backed builder's writable, which was a verbatim copy of connected. terminal.show now states writability explicitly as exactly what terminal.send's PTY gate enforces.
* test(runtime): type the payload-size fixture arrays for tsc
* fix(runtime): preserve terminal list compatibility
* test(runtime): guard terminal list optimization
* fix(cli): preserve agent access to terminal layouts
* fix(terminal): attach never-activated daemon sessions on remote subscribe and provider-read fallback
A daemon-backed terminal whose tab was never activated in the host UI was
never attached, so the daemon emitted no bytes: paired clients rendered
blank/frozen panes and `terminal read` returned an empty tail while the PTY
was alive.
- Runtime: first remote view subscriber of a known-but-unattached local
daemon session triggers an attach through the pty controller — attach-only,
no resize, no renderer mount/focus, headless-safe, deduped across
concurrent subscribers, and never detached on release. Excludes SSH-scoped
ids and sessions a local spawn already published this generation.
- Read path: withVisibleSnapshotFallback now falls back to the provider tail
for an empty-tail never-attached live local session; unprovable state stays
empty, never an error.
- pty controller: expose attach with getProviderForPty-style routing,
answering false on doubt; local daemon provider only.
- Daemon adapter: attach rides the session's applied size instead of a
hardcoded 80x24, sends attachOnly, and retires a pre-v31 daemon's
accidental spawn instead of publishing it.
Deterministic harness drives the real terminal.multiplex handler against a
real OrcaRuntimeService with an injected daemon-model controller whose data
events are gated on attach; covers snapshot-capable and snapshot-null
daemons, concurrency, release, replacement-spawn exclusion, and negative
safety. Red on base, green with the fix, red again with the fix reverted.
* fix(terminal): refuse degraded-provider attach fallback and surface failed legacy-spawn retire
Verifier follow-ups on subscriber-driven daemon attach:
- DegradedDaemonPtyProvider.attach routed unknown ids to the in-process
fallback, whose no-op attach resolves — the runtime then pinned a
subscriber-driven attach as succeeded while the stream stayed blank.
Attach now refuses any route that resolves to the fallback (a fallback pty
cannot own a daemon-surviving session), so the controller answers false,
no sticky success is recorded, and a later subscriber attaches once a
daemon adapter proves the id. Session-probe adoption moved to
degraded-daemon-session-routing alongside the new refusal.
- The pre-v31 attach-only TOCTOU retire (accidental legacy spawn kill) now
logs a warning with the sessionId on kill failure instead of swallowing
it, so an orphaned replacement shell is diagnosable.
Regressions: degraded provider refuses unowned/fallback-owned attach and
routes to a daemon once it proves the id (red on previous commit); runtime
harness pins refused-attach retry for a later subscriber; adapter test pins
the surfaced kill failure.
---------
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
* fix(terminal): seed list/read records from reattach restore payloads
After an app relaunch the PTY daemon survives and spawn silently
reattaches, but the restore payload (reattach snapshot, cold-restore
scrollback, relay replay, lastTitle) arrives as a spawn RPC result and
never passes through runtime.onPtyData — the only feeder of the terminal
records behind `terminal list`/`terminal read`. Every restart therefore
left connected terminals with empty title/preview/lastOutputAt and a
zero-line read tail, blinding orchestrators that poll terminals.
The spawn flow now calls runtime.seedTerminalRestoreTail with the restore
text and lastTitle, unconditionally of the renderer-authority emulator
gate (the records are main-side only). The seed reuses the live path's
normalize/tail/preview pipeline on a capped 256 KiB suffix (re-anchored
at a line boundary so a cut escape cannot leak), only fills records that
never saw output (a remount reattach cannot re-apply history), routes
titles through the applySeededAgentStatus precedent (state writes only —
no waiters, no side-effect facts), and never stamps lastOutputAt or
waitBlockedAt: restored bytes are historical, not fresh activity.
lastTitle is threaded from the daemon reattach snapshot and cold-restore
checkpoint into PtySpawnResult; relay replays seed preview only. SSH and
runtime-controller paths are unchanged — seeding is gated on the fields
existing.
* fix(terminal): seed restore records on the controller spawn path and prime the wait baseline
Follow-ups to the restore-record seed, from independent verification:
1. The runtime-controller spawn flow (createTerminal background creates —
headless `orca serve`/CLI — and pane splits) never consumed restore
payloads, so the exact orchestrator-blindness this fix targets survived
on the topology that needs it most. The extraction now lives in one
helper called from both spawn choke points (renderer pty:spawn and the
controller flow); the runtime's empty-record guard makes overlapping
seeds a no-op.
2. The throttled per-PTY wait scanner starts with a null baseline, so a
permission prompt visible only in seeded HISTORY read as newly gained
on the first benign live chunk and stamped waitBlockedAt "now".
Seeding now primes the scanner baseline from the seeded tail without
stamping; only a signal appearing in genuinely new output counts.
3. Cap re-anchoring accepts \r as well as \n (newline-free CR-redraw
streams), consuming a full \r\n pair so the seed does not start with
a phantom blank line.
---------
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
* fix(runtime): reject leaf terminal sends only on controller-proven PTY absence
orca terminal send to a leaf whose ptyId no provider in this process owns
was a silent no-op reported as success: the graph mirror answers
writable=true, every provider write to an unknown id is accepted
fire-and-forget, and bytesWritten is computed from the payload rather than
delivery. sendTerminal and sendTerminalAgentPrompt now consult a controller
liveness probe when the provider does not synchronously know the id
(hasPty), and throw terminal_not_writable only on an exact false — unknown
liveness, probe errors, SSH/remote scopes, and probe-less providers never
reject (#12393's rule: null is not absence), so a restored daemon session
still accepts writes before its pane remounts. Push-on-idle orchestration
delivery gains the same gate so a proven-dead leaf keeps its messages
queued instead of marking them delivered into a void.
The pty controller now exposes probePtyLiveness, routed like write: a
provider probe is preferred, the in-process local provider's refusal is
authoritative (sole owner), and remote-scoped or SSH ids without a probe
answer null after awaiting the cold-start daemon swap. Proven-absent
verdicts cache 15s per ptyId with in-flight dedupe, superseded the moment
the provider re-learns the id.
* fix(runtime): arm one probe-deferred delivery continuation per pty
Review (GPT verifier) confirmed: triggers arriving during one in-flight
absence probe each attached a continuation to the deduped probe promise, and
since Claude-target delivered_at stamps only after the delayed Enter, every
continuation re-read the same unread rows — double payload injection and two
armed Enters. Single-flight the deferred continuation per pty; the one armed
continuation re-reads fresh rows when it fires, so nothing is lost, and the
guard clears on settle so later triggers defer again. The narrower
pre-existing 500ms sync-path window is unchanged and out of scope.
* fix(runtime): single-flight the whole orchestration delivery window per pty
The probe-continuation guard cleared at probe settle, but Claude-target
delivered_at stamps only in the delayed-Enter callback ~500ms later — a
trigger landing in that gap armed a fresh probe cycle, re-read the same
un-stamped rows, and re-injected the payload. The identical window existed
on the pure sync path pre-PR (two triggers within 500ms double-deliver).
Hold a per-pty delivery-in-flight flag from before the payload write until
delivery settles: entry-checked before reading unread rows, cleared through
one settle point covering the failed write, the sync-stamped coordinator and
Cursor branches, any sync throw, and the delayed-Enter callback on submit,
refusal, and throw alike. A trigger arriving mid-flight is not dropped — it
parks the latest leaf per ptyId and re-runs delivery once on settle, so rows
inserted mid-flight deliver without waiting for the next idle event. The
probe single-flight stays; the new guard subsumes its post-settle gap, and
no trigger site bypasses it.
Both strengthened tests are red on the previous commit (first subject
injected twice) and green here: in-window re-trigger on the probe path and
sync-path double-trigger each deliver the first batch exactly once, with the
parked second row delivering alone after settle.
* fix(runtime): retire the armed delivery Enter on pty exit; guard fire-time on current state
Two variants of one root cause — the delayed-Enter callback outliving the
session it was armed for:
1. Cold restore respawns under the same session id. onPtyExit never
cancelled the armed Enter or the in-flight delivery state, and
onPtySpawned flips the same leaf writable again — so an exit + same-id
respawn inside the 500ms window let the stale callback inject \r into
the replacement session and stamp rows it never received, then settle
against a newer same-id flight.
2. Graph resync replaces leaf objects, so onPtyExit flips writable=false
only on the current replacement; a callback trusting its closed-over
snapshot still read writable=true and fired after exit with no respawn.
The flight record now carries its armed Enter timer and serves as settle
identity: onPtyExit clears the timer and drops the flight and any parked
re-delivery without stamping (rows stay unstamped and re-deliver on the
replacement's next idle — the existing contract), and settle no-ops unless
its own flight is still current, so a stale settle can never clear a newer
same-id flight or flush its parked trigger. At fire time the callback
re-resolves the leaf by key and requires the same ptyId binding and current
writability instead of reading the closure snapshot.
All three regressions are red on the previous commit: same-id respawn saw
\r plus a false delivered_at stamp, exit leaked the flight and parked
state, and the orphaned-snapshot resync variant fired Enter after exit.
---------
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
* fix(runtime): report terminal handles disconnected on controller-proven PTY absence
leaf.connected mirrors the renderer graph (ptyId !== null), so a restored
surface whose PTY died with a prior process was listed connected/writable
forever with empty title/lastOutputAt/preview — the exact signature automation
saw on run6 workspaces after a restart. listTerminals now threads the
controller inventory it already fetches into buildTerminalSummary and demotes
only on proven absence, only for locally-scoped ids; unknown liveness and
SSH/remote scopes never demote, and no session or pane is retired.
* fix(terminal): stop forking hidden restorable panes into replacement resume tabs
paneWillConnectOnActivation still assumed the pre-keep-alive mount model, but
every non-parked tab of the active worktree mounts and connects hidden at 0x0.
Activation therefore appended a replacement resume tab per non-group-active
agent pane and handed it the sleeping record, stranding the hidden pane as a
bare shell — or forking two live surfaces onto one provider session when the
old PTY survived in the daemon. The predicate now answers "will mount and
connect": any non-web-mirror tab of the active worktree qualifies; non-active
worktrees still answer false so background wake keeps its append-based resume.
Contract change: reverses the hidden-tab expectation from #6800, whose premise
(hidden panes never connect) no longer holds; that test is updated in place.
* test(terminal): pin the remote-scope exemption and the web-mirror ownership exception
CodeRabbit flagged both exclusions as untested: a remote-runtime-scoped leaf
absent from the local inventory must stay connected (its inventory lives on
the remote host), and a web-mirror tab must not own sleeping-session recovery
(it never mounts a local pane), so the appended replacement remains its
correct resume path.
* fix(terminal): rescue just-spawned ptys from absence demotion; unpark panes owning sleeping records
Review (GPT verifier) confirmed two gaps:
- listTerminals demoted a live just-spawned PTY when listProcesses snapshotted
before session registration (the sweep's hasPty rescue is leaf-gated), and
federation reads one connected:false as exited. The summary's proven-absence
check now also consults the provider's sync hasPty.
- Ordinary per-tab cold parking (30s hidden) kept a non-group-active pane
unmounted, so a sleeping record it owns under the new ownership predicate
could not cold-restore until the user revealed the tab. Per-tab parks now
exempt panes owning a sleeping-session record; worktree-level parks are
untouched (they clear on activation).
* fix(terminal): reconcile the daemon session cache on inventory; scope the park exemption to consumable records
Round-2 review confirmed two holes in the round-1 fixes:
- DaemonPtyAdapter.hasPty is cached activeSessionIds membership, and a
successful listSessions never removed ids the authoritative inventory
omitted — an exit missed while the socket was down kept hasPty true
forever, and the new spawn/list-race rescue would trust it, reopening
connected-forever for that pty. listProcesses now drops pre-request cached
ids the inventory does not list alive (ids spawned mid-flight are snapshot-
protected).
- The park exemption covered records a pane can never consume
(automaticResumeBlockedBy, passive-completed evidence), pinning hidden
panes mounted indefinitely. The exemption now lives in
sleeping-record-park-exemption.ts and requires a consumable record.
Also pins the web-mirror replacement's resume claim and startup command
(CodeRabbit round-2).
---------
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
* fix(status-bar): invalidate the CLI session count on kill and restart
`pty:management:killOne` / `killAll` / `restart` tear sessions down via `adapter.shutdown()` and broadcast nothing — unlike `pty:kill`, which ends in `sendPtyExitToRenderer`. The status-bar count is an event-sourced cache, so killing sessions from Manage Sessions or "Kill all terminals" left the `>_ N` chip frozen until the popover was opened, which itself triggers a refresh.
> [!NOTE]
> The dual-source split described in the issue text was already fixed by merged #9387. This closes a *different* remaining invalidation gap that produces the same reported symptom.
Broadcast the teardown so the chip updates without needing the popover opened.
Fixes#8372
Co-authored-by: Orca <help@stably.ai>
* test(e2e): add recordable proof for status-bar-cli-session-count
Fails on origin/main, passes on this branch.
Test: drops after Manage Sessions kills a foreign daemon session, popover never opened
Co-authored-by: Orca <help@stably.ai>
* fix(status-bar): avoid duplicate inventory refresh after kill all
---------
Co-authored-by: Orca <help@stably.ai>
* fix(setup): stop caching an unreadable orca.yaml as "no setup script"
`checkRepoHooks` returned `{hasHooks:false, hooks:null, mayNeedUpdate:false}` with no `status` field when the SSH filesystem provider was unavailable, and inside a blanket catch for any read error. The renderer only bails on `status === 'error'`, so that status-less false negative was cached as an authoritative "no setup script" and the prompt stayed on screen.
Mirror the `hooks:check` IPC twin exactly: `status:'error'` for a missing provider, ENOENT-aware in the catch, `status:'ok'` on the folder-repo, binary, SSH-success and local branches.
Fixes#8752
Co-authored-by: Orca <help@stably.ai>
* test(e2e): add recordable proof for setup-script-prompt-false-negative
Fails on origin/main, passes on this branch.
Test: recovers from an unreadable orca.yaml instead of pinning the failed verdict
Co-authored-by: Orca <help@stably.ai>
---------
Co-authored-by: Orca <help@stably.ai>
* perf(tabs): index tab agent status by tab instead of scanning the global map
resolveAnyCompletedTabAgent and its live/retained twins scanned the whole
agentStatusByPaneKey map and parsed every pane key, once per tab per render —
~10^5 parsePaneKey calls per render pass with 200 tabs. Cache a per-tab pane
index on the map's identity (the store replaces it on every write) so a render
pass scans once instead of once per tab. Insertion order is preserved because
the resolvers return the first match.
* test(tabs): lock agent status index scan count
Jinjing
Jinwoo Hong
Jihwan Kim
Neil
Shawn
PLANDER DEV 7
Brennan Benson
OrcaWin