mirror of
https://github.com/stablyai/orca.git
synced 2026-09-29 00:02:56 +00:00
ae426bfbeb024fc4d0f3c7ac0dafc84dfcdacc4e
381
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 Neil
|
8d8b9dad78 |
fix: keep macOS shell ownership proof within recovery budget (#18932)
* fix: keep macOS shell ownership proof within recovery budget * fix: parse the shell-proof column set with its own anchored parser The narrower macOS capture (`pid ppid pgid tpgid stat command`) was fed to the shared lenient parser, whose optional tty/start pair has no `tty=` column left to absorb it. It then eats the head of any argv shaped `python 3 app.py` (parsing command as `app.py`, tty as `/usr/bin/python`), and turns a command-less row into a garbage pid/stat pair. Either can flip a shell ownership verdict, which is what gates dead-TUI recovery. Give the column set a named constant and a parser anchored to exactly those six columns, beside its `CHEAP_PS_ARGS` sibling. A capture that yields no rows now raises `empty_capture` rather than reading as a machine with no processes. Update the `confirmShellForegroundProcess` fixtures from the 4-column legacy shape to the 6 columns the darwin reader actually emits; that describe block already forces `platform=darwin`, so the stale fixtures were failing. |
||
 Jinwoo Hong
|
0c33f58e8a |
fix(ssh-relay): daemon owns the endpoint credential; a losing start never rotates it (#19052)
<!-- orca-pr-loc -->
<!-- Programmatic LoC summary. Do not edit by hand; rewritten on every commit. -->
| | Files | Added | Deleted | Net |
| :--- | ---: | ---: | ---: | ---: |
| Test | 19 | $\color{#1a7f37}{\Huge{\mathbf{+}}}$962 | $\color{#cf222e}{\Huge{\mathbf{−}}}$136 | $\color{#1a7f37}{\Huge{\mathbf{+}}}$826 |
| Prod | 18 | $\color{#1a7f37}{\Huge{\mathbf{+}}}$295 | $\color{#cf222e}{\Huge{\mathbf{−}}}$116 | $\color{#1a7f37}{\Huge{\mathbf{+}}}$179 |
<!-- /orca-pr-loc -->
## Symptom
Live 2026-09-05 (Orca 1.4.198 client, Ubuntu host): both relay processes `kill -STOP`ped for 20 s, then `-CONT`. The client redeployed while the host was frozen. Its fresh daemon lost the socket bind (`Socket path already in use`) but had **already rewritten** `relay-<id>.sock.credential`. The surviving daemon kept its in-memory credential, so every later `--connect` got `Endpoint credential mismatch; closing socket`, then `Grace started … timeoutMs=0 … ptys=1, clients=0` every ~20 s, forever. Only a manual `kill -TERM` cleared it. Receipts: `review-archive/orchestration-v3-pr16904/smoke-receipts-t012b/E16,E17,E18,E24`.
Three independent defects kept the wedge alive; each is fixed at its own seam.
## Fix
**1. The relay daemon owns credential publication (race-free under two concurrent starters).**
`relay-daemon.ts` binds the socket first, then publishes via the new `src/relay/relay-endpoint-credential-publication.ts`: adopt a valid pre-existing file (older clients still pre-write), else mint 32 random bytes and write temp+rename at 0600. A start that loses the bind exits inside `listen()` and never reaches the file. Why this option and not restore-on-loss or a client-side write: the only process that can *prove* ownership is the one whose `listen()` succeeded, and that proof is atomic with the bind. The client-side pre-write (`ssh-relay-endpoint-credential.ts`) and the launch-command `chmod 600`/`icacls` are removed on POSIX and Windows. The racing test also exposed that macOS reports a mid-bind collision as `EEXIST` rather than `EADDRINUSE`; `relay-socket-ownership.ts` now treats both as "held or stale".
**2. The client distinguishes "no daemon" from "daemon present but not answering", and never rewrites.**
A credential refusal is now typed on the wire: the daemon replies `orca-relay-handshake-credential-mismatch` (same frame type, no new opcode) and the bridge exits **43**; `waitForSentinel` maps it to `RelayCredentialMismatchError`, which the takeover treats as handshake-refusal evidence exactly like exit 42. A relay that holds the endpoint but **never refused** (the stalled-host shape: kernel backlog accepts the probe, handshake gets no answer) is now `RelayEndpointUnresponsiveError`, routed to the relay-lost backoff instead of the terminal Reset Relay path. Silence is not a decision (`docs/reference/ssh-execution-boundary.md`).
**2b. Deploy honours the verdict.** The 40 s live run exposed that the `--connect` catch block in `deployAndLaunchRelay` predates the incumbent probe and swallowed both verdicts as "probe failed, launch fresh", so a fresh daemon was still launched over the live one (it lost the bind by luck, which is exactly the collision in the incident). Held and Unresponsive now propagate; the session backs off on Unresponsive and surfaces Reset Relay on Held. Red-first in `ssh-relay-deploy-incumbent-verdict.test.ts`.
**3. The daemon cannot be wedged by a rotated file, because nothing can rotate it.**
The credential lives in the content-hashed relay dir, and after (1) the only writer is the daemon that owns the socket, so the "file changed under a live daemon" state the incident depended on is no longer reachable in-product. The credential is therefore fixed for the daemon's lifetime, as a plain secret should be. A hand-edited file is refused with the typed reply until restored (tested). Startup adoption of a pre-written file applies an owner-only + same-uid rule (review finding): anything else is replaced by a fresh mint. An earlier revision of this PR also re-read the file on mismatch and adopted it; that was removed as unreachable machinery that turned the credential into a per-handshake file-ownership check.
**3b. Fail closed between bind and publication.** A client that arrives after `listen()` resolves but before the credential is set is refused, not admitted as `unproved`. Nothing can be delivered in that window today; the guard makes the boundary structural instead of an event-loop ordering fact. Red-first in `relay-reconnect-listener-credential-gate.test.ts`.
**Wire compat.** New optional handshake reply only; an old `--connect` hits `Unknown handshake type` and exits 1 pre-sentinel, which it already treated as a generic failure. New daemon adopts an old client's pre-written file; new client still passes `--credential-file` so an old daemon reads it as before. Absence of exit 43 is never used as evidence.
**Also.** `terminal create` on a reconnecting SSH host now says what to do instead of a bare `No PTY provider for connection "<id>"` (prefix preserved; the renderer matches it).
## Tests (red first)
- `src/relay/subprocess.test.ts`: two `--detached` starts race one socket + credential file → exactly one reaches the sentinel, loser exits 1 with `Socket path already in use`, file valid + 0600, a `--connect` reading it reaches `relay.status` and reports the winner's pid. Red before (both starters died: daemon required a pre-existing file), green 6/6 after.
- `src/relay/relay-endpoint-credential-publication.test.ts`: mints after bind; adopts a pre-written 0600 file; replaces a pre-written 0644 file with a fresh mint; refuses a stale credential with exit 43 while still serving the real one, and keeps refusing a rewritten file until it is restored.
- `src/relay/relay-reconnect-listener-credential-gate.test.ts`: a client in the bind-to-publish window is refused and never attached; after publication the right credential is accepted and a wrong one refused; a daemon launched without a credential file is not gated. Red without the guard.
- `ssh-relay-deploy-incumbent-verdict.test.ts`: live-but-silent incumbent → `RelayEndpointUnresponsiveError`, refused → `RelayEndpointHeldError`, and in neither case is `--detached` launched; a failed `test -S` probe still launches fresh. Red 2/3 without the deploy change.
- `ssh-relay-deploy-helpers.test.ts` (exit 43), `ssh-relay-endpoint-takeover.test.ts` (refused → Held even with no `lsof`; silent → Unresponsive, nothing unlinked or signalled), `ssh-relay-session-terminal-error.test.ts` (Unresponsive → `onRelayLost`, not terminal). Deploy/namespace/native-deps tests updated to assert the client writes **no** credential.
## Live proof
New `tests/e2e/ssh-docker-relay-stall-credential.spec.ts` (claimed in `run-ssh-docker-e2e.mjs` and PR source routing), two cases: `kill -STOP` every relay pid in the container, send input during the freeze, hold **20 s** (the incident's duration, which races the mux liveness timeout) or **40 s** (past it for sure), `kill -CONT`; assert status back to `connected`, same pty, same daemon pid, same credential inode and content, relay.log did not shrink (a relaunch truncates it) and has zero `Endpoint credential mismatch` / `Socket path already in use` lines, in-stall input delivered at most once.
Run output (local, fixture image `orca-e2e-ssh-relay:3a864c665ba2cefd`, `ORCA_E2E_SSH_DOCKER=1 SKIP_BUILD=1 ORCA_E2E_FORWARD_APP_LOGS=1 … --project electron-headless --workers=1`, head `c2c20fd994`; re-run identically on the final head after the credential-lifetime change, 2 passed (1.7m), same annotations, and the bind-to-publish refusal never fired):
```
✓ keeps the same daemon and credential across a 20s relay freeze (38.3s)
relay-processes-stopped: 2 relay-processes-continued: 2
bridge-pids-before-after: 480 -> 480
socket-clients-accepted-before-after: 1 -> 1
in-stall-input-delivered: 1
✓ backs off and reattaches, never relaunching, across a 40s relay freeze (57.5s)
relay-processes-stopped: 2 relay-processes-continued: 4
bridge-pids-before-after: 480 -> 1202
socket-clients-accepted-before-after: 1 -> 3
in-stall-input-delivered: 1
2 passed (1.6m)
```
Client log in the 40 s case shows the new path end to end: `Relay channel lost … reconnect attempt 1/6` → `Socket probe result: "ALIVE"` → `Socket reconnect failed … Relay failed to start within 10s` → `Relay endpoint incumbent: … verdict=live evidence=accepted-connection holders=unenumerable` → `Failed to re-establish relay … A relay still owns … but did not answer the handshake … Orca will retry` → `reconnect attempt 2/6` → `Reconnected to existing relay via socket`. The 20 s case never left the frozen bridge (same bridge pid, one accept), so it exercises the "silence is not death" side of the same race. The 20 s case passed 6/6 across the session; the 40 s case was red on the prior head (`Socket path already in use` + `Startup failed: listen EADDRINUSE` in relay.log from the swallowed verdict) and is green after 2b. Before the fix the same injection produced a fresh daemon that rewrote the credential and a survivor refusing every client.
The `relay-processes-continued` count exceeds `stopped` in the 40 s case because the timed-out client's `--connect` bridge and the loser-side processes are parked behind the frozen listener when `CONT` runs; they exit on their own once it resumes.
## Gates
`pnpm test src/relay src/main/ssh` 332 files / 3884 tests pass · `pnpm typecheck:tsc:node` clean · `check:code-quality:changed` 0 findings · `check:react-doctor:changed` 0 findings · `pr-e2e-gate-contract.test.mjs` 42 pass · no lint disables or max-lines bumps added.
## Noted, not fixed here
- `terminal list` `orphaned:false` / `terminal close` `ptyKilled:true` for a pane whose relay is gone (`orca-runtime-stop-explicitly-closed-tab-ptys.ts`): different seam, `@ts-nocheck` characterization-covered file.
- On a host with no `lsof`, a stalled relay still cannot be enumerated as the holder; it is now retried rather than declared held, but a relay frozen past the backoff budget still ends in the existing "reconnect manually" banner.
|
||
|
Jinwoo Hong
|
06a607a1d7 |
feat(orchestration): make multi-agent workflows durable (#16904)
<!-- orca-pr-loc -->
<!-- Programmatic LoC summary. Do not edit by hand; rewritten on every commit. -->
| | Files | Added | Deleted | Net |
| :--- | ---: | ---: | ---: | ---: |
| Test | 225 | $\color{#1a7f37}{\Huge{\mathbf{+}}}$21666 | $\color{#cf222e}{\Huge{\mathbf{−}}}$2820 | $\color{#1a7f37}{\Huge{\mathbf{+}}}$18846 |
| Prod | 348 | $\color{#1a7f37}{\Huge{\mathbf{+}}}$17107 | $\color{#cf222e}{\Huge{\mathbf{−}}}$4706 | $\color{#1a7f37}{\Huge{\mathbf{+}}}$12401 |
<!-- /orca-pr-loc -->
## ELI5
Orca now treats orchestration like a durable control plane instead of inferring success from terminal keystrokes. Agents can tell whether a prompt was accepted or a turn started, replay an ambiguous request without sending twice, and recover coordinator mail after a crash. Completed workers can be inspected, released, or retained, and their panes no longer auto-resume as if the work were still running.
## What changed
- **Run receipts** from `run-create/use/current/show/list` are the row without routing plumbing (`home_database`, `coordinator_pane_key`) and without the duplicate `binding` object.
- **`terminal send` receipts are honest and idempotent.** `input_accepted` and `turn_started` are the only stages; `--wait-submit` observes without resending; `--retry-request <uuid>` replays the exact request against the same process incarnation. A transport timeout keeps the retry ID; only a different runtime answering strips it. Value-less or non-UUID `--retry-request` is rejected on the CLI and the SSH shim.
- **Mailbox delivery is committed before wakeup.** Pointer writes are staged in the DB before any PTY byte, replayed once after restart, and never emit a naked Enter. The watermark that parks concurrent deliveries is released with the DB reservation. Restart rescans pointer-pending and `dispatch:` mailboxes.
- **Lifecycle is a guarded transition graph** (`lifecycle-transition.ts`) with a table-driven test over every caller edge. Task reopen/overturn stays in the public contract. A PTY exit during `worker-stop` is the stop succeeding, not a failure.
- **Worker lifecycle CLI:** `worker-start` (`--spec` creates Task + attempt in one call), `worker-show`, `worker-read` (provider transcript first, bounded terminal fallback with a typed reason, local/WSL/SSH), `worker-stop`, `worker-abandon`, `worker-release`, `worker-retain`, `worker-list` (rowid-fenced pagination, fleet liveness, `attention`, literal `nextAction`).
- **Release is an explicit ownership table** (`decideWorkerTerminalRelease`): only an `owned` resource can be settled, the archive is mandatory where reachable, and an owner whose process is proven exited can always get out of `retained` via `archive_status: unavailable`. User-taken-over, external, and transferred panes stay retained.
- **Settled-worker resume fence** (folds in #17651): a settled dispatch whose pane is still open is fenced at settlement, on stop/abandon/exit, and at startup; lifted on release, retain, takeover, and pane reuse.
- **Liveness is `live` / `unverifiable` / `exited` only**, from execution-host evidence. Fleet projection reads the evidence clock, not the relay delivery clock. A host-certified exit outranks the worker's settled state. `unverifiable` never authorizes stop, abandon, retry, or release, in code or in the guide.
- **Federation:** structured reads negotiate by `method_not_found` so every shipped host keeps transcript-first output; exited remote workers are closed before being reported closed; epoch fencing holds across peer restart, downgrade, and pairing rotation; no per-second forced capability probe.
- **Schema v35:** repairs databases stamped v34 by the pre-fix branch (mailbox_handle default, index predicates), drops the write-only `lifecycle_transition_receipts` ledger and five never-read v31 identity columns.
- **Schema v36:** `dispatch:<id>` mailboxes get a real consumer generation on `dispatch_contexts` and `remote_dispatch_attachments`, bumped and fenced in the same transaction on every re-attach (manual inject, worker-start, federated attach). A stale worker whose Dispatch moved to another process now gets `consumer_fenced` instead of silently acking the new worker's Delivery. Run mailboxes already worked this way.
- **Schema v37:** `dispatch_contexts` records its creator (`creator_handle`, `creator_pane_key`), so a coordinator's context-only self-dispatch is bookkeeping rather than a nesting parent; before this, one self-dispatch made every later `worker-start` from that coordinator fail the depth cap. Pre-v37 rows keep counting (fails closed).
- **Dispatch-mailbox ownership is checked, not inferred.** A `check` from a process whose pane no longer holds the Dispatch, or whose last Attempt was abandoned/failed and moved to another terminal, gets `consumer_fenced` instead of an empty inbox that reads as "no mail yet". `--peek`/`--all` stay readable. A paneless caller still gets `stable_pane_required` with the rebind recovery.
- **Liveness certification is stricter:** a `process_exited` stage whose termination reason is `unknown` (a stop that was issued but never observed) projects `unverifiable`, not `exited`. Federated `worker-show` carries the execution host's verdict and host kind instead of a local guess. A live, ready worker with nothing pending has `nextAction: none` rather than pointing at the `worker-show` that produced it.
- **Wire:** `workerShow` keeps `dispatch.task_id` next to `taskId` for shipped CLIs. `ask --json` uses the standard `{ok, result}` envelope like every sibling verb.
- **Migration start-version detection** treats the two v32 recovery columns as versioned. Before this, every shipped database stamped below 32 resolved to the v6 floor and replayed the whole chain (the v23 backfill synthesized 68 phantom retained workers on a real v30 profile). Verified on a copy of a real 62 MB v30 profile: starts at 30, no row delta, integrity ok, 11 ms.
- **Skill guide** rewritten as a ≤200-line kernel plus seven references, to the outcome-first standard (Result / Done / Safe failure first, conditions not case lists, one done bar, references loaded at the point of use). The canonical loop uses `worker-start --spec`, names `worker-list` for completion accounting, documents `--retry-request` / `request-show` / `--wait-submit`, and requires positive evidence before any stall action. The other seven guides get the same treatment in #18724, split out so this PR stays orchestration-only.
- **`rpc/methods/orchestration-*`** (126 flat files) regrouped into `orchestration/{worker,federation,messaging,runs,gates}/`.
## Why
User reports showed the same boundary failures: false `agent_prompt_stalled` causing duplicate sends (#15180), coordinators unable to trust screen scrapes, cold-parked terminals receiving a pointer without the submit, settled workers accumulating as live tabs and auto-resuming after restart, and no way to tell a stalled worker from a working one.
## Linked issues
Fixes #15180. Fixes #17935 (orchestration skill description is 866 characters; a guard now caps every bundled skill at 1,024). Supersedes #17651 (fence folded in). Advances #16660, #16522, #14907, #13047.
## Review record
This PR was reviewed adversarially after revival: eight independent lenses (lifecycle, mailbox, send, worker, federation, transcript, complexity, live ergonomics), each required to prove findings with a failing test. That produced 16 proven blockers, all fixed with red-then-green regression tests, followed by two re-review rounds and a third fix wave that caught 3 regressions introduced by the fixes and 7 fixes that missed their target; all closed. A final pass (five lenses incl. a live built-runtime smoke, then a re-review of the fix wave) found and fixed seven more, chiefly the stale-worker mailbox steal, the self-dispatch depth wedge, and the unproven-exit certification. Three independent Codex (gpt-6-astra) passes followed: the first found nothing new, the second found and fixed 3 defects (task-status reachability, WSL-local host classification, peer-capability epoch), the third found and fixed 6 (production PTY controller never installed settled writes, ambiguous in-flight pointer failures allowed duplicate replay, SSH/relay deadlines cut off a valid `--wait-submit`, stop-vs-exit race during inspection, and two release-recovery paths for vanished or exited terminals). The full record (findings, proof tests, triage, declines with reasons) is archived outside the repo.
**Rework after the live smoke.** A first live cross-host run on the shipped adhoc build (this Mac, a paired Windows host on the same build, a paired Mac on 1.4.195, and an SSH host) found a P1: a running local worker read `unverifiable`/`missing_status` because the fleet snapshot rows lacked the terminal handle the matcher keyed on. A 59-row failure table over every bug fixed during review showed the same two classes recurring: a fact dropped in transit through optional fields, and two authorities for one fact. Two blind designs (Opus, Codex) converged on the same mechanisms, and the scoped tranches landed here with red-then-green seam tests from the real producer to the real consumer, faults injected only at the transport or hook-ingest boundary:
- **Settlement (data-loss class):** one three-valued `WriteSettlement` (`accepted | refused{reason} | unverifiable{reason, bytesHandedToTransport}`) from the SSH multiplexer through daemon client, providers, controller, to pointer staging. No boolean, no rejection-as-third-state. The two silent degrades that fabricated a handoff are deleted; a provider that cannot settle refuses before any effect. Pointer text and Enter share the contract; a partial flush is `unverifiable`, never `refused`.
- **Evidence identity (false-liveness class):** fleet agent-status evidence is a tagged union (`binding: worker | pane | unresolved{reason}`, `clock: observed | delivery`) minted once at ingest, so a hook row captured on one process incarnation can never bind to a later dispatch on the same pane. The matcher's `!worker.paneKey ||` defaults are gone. One host-scope parser replaces two.
- **Small pre-merge items:** `capability_unsupported` from an old peer is no longer relabelled `host_unavailable`; a producer census test asserts every agent-status consumer path projects a pane-only hook row as `live`.
Two ergonomics defects the second live run surfaced on a real database are fixed here too: a pre-v3 dispatch already marked `completed` projected as `outcome_unknown` / `requiresAction: true` forever (three copies of the outcome ladder disagreed on legacy rows; now one resolver, legacy `completed` reads `succeeded` with nothing to act on, legacy `failed` stays actionable on the failure), and an unscoped `worker-list` enumerated the entire database (now defaults to the Run bound to the calling terminal, `--run` overrides, and the receipt's additive `scope` field says which).
A third live round on the shipped adhoc build of `b082443e1f` (same four hosts) plus an unscripted run in the user's own prompt style (a plain Claude Code shell, `/orchestration`, three workers, zero errors, bound-Run default confirmed) found two more branch defects, fixed with red-then-green tests: a worker freshly started on a paired server projected `unverifiable`/`host_indeterminate` with `requiresAction` for ~3 minutes, including after its own `worker_done`, because the host's federation observation returned `missing_liveness_verdict` for any PTY the liveness register had not yet swept (the host now reads a connected pane it owns locally as `live`; disconnected or SSH-scoped panes stay `unverifiable`); and six pre-v3 completed rows still carried an `input` category because settling through the task-status path or `failDispatch` never closed the Dispatch's pending question threads (both paths close them now, and schema v38 closes threads already pending on settled rows). The guide's `worker-start` examples now show `--model sonnet`, since an omitted model inherits the launcher's default.
A Codex adversarial pass on the tranche diff found one real design hole (identity minted at read time instead of ingest, now closed) and two daemon settlement paths that threw instead of settling (fixed). Two `@ts-nocheck` runtime mixins on these paths were extracted into checked modules; the repo-wide `@ts-nocheck` count is unchanged at 171.
Deletions during review: ~1,900 lines (write-only ledger, unread columns, dead v1 archive path, test harnesses shipped in prod, duplicated liveness and state-machine copies, self-capability checks that were compile-time true).
## Testing
- `pnpm typecheck:tsc:node|cli|web` clean
- `pnpm run check:code-quality:changed` 0 findings; `check:react-doctor:changed` 0
- `pnpm verify:bundled-skill-guides`, `verify:skill-bundle-manifest`
- full `pnpm test` on the integrated head: 72,332 pass / 292 skipped; the only failures were three non-PR files (two zsh live-shell suites hit a node-pty spawn-helper ENOENT while a concurrent native rebuild ran, 44/44 in isolation; `release-checkout.unit.test.ts` is a known 30 s load timeout that passes in isolation on `origin/main` too).
- CI on
|
||
  OrcaWinandOrca Worker
|
fc5fa16870 |
perf(windows): split the process table into two flag sets (#17866)
* perf(windows): split the process table into two flag sets
MDE flags "suspicious memory activity" on the process-table reader: it
opened a handle into every process on the box and read each one's PEB on
a repeating cadence. Two changes narrow that.
Drop `Memory` outright. It cost a second
OpenProcess(PROCESS_QUERY_INFORMATION | PROCESS_VM_READ) plus
GetProcessMemoryInfo per process, and nothing reads a working set off
this table -- the Resource Manager runs its own sweep, and the addon
stores WorkingSetSize into a DWORD so anything above 4 GB wraps.
Split the rest in two. `readWindowsProcessIdentityTable[Fresh]` is a
bare Toolhelp32 walk with zero per-process handles, and returns
`WindowsProcessIdentityRow`, which has no `command` to read.
`readWindowsProcessTable[Fresh]` keeps the command line for the callers
that match on it. PTY root identity and the owner start-time probe move
to the cheap reader; agent recognition, port attribution, codex turn
processes and structured-TUI matching all genuinely need the command
line and stay.
Two independently single-flighted caches, never one per caller: the
fan-out this module prevents is one scan per caller, and each reader
still serves every caller wanting its flag set. The wedge gate and the
3s deadline stay shared, because both readers call the same addon and
one wedged read latches its one `requestInProgress`. With no binding
there is only the 1.4s PowerShell scan to run, so the identity view
rides the detailed snapshot rather than forking a second one.
Measured on Windows 11, 492 processes (p50/p95): identity 6.3/7.0 ms,
detailed 12.3/13.4 ms, previous memory+commandLine 13.1/14.1 ms.
* fix(windows): serialize native process-table reads across flag sets
The two flag-set readers could both be in flight at once, and the
vendored wrapper does not tolerate that. `getRawProcessList` pushes the
callback onto one list and calls the addon only when no request is in
progress, so a second concurrent caller's `flags` are DISCARDED and it
is handed the first caller's rows. Measured against the real addon:
identity issued first, both callers got the same array, 0 of 541 rows
with a command line. A detailed read overlapping an identity read
therefore returned a table with every command line empty, which agent
recognition reads as "no agent" -- silently, and only under concurrency.
Nothing already here excluded that. Each snapshot cache single-flights
only within itself, and the wedge set latches only after a read misses
its 3s deadline, so through the healthy ~12ms of a scan neither reader
excluded the other. Overlap is the normal state: panes poll detailed at
750ms while a teardown takes identity snapshots.
`nativeReadGate` admits one native read at a time across both flag sets.
It also fixes the relay path, where `adaptAddon` has no queue at all and
two simultaneous CreateToolhelp32Snapshot calls are the crash the
vendor's queue exists to prevent. Every link settles, so a wedged read
never strands a waiter; the waiter re-checks the wedge and rejects. With
one call outstanding, retention stays bounded at one callback rather
than one per reader.
Also from review:
- The CIM fallback now belongs to the detailed flag set alone, and the
identity view projects that snapshot through `toIdentityRow`, so an
identity row carries no command line on a no-binding host either.
- The concurrency test modelled the wrapper's coalescing queue, which
the previous synchronous mock could not express; verified failing
without the gate and passing with it.
- `agent-session-process-identity-probe` early-returns when the
creation-time flag is unavailable, which no shipped addon build
provides, instead of scanning the table to produce null.
- Corrected the cost framing: Memory took an OpenProcess(...|VM_READ)
it never read through, so dropping it halves per-process handle opens
and leaves the PEB/ReadProcessMemory telemetry unchanged.
* test(windows): keep read exclusion across resets and flag each field
Two review follow-ups, both about tests passing for the wrong reason.
`resetNativeReaderState` replaced the read gate with a resolved promise,
so waiters still holding the old chain ran beside reads queued on the
new one. Reachable only from the `__set*ForTests` hooks, which is what
makes it worth fixing: it hands a suite two concurrent calls into its
own mock addon -- the exact condition the concurrency tests exist to
detect. Chain onto the gate instead; every link settles within the
deadline, so the bounded wait that costs is the right trade.
The coalescing mock shaped every field off the CommandLine bit, so an
identity read that did request CreationTime got `creationTimeMs`
stripped. The identity-side assertion was then only `!('command' in
row)`, which a correctly flagged read and a coalesced one satisfy
equally: a future regression losing identity flags under concurrency
would have kept the case green. Gate each field on its own bit and
assert `creationTimeMs` positively, inside the helper both orderings
share.
Concurrency assertions move to a new bare-addon mock. The coalescing
mock's own latch means it can never report more than one call in
flight, so measuring exclusion there proved nothing; the bare addon has
no queue -- like `adaptAddon` on a relay, where re-entering
CreateToolhelp32Snapshot is a real crash -- and makes re-entry visible.
Verified by deletion: restoring `nativeReadGate = Promise.resolve()`
fails the reset case with `expected 2 to be 1`, and restoring the
single-bit mock fails both overlap orderings on `creationTimeMs`.
* docs(windows): count the third test defect in the list that names them
The section opened "Two defects have now shipped", numbered two, then
described the third in its closing paragraph -- a list that reads as a
complete account while quietly omitting one, which is the exact failure
the section exists to warn about. Say three and number it, and note that
the third arrived inside the fix for the first two.
Also record why the creationTimeMs and flags-array assertions are not
redundant, in the doc and beside the assertions: the flags array catches
a read served another flag set's rows, the positional creationTimeMs
check catches field shaping (identity dropping CreationTime, or
toIdentityRow not forwarding it). Neither sees the other's failure.
* docs(windows): stop describing a PEB read this release removed
Every comment here that justified the flag split in terms of PEB reads became
false when the command-line reader moved to the kernel. Left alone, the
enumeration doc contradicted itself inside one file: the flag-set section
described three chained `ReadProcessMemory` calls per process while the
sections below it explained that the addon contains no such primitive and has
no PEB fallback.
The measurement is now attributed rather than merged. Dropping `Memory` halved
the per-process handle opens and nothing else -- both handles carried
`PROCESS_VM_READ` at the time -- and it was replacing the PEB walk that took
`PROCESS_VM_READ` and `ReadProcessMemory` out of the addon. Neither change
substitutes for the other, which is worth keeping straight: the split's
remaining value is the handle itself, not the memory access.
Also adds `relay/windows-port-scan.ts` to the caller table, the one caller this
effort introduced, and records that it reads only pid/name through the detailed
reader -- free while a pane is polling, not free on a headless relay.
* test(windows): pin the fresh links path against the identity TTL cache
The identity and detailed tables are separate snapshot readers with
independent TTLs, so the detailed path's existing freshness guard says
nothing about the ancestry walk's. Cover the identity reader on its own.
---------
Co-authored-by: Orca Worker <orca-worker@localhost>
|
||
|
Neil
|
d7767fb196 |
perf(worktree): remove redundant creation and terminal startup work (#18793)
* perf(worktree): remove redundant creation and terminal startup work * test(worktree): cover optimized creation call signatures Preserve explicit branch adoption, WSL callback routing and sparse cleanup expectations. * perf: preserve user Git checkout worker settings * perf(git): skip malformed remote base probes * perf(cli): avoid loading other agent hooks for Codex preflight * fix(build): retain Codex preflight entry for packaged CLI * test(ssh): wait for replacement PTY before lease recovery input * test(ssh): verify recovered shell execution and lease ownership * test(electron): reap isolated macOS crash reporters on teardown * test: allow either observed self-exit snapshot ordering * test: capture frozen-host input recovery evidence |
||
|
Neil
|
9927edd631 |
fix(ssh): let the host say whether it armed the ready marker (#18802)
#18796 made every SSH Codex background launch wait for the shell-ready marker, but the client cannot see the remote shell. On a host that never publishes one -- fish, sh, Windows, or a relay predating #18796 -- no marker arrives and delivery falls back at 1.5s where it used to write at 50ms. The relay already computes whether it armed the marker; publish that as an optional `shellReadyArmed` on the spawn reply and let the client skip a wait it now knows is pointless. Absent stays UNKNOWN and keeps the client's own guess, so an older host behaves exactly as before; false is only ever an answer a host gave. It rides every reply, false included, or absent would stop meaning "old host". A host that did not arm the marker did not arm bracketed paste either, so the released path still submits raw. |
||
|
Neil
|
e95d247be1 |
perf(terminal): cheap-tier process inspection for anchored local agent panes (#18780)
* perf(terminal): cheap-tier process inspection for anchored local agent panes Every idle local pane's completion cadence ran a full whole-host `ps` (with `tty=` and `command=`, 0.34-0.50s on a 1,900-process Mac, 1.15s on Linux) purely to build `foregroundProcessEvidence` that the renderer then discards for local ids. Add a cheap tier (same job-control columns, no tty/command, 0.03s) gated so that it introduces no user-facing trade-off: - Only a pane whose last FULL capture proved a recognized agent may take the cheap tier. Panes with no anchor always take the full capture, so start discovery keeps today's exact behaviour. - The cheap tick compares a per-pane fingerprint (root shell pid+start, tpgid, every descendant's pid+start+pgid+job-control state). Any change, a changed node-pty foreground name, an unreadable capture, or an incarnation mismatch escalates to the full capture. A recognized agent's exit is always a pid vanishing, which the fingerprint always sees. - A cheap answer OMITS evidence rather than fabricating a tty-less fence. Remote/restore consumers never send `steadyState`, so they keep the full capture unchanged. - `steadyState` is a new optional request field; an old daemon ignores it and answers with the full capture. Measured (8 idle panes, 60s, idle cadence, forks counted by column set): 30 full -> 1 full + 29 cheap. * fix(terminal): route the cheap ps capture through runProcess The cheap-tier reader imported node:child_process directly, which the child-process import-boundary and windowsHide ratchet tests reject (CI shards 1/8 and 3/8). Use Orca's single spawn entry point instead; it pins windowsHide and encodes argv. Map its result onto the capture-error vocabulary: outputTruncated -> capture_truncated, timedOut -> capture_timeout, non-zero exit -> ps_exit_<code>. Tests mock at the runProcess seam. * fix(perf): refuse a pane fingerprint when any descendant start marker is missing `buildPaneProcessFingerprint` rejected only a missing root start marker; a missing descendant marker was stamped as `?`. Two captures that both failed to read the same descendant therefore compared equal, which removes the pid-reuse protection the fingerprint exists to provide: a recycled pid could make a vanished agent look unchanged, and the cheap tier would keep serving its name instead of escalating. Reachable on Linux, where `readLinuxProcStartTime` legitimately returns null when a process exits between the `ps` capture and the `/proc/<pid>/stat` read. Every subtree member now needs a start marker or the fingerprint is refused, which sends the caller to the full capture — the same conservative default every other uncertain path takes. Reported by CodeRabbit on #18780. The two new tests fail against the previous code with `expected '4242@2400#4300:|4300@?:4300:+' to be null`. |
||
|
Neil
|
3f84c358f0 |
fix: recover from an alternate shell install, and close the relay duplicate-echo gap (#18796)
* fix: recover from an alternate shell install, and close the relay echo gap #18768: a startup profile that `exec`s a second install of the same shell keeps the pid but loses the wrapper's ready marker, and the recovery probe rejected the replacement's different canonical path -- costing plain Codex the full 15s barrier. The probe now also accepts an install that the pane's own PATH resolves, so a binary merely named bash/zsh outside it stays rejected. #18767: the SSH relay left plain Codex on early startup delivery, displaying the launch twice under a slow profile. The shell is the host's to know, so the relay now folds it into the same rule the daemon uses, and the SSH background client waits for the marker on any Codex launch. Bracketed paste is now gated on an observed marker rather than on the intent to wait, so a fallback release on a host shell that never publishes one submits raw. The non-daemon local provider needs no change: it hands Codex to the wrapper's own prompt hook and never writes it into the PTY. * review: correct an overclaiming comment, announce a silent skip, drop a shim Readiness review of #18796. The delivery comment claimed waiting "costs nothing", which is only true on a host that arms the marker -- fish, sh, Windows and hosts predating #18767 release on the fallback instead. Say so. The alternate-install recovery tests skip on usrmerge hosts, where /usr/bin/bash resolves back to /bin/bash; announce that rather than reading as coverage that does not exist. Import the line-editor predicate from shared directly instead of through a re-export left on daemon/shell-ready. |
||
|
Jinwoo Hong
|
30d7542bc5 |
fix(terminal): stop a hidden pane's unmeasured 80x24 from overwriting a live PTY's size on reattach (#18706)
* fix(terminal): stop a hidden pane's unmeasured 80x24 from overwriting a live PTY's size on reattach A pane that mounts while display:none (app relaunch or update with the floating terminal panel closed, a non-active floating tab, any background tab) cannot fit its container, so it reattaches with xterm's default 80x24. Main wrote those placeholder dims into `ptySizes` unconditionally, both before and after the attach. A daemon attach never resizes the live session, so the real PTY stayed at its wide grid while main's hidden headless model was created (or reflowed after renderer hydration) at 80 columns. Every byte the agent emitted while hidden was parsed 80 wide; reveal restored that image into the pane: rows clamped at column 80 with CHA fill, the status bar interleaved into response text, and for alt-screen TUIs the whole screen stuck in an 80x24 corner until a real resize forced a repaint. Scrollback damage was permanent. Fix, main-side only: - Pre-attach: seed `ptySizes` only for a genuinely fresh session id, or a measured request with nothing cached. A hidden reattach writes nothing. - Commit: on reattach, record the provider's proven grid (`attachedGrid`, set only by the local provider whose attach really resizes), then the reply's `snapshotCols/Rows` (the daemon emulator's grid), then the size main already held, and only then the request. - Reflow an already-created model to that grid after the seed block, so bytes that arrived before the reply no longer leave an 80x24 model. Both the ipc and runtime spawn paths take the same authority module. Renderer and wire formats are unchanged; `PtySpawnResult` is main-internal. Reproduced deterministically: close the floating panel with Claude Code streaming at 211x57, kill only the Electron main process so the daemon survives, relaunch. Main's cache read 80x24 against an applied 211x57 and the reveal snapshot was 80 columns wide; replaying the recorded bytes through an 80-column emulator reproduced the field screenshot. Relaunch with the panel open, and a fresh spawn, keep the wide grid. * fix(terminal): commit the adopted-claim reattach grid and reject non-integer provider grids Review follow-ups. The runtime spawn path's adopted-claim branch returned before the size commit, so an adoption attaching to a live session kept whatever the caller requested; it now commits and reflows like every other reattach. The grid validator requires integers so a malformed provider grid falls through to the cached size instead of reaching xterm. * fix(terminal): reflow main's headless model onto the committed grid for every spawn, not only reattaches A hidden attach whose daemon restarted comes back as a fresh session, and the pre-attach seed is now withheld for unmeasured attaches, so a live byte that created the model at 80x24 before the reply would have kept it there forever. * refactor(terminal): let the provider's reattach flag pick the adopted-claim grid source * fix(terminal): derive the adopted-claim reattach flag once for the size commit and the reservation The SSH relay's adopted reply carries no isReattach, so the size commit would have taken the request while the reservation was told it was an attach. Normalize once so both agree. |
||
  
|
a65332a8bd |
feat(claude): move structured native chat onto the Claude Agent SDK and enable it on macOS and Linux (#18560)
* Join structured attach teardown through journal bind * fix: restore structured chat parity * feat: add Claude structured session adapter * fix: harden Claude structured adapter * fix: close Claude adapter edge cases * fix: start Claude init deadline after launch * feat: wire Claude structured sessions * fix: harden Claude structured runtime * fix: fence Claude structured compatibility * fix: preserve Claude free-text prompt answers * fix: decode addressed Claude prompt text * feat: enable Claude structured chat on mobile * fix(mobile): keep structured chat provider-aware * fix(mobile): negotiate Claude structured tabs * fix: keep scoped RPC tests native-free * fix: secure mobile structured image delivery * fix: close structured session data-loss gaps * fix: prove real Claude structured startup * fix: consume pre-spawn proof before retry * feat(native-chat): add desktop structured sessions * fix(native-chat): satisfy structured session cleanup gates * fix(native-chat): keep structured renders pure * fix(native-chat): open composer pickers upward * fix(native-chat): use existing view for structured sessions * fix: harden structured desktop status projection * fix: close structured desktop lifecycle gaps * fix: fence structured AI Vault resumes * fix: fence structured AI Vault resumes * fix: preserve structured tabs during activation * feat: toggle structured sessions between chat and TUI * fix: harden structured session handoffs * fix: bind structured TUI before rollout proof * fix: complete structured chat round trips * fix: align structured TUI return readiness * fix(native-chat): make reverse handoff transactional * Add Claude structured TUI handoff seams * fix(native-chat): clear sticky handoff recovery * fix(native-chat): complete mobile reverse after TUI exit * fix(native-chat): keep TUI transcripts readable * fix(native-chat): recover TUI transcript gaps * fix(native-chat): recover claimed TUI owners * fix(native-chat): retain cold TUI proof authority * fix(native-chat): preserve Claude handoff authority * fix(native-chat): recover TUI transcripts read-only * fix(native-chat): harden Claude handoff recovery * fix(native-chat): serialize structured handoff recovery * fix(native-chat): close handoff admission races * fix(native-chat): validate pinned launch environment * fix(native-chat): revalidate restored and retried owners * fix(native-chat): gate restart recovery publications * fix(i18n): catalog Claude session controls * fix(native-chat): wait for structured TUI process proof * fix(native-chat): queue stale idle TUI handoffs * fix(native-chat): route structured Codex options directly * fix(native-chat): persist structured session options * fix(native-chat): hydrate resumed structured options * fix(native-chat): preserve options across structured handoffs * fix(native-chat): replay pending option mutations * fix(native-chat): rotate settled handoff operations * fix(native-chat): rotate refused send operations * test(native-chat): derive refusal retry state from host * test(native-chat): give the host-oracle matrix test an explicit timeout * fix(native-chat): keep Claude option controls idle * fix mobile structured first-send hydration race * fix(native-chat): preserve handoff launch authority * fix(native-chat): harden shared handoff recovery * fix(native-chat): serialize structured handoff recovery * fix(native-chat): close handoff admission races * fix(native-chat): validate pinned launch environment * fix(native-chat): revalidate restored and retried owners * fix(native-chat): gate restart recovery publications * fix(i18n): catalog structured session recovery control * fix(native-chat): wait for structured TUI process proof * fix(native-chat): queue stale idle TUI handoffs * fix(native-chat): keep structured recovery provider-neutral * fix(native-chat): drop local terminal topology from structured sync * fix structured outbox and tab restore races * fix(native-chat): preserve Claude question groups * fix structured provider visibility and request handling * fix structured session TUI handoff recovery * fix reverse structured session handoff * fix(native-chat): recover Claude outbox and resume state * chore(mobile): preserve the working-tree lockfile state before the main merge Carries the pre-existing uncommitted mobile/pnpm-lock.yaml modification into history so the main merge cannot overwrite it. Verified benign pnpm drift (babel 7.29.7->7.29.8 transitives plus deprecation metadata); drops no patchedDependencies (the mobile lockfile declares none). * test(native-chat): drop orphaned Claude handoff-auth test left by the main merge 'pins Claude handoff auth through the terminal provider boundary' is absent from main and its production counterpart preserveClaudeAuthEnv no longer exists outside this test - orphaned residue of the terminal/native handoff work this PR excludes by scope. Removed rather than repaired: the failure was a renamed field (providerHome -> providerRoot), and renaming it would have carried out-of-scope handoff code into the merge. Body preserved as evidence and logged in CLAUDE-STRUCTURED-DISPOSITION-TABLE.md. * Fix mobile structured turn state * fix Claude structured session blockers * fix claude structured lane blockers * fix Claude acquisition exit proof * fix(claude): route stream-json launch through process wrapper * fix(claude): gate structured chat support * Fix Claude structured launch gating * fix(claude): split session acquisition and prune mobile scope * test(claude): align structured session fixtures * fix(agent-session): preserve handoff launch arguments * fix(claude): open journals through the factory after origin/main split The journal opener moved to journal-store-factory on main; retarget the Claude structured tests that still imported the old path. * fix(claude): resolve Claude structured launch args, auth, and win32 proof The origin/main merge re-expressed the lane's Claude wiring onto main's split orca-runtime facade and dropped three wires past green typecheck and lint. - resolveLaunchArgs discarded its provider parameter, so structured Claude sessions were launched with Codex app-server flags; Claude exits on --dangerously-bypass-approvals-and-sandbox, and a Codex arg-parse throw could block Claude session creation outright. - resolveClaudeLaunchEnv was no longer supplied, so the launch resolver fell back to the whole process env as configuredEnv and buildClaudeChildProcessEnv re-applied every auth var it had just stripped. The resolver now merges the Claude overlay onto a strip-applied copy of the inherited env, which also keeps PATH intact for withCliRuntimeOnPath. - The windowsProcessStartTimeAvailable producer was gone while the contract field and both consumers survived, so the renderer gate fail-closed and structured native chat was unreachable on every win32 host. Separately, structured Claude pinned CLAUDE_CONFIG_DIR unconditionally. An explicit pin makes the CLI abandon the macOS Keychain even when it names the CLI's own default, so a default claude.ai account could not authenticate where the legacy Claude terminal could. Pin only a home the CLI would not resolve on its own, matching ClaudeRuntimePathResolver, and compare against the env the child would otherwise inherit so a diverging overlay cannot outrank the record's account home. Also await the now-async revealNativeSession in its regression test, and set the native status before revealing so a rejecting reveal cannot leave a session released but never marked native. Claude-Session: https://claude.ai/code/session_013UqKCRB6k5e8UaYhXUHeWY * fix(claude): scrub case-insensitive Windows auth env * fix(native-chat): settle handoff outcome-write failures instead of leaking them A store write failure while recording a handoff outcome escaped the flow runner's catch handler, so the client never received the failure and the flow surfaced as an unhandled rejection (seen as an intermittent agent_session_store_corrupt error in the proven-dead-retry suite, whose teardown raced the flow's trailing outcome write). Record the failed outcome best-effort, and drain the coordinator before that test's teardown removes the store root. Claude-Session: https://claude.ai/code/session_011aXkcHyeiRJuezupQdjZaM * fix(native-chat): make the structured close-failure toast provider-neutral The structuredSessionCloseFailed toast fires for any structured session, but its copy said 'Codex chat', so a Claude structured session that fails to close showed the wrong provider name. The launch-failure toast is only reachable behind the agent === 'codex' gate, so its copy stays as is. Claude-Session: https://claude.ai/code/session_013ugSpCx4AWkySaJb69BQax * fix(native-chat): wire structured handoff proof recovery * fix(native-chat): wire structured handoff proof recovery * fix(native-chat): correct the structured chat opt-in copy The one `experimentalStructuredNativeChat` toggle gates both providers — `useStructuredAgentSessionCreate` runs `canUseStructuredNativeChat` for `'claude'` as well as `'codex'` — but its description named only Codex. Its scope line also said Windows keeps using terminal chat, while the gate refuses win32 only until the host proves it can read a process start time. `structured-native-chat-availability.test.ts` already pins that Windows is allowed once the proof is cached, so the two contradicted each other. Claude-Session: https://claude.ai/code/session_01RJFsidQWmKYFmeoUuVu4Tp * test(claude): pin @anthropic-ai/claude-agent-sdk 0.3.251 contracts against a scripted CLI PR 1 of the SDK migration: dependency + test-only harness, no product wiring. - Pin @anthropic-ai/claude-agent-sdk to exactly 0.3.251 — not the newest release — because 0.3.251 (published 2026-08-28) clears the repo's 3-day minimumReleaseAge supply-chain gate with no exclusion, while the newest release was minutes old and would have required excluding a brand-new publish from the exact control built to catch brand-new malicious publishes. Every contract this design depends on was verified identical on 0.3.251: the full option surface, no pid on SpawnedProcess (custom spawner stays mandatory), env defaulting to process.env when omitted, and --replay-user-messages appearing only via extraArgs. - Exclude all eight bundled CLI platform binaries via ignoredOptionalDependencies. The setting lives in pnpm-workspace.yaml because pnpm 12 no longer reads the package.json "pnpm" field (it warns and ignores it; verified by install ablation). Excluding the binaries is what makes Orca's pathToClaudeCodeExecutable override mandatory rather than merely preferred. Note: pnpm 12.0.0 honors the ignore list when reconciling an existing lockfile but not on fresh resolution of a new dependency, so the lockfile's SDK entry was pinned surgically; both 'pnpm install' and 'pnpm install --frozen-lockfile' verify clean and stable against the committed lockfile. - Contract-pin suite drives the real SDK against a scripted fake CLI and pins: unknown type/field/content-block pass-through (and keep_alive interception), spawner env fidelity plus the omitted-env process.env inheritance sharp edge, extraArgs producing --replay-user-messages, argument parity for every CLAUDE_STRUCTURED_BASE_ARGS entry plus --session-id/--resume/ --resume-session-at, canUseTool wire request_id stability and abort on control_cancel_request, one spawn per query, pathToClaudeCodeExecutable honored by the default spawner, the exact SDK version, and the eight platform binaries staying uninstalled. Claude-Session: https://claude.ai/code/session_01FGCRfYUnb4hbvfTAHGtJKQ * feat(claude): drive the structured transport through the agent SDK Replaces the hand-rolled `claude -p --input-format stream-json` transport with @anthropic-ai/claude-agent-sdk 0.3.251, keeping the existing connection interface for this commit so the acquisition path changes minimally. The control-plane rewrite is a separate change. Orca still supplies the process. `spawnClaudeCodeProcess` routes through `spawnProcess`, retains the child and its pid — the triple the durable lease adjudicates on — drains stderr so exit errors keep their tail, and hands `.cmd` shims to Orca's Windows argument encoder rather than the SDK's plain spawn. `close()` keeps Orca's own bounded tree-kill and exit deadline, so it still resolves true only after an observed exit. Launch resolution emits an SDK options object instead of argv; durable `launchArgs` translate to a typed option where one exists and to `extraArgs` otherwise, refusing a token neither can carry rather than dropping it. The child env is always passed explicitly — omitting it would let the SDK inherit `process.env` and reintroduce the ambient `ANTHROPIC_*` leak. The stdout line parser is deleted; the SDK owns framing, and unknown frames still reach the translator verbatim. Claude-Session: https://claude.ai/code/session_01JMhFjh9HEnkcJ5YTfCdgD3 * fix(claude): settle the frame the SDK pulled but never wrote The SDK's input pump is `for await (frame of prompt) { await transport.write(frame) }`. When that write rejects — the child dies between Orca's liveness guard and the write — the for-await ends abruptly and calls the generator's `return()`, so the code after `yield` never runs. The frame was already shift()ed out of `queued`, so the later `fail()` from the exit path could not reach it and `send()` never settled: `dispatchClaudeTurn` awaits that send before it can return `unknown`, wedging the caller and the durable outbox. The pre-SDK transport rejected on the stdin write callback instead. Retain the in-flight entry and settle it from the generator's cleanup, and let fail() reach it too for the pump that never resumes at all. Claude-Session: https://claude.ai/code/session_01AobxxokqQ3qcxS7sy7ckum * fix(claude): keep the agent SDK behind the structured-Claude boundary The ordinary OrcaRuntimeService graph statically reaches the Claude adapter and so the transport module, whose first line imported @anthropic-ai/claude-agent-sdk. The SDK is evaluated whenever the regular runtime loads, before any structured Claude session is chosen: it sets process.env.NoDefaultCurrentDirectoryInExePath, changing Windows executable resolution for later subprocesses, and a missing or incompatible install would break normal runtime startup — for a user who never leaves the terminal/TUI path. Defer the SDK to the connection, memoized so it loads once per process, and add the import-graph ratchet: a walk from the Electron main entry that fails on any static import of the package, plus a clean-fork check that loading the runtime leaves the Windows search variable untouched and a child-process pin that the side effect is still real. Claude-Session: https://claude.ai/code/session_01AobxxokqQ3qcxS7sy7ckum * fix(claude): answer list_models so the picker stops serving the seed sendControlRequest had no list_models case, so every request hit the default reject; readClaudeStructuredSessionOptions swallows that with .catch(() => null) and falls back to the static catalog. Every structured session therefore served a hardcoded model list with no per-model effort levels, no resolvedModel and no default detection, and nothing surfaced the failure. The pre-SDK transport got the live catalog from the CLI. Route it through the SDK's supportedModels(), wrapped in the { models } envelope the existing parser reads. Claude-Session: https://claude.ai/code/session_01AobxxokqQ3qcxS7sy7ckum * fix(claude): reap the child's descendants before killing it The forced step of the exit ladder went through the Codex helper, which spawns `pkill -KILL -P <pid>` and SIGKILLs the parent in the same tick: the parent usually dies first, the descendants reparent to pid 1, and `-P` matches nothing. An MCP or launcher descendant of a stubborn Claude child was left running. The test named for that requirement declined to assert it and killed the survivor by hand instead, so it could not fail for the thing it was named after. Route the Claude reap through Orca's existing sweep, which snapshots descendants while their parent link still exists and signals them before the root goes, and on Windows uses the identity-gated `taskkill /T /F`. The test now asserts the descendant is dead; the manual kill stays only as a failure-safe. close() still returns true only on an observed exit. Claude-Session: https://claude.ai/code/session_01AobxxokqQ3qcxS7sy7ckum * fix(native-chat): merge the duplicated handoff type import CI's static-analysis lint (`oxlint --config config/oxlint-code-quality-native-plugins.json src config tests mobile --deny-warnings`) exits 1 on the two separate `import type` statements from the same module. Claude-Session: https://claude.ai/code/session_01AobxxokqQ3qcxS7sy7ckum * fix(claude): answer a permission callback whose signal already aborted settleFrom registered the abort listener and then delivered the request. A callback that arrives already aborted never fires that event, so the promise stayed pending behind a durable prompt with no cancel path. Check the signal first, emit the cancel, and resolve the SDK's null sentinel without registering. Claude-Session: https://claude.ai/code/session_01AobxxokqQ3qcxS7sy7ckum * test(claude): wait for the child to record the frame, not just for its report The scripted CLI writes its report at startup, so `until(readReport)` returned a report with no user messages whenever the child had not yet read the line. The assertion then failed under parallel load. Poll for the frame instead of for the file. Claude-Session: https://claude.ai/code/session_01AobxxokqQ3qcxS7sy7ckum * fix(claude): coalesce partial deltas onto one assistant item and stop painting result frames Under --include-partial-messages every stream_event frame carries its own uuid, and the final assistant frame for a block carries yet another; only message.id ties them. The translator keyed each delta by its frame uuid, so a reply painted as one bubble per delta chunk followed by a complete duplicate under the final frame's uuid. The block's first stream frame now mints the claude:(sessionId, uuid) identity, deltas coalesce onto it through the shared 60ms seam, and the final frame reconciles onto that same item. Known SDK bookkeeping no longer reaches the provider-fallback row: result subtypes are catalogued and settled by the turn lifecycle, an empty thinking block (redacted thinking) is a modeled kind, a string-content user replay is a text block, and an empty user frame paints nothing. An unmodeled result subtype or content kind still lands on the bounded fallback row. Claude-Session: https://claude.ai/code/session_01GaP5HpYQbvy2hYehVhwfEW * fix(claude): prove descendant exit at the close boundary instead of on an unref'd timer close() reported proven=true as soon as the direct child exited while the descendant sweep's SIGKILL sat on an unref'd 2 s timer, so a SIGTERM-resistant MCP server outlived the lease release. The reaper now composes the same shared primitives the Codex structured provider uses: snapshot, verified bounded descendant termination on POSIX, taskkill /T /F on Windows. The proof is false whenever descendants outlive the deadline, a retried close re-verifies the retained snapshot rather than trusting the dead root, and the raw pipe child no longer goes through the PTY job sweep it never owned a job for. Measured on macOS: a killed child of a SIGSTOPped parent stays a matching zombie row in ps, so the root is killed while verification runs rather than stopped first as the Codex non-group path does. Claude-Session: https://claude.ai/code/session_0161QFm3KVRNJKfdzWVGVNWk * feat(claude): replace the hand-rolled control plane with the SDK's native surface PR 3 of the Claude structured SDK migration removes the wire-frame scaffolding PR 2 kept, so Orca drives the SDK's typed control surface directly. Inbound permissions move from a rebuilt control_request dispatch to the SDK's canUseTool / onUserDialog callbacks. The prompt registry now carries the callback's own resolver: a decodable can_use_tool becomes a durable prompt whose answer settles the callback; a malformed one is denied without registering; the SDK's abort signal (fired on control_cancel_request, which the SDK matches and dedups itself) forgets the prompt and settles it null, and a late answer after abort finds no prompt and is refused. Closing settles every in-flight callback so no promise dangles. The claude-agent-sdk-control-bridge that rebuilt the wire frame is deleted. Outbound control maps to Query methods: interrupt() for cancel, setModel / setPermissionMode / applyFlagSettings for options, supportedModels for the model list, initializationResult() for init proof, each under Orca's own request deadline and error classification. Cancel is interrupt-receipt aware: a CLI advertising interrupt_cancel_queued_v1 gets cancel_queued in one round trip, otherwise the receipt's still_queued uuids are swept with cancel_async_message so a cancelled turn cannot spawn a later unexpected turn; older CLIs resolve no receipt. Init keeps the 10s deadline and the unauthenticated-startup guidance. Every behavior is failing-first and ablation-proven; the toggle-off import boundary and the accepted loss of unknown-control visibility rows are unchanged. Claude-Session: https://claude.ai/code/session_01Pqjduxt5G4rr9aYvtp7rNm * fix(claude): arm the descendant snapshot before stdin closes and make the tree verdict unproven by default A healthy Claude root leaves within the graceful window, and the close ladder only snapshotted descendants when the root was still alive after that window. So the common close never looked at the tree: `treeExited` stayed null, `!== false` passed it, and close() reported a proven exit with an MCP child still running. A root that died before the walk made the snapshot vacuous too. The proof is now unproven by default. The reaper holds one verdict in Orca's vocabulary (exited / live / unverifiable), assigned in exactly one place from the bounded verification, and close() returns true only on `exited`. The snapshot is armed before stdin closes, while the root can still be walked, and is verified after the root exits; a root that left before any snapshot could be armed stays unverifiable rather than vouching for descendants it never showed us. The shared verifier gains the three-way verdict behind its boolean face, and the connection reports the root and tree verdicts separately along with the child's exit status. One verification per close attempt: the retried close re-verifies, so the intra-attempt re-reap is gone from the teardown budget. Claude-Session: https://claude.ai/code/session_01BSmXgkWSsNHft8jFkdBFG9 * fix(claude): verify the Windows tree after taskkill instead of trusting that it ran `terminateWindowsProcessTree` resolves from taskkill's callback whatever the error says, so a timeout, an access denial, a recycled root and a surviving descendant all looked identical to the reaper — which then returned a proven exit unconditionally. close() reported true and the lease was released with an MCP descendant potentially still live. The Windows branch now snapshots the root's descendants while it is alive and, after taskkill, polls a fresh process table to a bounded deadline: a row still matching by pid AND creation time is `live`, an unreadable table is `unverifiable`, and only a table with no match is `exited`. Creation time is the PID-reuse guard the POSIX path gets from ps lstart, so a descendant that denied a creation-time query is omitted rather than signalled on a bare pid. A root already observed exited is never taskkilled: `/T /F` on a recycled pid would take an unrelated tree down with it. The captured tree is tagged by platform so neither verifier can be handed the other's rows. Claude-Session: https://claude.ai/code/session_01BSmXgkWSsNHft8jFkdBFG9 * fix(claude): release a reservation on a first-hand root exit instead of latching it into manual recovery Making close() strict about the descendant tree exposed a second defect at the same boundary. A create-time acquisition has no ownerProcess until publication, so an unproven cleanup mapped to handoffStage `manual-recovery`, and adjudication then refuses every later attach with agent_session_ownership_unknown. A user who was merely signed out, or whose --resume the CLI rejected, wedged the session id permanently. Each question now answers from its own evidence. close() is unchanged and stays strict about the tree. Separately, the lease is keyed on the root's pid and start time, so when Orca's own child handle observed that root exit and no descendant snapshot was ever admissible, the reservation is released and the CLI's exit code and stderr reach the user. A descendant observed still alive, or a root Orca never saw leave, stays unproven and keeps the reservation. The settlement records only what was observed: the released lease says the provider process exited and its descendants were not verifiable, rather than reusing the wording that claims cleanup proved no child remains. Claude-Session: https://claude.ai/code/session_01BSmXgkWSsNHft8jFkdBFG9 * fix(claude): surface an API error a result frame reports instead of settling the turn on it The SDK models an API failure as a SUCCESS-subtype result whose `result` string is the user-facing error text, with no assistant frame behind it. The translator suppressed every catalogued result subtype as turn bookkeeping, so that turn tombstoned its lifecycle and showed the user a completed, empty reply with no sign anything had failed. Suppression is now by meaning. A result reporting a failure routes to the bounded provider-error surface, leading with the provider's own sentence and keeping the raw frame behind the row's disclosure; ordinary successful results stay off the timeline as before. A turn the user aborted also stays suppressed: its interrupt frame already says so, and its execution diagnostic would only be noise on every stop. Claude-Session: https://claude.ai/code/session_01BSmXgkWSsNHft8jFkdBFG9 * fix(claude): drop the stream state of turns that never received their final frame Every streamed delta recorded its block's identity, latest text and checkpoint length. Only the final assistant frame removed them, so an interrupted turn left its whole accumulated reply reachable until the session was disposed, and a long session with repeated interruptions grew those maps without bound. The partial text was already journaled by the flush that precedes settlement, so the live copy was pure retention. That state now lives in its own module, named for what it does — grow a streamed block's journal row between its deltas and its final frame — and turn settlement drops every block still awaiting a final. The translator reports how many remain, which is the invariant: a settled turn leaves none. Also makes a timed-out process-table read retryable while the root is still alive. A loaded host can miss the table's one-second deadline, and latching that as "no descendants" both lost the descendant sweep and, on a busy machine, made the close ladder report unproven for a tree it never actually looked at. Only the root's death still makes a missing snapshot final. Claude-Session: https://claude.ai/code/session_01BSmXgkWSsNHft8jFkdBFG9 * perf(claude): capture the Windows descendant tree from one process-table read The capture walked the descendant tree and then read the table again for the creation times the walk's projection drops. Each read is bounded in seconds and both run inside the close ladder's budget, so the second one cost the worst-case teardown three seconds for data the first read already held. The walk is now exported from the module that owns it and runs over rows the caller has already read, which is also what lets the snapshot keep the PID-reuse guard the projection cannot carry. Claude-Session: https://claude.ai/code/session_01BSmXgkWSsNHft8jFkdBFG9 * fix(pty): spend the descendant verification window instead of surrendering on one slow table read The verification abandoned the whole check the first time a process-table read missed its own one-second deadline, with seconds of its window still unspent. On a loaded host that reported a tree unverifiable without ever having looked at it, which the Claude close ladder then turned into an unproven close and a retried teardown. It also made the descendant-exit tests flake under a parallel suite run, for the same reason and with the same honest-but-premature verdict. A read that missed its deadline is now simply not an answer: the loop waits and reads again until its own deadline, and only a window that ends without a readable table reports unverifiable. This can only turn a premature verdict into one backed by evidence; it never manufactures a proof. Claude-Session: https://claude.ai/code/session_01BSmXgkWSsNHft8jFkdBFG9 * fix(claude): never let a later failed look collapse an observed live descendant into unverifiable The reaper's single assignment site latched only 'exited', so a second reap whose table reads all missed their deadline overwrote an earlier completed verification's 'live' with 'unverifiable'. The acquisition release gate discriminates on exactly that pair, so a root exit after such a decay released the lease over a descendant that had been observed alive. The latch is now monotone in trust order: exited is final, and live is only ever raised to exited. Claude-Session: https://claude.ai/code/session_01HfdhsvSJucLw4cTZxzg2CP * fix(claude): never prove a Windows tree gone while a descendant denied identification The Windows snapshot dropped rows that denied the creation-time query, and an emptied snapshot was judged exited without any table read: a descendant Orca was refused information about was treated as one that had left. The snapshot now counts the unidentified rows it saw, and verification caps its verdict at unverifiable while any exist. Nothing is ever signalled on a bare pid, as before. Claude-Session: https://claude.ai/code/session_01HfdhsvSJucLw4cTZxzg2CP * fix(claude): classify cleanup after a first-hand exit as a root exit instead of a proven tree When the CLI died between a successful acquire and the host's commit or proof of the lease, handleExit had already removed the session, so releaseAcquisition found nothing and reported true. The attach flow then settled exit-proven with deathEvidence claiming cleanup proved no provider child remains, though the tree was never verified. The adapter now keeps the exit that removed a published session until the session is acquired again; acquisition cleanup runs that connection's close ladder and classifies its verdict exactly as a start-time failure would be, so the record reads root-exit-observed. The wire helper keeps that typed classification and its provider diagnostic instead of wrapping it as unproven, and the router gives up its owner even when the release throws. Claude-Session: https://claude.ai/code/session_01HfdhsvSJucLw4cTZxzg2CP * fix(claude): integrate SDK teardown and picker lifecycle fixes * fix(claude): preserve resume leaf and settle processless spawns * fix(claude): reacquire from persisted resume leaf * fix(native-chat): restore Claude grouped question handling * fix(claude): persist only resumable transcript leaves * fix(claude): recover structured session exits safely * fix(claude): close remaining structured session P1s * fix(claude): harden transcript branch proof * Remove superseded root fix reports * fix(windows): restore indexed descendant row walk * fix(router): forward force-close lifecycle * fix(claude): fence stale turn cancellations * fix(claude): fence cancellation after unknown dispatch * fix(claude): fence replay and option recovery races * fix(claude): block replay fallback after waiter eviction * fix(claude): fence evicted slash results * fix(claude): fence ambiguous results and restore options safely * fix(claude): scrub SDK child env and localize pending launch * fix(claude): pin transcript roots and exit recovery proofs * fix(claude): retain unproven SDK exits * fix(claude): settle retained exit before reacquire * fix(claude): resume from settled retained cursor * chore: remove tracked review artifact * fix: harden Claude SDK transport session cleanup * fix: close Claude sessions safely * fix(claude): close races with fresh child snapshots * fix(claude): fail closed on recycled child identities * fix(claude): gate root cleanup on process identity * fix(claude): fence same-second root identity reuse * fix(claude): restore the root SIGKILL fallback the identity gate took away The direct root kill goes through the handle Node owns, not through a pid: libuv drops that handle in the same turn it reaps, so the signal either reaches the process Orca spawned or reaches nothing at all. Gating it on a process-table probe therefore bought no safety and cost the tree its only fallback whenever the probe declined -- a first capture landing in the fork's own second, a recycled descendant pid voiding the snapshot, or a process table that could not be read on either platform. Identity verification stays where a bare pid is genuinely addressed: Windows `taskkill /T /F`, and the descendant sweep's own revalidation before it signals. Also stops a declined root probe from collapsing an observed `live` or `exited` descendant verdict into `unverifiable`, and stops a successful taskkill from reporting `unverifiable` because a later probe found the root correctly dead. * docs(claude): rewrap the root-kill ordering comment * Match the Claude structured launch to the terminal path's managed-account auth rules The SDK path stripped ambient Anthropic auth unconditionally, let an explicit agentDefaultEnv override beat a pinned managed account, and had no account-switch guard. Reuse the terminal preflight's own predicate and messages so both transports strip, refuse, and report identically, and cover the CLI transcript location that mobile native chat depends on. * Reach the Claude structured chat lane from the desktop UI The main process has had a complete, correctly gated Claude Agent SDK lane for a while, but no renderer ever asked for it: the launch route accepted only `codex`, and the create path was typed `agent: 'codex'` end to end. Widen both to the structured provider union that already exists (`AgentSessionHandleProvider`), and generalize the codex-named create path instead of adding a Claude twin beside it. The pending-launch registry is now keyed by agent as well as workspace — a shared key handed a second caller the first agent's intent, so a Claude and a Codex launch in one worktree collided. Windows, per agent. Codex's client-side win32 refusal is deliberate and settled elsewhere, so it stays exactly as it was. Claude's answer is no longer guessed from the client's platform: a structured session fences its provider child on that child's process start time, and only the executing host knows whether it can read one. `agentSession.createSupport` already answers precisely that, per agent, and had no renderer caller — so the Claude create path asks it before creating and turns a "no", or a probe it cannot get answered, into the definitive refusal the launch fallback already handles. Fail closed either way. That refusal mapping also closes a real gap: the host reports an unsupported location by throwing `structured_agent_session_unsupported`, which reaches the client as a transport rejection rather than a refusal envelope, so `StructuredAgentSessionCreateRefusalError` never fired. The launch would retry the create, strand itself in `visibilityUnknown`, run no legacy fallback, and show an error toast. Close a fail-open hole while Claude and win32 become reachable: `create` with a client-supplied location, and `ensure`, both skip the worktree-resolving support check. They now ask the executing host the same question directly, so a host that cannot fence a provider child no longer creates one on a client's say-so. Also deletes `structured-agent-session-provider-routing.ts`, a duplicate of `structured-agent-session-provider-support.ts` with no importers. WSL, SSH and paired hosts, floating workspaces, draft prompt delivery, explicit TUI customization and initial session options all keep refusing; folder workspaces keep working. * P1-1: make the structured Claude auth policy required and testable The optional dep plus a {stripAuthEnv:false} fallback meant a dropped wiring under-stripped silently. Required at all three hops, asserted at install time for the @ts-nocheck caller, and the settings-to-policy mapping is now a named tested function. * P2-3: mobile's default Claude transcript root must follow CLAUDE_CONFIG_DIR session-file-resolver's default ignored the variable the pinned account home follows, so a CLAUDE_CONFIG_DIR launch wrote one tree and mobile read another. The Task-4 test now resolves with no root override (mobile's own call) and checks the answer against the root the CLI itself reports, instead of mirroring the code under test's own expression. * P2-1/P2-2/P3: close the teardown window, join the live-auth gate, align the refusal P2-1: a switch beginning inside the acquire teardown left a dead chat and no replacement. Past that point the launch waits the swap out and refuses only if it never settles; the entry guard still refuses outright, because nothing is torn down there yet. P2-2: structured children now hold the same OAuth-refresh gate a Claude PTY does, so a managed refresh cannot rotate the token out from under a live turn. P3: the refusal now matches the strip it guards (case-folded on win32, presence not truthiness), and the dead structured-to-TUI builder states its auth policy instead of silently signing a system-auth user out. * Make the live-auth gate tests independent of sibling connection teardown order * Do not offer structured Claude under a WSL-only managed account Structured Claude launches against the ambient Claude config, which the account service keeps in sync with the selected HOST account. A WSL-bound managed account lives inside the distro and is never synced there, so on Windows a structured session would authenticate as whatever the ambient identity happens to be while the UI names the WSL account — the user is told one identity and given another. That was unreachable only because nothing offered structured Claude on win32. Enabling it makes it reachable, so gate it here rather than patching the auth layer: refuse the structured path when the active managed Claude account is WSL-bound, and let the terminal-backed path — which resolves the account per runtime — handle that account shape. The answer rides the agentSession.createSupport seam the renderer already consumes, so no new capability and no renderer knowledge of account internals. A create the host declines becomes the definitive refusal the launch fallback already turns into a legacy native chat tab, with no error toast. Unknown answers refuse. An install with no managed accounts claims no identity and is fine, but an active selection that cannot be resolved — or account state that cannot be read at all — is not evidence that the ambient identity is right. Claude only. Codex resolves its account through a different path and its createSupport answer is untouched, as is every Codex routing decision. * Read the structured Claude account gate through the auth policy's accessor The gate resolved the active account from the account-service snapshot's runtime map; the auth policy resolves it with getSelectedClaudeAccountIdForTarget(settings, { runtime: 'host' }). Those are two sources and two resolution rules, and they disagree on a legacy settings blob that carries the selection only in the flat activeClaudeManagedAccountId: the accessor falls through to it, a direct read of the runtime map does not. The gate would then refuse a launch the policy would have run under host-1 — and in the mirror case a session could be admitted under a policy computed from a different account than the gate approved. Read the same settings through the same accessor so agreement is structural rather than coincidental, and drop the controller accessor that existed only to reach the snapshot. No behaviour change for any state both already agreed on; Codex is untouched. * Round-3 review fixes: N-1 empty-value regression, N-2 gate leak window, N-4 lost history N-1: my presence-based conflict predicate refused a terminal launch that works today. 'ANTHROPIC_API_KEY=' is how a user blanks a variable and the settings pipeline preserves that empty value; an empty override cannot beat the pinned account and the strip removes the name anyway. Back to truthiness for the value, keeping the win32 case folding. N-2: enter the live-auth gate only after the exit/close handlers that release it, so no throw in between can leave an entry nothing reconciles. N-4: the Claude transcript resolver searches config-dir-then-default and de-dupes, matching the Codex sibling in the same file, so adopting CLAUDE_CONFIG_DIR no longer hides history written before it. * Run the managed-account gate on every Claude acquisition, not just create createSupport gates the create path, but a session's account state can change while it lives. A reacquire after an unexpected child exit re-resolves the launch and re-derives auth, with nothing re-checking the gate — so a session created while supported could come back up in the refused shape. With the strip predicate keyed on there being an active non-WSL account, the WSL-only user's normalized steady state (accounts exist, none active) does not strip, and that reacquire reaches the child with ambient auth while the UI names the account. Gate at resolveLaunch, the one choke point every acquisition passes through, refusing with the pre-spawn error the caller already handles. Same predicate as create-time, now sharing one settings reader so the two cannot drift. Claude only; Codex resolves its account on a different path and is untouched. The runtime class that wires this does not typecheck its own `this` calls — a missing hookup compiles clean — so the wiring is pinned behaviourally rather than trusted to the compiler. * Move the structured Claude gate out of the @ts-nocheck runtime files Both call sites of the managed-account gate sat in files whose first line is `// @ts-nocheck`, so neither was typechecked: three arguments to a one-argument function plus an undeclared identifier compiled clean. New auth-identity decision logic had no compiler behind it. Move the verdict into a checked module that takes the two facts the runtime owns — the adapter's answer and a settings getter — and decides. The runtime class now only forwards. Move the gate reader's construction into the checked installer too, so the nocheck file passes a plain settings closure and never names a gate symbol. Every reference to the gate predicate and its reader now lives in a checked file, so the ablation that used to pass silently is a compile error at both the create-support and reacquire sites. Removing the file-level @ts-nocheck is a separate, larger job and is not attempted here. * Derive the gate test's auth policy from the settings under test A hardcoded stripAuthEnv asserts a gate/policy pairing production cannot produce, and false additionally lets launch.env inherit the runner's real process.env. Derive via claudeStructuredAuthPolicyForSettings instead: the gate settings type is the same Pick the policy takes, and both resolve the account through getSelectedClaudeAccountIdForTarget. * Pin the absent-vs-empty distinction in the managed-account gate An empty claudeManagedAccounts array is a real answer: the user has no managed accounts, nothing claims an identity, and the ambient path is legitimate. A readable settings object with no such field is settings we failed to parse — the same unknown as unreadable — so it refuses. The two are one character apart in the code and the difference is invisible without the reasoning, so record it at the branch and pin both sides. The test fails under the obvious "consistency fix" of treating a missing field as empty. * fix(claude): keep command queue bookkeeping out of the transcript Claude Code 2.1.258 emits a `command_lifecycle` frame for every uuid-stamped command it starts, completes or cancels. The frame carries a command uuid and a state and no content, and the CLI keeps it out of its own transcript -- but it is absent from the SDK's SDKMessage union and so from Orca's frame catalogue, where an uncatalogued kind defaults to a substantive row. Every structured turn therefore painted raw JSON rows into the user-visible transcript. Catalogue it and disposition it as status chrome. The unknown-kind default stays `timeline-substantive`: a kind we have never seen is likelier to carry content than to be chrome, and a visible row we can catalogue later beats content we silently dropped. A lifecycle state that reads as a failure still surfaces, because the payload error check in `classifyProviderFrame` outranks the catalogue. * fix(claude): let a re-walked descendant become eligible for the forced sweep A descendant first observed by a capture inside its own birth second could never be SIGKILLed: `ps lstart` is second-resolution, so that capture cannot rule out a pid recycled later in the same second, and the merge pinned each retained row to the boundary of the walk that first saw it. SIGTERM-resistant children forked in that window were signalled and then never escalated -- they survived close, quit and restart, reparented to init, and had to be killed by hand. Advancing that boundary on any later capture would be unsound: a later capture matching pid, pgid and start-second is exactly what an impostor would also show. But a capture is not a match -- it is a fresh ppid walk from a root Node pins through its own handle, so a row it re-derives is proved ours at that instant without appealing to its start time. Chain the fence from there instead, and take that walk at the close boundary while the root certainly still lives: the root may leave inside the grace window, and the post-timeout refresh never runs. A row absent from the later walk still keeps its earlier boundary, and a row no walk has ever re-derived in a later second is still never escalated. * Treat an absent managed-account list as empty, not as unreadable An empty claudeManagedAccounts array and a missing one are the same answer: this user has no managed Claude accounts, so nothing claims an identity and ambient auth is the truth. Refusing on absence strands any profile that simply never wrote the key, and it disagrees with the auth policy, whose own predicate takes `(accounts ?? [])` for exactly this reason. Only settings that cannot be READ stay unknown, and those still refuse — as do a WSL-bound active account and a selection naming an account the list does not explain. The earlier reasoning treated a missing field as settings we failed to parse. That conflated "not present" with "not readable"; only the second is unknown. * Support structured Claude when accounts are registered but none is selected Registered-but-deselected Claude accounts were refused, which is behaviourally identical to having no accounts at all: the auth policy does not strip, ambient auth is the truth, and the UI names no host identity. A user who deselected their accounts silently got legacy chat with nothing explaining why. Nothing selected for the host runtime is two states the settings cannot tell apart after the fact, because pruneInvalidClaudeRuntimeSelection empties the host slot and persists null in the second one: honest deselection -> ambient auth, UI names nothing -> SUPPORTED the WSL-only steady state -> ambient auth, UI names the WSL account -> REFUSED The presence of any WSL-bound account in the list decides. Simplifying this to "none active -> supported" re-opens the auth-identity misrepresentation, so the tests fail loudly on exactly that: five of them, across the unit rule and the createSupport path. * Stop treating an unanswerable create-support probe as a refusal A worktree is not resolvable for a beat after createWorktree resolves, so a probe fired immediately after creation fails the RPC with selector_not_found instead of answering. The catch collapsed that into `supported = false`, so the composer refused and quietly built a terminal session — the gate never said no, it was never asked successfully. Elapsed time was the only input that decided whether a Claude launch went structured. "Could not answer" and "answered no" are different states and only the second is a verdict. Retry while the host cannot yet resolve the selector, with a bounded backoff that covers the measured window with margin, and keep refusing on the first ask for everything else. Fail-closed is unchanged: a probe that still cannot be answered when the budget is spent refuses. The retry is narrowed with the shared error-code matcher, which classifies a token that transports re-wrap into a longer message without matching prose that merely mentions it. Codex never probes, so this race has never been able to refuse a Codex launch — the race itself is identical for it. Recorded at the early return, because whoever gives Codex a probe inherits the bug. * fix(claude): fence the forced sweep on re-derivation, not on lstart's second A descendant forked in the same wall-clock second as every walk that sees it was signalled with SIGTERM and then never escalated, so a SIGTERM-resistant child survived tab close, app quit and a full relaunch. Two children of one parent 96ms apart across a second boundary took opposite paths. The leak predates this branch: it reproduces with the change reverted. `ps lstart` has one-second resolution, so a walk landing inside a row's birth second can never rule out a pid recycled later in that same second. But a walk is not a match: a ppid walk only reaches what the root actually parents, and the root is pinned by Node's own handle, so a row the walk re-derived is ours whatever second it was born in -- a stranger would have to have been forked into our tree, and then it is not a stranger. Fence the escalation on that. Rows a merge retained from an earlier walk are not re-derived and still answer to the start-time fence, which remains correct for them. Scoped to callers that revalidate identity before signalling, which is the Claude close path. Codex teardown reaches this same verifier and is unchanged; the argument holds there too, but widening it is its own deliberate change. Also reverts two changes from the previous attempt at this leak. Advancing the capture boundary on a later walk is inert once the sweep fences on re-derivation -- both key on the same set of rows, so the new term short-circuits for exactly the rows whose boundary it advanced. The extra ladder refresh was a duplicate full process-table read: close() already awaits tree.refresh() immediately before proveClaudeChildExit, on the only path that reaches it. Known property: the kill lands roughly a grace window after the walk that proved membership, so a pid recycled inside that gap could in principle be signalled. It is bounded -- matchingSnapshotRows already requires the live row to carry the same start-second and pgid, so an impostor must be born in the remainder of that one second, land on that exact pid, and sit in the same process group, and it has already received the unfenced SIGTERM from the same loop. * Run the Claude structured integration suite as a runtime client The suite exercises agentSession.* for Claude, not the mobile surface: nothing in it asserts anything mobile-specific and its sibling integration suites use 'runtime'. Mobile now additionally requires the experimental structured-chat setting, which structured-agent-session.test.ts pins in both states, so the stale 'mobile' fixture was claiming coverage it never had. * fix(claude): report effort from get_settings, which is the only frame that has it The composer's Effort pill rendered blank in every structured session. This is not a missing source: the publication reads `effortLevel` off the `system/init` frame, and that frame has never carried an effort of any kind, while the correct value is already fetched at acquisition and thrown away on the auth diagnostic. Verified two ways -- a live get_settings probe against Claude Code 2.1.258, and the shipped binary's own init frame construction, which lists `model` and no effort. So `reportedOptions.effort` was always empty, the options reader dropped the key, and the pill had no value. Model survived only because `currentModelId()` has a fallback chain. The get_settings call acquisition already makes reports the session's current effort as `effective.effortLevel`; pass that into the publication instead. Selecting an effort already worked, so this is the arrival value only. The legacy PTY path is unaffected and must not be "fixed" to match: it reads its effort by parsing the startup banner (`CLAUDE_MODEL_EFFORT` in src/renderer/src/components/native-chat/claude-terminal-session-options.ts), which is why it shows a value where the structured path does not. Also removes the fixture that hid this: the fake init frame invented `effortLevel: 'high'`, a field the CLI does not send, which is why every gate stayed green over a value that is always empty in production. The fixture's get_settings now returns the real {applied, effective, sources} shape instead of a bare `{env: {}}`, so the two adapter tests that asserted an effort keep asserting it through the path production actually uses. The reader returns null rather than defaulting: an effort nothing measured would repeat the fixture's mistake, and a blank pill is the honest degradation if the provider ever renames the key. * fix(claude): only record an effort the child confirms it adopted apply_flag_settings answers `success` for an effort it then ignores. Measured against Claude Code 2.1.258: applying `bogus-effort-xyz` returns subtype "success" with no error while `applied.effort` stays at its previous value, and a valid `low` moves it. The option write treated the absence of a throw as adoption and recorded the requested value unconditionally, so Orca would show and persist an effort the child was not using, with nothing anywhere reporting a problem. Read the effort back after applying it, through the same reader the arrival value uses, and reject when the child reports a different one. A readback that could not be taken is not evidence of a refusal -- the apply itself succeeded -- so it still records; only a readback that disagrees rejects. Not reachable from today's picker, which offers catalog values only, but the CLI's effort catalog is server-delivered and has changed before, so a retired id would otherwise become a pill confidently displaying a setting that never took. * test(claude): assert the effort contract against the real binary The blank pill survived every gate because the only tests that touched it were fixture-backed, and the fixture invented the field. A test that pins the shape we read cannot catch the provider renaming the key, which is the failure mode that produced this defect. Asserts both halves against a live authenticated CLI: that no frame it publishes carries an effort at all, and that the session's current effort arrives through get_settings. Which frame proves the session varies by host -- this machine proves it with a SessionStart hook rather than a system/init frame -- so the negative half asserts over every published frame rather than picking one. Skips with the rest of the file when no authenticated CLI is present. * fix(claude): stop the synthesised content-part kinds leaking into the transcript Sending an image put a bare `claude · message:user:content:image` row between the user's bubble and the answer. Two causes, and only the second is a family. An image part counted as modelled only when `source.type === 'url'`, but claudeDispatchMessageContent sends a local attachment as a base64 source and the CLI replays that shape back, so every attached image was classified unmodelled. Accept the base64 and file sources Orca itself sends. The family is the real defect. `message:<role>:content:<type>` kinds are synthesised at runtime from whatever `part.type` arrives, so unlike the top-level frame catalogue they can never be enumerated ahead of time -- the `?? 'timeline-substantive'` default then prints the synthesised name at a user who cannot act on it. That default is right for top-level frames, where "substantive" means show the frame; here it meant show our own vocabulary, which drops the content AND leaks the opcode. So an unrenderable part now renders a sentence saying exactly that, with the kind and payload still on the row's disclosure. A part that carries its own readable sentence keeps it -- the placeholder is a fallback, not an override. An unknown future part type is therefore visible, never silently dropped and never printed as a kind: the same principle as the effort readback, which records only what the provider confirms. * Declare agentSession.requestHandoff on the cross-version wire surface The manifest is a ratchet for cross-version reachability, so the method is declared with real HandoffParams rather than counted. requestHandoff is capability-gated through requireStructuredHost and has no client caller, so declaring it is the whole of the change. Also model two host capabilities the harness omitted: the stub host's supportsCreate, and the fake adapter's, without which adapterSupportsCreate falls through to a supportsLocation the fake also lacks. Every ensure was refused for the harness's silence rather than for its location. * Gate structured Claude session tabs on the client capability that names them The Claude structured lane deleted the projection's `agent !== 'codex'` filter and added CLAUDE_STRUCTURED_AGENT_SESSION_RUNTIME_CAPABILITY in the same commit, but never wired the constant to anything. Paired clients then received agent-session tabs for Claude, which no shipped client renders -- mobile's resolveMobileNativeChat returns null for every agent but codex, so the row listed and selected into a pane with neither chat nor terminal. Restore the filter behind the declared capability instead of the bare agent name. No client advertises it yet, so this matches main's behaviour today and becomes a negotiation a future client can opt into. * Confirm the structured Claude model against the model the CLI reports set_model answers success for any string, including a model it cannot resolve — the failure only surfaces when the turn runs — and get_settings reports the settings-file model, not the session's. The init frame that opens each turn is the only channel carrying the adopted model, so keep the session's reported model current from it instead of reading it once at acquisition. Also stop rejecting an effort the readback cannot represent: max is session-scoped and excluded from the persisted effortLevel, so a readback reporting the level underneath it is an absence of evidence, not a refusal. * Clear the session-option hedge when the provider confirms the value The pill claimed every option was unconfirmed for the life of the session: the renderer recorded each write as dispatched and nothing ever moved it, so a model the CLI had already reported back still read as unconfirmed. Carry the provider's own confirmation to the surface. Main reports which option ids the provider named rather than merely accepted, and the client re-reads options as a turn changes, because the frame that opens a turn is where the adopted model arrives. A value the provider has not reported stays hedged, including an effort whose readback could not be taken. The confirmed list is optional on the wire: a host that predates it sends nothing and the client keeps hedging, which is the behaviour it had. * Keep the model report current across an acquisition fence bump * Show the picked session-option value and let the provider report correct it The pill showed a "not confirmed" second tooltip line for any value we had sent but not yet seen reported back. Nothing acts on it, and for the PTY lane it was permanent — that transport has no report channel. The pill now shows the picked value immediately and the provider's per-turn report corrects it when the two disagree; a newer local write still outranks a report that precedes it. `dispatched` stays as a provenance member rather than collapsing into `applied`: it is produced independently by the PTY lane, and it is where the `confirmed` wire field lands, which would otherwise be unobservable. Effort keeps its readback and its rejection path. That matters more now, not less: with the hedge gone the rejection is the only user-visible failure signal on this surface, so a spurious one would be the loudest bug here. Skipping the readback for an effort the settings response structurally cannot echo is what prevents it — the response carries the persisted level, so reading it back for a session-scoped value would report the level underneath and fail a valid write. * Hedge a session-option value only when the terminal transport sent it Both lanes emit `dispatched`, so it could never say which one produced a value. The descriptor now carries the transport that built it, set once in the shared snapshot builder from a parameter that is required rather than defaulted — the builder is the only place a descriptor is constructed, so a new producer has to name its lane or fail to compile. The structured lane confirms every value from the provider's own per-turn report, which makes the hedge transient noise there. The terminal lane can only learn an outcome by parsing the screen back, and only for Claude: every other agent's `dispatched` value stays unconfirmed for the life of the session, so the line is the only signal that we sent something we never saw land. * Refuse an effort the session's model advertises no control for * Refuse tab mutations on a Claude row the client never negotiated The branch added a case asserting a client advertising only agent-session.structured.v1 may mutate a claude row. That is the same ungated behaviour the projection gate removes, encoded a second time — mutation authorization reads the projection, so hiding the row refuses the write. Assert that contract instead, and add the positive case for a client that does negotiate Claude rows. * Resolve the Claude session's current model in one place so the effort guard and the pill agree * Record an effort the child did not adopt instead of refusing the write apply_flag_settings answers success for an effort it then ignores, so the readback exists to detect that. Refusing on it made the detection a veto, and a veto is only correct if the readback can never be wrong about which model is current -- which it was, twice. The pre-flight guard already refuses a level the model advertises no control for, so the veto guarded a door that is now locked upstream. Keep the detection, drop the refusal: a disagreement records the child's own answer and omits the option from confirmed, so main stops vouching for a value the provider rejected without blocking the user's write. * Stop a slow whole-machine ps from being read as an absent process `ps -axo ...command=` pays a per-pid argv read: measured 1.15s for 1,948 processes (0.03s without `command=`), and CPU contention stretched the same capture to 6.0s. Two budgets sized for a cheap look then misreport a readable machine. The reader's 3s ceiling killed 6 of 20 consecutive captures at load 27, so every consumer answered "unverifiable" about a table it could read. Raise it to 15s, and stamp the capture instant at ps START so `capturedAgeMs` is the upper bound its contract promises -- a 6s capture used to report itself as freshly taken, understating staleness against a 5s kill gate. The TTL keys on completion so a slow capture still coalesces instead of forking ps per caller. `readStructuredTuiProcessIdentity` then spent its whole 5s wait inside one capture and concluded "no exact child" after a single look taken before the child existed (observed landing at ~3.5s). Absence needs a look that did not race the spawn, so require two captures before the deadline can end the loop. Both surfaced by the real-binary Claude TUI resume test, which failed ~1 in 5 under load; 14/14 now, 8 of those runs containing a capture the old 3s budget would have killed. * Let the desktop renderer negotiate Claude structured tabs The paired-client gate hides agent-session rows an agent the client cannot render. The desktop renderer's own IPC dispatches as clientKind 'runtime' advertising only agent-session.structured.v1, so the gate hid Claude rows from the surface this feature ships on. It renders them; it should say so. * Stop a slow process table from silently blinding every freshness gate Stamping `capturedAgeMs` at ps START made the number honest, and honest broke both consumers that read it. `ps -axo ...command=` measured 2.5-9.0s on an idle 2,002-process laptop and 4.0-18.6s at load 46, so the age it now reports lands past every budget: `planRelayPtySweep` refuses the stop as "too old", and the renderer's `admitRemoteForegroundEvidence` refuses the record outright. That second one is the expensive half and was outside the diff -- a refusal bumps `consecutiveInspectionErrors`, the poll scheduler backs off to its 10s floor, and agent-completion detection stops for the pane. The subsystem went blind on exactly the loaded hosts the honest stamp was meant to serve. The evidence-publishing read now gives up at 1,200ms instead of waiting out `PS_TIMEOUT_MS`. It is one budget for one question: these consumers ask whether an observation describes NOW, and past this it does not -- a late answer is refused by the age gate anyway, having first blocked a polled path for the whole capture, so a prompt `unverifiable` is both the truthful verdict and the cheap one. Both relay call sites already produce it from a rejection, and an admitted `unverifiable` costs a poll where a refusal costs the cadence. Identity proof keeps the full 15s through `getFreshProcessTableSnapshot`, because it asks whether a process EXISTS and must never read slow as absent. The budget bounds the wait, never the capture: the reader coalesces, so an abandoned wait leaves its capture running to fill the cache rather than forking a second whole-machine `ps` on the host that can least afford one. 1,200ms is bracketed rather than picked. The floor is the capture's own cost -- `command=` measured 1.15s for 1,948 processes on an idle host, and a budget under that answers `unverifiable` about a machine nobody is straining. The ceiling is the consumer's: 2,000ms, less the 500ms a TTL-shared capture may already have aged, leaves 1,500ms, and transit takes the rest. That ceiling only fits once the capture stops being charged twice. `ps` runs inside the RPC round trip, so its duration is already in `receiveDelay`, and `capturedAgeMs` is that same duration on the host's clock; summing them halved the budget this gate grants a host from ~2.0s of `ps` to ~1.0s, which is why a 1.2s capture arriving at 1.3s read as 2.5s old and was refused. Admission now takes the larger of the two. The sweep's gate keeps its sum, which is correct there: `evidenceAgeSinceListingMs` is stamped after the listing ARRIVES, so it measures planning time and overlaps nothing. A stated limit rather than an assumed one: 15s is not proven sufficient for identity proof. The same capture reached 18.6s at load 46, so that path can still time out and answer "no exact child" about a host it simply could not read in time. Narrowing it needs a cheaper question than a whole-machine argv read, not a larger number. The one test guarding this field could not fail. `beginPtyHandlerTest` installs fake timers, so `Date.now()` is frozen, the real reader reports exactly +0, and `0 <= 500` held identically for a hardcoded zero, for completion-stamping and for start-stamping -- while the real reader on that host returns thousands of ms. It now drives a measured age in and asserts the handler publishes it rather than restamping; that the reader MEASURES it correctly stays pinned separately, against a controllable clock. Both consumers get boundary coverage either side, and each new gate was ablated red before it went green. * Keep the compatibility fields off the capture the budget just abandoned inspectProcess falls back to processHasChildren and listProcesses to getForegroundProcessName, and both read the same TTL-shared capture with no budget of their own. On a slow host they joined the in-flight capture the budgeted evidence read had just given up on, so the call still blocked for the full 6-18s and the budget bought nothing -- once for inspectProcess and once per managed PTY for listProcesses. Use the degraded answers those helpers already give for an unreadable table, reached promptly. pty.hasChildProcesses keeps its unbudgeted fresh probe: it is a one-shot destructive gate that can afford to wait. --------- Co-authored-by: Merge Sim <merge-sim@local> Co-authored-by: Merge Sim <sim@local> |
||
|
Neil
|
637dc30a32 |
fix(relay): observe Windows PTY child processes instead of answering false (#18591)
* fix(relay): observe Windows PTY child processes instead of answering false `processHasChildren` returned a hardcoded `false` on Windows, and a hardcoded negative is indistinguishable from a measurement. Every close guard reads it as "nothing is running in this pane", so an SSH-to-Windows tab running a build closed with no prompt. Measured on a real Windows SSH host: a live `PING.EXE` under the pane's `cmd.exe` still reported `hasChildProcesses: false`, while the identical harness on Linux reported `sleep` / `true`. Windows has no `ps`, but it does have a process table, and the pane walk over it already existed for the foreground reader. The answer now comes from `queryWindowsPaneProcessInventory`; a table it could not read reports `unverifiable` rather than a fabricated negative. `hasChildProcesses` is a boolean, which cannot hold the third answer, and it is read both as "busy, do not close" and as "the agent took the PTY, safe to type into" — so no single mapping of `unverifiable` is safe for both. The verdict moves to a new optional `childProcessEvidence` member that the close paths read; the boolean keeps its exact meaning for every client that cannot. Cost: `pty.inspectProcess` is the polled path and a relay host has no `@vscode/windows-process-tree`, so its table read falls back to the 1.36s CIM scan. Polling that would reinstate the fork storm the shared table exists to prevent, so only a caller whose answer decides something asks for the scan. * fix(runtime): forward scanChildProcesses through the environment inspection RPC `guardRunningTerminalClose` asks the host to pay for a real child-process read, but the environment path dropped the option before it reached the wire: the renderer sent only `expectedIncarnationId`, and the RPC schema — the shared `TerminalHandle` — silently stripped anything else. A host routing that pane through an SSH relay then declined to scan and answered `unverifiable`, which `inspectionReportsRunningWork` reads as running work. The result was a close confirmation on an idle pane, which is the nag this PR exists to avoid. Forwarded through all four layers: renderer payload, RPC schema, method handler, and the runtime/controller signatures. The schema is a dedicated extension rather than a field on `TerminalHandle`, so `clearBuffer`/`agentStatus`/`isRunningAgent` keep refusing an option they have no use for. The silent strip is not itself the defect — it is what makes a new optional member safe to send to an old host, per docs/reference/remote-wire-compatibility.md. The defect was the schema and its caller drifting inside one version, so the tests pin the registered method rather than the schema alone: pointing it back at `TerminalHandle` compiles, parses, and drops the option. Found by review on #18591. * fix(terminal): teach the shared running-work probe the third child-process answer Rebasing onto main landed `probePtyRunningWork`, which is a better home for this than the close guard: it already speaks `live` / `unverifiable` / `exited`, and it exists so the tab-close and window-close guards cannot drift. The child-process verdict belongs there, not in a parallel predicate beside it. So the mapping moves into the probe and `inspectionReportsRunningWork` is deleted rather than kept alongside. The probe now asks for the scan, and a host that could not observe the pane reports `unverifiable` instead of collapsing onto `exited` -- which is what `hasChildProcesses: false` meant on every Windows relay. The pane-close path is routed through the same probe for the same reason; it was the third caller asking this question through a direct inspect of its own. |
||
|
Neil
|
8854b5ded5 |
perf(ssh): coalesce concurrent git.listWorktrees reads (#18419)
* perf(ssh): coalesce concurrent pty.inspectProcess and git.listWorktrees reads Both were the only reads in their provider class with no in-flight dedupe while their siblings already had it. Route them through the existing InFlightPromiseDedupe, keyed per (relay pty id, incarnation) and per repoPath, scoped to the provider instance so two hosts never share an entry. The worktree listing clears from invalidateGitReads(), and a signalled read keeps its own request so one caller's abort cannot cancel its joiners' scan. In-flight only, no TTL: the relay does answer inspectProcess from a 500ms TTL-cached process table, but a client TTL would compound with it rather than match it, so it is not a free win. * perf(ssh): drop the inspectProcess half, ratchet per-read host observations The `git.listWorktrees` dedupe ships unchanged. The `pty.inspectProcess` dedupe is reverted: the host mints one `observationEpoch` per request and the pane foreground reader commits it per read, so two overlapping probes sharing one reply make the second read a stale replay and `admitRemoteForegroundEvidence` rejects it -- a would-be `live` identity read becomes `unverifiable`. The pane foreground tracker overlaps its own probes by design (cancel-and-reissue after a 350 ms settle), so that path is reachable. Adds a ratchet that fails when the dedupe returns, driving the real reader through the real provider operations. * fix(ssh): move the inspect ratchet to the provider it guards The ratchet lived under src/renderer and imported src/main/providers/ssh-pty-provider-rpc-operations, dragging the whole main-process graph into config/tsconfig.tc.web.json (TS6307). Split it: the request-counter ratchet moves next to the provider it pins, and the renderer file keeps the why -- a shared host observation degrades the second overlapping read to unverifiable -- against the real reader with no cross-project import. * docs(ssh): document the worktree-list coalescing contract |
||
|
Neil
|
a5d6114baf |
fix(ssh): stop pane adoption certifying a death from the relay's not-found union (#18531)
* fix(ssh): stop pane adoption certifying a death from the relay's not-found union
`attachStablePaneOwner` was the last reader that synthesised a runtime exit
from a reattach refusal, and it published code `0` — which
`orca-runtime-on-pty-exit` records as `rememberPtyLivenessVerdict(exited)`, a
death certificate whose only legitimate writer is a host-delivered exit frame.
The refusal it acted on is a union. `pty.attach` answers `PTY "<id>" not found`
both for a pid the relay probed with `isProcessAlive` and for an id its session
map simply never had — which, because ids carry a per-start mint epoch, is every
id minted before a relay restart, checked against nothing. So a relay restart
plus a reconnect certified a shell that was still running under the old daemon's
orphaned process tree, retired the pane binding, and cold-started a second agent
onto the same transcript. The sibling `handlePtyReattachFailure` has always
refused to certify from that union; this path did not.
- The relay marks the one refusal it backed with a liveness check
(`PTY_ATTACH_PROVEN_EXITED_MARKER`). The marker is additive, so an unmarked
answer — including an older relay's — stays ambiguous, which is the safe
direction.
- The client mints that half as `SshPtyProvenExitedOnRelayError`, a subclass so
every existing `isSshPtyAbsentFromRelayError` consumer is unchanged.
- Pane adoption publishes `UNVERIFIED_PROCESS_EXIT_CODE` (-1), the sentinel its
sibling publishes, and passes `hostExitConfirmed` only for evidence that
observed the process: the marked relay refusal, or `SessionNotFoundError` from
the registry that owns the PTY. The ambiguous half now records `unverifiable`
instead of `exited`.
- The gone-branch keys on the error type rather than the bare `PTY ".+" not
found` text, so an untyped string can no longer authorise abandoning a
binding — the discriminator `pty-connect-limits.ts` already documented.
Refs docs/reference/ssh-execution-boundary.md
* test(pty): make the pane-adoption fixtures throw what real providers throw
These four fixtures rejected with bare `new Error('Session not found: ...')` and
`new Error('PTY "..." not found')`. No provider produces either untyped:
`local-pty-spawn` and `decodeDaemonResponseError` both mint
`SessionNotFoundError`, and the SSH reattach path types the relay's wire text
before any pane sees it. Fixtures that skip the type were the reason a
message-shaped gate looked adequate.
The exit-code expectations move with it: the pane path now publishes the -1
stop sentinel plus `hostExitConfirmed`, so a certificate follows the evidence
rather than a synthesized zero.
|
||
|
Neil
|
f35015d0c8 |
fix(ssh): measure pane idleness in the unit the sweep's kill operates on (#18415)
The orphan-relay-PTY sweep authorizes `pty.shutdown { immediate: true }`, which
runs `forceKillPosixPtyProcessGroups`: collect every process group on the pane's
tty, then `killpg` each one. The blast radius is therefore (groups on the tty) x
(members of those groups, wherever they are). The idleness evidence measured only
the first factor, so three shapes read as idle and were SIGKILLed:
- with job control off (`set +m`) a background job keeps the SHELL's pgid, so the
tty carries exactly one process group and that group is running the user's build;
- a child that drops the controlling terminal (`ioctl(TIOCNOTTY)` without `setsid`)
keeps the pgid, reports `tpgid == -1`, and never appears in `ps -t <tty>`;
- a double-forked grandchild keeps the pgid and tty but reparents to pid 1, so the
`ppid` walk cannot reach it and the named-process backstop never fires.
`shellOwnsEveryTtyProcessGroup` now also requires the shell's own process group to
hold no other member anywhere in the table, indexed in the same single pass. A
pids-per-tty set would catch the first and third but not the second, which is why
the count is pgid-wide rather than tty-scoped. The wire field keeps its tty-shaped
name: the value only ever became stricter, so an old client skips more, never less.
Second, unrelated-in-mechanism but same file family: `foregroundSkipReason` summed
`capturedAgeMs + evidenceAgeSinceListingMs` without validating either. A non-numeric
`capturedAgeMs` makes the sum `NaN`, and `NaN > 5000` is false, so a malformed record
PASSED the freshness gate and proceeded toward the stop — the one place in the file
that defaulted toward kill. Nothing validated it on this path
(`mapSshPtyProcessList` checks the ownership fields and spreads the rest through;
`PtyProcessListAdmission` is not on the sweep path). It now runs
`isForegroundProcessEvidence` and fails closed.
Verified on real Linux, not only in mocks: a container drives `bash -i` on a real
pty, builds each construction, runs the real publisher and planner, and then calls
the real `forceKillPosixPtyProcessGroups`. Before, all three published
`shellOwnsEveryTtyProcessGroup: true`, planned SWEEP, and the planted pid was gone
after the signal. After, all three skip and survive, and an idle shell is still
reclaimed.
Residuals are written down at the predicate and in ssh-execution-boundary.md: the
capture is a snapshot (bounded by the evidence-age budget, not removed), and a
process the host's own `ps` cannot enumerate stays unobservable while `killpg`
still reaches it.
|
||
|
Brennan BensonandMerge Sim
|
d5803bdbc4 |
feat(ssh): host-stamped remote foreground identity (#18078)
* docs: add SSH agent identity implementation plan * feat(ssh): host-stamped remote foreground identity * fix(runtime): preserve unfenced inspect call shape * perf(ssh): traverse foreground descendants linearly * fix(ssh): bound retired PTY evidence records * test(ssh): cover retired incarnation retention * fix(ssh): make remote process inspection total * Split SSH identity build hot spots * Fix process table snapshot module split * test(ssh): update process inspection expectations * docs: drop the SSH identity plan from the PR The design doc does not belong in the product repo; it stays out of the shipped tree while the implementation carries its own comments. --------- Co-authored-by: Merge Sim <sim@local> |
||
|
Neil
|
08c7152ab6 |
fix(ssh): compare a lease's relay pty id against the pane's app id (#17969)
`getRecentExpiredSshLease` compared the stored lease ptyId (relay form, written through `toStoredPtyId` -> `toRelaySshPtyId`) raw against the runtime's app-form `pty.ptyId`, so `'pty-3' === 'ssh:target@@pty-3'` never held and `recoverTerminalPane` refused every real SSH pane. Normalize with the same tolerant helper the binding reader already uses, now shared as `toComparableRelaySshPtyId`. Switching the path on is only safe on top of #17957 (respawn gated on the runtime liveness verdict), #17965 (`expired` no longer withdraws bindings) and #17966 (supersession and id recycling carry their own marks). `recoverTerminalPane` additionally refuses a lease those marks disqualify, so it acts only on an `expired` lease that means "reattach gave up". The path's outcome is a reattach, not a respawn: `createTerminal` calls `adoptStablePane` first, which attaches attach-only to the retained binding and only falls through to a fresh shell once the host itself answers that the PTY is absent. |
||
 NeilandNeil
|
76836b30ea |
fix(ssh): stop respawning an agent onto a PTY the relay just proved alive (#17951)
* fix(ssh): stop reporting live relay PTYs as expired sessions A `pty.attach` reply carrying `sourceRecovery: restoreRequired` is the relay answering for a PTY it just found in its pool and proved alive with `isProcessAlive`; only the stale output delivery was retired. Main converted that into `SSH_SESSION_EXPIRED`, which is the token every caller uses to retire the pane binding and cold-restore the agent, so a transient reconnect started a second `claude --resume` over a running one's transcript and left the previous remote PTY detached — one more per reconnect until the host refused to fork. Retry the attach once (the relay retires the stale delivery as it answers, so the next attach opens a fresh one with full replay), then fail with a restore-required verdict that makes no claim about absence. Callers already route anything short of absence to the unverifiable pane-recovery path. Also tighten the renderer's expiry verdict, which was a bare substring test: an identity mismatch names a LIVE PTY owned by another pane and observes nothing about this one, and main's own gate already refuses to respawn on it. Refs #11006, #9034 * fix(lint): merge the duplicate pty-connect-limits import * test(ssh): stop pinning the expiry token on a restoreRequired refusal The refusal now carries SSH_PTY_SOURCE_RESTORE_REQUIRED, so the ratchet asserts the discriminating token instead of the one it no longer shares. --------- Co-authored-by: Neil <neil@example.com> |
||
|
Neil
|
d5750648c2 |
fix(runtime): route runtime Git by resolved execution host, not repo connectionId (#18307)
`RuntimeGitTarget` carried `connectionId?: string` and no host id, so `undefined` spelled three different answers at once — "runtime: host", "unresolved", and "genuinely local". Its sole resolver read `store.getRepo(worktree.repoId)?.connectionId` and never looked at `worktree.hostId`, which outranks every repo row, so one arbitrarily chosen row decided the execution host for 36 downstream dispatches. The target now carries `executionHostId: ExecutionHostId` (never null, never optional), resolved through the shared rule that landed with #17909/#17919 and dispatched through the host-keyed routes from #18296. Dispatch sites call `requireRuntimeGitProvider`, where `null` means exactly one thing: the host is `local` and the command runs here as free functions. Four answers that used to collapse into one: - `ssh:x` with a rival row on `ssh:y` — routes to x. Previously the first row won, which is the reproduced cross-host leak. - `local` with a surviving `connectionId` — a row contradicting itself; no SSH connection is handed out. - `runtime:<env>` — throws `ExecutionHostNotDispatchableError`. Its repo row's connection names a target in the *server's* namespace; dialling it here reaches a same-named target on this client. - rival rows disagreeing with no worktree host — `worktree_execution_host_unresolved`, matching the launch path rather than guessing a row. An unreachable SSH host still throws `SSH_GIT_PROVIDER_UNAVAILABLE_MESSAGE`; loss of contact is never evidence of locality (docs/reference/ssh-execution-boundary.md). `resolveWorktreeLaunchHost` keeps its exact signature and now delegates to `resolveWorktreeHostRouting`, the same resolution answering "which host is this on" rather than "what may this client dial" — the git target needs the first question because `local` and `runtime:` are two different non-SSH answers. No wire change: `RuntimeGitTarget` is main-process internal, and the SSH and local model-discovery host keys are byte-identical to before. `RuntimeFileTarget` has the same defect in ~30 filesystem dispatches and is deliberately left for a follow-up. |
||
|
Neil
|
953df47fc4 |
feat(providers): dispatch git and filesystem providers by execution host (#18296)
`const c = repo.connectionId; c ? sshProvider(c) : local()` overloads `null` to mean both "resolved: local" and "could not resolve", so every path that cannot determine the host silently runs remote work on the client (#11163). It also cannot express a `runtime:` host at all. Add a host-keyed dispatch whose input is an `ExecutionHostId` — never null — with `local`, `ssh` and `runtime` as three symmetric entries, and which throws on an id that names no host instead of degrading to this machine. `ssh` carries `provider: null` for "remote, currently unreachable", which is now a different answer from "local" rather than the same one. `runtime:` is a distinct entry rather than a provider because main does not execute runtime hosts at all: they are forwarded over the environment transport, and a runtime row's `connectionId` names a target in the server's namespace. Dialing it from this client's SSH table would trade a silent-local bug for a silent-wrong-host one. First migrations, both to rows resolved via `getRepoExecutionHostId`: - repo-worktrees: an `executionHostId: 'ssh:*'`-only row no longer lists, root-matches, or strict-lists against a same-named local path. - workspace-space-repo-scan: same for the size scan, and `isRemote` no longer contradicts the `executionHostId` emitted beside it. |
||
|
Neil
|
31007c0d86 |
fix(ssh): reclaim relay PTYs the client has provably lost, on host attestation only (#17831)
* fix(ssh): reclaim relay PTYs the host attests this client orphaned (#9819) Orca could lose track of terminals running on an SSH relay until the 50-slot cap refused to open any more. This reclaims them, and the whole design is built around the fact that getting it wrong destroys a user's running process on their remote machine: the failure mode is leak, never kill. A stop requires all nine of: 1. the relay published an `ownerClientInstanceId` read from the live authenticated consumer grant of the connection that requested the spawn — never from a spawn parameter, since an echoed claim is no evidence; absent means skip 2. that id equals this client's persisted consumer identity 3. this connection holds the negotiated `session-owner` grant 4. `paneBound === true`, host-published 5. no `agentSessionOwners` — the host still advertises it as adoptable 6. `hostAgeMs >= 30s`, measured on the host's clock 7. this client has no route: not reattached, no lease outside terminated/expired, no pending kill, and no `expired` lease either — an expired lease is the record of a process deliberately left running, never a licence to kill it 8. every stop is fenced on the incarnation the same listing published, and on the owner identity, both re-checked by the host 9. a pass wanting to stop more than 8 refuses entirely Absence from a client-side set is `unverifiable` by construction (docs/reference/ssh-execution-boundary.md): a second machine attaches to the same relay and displaces the session owner, and its live agents are missing from this client's store for exactly the reason a genuine orphan is. So the host has to attest ownership, and the host has to attest that nothing is running. That second attestation is measured over the pane's whole tty, not its foreground process group. `tpgid == pgid` is foreground-only: on a real `bash -i` on a real pty, a shell holding `sleep 300 &` and a shell holding a Ctrl-Z'd job both read `pgid == tpgid`, `Ss+` — byte-identical to an idle prompt, with only the job's own row differing. A foreground-only gate therefore attests `pnpm build &` and a suspended editor as idle, and the stop that follows SIGKILLs every process group on the tty. `shellOwnsEveryTtyProcessGroup` is measured over that same set of groups, so the evidence and the kill describe the same thing. No new probe: `tpgid` already identifies the terminal, because a process group belongs to one session and a session to at most one controlling terminal. The freshness field is real rather than decorative. `capturedAgeMs` is stamped from when the capture was taken, deliberately as an upper bound since the process table is TTL-shared, and the sweep refuses an observation older than its own pass budget, counting its own elapsed time since the listing arrived. Stale evidence degrades to "do not sweep", never to "sweep". The display consumer of the same measurement keeps no age budget, as a stated decision: a stale pane title costs a redraw and self-corrects. `pty.shutdown` is authorized on the host that owns the process. `pty.spawn` and `pty.attach` both take a request context and check it; the one irreversible call took none, so the rule above lived entirely on the client that decided to make the call. It gains an optional `expectedOwnerClientInstanceId` and refuses unless the connection still authenticates as that identity AND this host recorded it at spawn. Finally, a reattach refusal now says whether it observed the process. Three refusals carry the same `SSH_SESSION_EXPIRED` text and only one is absence; `restoreRequired` means the PTY is live and only its source stream is not. Testing that text with `.includes()` expired the lease and deleted ownership for a running process, erasing this client's only record of it — and a PTY with no record is one the sweep may stop. Wire compatibility: four new optional fields and one new optional param on existing methods, no new method and no new stream opcode (Rule 1, and Rule 2 does not apply). Rule 1's caveat is discharged explicitly — no reader requires any of them, each absence is a named skip reason, and an ordinary pane teardown must omit the owner fence because a revived PTY carries no attested owner at all. New client plus old relay stops zero PTYs; old client plus new relay never reads the fields. Windows relay hosts publish no evidence and therefore never sweep. Verified by joining the real publisher to the real client reader over `ps` captured verbatim from a Linux container, and by driving a real group-for-group SIGKILL against a real pty: backgrounded and suspended jobs survive by pid, and an idle shell is still reclaimed, so the narrowed predicate is not a silent no-op. Squashed deliberately. The sweep is unsafe at every intermediate commit of its own history — before the foreground gate it reaps a hand-launched `claude`, and with a foreground-only gate it reaps a backgrounded build — so this ships as one commit with no bisectable state that kills live work. Refs #9819. Folds in #17939. * fix(i18n): restore the activity-options key the rebase dropped * fix(i18n): union en.json with main so the rebase cannot drop keys |
||
|
Neil
|
0886db2b90 |
refactor(process-table): extract the correlation indexes into their own module (#18246)
`src/shared/process-table-snapshot.ts` is 308 code lines against the 300 cap for `**/*.ts`, so `static analysis` is red on `main` and every open PR inherits it. Neither PR that grew the file crossed the cap alone. #18151 took it to 427 raw lines; #18166 added ~35 more. #18166's branch predated #18151, so the head CI linted was 428 raw lines and passed, while the squash onto main is 463 -> 308 code lines. The gate lints the PR head, not the merge result, so nothing linted the sum until it was on main. Pure move, no behaviour change: the generic index machinery (ProcessIdentityRow, ProcessTableIndexOf, buildProcessTableIndex, collectDescendantsFromIndex, lookupProcessTableIndex, getProcessTableIndex and its WeakMap) moves to process-table-index.ts. `ProcessTableIndex` and `scoreForegroundCandidateRow` stay behind because they need `ProcessTableRow`, which keeps the new module free of any import back and so introduces no cycle. |
||
|
Neil
|
53f105827b |
perf(windows): stop asking the process table for memory, and share one projection per snapshot (#18151)
Two costs on the Windows process-table hot path, plus the EDR doc that described neither of them accurately. 1. The snapshot set `ProcessDataFlag.Memory` and surfaced `memoryBytes`, which nothing read. The addon serves that flag with a second `OpenProcess(PROCESS_QUERY_INFORMATION | PROCESS_VM_READ)` and a `GetProcessMemoryInfo` per process (process.cc:47-63), so the flag was one wasted handle per process per snapshot. 2. The shared TTL cache gave every pane the same native rows array, but each pane still ran `native.map(toProcessRow)` over the whole table, rebuilt a `childrenByPpid` Map from scratch, and did two linear scans. The `.map()` also handed `getProcessTableIndex` a new array each call, defeating the POSIX memo by construction. Both now cache per snapshot identity, and the POSIX resolver drops its duplicate descendant walk. `getProcessTableIndex` / `buildProcessTableIndex` are generic over the row shape so the Windows rows reuse the existing pass instead of a parallel one. No behavior change: same rows in, same rows out, same descendant ordering and same has-children answers. |
||
|
Brennan BensonandMerge Sim
|
1d94ebee3f |
fix(agents): stop a deeper vendor helper from stealing a pane's agent identity (#18062)
* fix(agents): keep outer agent identity over vendor helpers * fix(agents): preserve outer identity across relay scans --------- Co-authored-by: Merge Sim <sim@local> |
||
|
Neil
|
f737f3499f |
fix(relay): stream an oversized fs.listFiles reply instead of refusing it (#17954)
Opening Orca's own checkout over SSH cannot list its files in one response frame. 22,617 tracked paths average 58 characters, so the 20,001-row page the client asks for serializes to 1,223,415 bytes — past `DISPATCHER_CONTROL_QUEUE_MAX_BYTES`, so `sendResponse` demotes it to the `legacy-response` lane, where an unrelated producer backlog can refuse it as an opaque `ResponseOverCapacity`. Break-even is around 49 characters of average path; any `packages/<name>/src/...` monorepo is over the line. Picking a ceiling to refuse at does not fix that, it just moves where it shows up and refuses listings that would have been delivered. `__streamResponse` already exists for exactly this on the git methods, and it is its own negotiation in both directions: an old client never sends it and gets the plain array on the legacy-response lane as before, and an old relay ignores it and answers plainly, which the client detects by the sentinel marker being absent. So fs.listFiles opts into it — no new method, no new opcode, nothing to advertise — and the size of a listing stops being a correctness question. The response-stream registry becomes one per relay, shared by FsHandler and GitHandler. A second registry is not an option and the header of git-response-stream.ts says why: a client keys reassembly on `streamId` alone, so two would hand out the same id and cross-feed chunks, and only the handler that registers `git.responseAck` can credit the window a pump parks on. Also declares `maxResults` on the runtime-RPC `files.listAll` and forwards it. The mechanism "the client names its cap, so a full page reads as truncation" was wired only on the Electron IPC hop; web and mobile were saved incidentally by `remoteFileContentBudget` defaulting the cap inside `listRuntimeFiles`. A new optional field is additive in both directions (wire rule 1). The new Docker-gated spec is claimed by run-ssh-docker-e2e.mjs. The sharded e2e lanes set no ORCA_E2E_SSH_DOCKER, so a Docker-gated spec that no runner names self-skips everywhere and still reports green — pr-e2e-gate-contract enforces that. Closes #12547 |
||
|
Neil
|
34999e328e |
fix(orcad): stop demanding a spawn-helper only macOS builds (#18122)
node-pty declares the spawn-helper target inside binding.gyp's OS=="mac" block and pty.cc execs it only under __APPLE__. Asserting it on `!== 'win32'` made every Linux orcad boot degraded with spawn_helper_missing while its terminals worked fine. Route all four sites through one shared `usesNodePtySpawnHelper` predicate: the precondition verdict, the prebuilt slot install, the +x repair, and the prebuilds build script (which threw outright on a Linux slot build). Fixes #17844 |
||
|
Neil
|
7eb13c184c |
fix(ssh): keep remote PowerShell commands inside what sshd's cmd.exe accepts (#17947)
* fix(ssh): keep remote Windows commands inside cmd.exe's command-line limit Windows OpenSSH runs every exec request through sshd's DefaultShell, which is cmd.exe on a stock install, and cmd.exe refuses a line over 8191 characters with exit 1 and a localized "The command line is too long". `-EncodedCommand` spends 2.67 characters per script character, so five commands on the first-connect path were already over: the stale upload-stage recovery that opens a fresh install (23,210), promote (20,646), cleanup (19,798), the install-lock steal (11,798) and reserve (9,434). A Windows-to-Windows `ssh:connect` died on the first of them before the relay was ever uploaded (#16126). powerShellCommand now falls back to a gzip self-extracting bootstrap once the inline form passes the budget - these scripts are repetitive enough that the worst one lands at 6.5KB - and throws a message naming the limit if even that cannot fit, rather than letting cmd.exe answer in the host's locale. Commands that already fit are byte-identical. The real-binary PowerShell suite in ssh-relay-upload-stage-commands.test.ts exercises the bootstrap end to end, including `exit` and here-string semantics through Invoke-Expression. * fix(ssh): cite the real command-line budget and reuse the cmd.exe ceiling |
||
|
OrcaWinandOrca Worker
|
c3c764fa03 |
docs(win): record why the PTY OSC 133 bootstrap keeps -EncodedCommand (#17875)
The MDE report lists src/main/daemon/shell-ready.ts as a contributing "suspicious PowerShell" site and cites VS Code as using -Command. Measured against a real VS Code fork install, VS Code's -Command payload is a one-liner that dot-sources a *file*; that shape is execution-policy gated and is blocked under Restricted and AllSigned, so it would silently drop OSC 133 -- and with it foreground-process and exit-code tracking -- on the managed fleets MDE runs on. Inline -Command does carry the payload intact through node-pty/ConPTY (powershell.exe 5.1 and pwsh 7.6.5), so the switch is feasible. It is declined because no PTY site spells -ExecutionPolicy Bypass, AMSI and script-block logging decode the payload either way, and the swap would put $ExecutionContext.SessionState.LanguageMode, a Global:prompt override and [char]27-assembled control sequences in clear text on every terminal's command line -- higher-signal than the token it removes. No behaviour change. Adds the rationale at the payload's source of truth, one-line pointers at the three PTY launch sites, and a ratchet that both launch builders must deliver the bootstrap byte for byte. Co-authored-by: Orca Worker <orca-worker@localhost> |
||
|
Neil
|
f9587f74f5 |
fix(ssh): repair a rebuildable node-pty failure once, instead of asking the user to reconnect (#17907)
* fix(relay): diagnose why node-pty will not load instead of hedging The relay could only say "terminals are unavailable" and then list three remedies for four different faults, none of which the user could verify (#17830). Two things were destroying the evidence: - `loadPtyUncached` caught the load error into bare `catch {}` blocks (pty-handler.ts:539, :551) and returned null. The only cause anyone had was discarded on the spot. - node-pty's own loader walks three directories and rethrows only the LAST failure, so even an uncaught error arrives as `Cannot find module '../prebuilds/...'` — the GLIBC/ABI/arch sentence is already gone. The relay now keeps the load error, recovers the real dlopen message with an out-of-process load of the file node-pty would have opened, reads what node-gyp configured the binding for (`build/config.gypi`), captures the host's Node ABI, arch and glibc, and probes the toolchain only when nothing was compiled. Each fault gets its own message naming values the user can check: toolchain_missing, dependency_missing, abi_mismatch, arch_mismatch, libc_floor, shared_library_missing, load_crashed, and load_failed which quotes the loader verbatim. A probe that did not answer stays `unverifiable` and prescribes nothing. The classification is now also structured data on the error, so a client can repair the host instead of printing a paragraph: an additive, schema-validated `data` field on an existing JSON-RPC error, with `repairable` true only for a proved fault that recompiling on the host actually fixes. Reuses orcad's loader-message parsers and out-of-process probe rather than adding a second copy; `classifyLoaderMessage` moves to a shared module and gains architecture and missing-shared-library cases, which the orcad boot precondition picks up too. * fix(ssh): repair a rebuildable node-pty failure once, instead of asking the user to reconnect |
||
|
Neil
|
b4ba3e97ff |
perf(worktree): defer fork-PR remote creation from create-time to first use (#17922)
* perf(worktree): defer fork-PR remote creation from create-time to first use Fork-PR review worktrees eagerly ran `git remote add` + `git fetch` for the contributor's fork (and pinned branch.<x>.remote) at create time, even for a read-only review. That grows remote count unboundedly with review volume and pays a network fetch nobody asked for yet. Defer prepareWorktreePushTarget(Ssh) and the --set-upstream-to configure step at create time (local + SSH, IPC + runtime create paths); persist the pushTarget metadata untouched. Materialize the remote on demand the first time push/pull/fetch/fast-forward actually needs it, via two shared functions (materializeWorktreePushTargetRemote(Ssh)) reused across the legacy IPC handlers and the RPC runtime sync commands. A cheap `remote get-url <name>` probe keeps steady-state calls down to one extra subprocess once materialized, instead of repeating the O(remotes) scan. Add repo-local `remote.<name>.orca-created` config provenance, written when the remote is added, so cleanup can recognize ownership of a remote that was lazily materialized (and therefore never round-tripped through the store's `remoteCreated` flag). Refs #17828 * perf(worktree): materialize a deferred fork-PR remote on terminal spawn An agent running raw git in a freshly opened fork-PR review terminal has no usable upstream until an Orca-driven sync happens -- "sync through Orca first" isn't available mid-task, and git pull/log @{u}.. hard-fail without one (verified against real git). Fire the same on-demand materialization used by push/pull/fetch/fast-forward from the single terminal-spawn resolver (resolveTerminalWorkspaceLaunchTarget), fire-and-forget, so a newly opened terminal gets a working upstream without blocking spawn. * fix(worktree): retest deferred fork-remote CI failures, fix SSH provenance-marker RPC Rewrites the 5 CI failures on the deferred fork-remote change (#17828) as evidence, not fixtures: the SSH relay-upgrade/rollback/sibling-ownership tests move to materializeWorktreePushTargetRemoteSsh, where that unchanged logic now actually runs (create defers it to first sync). While writing a stricter test that routes its mock exec through the relay's real validateGitExecArgs, found that the SSH provenance-marker write (`git config remote.<name>.orca-created true`) was unconditionally rejected by the relay's generic git.exec (it blocks all non-read-only config writes) -- a real bug that would break every SSH fork-remote materialization against a live relay. Fixes it with a narrow git.markRemoteOrcaCreated RPC, mirroring renameCurrentBranch, with a graceful no-op fallback for relays that predate it. * fix(worktree): scope post-#17887 test assertions past narrow-refspec config calls Rebasing onto #17887's narrow-refspec `remote add` broke two broad `['config']` call-filters into false positives/negatives, and the local materialize test still asserted the pre-#17887 wide `remote add`/fetch-refspec forms. * fix(worktree): restructure upstream restore, persist provenance, widen short-circuit refspec (#17828 review) - Move upstream restoration to the materializer level so it runs on both the remoteAlreadyMatchesUrl short-circuit and the full-prepare path, not just buried inside prepare*. - Persist {remoteCreated, remoteName} to the store on materialize so #17842's orphan sweep can see a lazily-created remote, including via desktop IPC, terminal-spawn, and the RPC host-callback paths. - Widen the refspec on the local short-circuit path too (SSH's bare `remote add` refspec gap remains a documented, pre-existing limitation). - Fetch the branch's tracking ref before restoring upstream when the short-circuit widens onto a *new* branch on an already-existing remote -- a bare refspec-config widen never itself imports anything, so `branch --set-upstream-to` was hard-failing for a sibling worktree's first materialize (found via a real-git fixture, not just mocked unit tests). Skipped when the ref already exists so the common repeat-call case stays a local-only probe with no network round-trip. * fix(worktree): merge duplicate shared/worktree/types import oxlint --deny-warnings flags the split import as no-duplicates; full pnpm lint was failing on it after the #17828 review restructuring. * fix(worktree): scope the deferred fetch timeout to fetch calls, retarget stale create-time assertions CI on the previous push failed 3 shards, all argument-shape mismatches: - worktrees-wsl-runtime-routing.test.ts: the "restructure upstream restore" commit wrapped every call `prepareWorktreePushTarget` makes (remote, remote add, config, fetch) with DEFERRED_PUSH_TARGET_FETCH_TIMEOUT_MS, not just the network fetch. Local git subprocesses never need a timeout; scope it to `args[0] === 'fetch'` only, matching the short-circuit path's existing pattern. Updated the test to expect the timeout on the fetch call specifically (point 5 legitimately adds it there), while every other call stays untimed. - worktrees-create-metadata-persistence.test.ts (2 tests): stale from before this session -- create no longer mints a fork remote at all (#17828 deferred that to first sync), so asserting `remote add`/`fetch`/`remoteCreated: true` at create time no longer matches reality. Retargeted both tests to assert the deferred contract (no remote add at create, pushTarget persisted unmaterialized); minting itself stays covered by worktree-remote-push-target-materialization.test.ts and worktree-push-target-setup.test.ts. Re-verified all 5 fixture points (mint upstream, store persistence, single-flight, short-circuit refspec widen + fetch-missing-ref for local and SSH, finite timeout) against a real git fixture after this fix -- all still pass. * fix(worktree): hook pty:spawn into deferred push-target materialization (#17828) triggerTerminalSpawnPushTargetMaterialization only fired for agent/background/ mobile terminals; the desktop GUI's own pty:spawn path (new tab, split, reattach) never materialized a deferred fork-PR remote before raw git commands could run there. Add a small wrapper that resolves the worktree's push target and owning repo from args.worktreeId via the store, and fire-and-forget delegates to the existing materializer, wired as the first statement of runPtyIpcSpawn. Degrades silently (optional chaining + catch) so a partial/fake Store in existing spawn tests can't turn this into a spawn-blocking throw. * test(worktree): retarget stale editor-remote-branch assertions for worktreeId threading runtime-git-sync-client's local-path fetch/pull/fastForward/push calls now forward context.worktreeId (needed by the main-process handlers to key deferred push-target materialization). Update the 17 call-site mocks across 15 tests in editor-remote-branch-actions.test.ts to expect worktreeId: 'wt-1', matching the already-correct source behavior -- no assertion was loosened. * fix(worktree): give a materialize joiner its own branch wiring The materialize single flight is keyed on the remote, but everything after the remote add is per-branch. A sibling worktree joining an in-flight mint for a different branch received the minter's target and skipped its own refspec widen, tracking-ref fetch, and upstream link, so its branch ended with no upstream at all. Wait for the remote, then run the per-branch work against the joiner's own target -- the same path the already-exists short-circuit takes, now shared rather than duplicated. Adopting a remote a sibling minted also stamps ownership, so removing the minter cannot strand the survivor's metadata outside the orphan sweep's reach. * fix(worktree): stop a failed mint from leaving a config-only fork remote Review of the joiner fix found it made things worse in three ways. Swallowing the mint's rejection let a joiner adopt a remote the rollback had already removed, writing remote.<name>.fetch with no URL. Verified on real git: that ghost section breaks `git fetch --all`, forces every later mint to a `-2` name, and cannot be removed by `git remote remove`. Propagate instead; the in-flight map is already cleared, so a retry re-mints. The SSH twin still returned the minter's target to a joiner, so the original per-branch bug survived there. It now adopts against its own target through a twin helper. The ownership stamp was unreachable: it required both a store and a repo id, and no caller passes both. Derive the repo id from the worktree id. Adopters also write remote config, and concurrent `git config --add` has no lock retry -- 135 of 160 writes failed at 8-way concurrency, and equal values duplicate the refspec. Chain adoptions per remote. |
||
|
Neil
|
058e618bb4 |
fix(ssh): stop a failed worktree scan from publishing authoritative emptiness (#17833)
* fix(ssh): keep an unreadable worktree catalog from authorizing teardown #14004: the relay's worktree-list fallback caught every failure and returned `[]`, so `SshGitProvider.listWorktrees` resolved as a success with an empty list. Downstream reconciliation treats a resolved listing as authoritative, which reaches `teardownMissingWorktreeTerminalsBestEffort` and the unregistered-worktree removal paths — a data-loss path from a failed scan. - relay: the `-z`-unsupported fallback lane propagates its failure instead of swallowing it to `[]`. - provider: an empty or malformed `git.listWorktrees` response is refused as `WorktreeCatalogUnavailableError`. A Git repo always lists its own checkout, so a zero-row listing can only be a scan that never answered — this is the mixed-version guard against relays that still swallow. - `listRepoWorktrees`: an unreachable SSH host reports unavailable instead of an empty catalog. #12661: `ssh:terminateSessions` now returns `{ terminated, unverifiable }`, so an offline sweep that only tore down local transport cannot be mistaken for a remote kill. The Manage-hosts toast warns instead of claiming success. * chore(i18n): register the unreachable-terminal terminate message |
||
|
Neil
|
704167197a |
perf(relay): serve one ps capture per window and pin the batched inventory path (#17763)
Follow-up defect fixes for the batched PTY-inventory evidence path (#17525), now on main. - One memoized `ps` capture serves both the lenient and strict views. The two readers ran byte-identical argv behind separate caches, so a relay serving both forked `ps` twice per 500ms window — the doubling issue #6288 removed. - Drop the `byPgid`/`byTpgid` indexes no resolver reads, plus the zero-caller `parseProcessTableRowsStrict` and `getFreshStrictProcessTableSnapshot`; the batch resolver now reuses the shared index lookup and candidate score instead of private copies. - Restore `getForegroundProcessName`'s ladder contract: the extracted table scan answers null again, so an unconfirmed wrapper fallback publishes the recognized (normalized) name rather than node-pty's raw one. - Pin the SHIPPED `pty.listProcesses` path: one capture and one linear row pass for N panes, and node-pty's own name (never "shell") when the capture cannot disambiguate a `node`/`python` wrapper. - Pin the hidden-pane cadence gate in the production option shape, and move the strict-parser coverage next to the parser it tests. |
||
|
Jinwoo Hong
|
8bebcf5def |
fix(terminal): preserve large agent prompt pastes (#17718)
* fix(terminal): preserve large agent prompt pastes * fix(terminal): guard oversized SSH PTY writes * fix(terminal): avoid timer delay for generic sends * fix(pty): propagate provider write refusals |
||
|
Neil
|
dc5db4b01c |
fix(lint): merge duplicate process-table-snapshot imports (#17724)
main is red on `static analysis`: oxlint's code-quality pass runs with --deny-warnings, and agent-foreground-process-batch.test.ts imports '../../shared/process-table-snapshot' twice (lines 5 and 13), tripping "Modules should not be imported multiple times in the same file". Introduced by #17525. It blocks every open PR, none of which can go green until this lands. |
||
|
Brennan BensonandMerge Sim
|
9477b5fcbb |
feat(ssh): batch process evidence in PTY inventory (#17525)
* feat(ssh): batch process evidence in PTY inventory * fix(ssh): accept Linux kernel process rows and make no-evidence polling push-driven * fix(ssh): preserve process evidence polling semantics --------- Co-authored-by: Merge Sim <sim@local> |
||
|
NeilandBrennan Benson
|
fbe94ceff6 |
fix: close readiness gaps found by merged-change audit (#17159)
* fix(ssh): fence stale kills and retired pane replay * fix(ssh): support cancellable interactive authentication * fix(ssh): await remote catalog before snapshot adoption * fix(pty): contain Windows ConPTY input failures * fix(power): avoid redundant macOS display blocking * perf(editor): narrow markdown override subscriptions * fix(quick-open): close directory handles after reads * refactor(linux): remove unused proc socket scanner * fix(usage): apply flat Sonnet 4.6 pricing * ci: prime Node next native test cache * docs(skills): resolve snapshot cleanup data path * fix(ssh): recover install locks after host reboot * test(ssh): recognize boot-aware install locks * test(ssh): prove previous-boot lock recovery live * test(wire): pin pre-metadata release coverage * fix(terminal): preserve remote tab ownership through recovery races * test(runtime): fence replaced terminal handles in agent guard * fix(ssh): preserve remote snapshot authority across polls * fix(pty): contain late ConPTY output EPIPE * test(pty): register Windows exit watcher before kill * fix: close SSH and tab readiness race gaps * fix(tabs): retain headless order and placeholder titles * fix(build): avoid parallel electron-vite config race * test(windows): avoid MSYS temp path rewriting * test(windows): avoid killing exited PTY * fix(pty): avoid late ConPTY input teardown race * fix(terminal): sync reconnect error ownership after commit * fix(runtime): use canonical worktree identity comparison * test(ssh): assert complete cold-hydration baseline * test(windows): invoke quoted retention fixture via PowerShell * test(windows): read ConPTY grid through mode con * fix(terminal): publish PTY replacements atomically * fix(terminal): infer stale identity on reattach * fix(terminal): fence stale pane PTY callbacks * fix(terminal): fence stale pane binds after rebind * fix(terminal): reject stale pane transport callbacks * fix(terminal): fence mirrored reattach spawn callbacks * fix(terminal): replace stale pane PTYs on remount * fix(ci): size the Windows launcher-compile test budget from measurement `native-smoke (windows-latest)` fails ~4.5% of runs on `preserves a multiline argument through the compiled remote launcher` with "Test timed out in 15000ms" — on unrelated PRs, for reasons that have nothing to do with them. Across 176 sampled attempts it is the only red that job produced, and it hit seven different PRs in two days: #16900, #16904, #16915, #16955 (twice), #16979, #17014, #17085. The test is six process creations: powershell.exe forks csc.exe, then the freshly compiled orca.exe forks node.exe, twice. Hosted Windows runners periodically slow process creation down, and this test amplifies that far harder than anything else in the job. Comparing the 80 attempts where it ran under 3s against the 12 where it ran over 12s, its own median goes 2198ms -> 15917ms (7.2x) while the same file's powershell-only test moves 556 -> 686ms (1.2x), the cmd.exe and Git Bash process tests in the neighbouring file move 1.4x, and the other 35 files put together move 1.5x. Measured across those 176 attempts: 1881ms to 35438ms, p50 4264ms, correlation +0.881 with the job's total Vitest duration. 8 of 176 (4.5%) exceeded the 15s cap; 2 of 176 (1.1%) also exceeded the shared 30s testTimeout, so deleting the override and inheriting the config is not enough on its own. 60s clears all 176 with 1.7x headroom on the worst. This is slow, not hung. Every body here is synchronous spawnSync, so Vitest cannot interrupt one — the timer fires only after the body returns and the reported duration is real elapsed time. That is why a failure reads `× ... 22464ms` under `Test timed out in 15000ms`. The work finished; the stopwatch was short. Seven reruns at one identical head measured 2053 / 4680 / 5551 / 8732 / 13506 / 14868 / 21937ms — the last of those would have been red on code that had not changed. The 15s came from #8897, which raised this test off Vitest's built-in 5s default because the job then ran bare `pnpm vitest run`. #8909 landed 3h27m later and pointed the job at config/vitest.config.ts, which is the real fix for that. The constant stayed behind and has been the binding budget ever since. * fix(terminal): fence stale remount reattach ownership * fix(terminal): reconcile mounted pane identity after replacement * fix(terminal): fence stale reattach fallback ownership * fix(terminal): fence deferred SSH reattach ownership * fix(terminal): fence stale split pane ownership callbacks * fix(terminal): keep stale spawns from consuming startup --------- Co-authored-by: Brennan Benson <79079362+brennanb2025@users.noreply.github.com> |
||
|
Neil
|
251c9a0f7d |
perf: remove shell wrapper rebuilds from path checks (#17548)
Avoid rebuilding shell-wrapper contents when only their paths are needed, preserve dispatcher lint boundaries, and isolate updater test timers. |
||
|
Neil
|
1f20a53d22 |
Fix duplicate Codex startup command echo
Deliver local POSIX Codex startup commands through the shell wrapper at shell initialization, preventing duplicate PTY echo. |
||
|
Brennan BensonandMerge Sim
|
b5a85890ac |
perf(git): bound git subprocess execution with an atomic admission scheduler (#16874)
* perf(git): bound git subprocess execution with an atomic admission scheduler Field traces (#16038, #11363) show Windows freeze storms driven by unbounded concurrent git children (12+ at once, 50-65s status convoys for 25+ minutes). Admit every main-process git child against atomic per-budget base+headroom counters (general / network / per-route), with reserved interactive capacity, ordering-only aging, close-bound permit release, a 120s fail-safe read timeout that feeds scheduler backoff, tier plumbing through every option carrier, and coalesced+jittered visibility pollers. Killswitch: ORCA_GIT_ADMISSION_DISABLED=1. Storm harness A/B: max concurrent children 65 -> 6, interactive p95 791ms -> 88ms; output-parity battery byte-identical with admission on vs off. * test(git): run the admission output-parity battery on every platform Parity needs real git, not the storm harness's PATH stub, so it must not share that file's POSIX gate - Windows is the platform where parity evidence matters. * fix(git): preserve interactive admission invariants * perf(git): keep admission queue drains linear * fix(git): close final admission gaps * perf(git): bound eligible route selection * fix(merge): remove unrelated stale snapshot changes * fix(git): preserve refresh lifecycle authority * test(git): align admission lifetime contracts * fix(git): harden admission across runtime paths * fix(git): restore freshness for bulk status reads * test(git): repoint delete-dialog source pins after admission plumbing The hydration effect now orders its targets through orderDeleteWorktreeStatusHydrationTargets and passes includeLineStats alongside the abort signal, so both literal anchors stopped matching. The invariants are unchanged and still pinned: dropping the signal, the main-worktree/folder filter, or getState-instead-of-subscribe each still reddens this test. * Fix git admission tier propagation and lock ordering Decode optional Git status tiers permissively and default runtime RPC status reads to the status lane while preserving renderer caller intent. Acquire the FETCH_HEAD mutex before atomic admission so same-repository fetch waiters hold no global or route permits. Preserve automatic pull-request refresh reasons, keep explicit hosted-review refreshes interactive, remove the dead candidate tier, and keep relay scheduling unchanged. Use tier-aware status lease keys because a shared lease cannot be safely promoted after its admission request is queued or granted. * test: align expectations with admission plumbing * refactor(child-process): move the process contract types to process-spec run-process.ts crossed its line cap after gaining the termination observer; the public types and defaults move out with re-exports so no caller changes. * chore: restore pnpm-lock.yaml to main (unintended local drift) --------- Co-authored-by: Merge Sim <sim@local> |
||
|
Neil
|
5650ec8c6b |
Split local PTY provider responsibilities (#17206)
* Split speech session lifecycle * Split terminal output scheduler pipeline * Split mobile browser pane modules * Prune resolved max-lines suppressions * Split pane tree equalization logic * Extract mobile troubleshoot screen styles * Split external automation manager * Split main window service attachments * Split hosted review creation checks * Split automation dispatch event handling * Split settings navigation metadata * Split daemon initialization lifecycle * Split GitLab item dialog * Split relay dispatcher layers * Split mobile host screen * Retarget mobile view settings source test * Split runtime file client layers * Split ports panel layers * Split runtime environments pane layers * Split local PTY provider responsibilities * Fix F3-speech for #17123 * Fix F1-cycle for #17131 * Fix F4-navtest for #17157 * Fix F2-allowlist for #17161 |
||
|
Neil
|
7879dead69 | perf(ssh): cache legacy worktree clean fallback (#17213) | ||
|
Brennan Benson
|
fd9125ea8c |
feat(native-chat): Codex structured native chat restructure (#16729)
* feat(native-chat): port structured Codex sessions from restructure-recovery Rebuilds the desktop structured native-chat implementation from brennanb2025/native-chat-restructure-recovery (tip 4e31c08db3) on top of current main as a single commit, scoped to the local Codex path. Ported: - Structured agent-session core: durable record store + single-writer lease, canonical journal, agent-session wire host/attach/eviction/subscribers, `agentSession.*` RPC surface (registered via ALL_RPC_METHODS; host-side mobile allowlist included for wire compat), pty write gate, transcript additions, and the Codex app-server adapter/launch resolution. - Renderer: NativeChatStructuredSession view/composer stack, structured launch path with the single-flight guard, local structured session tabs sync, activation gate + structured inventory (read-only `agentSession.handoffStatus` probe), agent-session tabs in the tab strip, AI-vault structured session activation, and the settings pane with the parent Experimental Chat UI toggle plus the nested "Use updated structured native chat" toggle. New sessions require both flags, agent codex, no prompt, and a local non-WSL, non-Windows-host execution host (structured-native-chat-availability). - Fixes 72c013cea6 (verified Codex launch recovery), 8ddbaf5e3d (defer native terminal view switching affordances), and 4e31c08db3 (release the launch gate after a visibility retry) with their regression tests, including the third-launch-after-retry guard case. - Cross-version agent-session wire test + CI lane, packaging entries (proper-lockfile, agent-tooling asar excludes), and the wire-compat doc section. Deliberately not ported: mobile/ changes, the Claude structured runtime (only the claude-transcript-branch-proof and claude-structured-owner-identity leaf modules remain, backing the kept TUI-recovery arms), the terminal↔chat adoption/handoff flow (`agentSession.adoptTerminal`/`requestHandoff`, the handoff request engine, TUI adoption machinery, orca-runtime adoption methods), renderer switching affordances and their dead leftovers, the hook/subagent-status refactor cluster, and unrelated branch changes. The crash-during-acquisition recovery path (restart handoff adjudication, restore/reverse re-acquire, lease schema handoff keys) is kept because every plain direct launch depends on it; a trimmed handoff coordinator exposes only status/restore/close. Branch edits that targeted files main has since split (ipc/pty.ts, worktrees.ts, rpc/methods/terminal.ts, useIpcEvents, pty-connection, store/slices/terminals.ts, runtime-types, web preload) were re-applied to the split modules, preserving main's newer logic (Windows CIM fallback, browser tab close rework, cold-restore resume flow, dispatcher threading). Known seam: the mobile clipboard image-provenance CONSUMER gate ships (agentSession.send refuses unproven mobile image refs with agent_session_image_untrusted) but the producer hunk in rpc/methods/clipboard.ts stays with the unported mobile cluster, so mobile image sends into structured chat fail closed until that side ports. * fix(native-chat): trust only authenticated local image uploads * fix(build): preserve Windows process-tree patch application * test(windows): include process creation time in addon fixture * fix(build): run windows-process-tree node-gyp from the physical package dir gyp expands the node-addon-api dependency by probing node, whose cwd resolves to the package's physical directory in the store, so the emitted target is a store-relative ../../../../node-addon-api@... hop. gyp then resolves that hop against the rebuild cwd; from the node_modules symlink/junction it escapes the store and configure fails with "node_addon_api.gyp not found" (run 32999886072). Rebuild from realpath(package dir) so both bases agree, matching how the package manager itself runs native install scripts. The regression test replays gyp's expansion+resolution against the planned cwd and fails without the fix. * fix(native-chat): keep chat tabs visible through terminal closes and empty-worktree launches Two proven blockers in the native Codex tab contract: closeTerminalTab pre-empted the canonical unified close. With one terminal left it deactivated the worktree on a terminal/editor/browser-only check, blanking a workspace that still held a renderable agent-session tab; with two or more it pre-picked a successor from terminal entities only, re-stamping the group active before closeUnifiedTab's MRU/neighbor repair could land on the chat tab. Successor choice now defers to the unified contract whenever the terminal has a unified row, and deactivation is gated on the unified renderable count (matching leaveWorktreeIfEmpty), with the legacy pre-pick kept only for terminals without a unified row. A structured session created on an empty worktree was published into the host's headless group while preserveLocalLayout froze the local layout, leaving the tab in store but permanently off screen. A preserveLocalLayout owner now always takes client-owned placement — repairing a rendered leaf whose group record is missing, or materializing a rendered group on a truly empty worktree — and applies the client-derived layout repair while still rejecting host-authored layout. Regression tests drive the real store through closeTerminalTab (git worktree and folder workspace) and the real snapshot applier for the empty-worktree adoption states; all fail without the fixes. * fix(native-chat): close stale turns and retry rejected sends * fix(native-chat): retire hosted rows on structured tab activation * fix(native-chat): preserve rpc defaults across main merge * chore: format remote wire compatibility guide * test(native-chat): cover retry after unconfirmed send * fix(native-chat): reload outbox on session switch * docs(settings): disclose structured chat platform limits * fix(native-chat): await Codex launch-home preparation * fix(codex): align child-process allowlist with async trust bridge * test(identity): update inventory for tab surface refactor * fix(windows): preserve process-tree CRLF patch sources * fix(native-chat): anchor an unmatched chat echo where it was sent (#16117) * fix(native-chat): anchor an unmatched chat echo where it was sent The reported symptom was old user messages replaying below every new turn, so the conversation read as scrambled. The cause was not that the echo failed to match a transcript row. Claude consumes a mid-turn send through a `queued_command` attachment and writes no `type:"user"` record for it, so some echoes can never match, and no amount of matching will change that. The cause was WHERE an unmatched echo rendered: buildMobileNativeChatTransientData appended every pending item after the entire transcript, so it re-read below each turn that landed afterwards. Render each echo directly after the transcript row it was sent against, using the baseline the send already captures. An unmatched echo is then at worst a duplicate in the right position rather than a scrambled one, and it stays visible. Echoes sharing an anchor keep send order; a send with no baseline, or one whose anchor folding dropped, still falls back to the tail. Deliberately NOT fixed by deleting the echo. Inferring from send ordering that an echo can never match, then removing it, loses the user's own text for a message the agent did receive, and it cannot fire in the common case anyway - measured drain groups are 1,017 of size 1 against 55 larger. It also escalates an existing gap: the count pass has no baseline-tail guard, unlike the glue pass, while `messages` is a 40-row window that head-trims, resets on reconnect and grows at the front on loadEarlier, so a false landing there would license deleting a DIFFERENT outstanding message. That count-pass gap is real and left for a separate change; anchoring makes its worst case a duplicate in place rather than a scrambled conversation. * fix(native-chat): preserve folded echo anchors * fix(native-chat): preserve forward-folded echo anchors * fix(native-chat): keep leading folded echoes in place * fix(workspace-cleanup): show git status for every row (#16690) * fix(native-chat): refuse structured chat on every Windows execution path canUseStructuredNativeChat only refused win32 when a project runtime resolved, so folder-workspace keys (and other keys with no project runtime) failed open into structured chat on Windows. Fail closed on win32 unconditionally after the host check, matching the settings copy: local macOS/Linux only; Windows/WSL/SSH stay on terminal chat. * fix(native-chat): restore runtime refusals behind the win32 gate |
||
|
Jinwoo Hong
|
fc8c981103 | fix(browser-preview): enforce canonical runtime grants (#16975) | ||
|
Neil
|
25fa6a9920 |
fix(ssh): a relay that reports a PTY absent must authorise a fresh spawn (#16763)
* fix(ssh): a relay that reports a PTY absent must authorise a fresh spawn reattachSshPtySession flattened the relay's `PTY "<id>" not found` into a bare `SSH_SESSION_EXPIRED`, discarding the evidence. isPtyAlreadyGoneError classifies on that message via /PTY ".+" not found/, so it could never match — and attachStablePaneOwner's existing already-gone fallback (retire the pane binding, onPtyExit, spawn fresh) is gated on exactly that predicate. The recovery machinery was already correct; only the classification was broken, which is why `pty:spawn` rejected outright after a relay restart renumbered its PTYs from pty-1. SshPtyAbsentFromRelayError carries the verdict the message cannot: a reachable relay answered for this exact id and reported it absent. Both thrown messages stay byte-identical to the previous ones, so all six message-based consumers are untouched; only callers that can act on the stronger verdict test the class. Deliberately not raised for a transport failure, timeout, disposed multiplexer, identity mismatch, or restoreRequired — none of those observe the process, and treating them as absence would orphan live remote work. * test(ssh): cover the absence fallback against real persistence The first cut passed no store/worktreeId, so attachStablePaneOwner's guard short-circuited and retireTerminalSurfaceFromPersistence — which deletes the parent tab and its layout when the retired leaf is the last one — never ran. retirePersistedStablePaneOwner and attachStablePaneOwner had no coverage at all. |
||
|
Neil
|
f48c038a93 | Extract local PTY launch helpers (#16778) | ||
|
Jinwoo Hong
|
d60a3c900b | Reset stale terminal modes after dead TUI replay (#16379) | ||
|
Neil
|
26721bd632 |
fix(codex): stop blocking the main thread on trust grants (#16441) (#16594)
* fix(codex): stop blocking the main thread on trust grants (#16441) Codex hook trust was granted by blocking the Electron main thread on `spawnSync` of a bundled ELECTRON_RUN_AS_NODE entry for the whole app-server deadline: 15s native, 35s WSL, ~45s on the real-home path (rebase inspect + repair + grant). Cold start and every Codex pane launch showed "Not Responding"; the reported event-loop gap was 15,049 ms. The subprocess only ever existed to donate an event loop to a deliberately blocked parent — `runCodexHookTrustGrantSession` was already the real async implementation. Make the callers async and the fork is unnecessary, so the bridge, the forked entry and its envelope are deleted along with their build/knip/tsconfig registrations. The CLI `agent hooks prepare-codex` handler is already async, so it awaits the in-process session and saves a process spawn per managed-home shell. `resolveCodexTrustGrantHost` is async too; the WSL identity probe moves from `execFileSync` to `runProcess`, dropping that file from the child-process import allowlist. Status reads keep a synchronous native-only stamp path. Two invariants that held only because the lane blocked: - Overlapping capability probes were impossible by construction. `GitCapabilityCache`'s dedupe engine is extracted to a shared `CapabilityProbeCache` and `CodexAppServerCapabilityCache` now inherits it, so concurrent launches against a cold host share one app-server session instead of one each. - Two grants on one `config.toml` could not interleave capture and restore. A reentrant per-file lane now serializes the whole install sequence (managed, WSL runtime, real-home ensure, legacy sweep) and the grant and rebase inside it. Cold-start work moves off the critical path: retained-home reconciliation (N sequential sessions) is fire-and-forget behind the daemon provider, and the startup real-home ensure chains into managed hook reconciliation instead of blocking app init. Every preserved semantic is unchanged: never throws, the ORCA_DISABLE_CODEX_TRUST_RPC kill switch, ledger hits, backfill-pending and cooldown fallbacks, config rollback on every failure path, pre-grant self-computed trust removal, the verify-failure taxonomy, diagnostics and telemetry. * fix(codex): widen the trust-config lane to every config.toml writer Review follow-ups on #16441's async trust grant: - `markCodexProjectTrusted` now runs inside the runtime+system config.toml lanes, so a project-trust write can no longer land inside a hook grant's capture->restore window and be silently reverted. Its callers await it. - `install`/`refreshRuntimeUserHooks`/`remove` hold the system config.toml lane as well as the runtime one — they promote approvals into ~/.codex/config.toml and mirror it back. Lock order is runtime-before-system everywhere. - The real-home ensure chain resumes after a rejection instead of returning the same rejected promise to every later pane launch, and resolving the real home is now inside the module's never-throws boundary. - `buildSpawnEnv` awaits inside a cancelable pending-spawn registration, so shutdown during the (now long) env build stops the PTY from launching. `prepareLocalPtySpawn` generalizes into `awaitCancelableLocalPtySpawn`. - CapabilityProbeCache drops the test-only `nowMs` passthrough; its probe backstop comment now describes what it actually guards. - Preflight is a plain async function; the trust dispatch in orca-runtime collapses into one `markWorkspaceTrustedForAgent`. * test(codex): exercise the trust-config lane under real concurrency The async grant makes two pane launches overlap for the first time. These drive the real modules end to end on real files: a rollback swallowing a sibling's grant, a markCodexProjectTrusted write landing inside a capture -> restore window, shared capability-probe dedupe on a cold host, the host-scoped transient cooldown, and reentrancy from inside an installer. Each was verified to fail against a deliberately broken implementation (lane removed, dedupe disabled, cooldown made global, reentrancy pass- through disabled). * test(codex): stop hook-service suites spawning the developer's real codex The forked grant bundle never existed under vitest, so the RPC lane was unreachable in tests on main. Running it in-process makes these suites spawn a real `codex app-server` when one is installed: 38 spawns and two failures in hook-service-runtime-trust-repair on a machine with codex, green in CI where there is none. Stand in for the missing binary so both environments exercise the same fallback lane. * docs(codex): scope the trust-RPC kill switch comment to what it actually gates The comment read as though the flag forces the fallback lane everywhere. It gates the managed grant only: the real-home rebase still runs its own inspect/repair app-server sessions when Orca's insertion shifts a user's hook positions, and never reads the flag. Verified by exercise, not by reading — with the flag set, both inspect-user-hook-trust and repair-user-hook-trust still ran. Pre-existing: main has no check there either, it just blocked the main thread while doing it. Widening the flag to cover the rebase is a follow-up; this only stops the comment promising something the constant does not do. |
||
|
Neil
|
5e900b10b3 |
fix(windows): let a build with no job exports still retire a dead agent (#16563)
* fix(windows): let a build with no job exports still retire a dead agent
#16419 (
|
||
|
Neil
|
a1ec0479e2 |
fix(windows): revalidate PTY liveness from the job object, not a forked helper (#16419)
* fix(windows): answer console membership from the job object, not a forked helper
node-pty answers "which processes are attached to this pane's console?" by
FORKING a helper, because GetConsoleProcessList must run from a process
attached to that console. Orca asked on a foreground poll, per pane, so each
read spawned a conpty_console_list_agent -- hundreds of hidden processes
exhausting RAM within minutes, respawning as fast as they were killed (#10857).
QueryInformationJobObject has no console-attachment constraint: any process
holding the job handle can ask. Orca already creates that job per PTY, and
listPtyJobProcessIds has exposed it since the W1/W2 work with zero callers.
One syscall, no children.
Semantics the three call sites rely on are preserved: a root-only set still
proves the shell is alone (so a stale agent can be retired), and size > 1 still
proves something is running under it. The single difference is that a
descendant detached from the console stays in the job -- which widens the set,
the conservative direction for every caller.
Also fixes the third call site, which returned { available: false } whenever
membership was unavailable AND a recognized agent existed -- i.e. exactly while
an agent was running. Membership only ever narrowed the candidate list, so an
unavailable answer now leaves it unfiltered instead of failing the whole
resolution.
The no-fork test is asserted through a module-level vi.mock of
node:child_process. A vi.spyOn of a require()'d child_process does not
intercept the module's own import binding: the first version of that test
passed with a fork() deliberately reintroduced.
* fix(windows): keep console attachment for the candidate filter
Readiness review caught that this PR changed two different questions as if they
were one, and the repo's own plan doc had already said so:
"The job is the wrong set here -- it would re-admit precisely the detached
process the filter exists to drop." (windows-wsl-root-cause-plan.html, Use B)
The two uses:
- Use A, `size > 1` at local-pty-provider and the daemon tracker -- "is anything
in this pane besides the shell?". The job answers this, in-process and with no
fork. Unchanged from the previous commit.
- Use B, the candidate filter -- "which of these are ATTACHED TO THIS CONSOLE?".
Its whole job is dropping a descendant that detached, and the job object keeps
those, so answering it from the job makes the filter a no-op in its motivating
case: a detached `Start-Process droid` would be granted byte authority, and a
detached sibling would make an attached agent look ambiguous.
Use B goes back to GetConsoleProcessList, in its own module named for what it
answers, with its fail-closed null restored. That path is not the #10857 storm:
it runs only when a recognized agent candidate already exists, not on every
foreground poll. Bounding it to one pooled supervised helper is the remaining
half, and per the plan doc either half alone takes #10857 from unbounded to one.
My earlier claim that widening membership is "the conservative direction for
every caller" was wrong -- true for Use A, backwards for Use B. The hardware run
did not catch it because I measured a WSL pane, where the superset is harmless,
and never a detached GUI child, which is the divergence.
* fix: restore the coverage and ratchets the module split dropped
Round 2 of review. Two blockers, both from moving the forking code to a new
file without moving what guarded it.
- The child_process import ratchet was RED: windows-console-attached-processes.ts
imports node:child_process and was unlisted, and the old entry was stale. I
never ran that suite -- lint and the providers/daemon tests both pass without
it, which is exactly the gap the ratchet exists to close. Entry repointed;
count unchanged at 159.
- The forking module had ZERO tests. Its 11 assertions -- bounded timeout,
single kill, spawn error, malformed message, helper-pid removal -- were in the
file that now answers a different question, so the module that actually caused
#10857 was shipping untested. Moved with the code.
Also: nothing pinned the round-1 fix itself. No test drove console attachment to
null and asserted the fail-closed result, so re-deleting that branch would have
gone green. Now covered, and verified to fail when the branch is removed.
Cleanups the split left behind: `consoleMembershipUnavailable`/`consoleProcessIds`
renamed to `pane*` where they now hold job membership, the duplicated
`WindowsConptyMembershipDeps` type name, comments still describing the console
on the job path, and eight reliability-gate paths pointing at the moved tests.
* fix(windows): let a superset job answer expire instead of vetoing retirement
Round 3. The job read had reintroduced #9258's bug by a new mechanism.
`size > 1` returned unconditionally, so any pane holding a console-detached
descendant never retired its cached agent. A WSL pane always holds some: the
measurement in this PR's own test recorded job [40980,104068,4888,69908] against
console [69908,40980], i.e. console said "shell alone, retire" while the job said
"three others alive, keep". #9258's third commit describes the identical failure
from the other direction -- a bare shell reading as [helper, shell] "looked like
it still had a child ... the foreground refresh held the exited agent's identity
indefinitely" -- and that is what came back.
It bites because the read branch that serves the cached name across a Windows
shell fallback is deliberately untimed: #9258 made it so on the stated assumption
that "the background refresh authoritatively retires it". Removing the retire
authority left the identity with no bound at all. Second-order: a non-null cache
makes idleNoEvidenceShell false, which pins the refresh at the 1s TTL, so an idle
WSL pane also scanned the process table every second forever.
A TTL on the read would have been the wrong fix -- untimed is deliberate, because
on Windows the fallback name is structurally uninformative. Instead the job answer
is treated as what it is: a SUPERSET of the console, which cannot tell a working
agent from a leftover. Proof of absence retires immediately (size 1, unchanged);
an inconclusive answer ages out at 30s; unverifiable (null) still holds forever
per ssh-execution-boundary.md. Only successful scans that found no agent advance
the clock -- a degraded scan returns before this -- so the fix cannot expire an
agent it simply failed to see.
Also from review:
- Restore the root requirement the forked probe had. Without it a set of one
non-root pid -- shell gone, descendant alive -- read as "shell alone, retire",
inverting the truth.
- Rename to windows-pty-job-membership.ts / readWindowsPtyJobProcessIds. The old
name still said ConPTY console while reading the job, and conflating those two
sets is precisely the bug
|
||
|
Neil
|
07b13c8468 | refactor git repository host boundaries (#16177) | ||
 Jinjing
|
7e76bb3aec |
Fix rebase race by fetching to private ref before rebasing (#15990)
* Fix rebase race by fetching to private ref before rebasing
`git pull --rebase` is vulnerable to concurrent fetches modifying remote-tracking refs during execution. Fetch to a temporary private ref (refs/orca/rebase/*) first, then rebase from that stable ref to avoid the race condition.
* Fix rebase race by fetching to private ref with timeout
Concurrent fetches can interfere with remote-tracking refs between
fetch and rebase. Use a unique private ref and 60-second timeout to
isolate each rebase operation and prevent hangs on stalled remotes.
Extract gitPullRebaseFromBase to a dedicated module.
* fix rebase race by fetching to private ref with timeouts
Concurrent fetches can replace FETCH_HEAD and remote-tracking refs between
fetch and rebase, causing the rebase to fail. Fetch to a temporary private
ref instead, use --no-write-fetch-head when available (Git 2.29+), and
serialize FETCH_HEAD access for older versions. Add process termination
barriers to ensure proper cleanup and extend timeouts for SSH operations.
* Fix rebase race by fetching to both private and tracking refs
Concurrent fetches between source and rebase can replace remote-tracking refs,
causing rebases to use stale bases. Now fetch to both a private ref and the
remote-tracking ref simultaneously, ensuring the tracking ref stays current.
Also improves process termination for WSL guests with process-group tracking,
fixes process-tree termination timeouts on POSIX, and serializes FETCH_HEAD
operations for linked worktrees through their shared Git directory.
* Add WSL setsid --wait probe and barrier termination timeout
Probe for `setsid --wait` support and fall back to unwrapped execution for BusyBox compatibility. Add a deadline for process termination barriers to prevent hanging when tree termination cannot be verified. Update tests for cross-platform compatibility.
* Add wsl-process-group-termination to WSL invocation allowlist
* Serialize per-worktree git mutations to fix rebase race
Introduce operation locking for each worktree to prevent concurrent
mutations (like rebase) from interfering with each other. Ensures
rebasing a linked worktree doesn't affect the source worktree state.
Add SIGKILL fallback if process termination barriers cannot verify
tree termination.
* Serialize pull and fastForward operations per-worktree
- Extract generic git operation lock to reuse locking pattern
- Refactor existing locks to use the generic implementation
- Apply per-worktree serialization to pull and fastForward to prevent races
* Route WSL group termination through runWslProcess
|