* Add daily macOS dev build release channel
Publish once-daily signed macOS builds from main at a dedicated cadence,
separate from hourly (too noisy) and release branches (too infrequent).
Builds are notarized and installable via the updater, but unvetted —
published to stablyai/orca-daily rather than the main repo to avoid
evicting stable/RC entries from the releases feed.
* fix lint
* fix commit
* Add third token mint to daily macOS build workflow
The upload step's 2x45m retry budget can outlive the one-hour token, so a third
is minted after it for verify and cleanup operations. Release notes are moved to
a file to ensure consistency between draft creation and publish. Daily channel
description updated with specific UTC release time.
* fix(terminal): extend the shell-ready startup barrier to fish (STA-3417)
Fish never emitted the OSC 777 shell-ready marker, so agent launch
commands were written into the PTY while fish/Starship were still
initializing: the daemon path wrote them synchronously at session
create and the local path blind-wrote ~30ms after the first output
byte. The command was echoed by the kernel but never executed.
- shell-templates: shared fish --init-command that emits the marker
once on the first fish_prompt event (the earliest point fish's own
reader owns the PTY, mirroring zsh's zle-line-init marker)
- daemon shell-ready: fish joins the startup barrier so the launch
command queues until the marker (timeout fallback unchanged)
- local-pty-shell-ready: fish launch config gains the marker wrapper
- codex-startup-delivery/tui-agent-startup: omp/pi/opencode plans now
request shell-ready delivery (codex parity) so the SSH renderer path
also waits for the prompt; plain payload-free codex stays on the
markerless fast path
* fix(terminal): answer DA1 past the shell-ready barrier
The barrier queues all inbound input until the ready marker, including the
renderer's DA1 reply. A shell that withholds its first prompt until DA1 is
answered — fish waits 10s — therefore never emits the marker that would
release the reply it is waiting for. Measured: 10.37s to launch an agent,
versus 0.35s once the reply lands.
Answer DA1 from the daemon while the barrier holds, writing straight to the
subprocess so the reply bypasses the queue, and consume the query so the
renderer's xterm cannot also reply. Released on ready, timeout, or dispose,
handing DA1 back to the renderer for steady state.
Consolidates the identical DA1 handler the ConPTY override already used.
* fix(terminal): prevent duplicate startup DA1 replies
* fix(ci): run the root-directory guard on stock macOS bash 3.2
The guard script builds its base-tree lookup with `declare -A`, which
needs bash 4+. Its test spawns plain `bash` from PATH, and stock macOS
has shipped /bin/bash 3.2 since 2007, so on any Mac without a Homebrew
bash the script exits 2 before asserting anything and the default
`pnpm test` suite fails 3 of the guard's 4 cases. Machines with a
Homebrew bash on PATH never see it, which is why it went unnoticed.
Replace the associative array with a plain-array linear scan. Root
directories number in the dozens, so the O(n^2) membership check is
negligible, and the NUL-delimited reads that protect unusual filenames
stay as they were. The empty-array expansion is guarded for
`set -u` under bash 3.2.
All four guard tests now pass with /bin/bash 3.2; behavior under CI's
bash 5 is unchanged.
* fix(ci): run the root-directory guard under node instead of bash
The guard is the only check in the repo written in shell, and it used
`declare -A`, which stock macOS `/bin/bash` 3.2 does not have — so the
guard's own test suite failed 3 of 4 cases on any Mac without a Homebrew
bash. CI never noticed because runners ship bash 5.
Porting it to node removes the interpreter-version variable instead of
working around one construct: node is what the sibling script in this
directory already uses, it is the runtime that runs the test, and the
NUL-delimited read is the same shape as check-changed-code-quality.mjs.
It also drops a latent false pass — a failing `git ls-tree` inside the
shell's `< <(...)` was not caught by `pipefail`, so the read loop saw
nothing and the guard reported success. `execFileSync` throws instead,
which is why the two `git rev-parse --verify` probes are no longer
needed.
Output and exit codes are otherwise unchanged; the usage line now prints
node's script path where the shell printed `$0`.
Tests pin each guarantee and fail when it is reverted: NUL-delimited
reads so odd paths are reported unmangled, exit 2 on bad usage, and
git's own 128 with no node stack trace when a sha does not resolve.
* fix(ci): keep root entry bytes intact and fence guard output
git pathnames are arbitrary bytes, but the guard read ls-tree with
encoding 'utf8', so every invalid sequence collapsed to U+FFFD. That
mangled the reported name and, because the replacement is not
injective, let two different entries compare equal — a genuinely new
root entry could be waved through as pre-existing. Read the bytes as
latin1 and write them back unchanged.
The blocked-entry list is also attacker-controlled and went straight to
stdout. The runner trims leading whitespace before matching '::', so an
indented entry name still parses as a workflow command, and a pathname
may embed a newline. Wrap the list in ::stop-commands:: with a random
resume token so only the guard's own annotation is acted on.
---------
Co-authored-by: Brennan Benson <79079362+brennanb2025@users.noreply.github.com>
* Revert "test(ime): restore coverage the composition-ownership change removed (#13168)"
This reverts commit 25a8c517e1.
* Revert "refactor(terminal): return IME composition ownership to xterm (#13128)"
This reverts commit 17b3dff3c4.
* test(ime): keep the architecture-neutral Korean trace coverage
The recorded IBus/fcitx5 and Windows MS-Korean traces from #13168 assert PTY
byte order, not composition ownership, so they still hold once the terminal
composition layer is restored. The mobile accessory-order test pinned the new
handleLiveInputChange signature and does not.
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): keep the macOS Backslash bypass through the revert
The restored native-text forwarder only claims keys for input sources in its
hardcoded CJK allowlist, so third-party IMEs off that list (Qingg, #10896) still
get a raw backslash. #13128 added this bypass as a partial replacement; keep it
rather than trade the open issue back.
Scoped to the bare backslash key. The rest of shouldBypassXtermForMacNativeText
bypassed all unmodified non-ASCII text, which would race the restored forwarder.
Co-authored-by: Orca <help@stably.ai>
* fix(mobile): move the mirror-step ref write out of render
The restored hook assigned runMirrorStepRef during render, which is not
replay-safe — React can discard render work, so the mutation can leak from UI
that never commits. Its only read is inside the held-commit timer, which fires
long after commit, and the ref has a safe default, so an effect is soon enough.
Surfaced by the changed-lines React Doctor gate: the rule postdates this code,
so restoring the file re-introduced it as a new violation.
Co-authored-by: Orca <help@stably.ai>
---------
Co-authored-by: Orca <help@stably.ai>
* fix(ci): stop hourly prune from deleting the just-published release
Hourly prune sorted non-draft releases by createdAt, but nearly every
orca-hourly release shares one createdAt from bulk import. At the retain
cap, stable sort + reverse put the newest release past the window and
immediately deleted it with --cleanup-tag.
Sort by publishedAt (tagName as tie-break) and hard-skip the tag this run
just published so prune cannot self-delete.
* fix(ci): harden hourly prune protect without retain+1 drift
Review found that skipping only in the delete loop under-prunes when sort
is wrong, and excluding TAG before the retain slice would permanently keep
retain+1 releases. Force this run's tag to the front of the sorted list
before slicing so it always has a retain seat and oldest builds still prune.
Also gate prune on publish_live success and warn if TAG still appears stale.
* fix(terminal): return IME composition ownership to xterm
* fix(mobile): derive terminal input from native replacement ranges
* test(mobile): record iOS Japanese IME traces
* fix(mobile): preserve native IME replacement ranges
* fix(xterm): flush queued application input after IME commit
* test(terminal): pin Korean intermediate commit
* test: pin Windows IME shortcut ownership
* test: replay IBus number candidate commit
* fix: preserve native macOS input-method punctuation
* refactor(terminal): remove stale mac focus override
* fix(mobile): preserve soft keyboard deletion ranges
* fix: keep IME-owned palette chords in renderer
* fix: stop carried IME shortcuts at renderer owner
* fix: preserve carried IME shortcut dispatch
* fix: narrow main-owned shortcut actions
* test(mobile): pin Japanese IME replacement traces
* test(terminal): retain paired native IME trace
* fix(chat): preserve browser IME composition ownership
* fix(chat): retain macOS IME confirm gesture
* fix(chat): expire unmatched IME confirm carry
* fix(chat): isolate IME confirmation expiry
* fix(chat): retain active IME confirmation
* refactor(terminal): remove dead composition handler
* feat(ime): add shared Enter-ownership seams for CJK composition
The confirming Enter of a CJK composition arrives as two keydowns and the
orderings differ by platform: Windows/Linux redispatch the unmarked Enter/13
before keyup, macOS delivers keyup first. A guard reading only isComposing or
keyCode 229 misses the redispatch, so surfaces submitted on a confirm.
Adds useImeEnterGestureOwnership (carry token, next-frame expiry), a shared
ImeEnterGuardedForm for native implicit submission, and the cmdk seam covering
18 CommandInput surfaces at one site.
A chorded Enter arms the carry but is never swallowed — the reverse would eat a
user's deliberate Cmd/Ctrl+Enter. Both failure modes are pinned by
ime-enter-gesture-ownership-contract.test.ts.
Co-authored-by: Orca <help@stably.ai>
* refactor(terminal): consolidate native input listeners and parked-screen owner
Extracts the shared native-input listener installer and renames the parked-screen
detector for what it actually does, replacing per-call-site duplication. The
listener installer keeps a forgetOptionKeyLocationOnBlur flag so per-window
semantics are preserved rather than flattened.
Net deletion; no behaviour change intended.
Co-authored-by: Orca <help@stably.ai>
* test(terminal): pin recorded IME shapes as regression tests
Nine regression tests built from hashed affected-platform captures, each with a
paired ordinary negative and a discriminating mutation verified to take the file
from all-passing to exactly one failure.
Covers the Windows MS-Korean Shift family (#12179, #11878, #12151, #11946,
#12152) and the Korean TUI line-break rows (STA-3237, STA-3222, STA-3129).
STA-3237 pins the empirical 3-Shift / 2-active-composition / 2-newline ratio the
device run established — the third Shift produces nothing because Space has
already committed. That ratio is not derivable from a static capture.
Co-authored-by: Orca <help@stably.ai>
* fix(ime): guard Enter-commit surfaces against CJK confirm
Applies the Enter-ownership guards across the surfaces whose Enter commits
something: publishes, clones, pairs, installs, posts, or persists.
Tiered deliberately rather than uniformly. Irreversible and remote-effect sites
take the carry token, which also blocks the unmarked redispatch. Locally
reversible sites take the oracle check with a one-line comment naming the
residual, because a spurious commit there costs one undo.
Three numeric fields are left unguarded with the reason in-code: Chromium blanks
number inputs at compositionstart, so a confirm-Enter only ever reaches an
empty-draft reset. Measured with a CDP probe rather than assumed — a guard that
cannot fire is noise.
Co-authored-by: Orca <help@stably.ai>
* test(ime): teeth-check the Enter guards on every guarded surface
One suite per guarded surface, each verified by deleting the guard and
confirming the test fails. A green guard test without that check is unverified,
not verified.
Two shapes pass vacuously in happy-dom and are avoided here: native implicit
form submission never fires, and blur() is inert on an unfocused element. Both
made "the commit did not happen" assertions pass with the guard removed, so the
suites assert the guard's contract directly instead.
Co-authored-by: Orca <help@stably.ai>
* fix(mobile): keep iOS Korean commits whole through the live-input path
iOS Korean reports isComposing: false on every event, so it bypasses the
composition guard entirely. The strict owner rejected UIKit's transformed
post-change field and sent only the leading jamo — the reported symptom.
Prefers the authoritative same-event field text over the predicted text when the
supplied operation cannot produce it. Generic: no Korean special-case, no locale
classifier, no normalization. Adds the RN-target-keyed submit carry alongside it.
Co-authored-by: Orca <help@stably.ai>
* test(e2e): make IME capture harnesses fail loudly instead of silently
Four instruments recorded silence as success, so a void run scored as a clean
one:
- readTerminalImeBoundaryTrace returned an empty trace when the probe never
installed, making every "nothing leaked" negative pass vacuously
- summarizeLatencies([]) returned a perfect zero distribution that passed all
three latency thresholds
- the macOS Vietnamese spec pinned an input-source ID that does not exist, and
failed as though the operator had chosen the wrong source
- the expectedLineCount=1 prefix property was undocumented and one edit from
silently downgrading a PTY assertion
Input sources now resolve by enumeration and name the near-matches on failure.
Co-authored-by: Orca <help@stably.ai>
* test(terminal): cover Cangjie cancellation and fix a cross-namespace assertion
Adds #11951's recorded Cangjie cancel shape to the existing cancellation suite,
which covered Pinyin and Sogou but not Cangjie. One keystroke then Backspace
arriving as deleteContentBackward with data: null, so the stale preedit is the
only thing a fallback could replay.
Verified against the historical pre-6cd944c62b3 bundle: the positive fails with
['尸'] where [] is expected, while the ordinary negative stays green.
Also fixes the Vietnamese spec, which asserted a TIS-space input-source ID
against getKeyboardInputSourceId(). Those two Orca APIs report the same source
in different namespaces — TIS nests it under VietnameseIM, the app API does not.
The resolver stays as an installation precondition; the assertion matches the
leaf.
Co-authored-by: Orca <help@stably.ai>
* test(e2e): add a real-IME macOS arm for the Korean chord commit
The existing korean-ime-terminal-shift-enter-commit spec synthesizes composition
over CDP: Input.imeSetComposition sets the preedit directly and Input.insertText
performs the commit. Asserting the IME produced events you injected yourself is
circular, so that spec cannot certify real-IME behaviour.
This arm selects 2-Set Korean via TIS, reads it back live, and injects through
System Events key codes, so the OS owns the preedit, the commit instant, and
isComposing. PTY byte expectations are preserved verbatim.
Covers 2 of the original 4 cases by design. The other two are the Windows/Linux
redispatch-before-keyup ordering, which macOS cannot produce and which cannot be
selected -- the OS decides it. Reintroducing synthesis to "restore coverage"
would reintroduce the circularity.
Co-authored-by: Orca <help@stably.ai>
* test(e2e): assert the macOS chord arm at the PTY boundary, not the renderer
The byte expectations were transcribed from korean-ime-terminal-shift-enter-commit
:364/:383, which assert against onData -- a renderer boundary where the terminator
is CR. This spec reads the PTY child, where the tty has already converted CR to LF.
Names both forms per row rather than swapping the constant, so the conversion reads
as evidence that the capture reached past the renderer, as #11936 and #11951 record.
Ctrl+Enter's CSI-u sequence is unaffected and is identical at both boundaries.
Co-authored-by: Orca <help@stably.ai>
* test(e2e): measure composer-to-onData latency and stop dropping IME keystrokes
Two defects in the echo latency probe.
It hooked onWriteParsed and onRender but never onData, so it measured
key->parse->render echo rather than the composer-vs-onData delta the latency rows
need. Adds a third hook feeding its own sample set.
And `event.key.length !== 1` silently dropped IME keystrokes: Pinyin and Cangjie
keydowns arrive as key:'Process' (length 7). Replayed over the captured corpus,
the old filter accepted 580 of 4137 Chinese IME keydowns -- it was discarding 80%
of them. The new filter matches the shape the owner itself branches on.
Attribution charges each onData to the latest keydown rather than a FIFO head,
because composing jamo emit no onData at all and a queue would credit a whole
composition to its first keystroke. The consumer now asserts sample count before
any percentile, so a zero-sample run cannot render as a flawless distribution.
Co-authored-by: Orca <help@stably.ai>
* test(terminal): pin the WSL shifted-jamo newline shape for #11919
In Korean 2-set, Shift types ordinary letters -- the double consonants and the
compound vowels. Each such keystroke reaches Chromium as key='Process',
keyCode=229, shiftKey=true.
The v1.4.163 classifier matched exactly that pattern with no code guard, so it
called those keystrokes Enter, rewrote them to a synthetic Shift+Enter, and
injected a newline into the middle of the word -- with no Enter key pressed.
That is why the reporters said "no modifier key pressed": they had not chorded
Shift+Enter, but they had pressed Shift, to type the double consonant.
Asserts the row's own recorded capture: 40 immediate keydowns, exactly 3 of them
Shift-carrying inside a single syllable, and an onData stream with one newline
per Enter press and none mid-word. Two ordinary negatives keep it from being a
blanket mute -- the same session's non-IME keydowns still reach shortcut policy,
and an ordinary Shift+Enter still resolves through the real policy.
Co-authored-by: Orca <help@stably.ai>
* test(terminal): pin the composition commit lag that made Korean type one behind
macOS Korean 2-Set commits syllable N only when the first jamo of N+1 arrives, so
compositionend and compositionstart land in the same task. A composition-start
handler cancelled the pending finalizer that was the only path to triggerDataEvent
and ended the session without emitting bytes, so every committed syllable reached
onData exactly one syllable late and the backlog cleared only at a Space or Enter.
Types continuously with no Enter and no Space -- either would flush the backlog and
hide it -- and samples onData at every syllable boundary. Paired with a
length-matched ASCII arm that stays green throughout, so the positive is a fact
about composition rather than about timing in general.
Bisected to a single call site across five builds: pristine, 1.4.155 and 1.4.162
pass, 1.4.163 fails, removing the one call repairs it, restoring it fails
identically. That window is exactly the reporter's "started immediately after
updating".
Co-authored-by: Orca <help@stably.ai>
* test(mobile): cover the send-queue abort that silently drops queued keystrokes
One failed send in use-terminal-live-input-commit aborts every keystroke
queued behind it, with the error swallowed by .catch(() => false). The
existing test resolves(true) on every send, so the failure branch was
uncovered.
Four arms: the abort itself, an ordinary negative on the healthy path, a
throwing sender, and a liveness control proving the queue recovers once
the chain settles. Deleting the abort takes 4 passed to 3 failed, with the
ordinary negative correctly surviving.
Scope is stated in the docblock: this is a transport send-queue abort,
reachable only via a real disconnect or RPC error. REQUEST_TIMEOUT_MS is
30s, so latency alone cannot reach the branch — consistent with #7094's
symptom class, not proven to be its cause.
* test(terminal): pin that daemon snapshot/restore cannot disturb a composition
Two independent reporters attributed broken Korean composition to the
always-on PTY daemon repainting terminal state over the preedit. The
attribution is wrong on ancestry — the daemon shipped three months before
the version both call good — but the boundary was never actually tested.
Runs the real applyMainBufferSnapshot choreography against a live
composition, including the full 2J/3J/H wipe plus the resize and
alt-screen branches. textarea.value, selectionStart/End,
compositionView.textContent and .active all survive byte-identical, and
interleaving a restore between every jamo of 문제 still commits 문제 at
onData. Also pins that the uncommitted preedit is absent from the captured
snapshot: it lives in the textarea, never the buffer, so a restore has
nothing stale to echo back.
Injecting one textarea.value = '' into the restore fails exactly the three
restore-boundary tests.
* test(terminal): pin that Cmd tears down a composition where Ctrl and Shift do not
xterm's composition keydown exempts only keyCode 16/17/18 (Shift/Ctrl/Alt)
plus 20/229. macOS Meta — 91/93/224 — is absent, so a Cmd press mid-composition
takes _finalizeComposition(false): the overlay goes dark and never recovers,
because compositionstart is not re-fired. The user composes the rest of the
word blind. Linux and Windows users press Ctrl and are exempt.
xterm already has a Meta-aware modifier predicate in wasModifierKeyOnlyEvent,
so this is an internal inconsistency rather than a deliberate choice.
Owns no reported row and is version-neutral: 5/5 on both 1.4.162 and 1.4.163.
The branch is unexercised in all 328 recorded traces, so this is a hazard pin,
not a regression guard. Only the teardown is asserted; the likely duplicated
commit needs a compositionend the IME kept alive across the Cmd, which no
capture contains.
Deleting the exemption fails exactly the three paired negatives; adding Meta
to it fails exactly the two Cmd arms.
* test(native-chat): characterize preedit loss when a question card replaces the composer
An AskUserQuestion card fully replaces the composer by design, but the
in-flight composition goes with it: the composer unmounts before
compositionend reaches it, so the preedit is never committed to the draft.
The committed text survives only because the draft is cached and restored
via defaultValue. Node identity changes, value 'abc' is preserved, the 가
is gone.
Drives the real NativeChatView -> SessionGate -> InteractiveCard ->
questionActive swap -> Composer -> ComposerField, flipped by writing the
same store field an AskUserQuestion hook event writes. Flipping
questionActive to false fails exactly this test and nothing else across
639 native-chat tests, so the path was entirely unguarded.
CHARACTERIZATION TEST: it asserts the loss. Fixing the defect — committing
the preedit before the swap, or keeping the composer mounted — will make
this file fail. Update the expectations to the new contract rather than
working around them.
Owns no reported row. #12118/STA-3219 flicker is keyed to token counters,
which provably do not remount, and a question card arrives once per
question.
* test(terminal): pin the duplicated commit when Meta interrupts a composition
_finalizeComposition(false) sends textarea.value.substring(start, end) but
cannot clear the IME-owned textarea, so a later compositionend re-sends the
same range. Meta reaches that path because CompositionHelper exempts only
Shift/Ctrl/Alt; xterm's own wasModifierKeyOnlyEvent covers Meta four ways,
so the omission is an internal inconsistency rather than a choice.
Companion to the modifier-exemption guard, which deliberately pins only the
overlay teardown. This pins the data consequence.
HAZARD PIN: owns no reported row. The trigger is unverified on hardware —
no capture in the corpus contains a Meta-during-composition gesture, and
whether macOS keeps the composition alive across it is unmeasured. The
duplication follows from the code given that sequence; whether users reach
the sequence is the open half.
An earlier premise that Space (keyCode 32) reaches this path was refuted by
a corpus scan: 0 of 731 evidence files carry a keyCode-32 Space while
composing, against 171 at 229, and 229 returns early.
* test(terminal): characterize the syllable lost when the textarea blurs mid-composition
CoreBrowserTerminal._handleTextAreaBlur clears the helper textarea
unconditionally — "Text can safely be removed on blur" — while
CompositionHelper._finalizeComposition reads the committed text back out of
that same value from a deferred timeout. By the time it runs the value is
empty, the substring is '', and triggerDataEvent never sees the syllable.
xterm checks composition state in _syncTextArea and omits the same check
here.
Six cases. Blurring mid-composition loses the syllable in every ordering,
including compositionend-before-blur, which is Chromium's real order — so
it is not an ordering artifact. A bare textarea.blur() with no Orca code
loses it too, which places the owner upstream: Orca's unguarded release on
outside pointerdown is one trigger, not the cause. Committing 한 then
blurring mid-가 yields ['한'] where ['한','가'] is correct: one syllable
gone, surrounding text intact.
Teeth checked by inverting — adding an Orca-side composition guard flips
exactly the three cases that route through the release path and leaves the
bare-blur and no-blur cases green, which is the scope split: a fix in
regular-terminal-focus-ownership alone would not close this.
HAZARD PIN, but unlike the others this one has a real production injector —
clicking outside the terminal mid-composition. Owns no reported row. The
shape matches #9738's report; the injector does not, and a shape match with
a mismatched injector is not an owner.
* test(terminal): say which arm the STA-3237 fixture came from
The recorded keydowns are wave 4's A-shift-unmarked-only — the arm that
emits no PTY bytes. Nothing in the file said so, so two readers concluded
the row's events fail the owner's predicate and that STA-3237 and STA-3222
were different defects. They share an owner; the arm that fires is
Process/229+Shift, absent from this bubble-phase trace because the owner
claims it in the capture phase.
Also corrects "code-blind": the v1.4.163 policy emits \x1b\r only for a
shift-only key:'Enter', and a jamo keydown reaches that branch solely via
the isTerminalImeProcessEnter rewrite. The mock is deliberately wider so
the ownership guard stays under test if that rewrite moves.
Comments only — no assertion, fixture value, or mock behaviour changed.
* test(e2e): track the input-source selector the macOS specs shell out to
Five tracked macOS IME specs ran `swift .tmp/select-input-source.swift`, a
file that is gitignored and existed only on one machine. Anyone else
checking out the repo — or the same machine after .tmp is cleaned — could
not run them, and they are the capture drivers for the macOS rows that are
blocked waiting for exactly those runs.
Moves it to tests/e2e/ beside its callers. The chord spec now resolves it
from __dirname rather than reaching two levels up into .tmp.
* test(terminal): pin the CJK repaint decision against the reporter's own output
#12164 comment 1 and #5921 report agent output with double-width glyphs
rendering duplicated character-by-character while ASCII in the same line
stays clean. No IME, no composition, no keystroke — the user never types
the CJK.
Segmenting all three verbatim samples into maximal same-risk-class runs
gives 33 runs and zero violations of "this run is corrupted iff the
production detector flags it": 17 wide runs all corrupted, 16 narrow runs
all byte-identical. The paired negative is co-located in the same line
rather than in a separate run — the reporter supplied it without knowing.
Doubling is asserted as present, not uniform: 자바스크립트 and 시스템 each
leave a jamo undoubled, which is a repaint-region boundary artifact rather
than a per-character transform.
The discriminating arm is in the test rather than a source mutation:
be3f30e2f8 (#6890) elects a repaint for all 17 corrupted runs when the
agent types nothing, and reverting its disjunct elects none. Both
predicates agree once the user has recently typed, which is the pre-#6890
condition.
Samples inlined with per-sample SHA-256 because .tmp is gitignored and
cannot back a landed test.
* test(terminal): pin macOS period substitution landing after the composition
#11504's reporter published a DOM trace showing insertText ". " arriving
149ms after compositionend, when two spaces are typed with a CJK input
source and NSAutomaticPeriodSubstitutionEnabled is on. This replays that
trace against a real Terminal and asserts what reaches onData — bytes to
the PTY, not anything visual.
The owner is stock upstream CoreBrowserTerminal._inputEvent, not an Orca
module, confirmed at the resolved install and in the shipped bundle Vite
loads rather than in the TypeScript source.
Three mutations against that install, predictions written before the runs,
each failing exactly the arms predicted: dropping the composed/keyDownSeen
guard fails two, dropping Orca's intercept fails the one arm where the
payload arrives before the send window drains, and flipping || to && —
the candidate-fix shape — fails the arm that pins the defect itself.
CHARACTERIZATION: arm 1 asserts the broken behaviour and will fail the
moment #11504 is fixed. Update it to the new contract rather than working
around it.
composed is absent from every recorded bundle, so composed: true is the
spec-required value rather than a captured one; the test asserts it before
dispatching so a harness that dropped the field fails loudly.
* test(terminal): replay the recorded Windows Shift sessions through the IME guard
STA-3179 reports a Shift release sending Enter; #12171 reports delayed
Hangul plus doubled newlines. Both replay their own recorded Windows
MS-Korean keydowns through resolveTerminalKeyboardShortcutAction with the
shortcut policy mocked, so the assertions are about which events reach the
policy and what reaches terminal input.
STA-3179's held-Shift gesture yields exactly one newline, from the unmarked
Enter alone; its release arms nothing for the next composition, asserted
after a precondition check that the release really is keyups with shiftKey
already dropped; and an ordinary Shift press-and-release still routes every
keydown, which is the paired non-IME negative.
Teeth, verified by mutation: bypassing the isImeOwnedKeyboardEvent guard in
keyboard-handlers takes STA-3179 from 3 passed to 2 failed / 1 passed — the
survivor being the ordinary-session negative, which is correct, since a
non-IME session should not depend on that guard — and #12171 from 2 passed
to 2 failed. Source restored byte-identical.
Recorded shapes are inlined and the bundles cited in comments; nothing is
imported from .tmp, which is gitignored.
* test(native-chat): correct 61977d45177 — the preedit survives the question card
61977d45177 claimed a composed syllable vanishes silently when an
AskUserQuestion card replaces the composer, and characterized that loss.
The claim was false. Its premise was an artifact of the harness: the test
simulated a preedit with a silent textarea.value assignment and no input
event, which no IME does.
Real composition fires input with insertCompositionText on every keystroke
— the shape this repo already records in its own observed-event capture —
and React's change handler returns on input/change with no composition
gate, so onChange runs for each frame. The draft cache is written
synchronously inside the updater, so the preedit is already committed
before the card can arrive. Driven that way, it survives.
Renamed to match the contract that actually holds, and extended: Hangul
jamo-per-frame, Japanese kana accumulation followed by per-segment
conversion asserting the candidate the user was looking at survives, and a
pin on the mechanism itself — the draft cache holds the preedit while the
card is up.
Teeth: there is no fix to revert, so the mutation is the plausible wrong
one — gating onChange on isComposing(). That takes 4 passed to 3 failed,
with the English negative correctly surviving, since it has no composition
to gate.
Two consequences remain, recorded rather than fixed: the OS aborts the
composition when the field disappears, so a lone jamo returns as a
compatibility jamo the user cannot compose onto, and the remounted
composer is unfocused because the card owned focus.
* docs(native-chat): name the corrected commit and the degraded-jamo consequence
Records in the file itself that 61977d45177 is pushed and wrong, quoting
the two claims that are false, so a reader who finds it in git log reaches
the correction from the file that replaced it.
Also states the residual as a consequence rather than a curiosity: a lone
leading jamo returns as a standalone compatibility jamo (U+3131), which is
not a composable state — the user cannot resume the syllable, only delete
and retype. Preserved, but degraded into something unusable. That is the
note to find if a reporter ever describes exactly that.
The invariant these tests pin is not "the composer commits on unmount" but
"composition input events must reach React" — which is what a future IME
change would break, and is not visible from the swap site at all.
* test(terminal): replay the recorded macOS Telex commit boundaries
#6905 reports Vietnamese composed characters breaking in the terminal.
Replays the retained macOS built-in Simple Telex capture — recorded
selection and value set before each dispatch, since that is what the
commit range reads — and asserts what reaches onData: the first commit
alone, then through the real Enter, then the ASCII tail of the same run.
A code-point count would catch NFD normalisation.
ENGINE CAVEAT, stated first in the docblock: this is macOS built-in Simple
Telex, Telex only. The reporter's three named engines cannot run on the
platform they declared, and which macOS Vietnamese engine they used is
unconfirmed. This file certifies no engine, and does not imply VNI.
The owner is upstream's — CompositionHelper._finalizeComposition's
waitForPropagation branch — so the arms are copies under .tmp aliased by a
scratch config, with node_modules verified unchanged by shasum after every
run. Collapsing the range end onto its start fails all three; collapsing
the start to zero re-emits the first word into the second commit, which is
the reporter's "duplicated" direction. Different failure sets, so the
mutants are distinguishable rather than merely detectable, and the ASCII
assertion passes under both.
Falsifiability here is by mutation, not by a defective build: #6905 does
not reproduce at HEAD, so this has never been watched going red on a real
reproduction.
* docs(terminal): lead the #6905 test with its engine caveat
Comment-only. Moves the caveat above the source line so a reader meets what
the file does NOT establish before what it does — the capture is macOS
built-in Simple Telex, the reporter's named engines cannot run on the
platform they declared, and which engine they used is what gates this row.
Co-authored-by: Orca <help@stably.ai>
* docs(terminal): record that the swallow eats a keystroke after Japanese conversion
This pin framed the swallowed insertText around Cmd interrupting a
composition. A differential through Japanese multi-segment conversion shows
it is broader: type a segment, convert, then press `a`, and the `a` is
lost. No modifier, no exotic gesture. Korean surfaced it first only because
2-Set composes on nearly every keystroke.
Also records why it cannot simply be fixed. The suppression de-duplicates
IMEs that deliver their commit a task after compositionend, which a sibling
test pins; this swallow is that dedup's false positive, and the two events
differ only in payload, so no flag-timing change separates them. Both a
smaller redesign and a content-aware variant were built and measured — the
first duplicates on IBus, the second costs a reported row's test and is
blocked while the patch cannot be regenerated.
The Japanese arrays behind this are authored, not observed: no Japanese DOM
composition trace exists in the corpus.
* docs(terminal): a Japanese capture does exist — correcting dbeecb11bee
That commit said no Japanese DOM composition trace exists in the corpus.
False. One does, filed under the Linux bundles rather than the bundle named
for Japanese: 30 DOM events, two にほんご->日本語 conversions, with full
selection state per event. It is retained byte-identically in three further
bundles — one capture copied four times, not four observations, checked by
hash rather than by counting files.
The claim came from checking the bundle named for Japanese, finding nothing,
and generalising to the corpus without querying the rest of it.
Replaying it emits 日本語日本語 under both sequencing extremes on all four
arms, matching its own recorded onData. So "repeated conversion is
undisturbed" is now captured rather than authored. It carries no
post-compositionend insertText, so it cannot speak to the swallow: the
a-after-conversion figure stays authored and unobserved.
Also rewords the paragraph opener. It claimed to broaden a Cmd framing, but
hazard 2 was never Cmd-framed — the lines above already say Cmd does not
reach it. The real gap was that hazard 2 named no trigger at all, which
reads as exotic when it is ordinary.
* build(xterm): land the patch regeneration harness
The five dependency patches under config/patches/ shipped with no tracked
way to regenerate any of them. The xterm one is the hard case: it is derived
from an upstream build, so no fix could be made without rebuilding, and the
tooling to rebuild lived only in one machine's scratch directory. That
blocked a measured fix for a live keystroke-loss bug, and the EditContext
reduction an OSS survey identified as the only real one available.
Adds the regenerator, the upstream pin, the hand-written source patch the
bundle hunks derive from, tests, docs, and a PR job that verifies the
shipped patches still match the pinned build. The job caches the shallow
clone keyed on the manifest, so a cold run is minutes and a warm one under
one. Round-trip verified: regenerating from a clean checkout reproduces the
shipped patch byte-for-byte.
Marks the emitted patch -diff -text. pnpm hashes it byte-for-byte, so a
CRLF checkout would break install on Windows, and its minified bundle lines
make a diff nobody can read — review the source patch instead.
Also rejects unknown flags. --check was the fallback for any unrecognised
argument, so a typo, or --help, silently triggered a full upstream build
instead of what the caller asked for.
* fix(xterm): stop swallowing a keystroke typed after an IME commit
Type a Japanese segment, convert it, then press a key one macrotask later
and that key was lost. No modifier, nothing exotic — every user who keeps
typing straight after converting. Korean surfaced it first only because
2-Set composes on nearly every keystroke.
handleCompositionInput discarded the payload unconditionally in the window
after the deferred send: _isSendingComposition stays true for one macrotask
after the timer cleared _pendingCompositionStart, and the branch substituted
'' for whatever arrived. The suppression is not itself wrong — it
de-duplicates IMEs that deliver their commit an event-loop turn late, which
terminal-stock-composition.test.ts pins. It just could not tell a duplicate
from new input, because the two events are identical apart from payload.
Now it compares against _sentComposition, the text the deferred send
actually emitted, and discards only a match. A flag-timing redesign was
measured first and rejected: it fixed this and duplicated on IBus, because
no timing change can separate events that differ only in content.
Edited in config/patches/xterm-src/ and regenerated through the harness, so
the emitted patch and the lockfile hash are derived, not hand-written.
The commit-overlap pin's swallow arm now asserts the repaired contract —
the value its own comment already named as correct and as what stock
beta.287 emits. #11504's arm at :184 flips too; it never covered that
report, as its own prior note recorded, and the reporter's +149ms arm is
untouched and still asserting the defect. Provenance hashes in three test
docblocks are updated, since regenerating changes the patch hash and with it
the resolved install directory.
* docs(terminal): re-measure the #6905 mutation citations against the new bundle
Regenerating the patch moved the resolved install, so this docblock's
patch_hash, line count, two line numbers and three mutation outcomes all
described a bundle that no longer exists. The deferred branch is one the
fix writes into, so the outcomes could not be re-pointed on reasoning.
Line numbers read off both files by diffing anchors rather than derived by
arithmetic: 201 to 205, 159 to 163. Outcomes re-run through the retained
rig, which re-resolves through the module loader and re-derives each arm
from a unique minified anchor: pristine 3 passed, m1 3 failed, m2 2 failed,
m3 3 passed — identical to the old bundle. Guard controls in both
directions exit 1, so the counts are falsifiable.
Comment-only; the assertions and expectations are unchanged.
* fix(xterm): size the preedit overlay to the cells its text will occupy
updateCompositionElements computed the overlay's left edge from the grid
but never its width, so the preedit rendered at the font's natural advance
while the committed text takes two cells per wide glyph. Measured in
Chromium 150: 가나다라 drew 48.45px as a preedit and 69.20px once committed
— the same characters, same font, 30% narrower, and drifting further with
each syllable. Every macOS mono font carrying Hangul measured 0.49–0.72 of
two cells; never 1.0.
Deriving the width from wcwidth and the cell measure moves Korean, Japanese
and Chinese to 1.000 and leaves ASCII at 1.000, which it already was:
한 12.125 -> 17.297 (17.30 expected)
가나다라 48.453 -> 69.188 (69.20)
안녕하세요 60.563 -> 86.500 (86.50)
日本語 42.000 -> 51.906 (51.90)
abcdefgh 69.234 -> 69.203 (69.20, unchanged)
Edited in config/patches/xterm-src/ and regenerated through the harness, so
the emitted patch and lockfile hash are derived rather than hand-written.
The unit test asserts the arithmetic, which is what CI can run. The pixel
consequence was measured on macOS with SF Mono in an Electron harness, not
on the Windows font stack STA-3232 reports from — so this demonstrates the
mechanism and does not stand as that row's platform evidence.
* test(e2e): pin the macOS Korean preedit as visible only while composing
#11914 reports the composing text invisible until Space. Its c3 was recorded
as unobtainable, and the reason on file was wrong: the boundary IS
assertable, but not in happy-dom, which reports display:block in BOTH the
active and inactive states and zeros for every rect. A test there passes
with the defect present.
Captured on real hardware instead: hidden and 0x0 before, .active with
display:block, a 15.84x16 rect and checkVisibility() true while composing
그, hidden again after. 39 DOM events, 2 composition starts, onData
["한","그","\r"].
Two mechanism findings are carried in the setup because both are invisible
in the result and fatal if removed. The input source must be selected AFTER
the app takes focus — focusing resets it to ABC. And the IME must be warmed
until an observed keyCode 229; typed cold it emits raw QWERTY (g k s r m)
with no composition at all, which is indistinguishable from an IME that is
not installed. Two runs were voided on exactly that signature before the
warm-up was found.
The has229 and compositionStarts assertions exist to make such a run fail
loudly rather than pass as a clean negative.
Gated on darwin plus ORCA_E2E_NATIVE_MACOS_KOREAN, like its siblings. The
final spec form has not itself been executed — the machine became
unavailable — so it carries the probe's measured values as literals rather
than a run of its own.
* docs(e2e): correct 19a8d133db7 — the Korean preedit spec has been executed
That commit said the landed form had never run and carried the probe's
values as literals. It has now run on real hardware: 1 passed, 9.1s, rc=0,
with the capture and log sealed under a verified hash manifest.
The teeth check was also run rather than reasoned about, and it changes
which assertion matters. Forcing the active overlay to max-width:0 with
overflow:hidden — invisible on screen — leaves the active class, the
textContent, display:block AND checkVisibility() all passing. Only
during.rect.width fails. checkVisibility() is not sufficient against this
defect; the bounding rect is the single load-bearing assertion, which the
docblock already said and this run confirms.
An earlier teeth attempt injected the CSS mid-run and tripped the
hasActiveClass poll instead, failing at the wrong assertion. It is
inconclusive and excluded from the seal rather than counted.
* test(terminal): add #12171's ordinary-English arm from a real Windows capture
c4 was recorded as unmet and the ledger sourced its control to
evidence/windows-current/, which holds 12 captures and not one English one.
The arm here comes from windows-9803-final instead — same probe, same host
geometry, same injector, en-US with no IME, replayed keydown for keydown.
Two limits are stated in the file rather than left for a reader to find. It
is a different bundle and a different run about 3.6 hours later, so it is
not a same-run arm. And it is #9803's range-active MUTANT arm: ordinary
English stays byte-exact even with that saved-range mutation live, which is
why it reads as a negative rather than as a baseline.
Bundle cited by directory with its file SHA-256; MANIFEST.sha256 verifies
21/21, rc=0. Nothing imported from .tmp.
* docs(terminal): correct #12164's grounds — the cited comments say no such thing
The rejection of #12164 from this file's family was recorded as resting on its
comment 1 (output doubling) and comment 2 (filed against 1.4.163). Checked
against the API: the issue has exactly two comments, neither of which says
that, and the string 1.4.163 appears nowhere in the thread.
The conclusion survives on better grounds. The issue BODY's repro is "Run any
CLI agent (Codex, AGY, Claude, etc.) that outputs Korean text into the Orca
terminal" — untyped output, no keystrokes, no composition — so excluding
CompositionHelper is right, and the input-path hunt was looking in the wrong
place. The body is also LLM-authored (it still contains a literal
"## 5. GitHub Submission Draft (Ready to Post)") and its Root Cause section
blames a CJK IME preedit buffer its own repro never engages, so it should not
be read as observation.
Comment-only; suite unchanged at 5/5.
* test(native-chat): make composition frames carry isComposing, not just inputType
This suite's comment claimed "Gating onChange on `isComposing` breaks here."
It did not. composeFrame() fired `input` with `inputType` but never set
`isComposing`, so a gate on `isComposing` passed all four tests untouched —
the suite asserted a discriminator it did not exercise.
Composition frames now carry both, so neither gate is exempt. Verified by
pointing the mutant at it: with an `isComposing` gate on the composer's
onChange, this suite now fails 3 of 4 (it passed 4 of 4 before), and the
ordinary-English arm correctly survives, since a composition gate should not
touch it. Production code is unchanged and stays gate-free; the mutation was
applied, measured, and reverted.
Found while excluding NativeChatView's question-card remount as the owner of
#12118 / STA-3219: the remount is real, but the preedit survives it precisely
because this write path has no composition gate.
* fix(mobile): ship the patched xterm build, matching desktop
mobile pinned @xterm/xterm 6.1.0-beta.285 while the patch is keyed to
6.1.0-beta.287, so mobile shipped stock xterm and neither IME defect fix
reached it: the swallowed keystroke after an IME commit (9506039de72) and
the preedit sized to the font rather than the grid (e04e0c88da5).
Bumps the three xterm packages to the desktop versions and adds the patch
to mobile's own pnpm.patchedDependencies. No copy of the patch: pnpm
accepts the parent-relative path and records it in the lockfile against
hash 8d63166272e9040a…, byte-identical to what desktop resolves, so the
two stay in step by construction rather than by a drift check.
The workspace separation is untouched — root pnpm-workspace.yaml still
declares `packages: []` and mobile keeps its own lockfile, which is what
keeps the root's patches from failing as ERR_PNPM_UNUSED_PATCH.
Verified in the generated webview bundle rather than at the install:
alignPreeditToGrid 0->2, sentComposition 0->3, pendingInput 0->11, and the
stock-only _handleAnyTextareaChanges 2->0 and dataAlreadySent 4->0. pnpm
applies patches during linking before postinstall regenerates the bundle,
confirmed by a revert/reinstall/re-apply cycle in both directions.
Mobile suite 2971 passed, 3 skipped — identical before and after. Bundle
+1,514 B (+0.24%). Lockfile churn is xterm-only; --frozen-lockfile passes.
mobile/src/ime/ime-submit-carry.ts is NOT made redundant and is untouched:
it handles iOS firing onSubmitEditing on a React Native native TextInput
after unmarking a composition, which is outside the WebView entirely.
Known divergence left alone: desktop also patches @xterm/addon-webgl and
mobile now runs that version unpatched. That patch is glyph/texture-atlas
rendering with nothing IME-related, so it affects neither fix.
* test(terminal): pin the preedit overlay against already-committed cells
STA-3132 (arm A), STA-3170 and STA-3232 report a Korean preedit painted on
top of text already on screen. Builds v1.4.163-v1.4.166 cancel the pending
finalizer in compositionstart, so a committed syllable reaches onData one
syllable late and buffer.x is stale — the overlay lands on the cell the
flushed syllable is about to occupy.
Replays a recorded hardware trace rather than an authored one: the ordered
DOM event stream captured on Windows + MS Korean (wave5-r2 evidence, 64
events), echoing onData back as PTY output.
The load-bearing assertion is deliberately not the obvious one. Comparing
overlay style.left against cursorX is tautological — left is computed from
buffer.x. This counts committed syllables from the compositionend events
the IME fired, so the two sides are independently derived.
Discriminated by a historical re-add across seven real bundles, since the
owner is deletion-shaped: pristine beta287, v1.4.155 and v1.4.162 pass;
v1.4.163 fails; v1.4.163 with that single call removed passes; the byte
identical baseline restored fails again; head passes. Every failing arm
fails only this case — the ordinary negative stays green in all seven.
The negative asserts its own category rather than claiming it: zero
composition events, zero isComposing, zero keyCode 229, exactly 16 events,
paired against the Korean arm's 4 starts / 3 ends / 11 updates / 64 events.
Scope: cell indices, not pixels. happy-dom has no layout, so the recorded
8x16 cell metrics are supplied to the render service. This makes no claim
about pixels visually overlapping; that is affected-OS confirmation and
stays open. Covers the overlap arm only — STA-3132's auto-line-break arm
and STA-3232's half-line-capacity and a11y arms are untouched.
* test(e2e): matrix macOS period substitution against the OS preference
#11504 reports macOS inserting ". " after a Hangul Space commit. This
sweeps six arms across both states of NSAutomaticPeriodSubstitutionEnabled,
reading the preference live per run rather than asserting a literal.
Two results worth having on record.
The reporter's stated trigger did not reproduce. Their words are "There is
no second press at all. One space is enough", but korean-single-space emits
zero insertText with the preference on or off. So does word-space-word-space.
Their timing does reproduce, with different content. korean-double-space and
korean-longer-word-double-space emit a delayed insertText at +122.5-122.7ms
after compositionend — squarely the reported +149ms — but the payload is a
space, never ". ". Consistent with the double-space rule seeing two slots
under ABC and only one under Korean, where the IME commit consumes the first.
The substitution itself is real and preference-bound: latin-double-space
gives "ab . " with the preference on and "ab " with it off, on one build
with the preference as the sole variable, reproduced across two runs.
That also refutes a claim in PR #11506, which states the substitution "is
enforced outside the renderer and never reproduces in dev builds, so changes
here must be verified against a packaged app". It reproduced in the dev build
twice and did not reproduce on the signed packaged app. That claim should not
be used as a verification gate.
Gated @headful behind ORCA_E2E_NATIVE_MACOS_PERIOD, same shape as the Korean
preedit spec, so it does not run in ordinary CI. Evidence is onData and DOM
only — the PTY-child reader aborted and no packaged-app arm was stable.
* test(terminal): actually enforce the recorded jamo progression
The preedit assertion compared sample.overlayText against sample.overlayText
— the same expression on both sides. A lane proved it by mutation: corrupting
seven of the eight recorded preedit values left the suite fully green. So the
docblock's ㄱ→가→간→나→낟→다→달→라, which the matrix also cites as this row's
recorded shape, was cited and unenforced.
The first attempt at a fix was insufficient and is worth recording. Threading
stroke.preedit through to the expectation still passed on a corrupted fixture,
because that value both drives the rig and was the expectation — corrupting it
moved both sides together. Same tautology, one level down.
The expectation is now an independent literal. Verified by mutation rather
than by reading: corrupting two recorded values fails one arm; restoring them
passes 3/3.
overlayCell was never affected — it is compared against a count derived from
the compositionend events, not from the buffer, and remains the load-bearing
assertion for the overlap.
* fix(e2e): select the selectable input source, not the first match
TISCreateInputSourceList can return several entries for one input source
id. A third-party IME publishes a non-selectable parent alongside the
selectable mode, and taking sources.first can return the parent — after
which TISSelectInputSource fails with paramErr (-50) while the caller
reports success from the enable step.
Found with Qingg (com.aodaren.inputmethod.Qingg), which exposes exactly
that pair under one id. Its mode id equals the bundle id, so filtering by
name would not have helped; selectability is the discriminator.
Now filters on kTISPropertyInputSourceIsSelectCapable and falls back to
the old behaviour when nothing advertises it, so single-entry sources are
unaffected. Also enables every entry for the id rather than only the one
being selected: selecting a mode whose parent is still disabled fails the
same way.
Compile-checked, and selecting com.apple.keylayout.ABC still exits 0.
Unrelated to the enable path: on macOS 26.5.2 third-party IMEs are gated
behind a consent sheet in System Settings. TISEnableInputSource returns
noErr immediately regardless, and the enable only lands if that sheet is
answered while the requesting process is still alive.
* test(terminal): discriminate #12171 against the real shortcut policy
The prior candidate mutation for this row was correctly refused: its suite
mocked shortcut policy so Process/229 became actionable, while the real
resolveTerminalShortcutAction has no Process branch — so the kill measured
the mock. This does not mock it.
Replays a capture of this row's own gesture (d, l, Shift+T, e, k, Space,
Enter under MS Korean, committing 있다) taken on Orca 1.4.164, through the
real useTerminalKeyboardShortcuts hook, capturing bytes at terminal.input.
The earlier capture could not discriminate at all because it recorded no
shiftKey; this one records it on 10 of 10 keydowns with code populated.
One physical Shift+T produces two shifted Process/229 keydowns. Under the
pre-#12265 classifier each synthesizes {key:'Enter', shiftKey:true}, which
the real policy resolves to sendInput '\x1b\r' — twice, giving 1b0d1b0d,
the two escapes the known-bad ed96881b0d1b0d contains.
Mutation is the retained pre-12265-process-shift.patch applied to HEAD, not
an authored one: patch -p1 applies clean and diffs identical to the mutant
copy. Arms are copies; shared source hashes the same before and after.
The English arm stays clean under both modules, so the mutation
discriminates by language rather than by harness — and a real Shift+Enter
through the same rig yields exactly ['\x1b\r'] in every arm, so a silent
pristine result means the code is quiet rather than the harness dead.
Scope: 1b0d1b0d is measured at the renderer boundary. The capture recorded
no PTY bytes — window.api.pty is frozen on shipped builds and the onData
channel needs a build-time flag — so this shows the renderer producing the
two escapes that payload contains, not a re-observation of the payload.
* docs(native-chat): narrow this file's disclaimer to what is now true
It said "THIS OWNS NO REPORTED ROW". Half of that is stale: the remount site
is now the attributed owner of #12118 and STA-3219. On real Windows TSF the
questionActive swap aborts a live composition — the old node gets only a
blur and no compositionend, the text returns as committed, and the next jamo
yields 아ㄴ rather than 안.
The other half holds. This file pins the opposite property, that the text
survives, which is the half those reporters already agree with. Mutation
shows the gap rather than asserting it: deleting the unmount entirely leaves
three of four tests green, because every substantive assertion is
after.value === … and a composer that never unmounts keeps its value.
Also records why the abort cannot be asserted here. The DOM exposes no
observable separating committed text from a live preedit — value is the same
string either way, there is no EditContext, and the only composing-ness
state is a per-instance ref discarded with the node. A test pinning "no
compositionend fires" would be an anti-guard: red the day it is fixed.
The cadence objection is kept, since it is now the open question rather than
the reason for exclusion.
* refactor(terminal): drop the unread isComposing field from XtermBypassEvent
Added by #6396 for terminal IME candidate handling that this branch has since
removed. No production or test code reads it, and the policy is safe without it:
during composition `key` is 'Process', so the non-ASCII printable checks that
would care never match.
Co-authored-by: Orca <help@stably.ai>
* fix(native-chat): keep the composer mounted through an in-flight IME composition
A question card replaced the composer outright
(`{questionActive ? null : <NativeChatComposer/>}`). Unmounting the field
mid-composition aborts the composition in the OS: the node is detached before
`compositionend` can fire, the preedit returns as committed text, and a resumed
Hangul syllable degrades — 아 then ㄴ yields `아ㄴ`, never `안`.
Confirmed in rasterised pixels on Windows with a real MS Korean IME, at both
v1.4.171 and the reporter-era v1.4.164 (the swap block is byte-identical
across them): the preedit underline present before the swap, the composer
visibly absent during it, and the same glyph back afterwards WITHOUT the
underline — committed, not composing.
The swap is now deferred while a composition is in flight, which is what
editors that survive IME do: ProseMirror gates DOM work on `view.composing`,
CodeMirror protects the composing subtree from redraws. Hiding instead of
unmounting does not work — `display:none` and `visibility:hidden` both blur the
focused element and abort the composition the same way.
The hold releases on `compositionend`, which browsers also fire on blur, so
clicking into the card's own answer input yields the input region immediately;
with nothing composing the card still replaces the composer at once, so no
stray "Send a message" appears beside a question.
The existing characterization test flips to a regression guard: it pinned the
node being destroyed, which was the defect. Node identity is the load-bearing
assertion — value-only checks are trivially satisfied by a composer that never
unmounts and cannot tell a held composition from a destroyed one.
The typing-redirect handler moves to its own hook. That is not cosmetic: both
touched files sat at the 400-line cap, and `max-lines` suppressions are
forbidden, so the room had to come from a real extraction.
* fix(macos): opt Orca out of AppKit automatic period substitution
macOS "Add period with double-space" (`NSAutomaticPeriodSubstitutionEnabled`,
on by default) is applied by AppKit's text input system. Native terminals never
join that system; Chromium text fields do, so xterm's helper textarea inherits
it and a double space arrives as `". "` — a period nobody typed, handed straight
to the PTY (#11504).
Chromium answers AppKit for quote and dash substitution and defaults both off,
but declares no period accessor at all, so AppKit applies that one without
asking. This user default is the only lever: there is no per-field or
per-webContents opt-out to prefer over it. Writing the key into Orca's own
defaults domain overrides the global value for this app alone and leaves the
user's system-wide setting untouched. It necessarily covers every Orca text
field, not only terminals — AppKit offers no narrower scope, and that tradeoff
is deliberate rather than accidental.
Measured on the reporter's own build v1.4.161: with the preference ON, typing
a,b,space,space yields `onData ["a","b"," ",". "]`; with it OFF the same arm
yields two spaces.
Note the issue's causal model is wrong and this fix does not follow it. It
claims the substitution only fires with a CJK input source and never with ABC.
The measurement is the inverse — every Korean arm is clean and the ABC arm is
the one that fires — so the fix is not conditioned on input source.
NOT YET VERIFIED ON HARDWARE. The unit tests inject the writer, so they prove
the call is made on darwin and skipped elsewhere; they do not prove AppKit
honours an app-domain override for this key. That check is outstanding.
* fix(xterm): keep a live composition across a lone Cmd press on macOS
CompositionHelper.keydown exempted keyCodes 16/17/18 from tearing a composition
down, which covers Shift/Ctrl/Alt but not macOS Meta — 91/93 in Chromium, 224 in
Firefox. A lone Cmd press mid-composition therefore reached
_finalizeComposition(false), which dropped the preedit overlay's `active` class
and committed the live syllable early. macOS keeps the marked text alive across
that press, so no later compositionstart re-arms the overlay and the rest of the
word composes invisibly.
Measured on hardware (m4air, macOS 26.5.2, Apple M4, 2-Set Korean) with the Cmd
posted as a CGEventType.flagsChanged, which is what a physical modifier emits.
AppleScript `key code 55` posts nothing a browser can see — a bare `key code 56`
for Shift is equally silent — which is why no capture in the corpus ever reached
this branch. Three arms, same build otherwise: overlay live throughout with the
exemption, dark and prematurely committed without it, live again with it
restored. Evidence under
.tmp/ime-handoff/swarm-scratch/wave31-cmd-preedit/evidence/.
The fix cannot widen past a lone modifier: only a standalone press reports these
keyCodes, and a Cmd chord during composition is reported by Chromium as 229,
which was already exempt. Cmd+A still ends the composition, via the IME's own
compositionend. Ghostty draws the same line, returning early from flagsChanged
under hasMarkedText() for every modifier including Super.
Orca's terminal pane was never affected — shouldSuppressTerminalModifierKeyboardEvent
drops a standalone Meta keydown before xterm sees it, and deleting only 'Meta'
from that set is what flipped the hardware arm to broken. The popout preview
terminal and mobile's webview install no such guard and did reach the teardown.
terminal-ime-xterm-composition-commit-overlap.test.ts asked its fixer to update
the two Cmd arms to the values it named as correct; both now emit a single ['한'].
* test(native-chat): drop two byte-identical duplicate cases
`4632b86919d` copy-pasted two cases twice into the same describe block:
`retains carry across a same-frame non-Enter keyup before redispatch` and
`expires carry before a deliberate Enter after the next frame`. Each pair is
byte-identical — same title, same body — so the copies asserted nothing the
originals did not.
This is what has been failing `static analysis` on this branch since 2026-08-06:
`oxlint vitest(no-identical-title)` reports both under `--deny-warnings`, and
`verify` fails solely because it requires static analysis to pass. Every other
gate in `verify` was already green, including typecheck, xterm patch sync, the
full test shard set, and both package jobs.
12 cases still pass in the file.
* test(e2e): skip the WebGL arm when no WebGL renderer exists
The #12164 probe runs two arms, webgl and dom, and closes by asserting the
active renderer is the requested one. That assertion is right for the dom arm —
it is what proves the pane actually left WebGL, without which the arm is
meaningless — but headless CI has no GPU, xterm falls back to DOM silently, and
the webgl arm then fails.
The failure reads as a Korean rendering defect and is not one, so the webgl arm
now skips with the active renderer named. The dom arm keeps the assertion
unchanged.
This is the third of three checks that have been red on this branch since
2026-08-06. `static analysis` and `verify` were fixed in c51c6b5837e; the CI log
shows this job as 1 failed / 1 passed, the pass being the dom arm.
* chore(lint): drop five unused no-console disable directives
`check-changed-code-quality` reports unused eslint-disable directives as errors,
and these five sat above diagnostic `console.log` calls in IME test and spec
files where `no-console` is not enabled — so each suppressed nothing.
This is the second of the two static-analysis steps. `c51c6b5837e` fixed
"Enforce focused code-quality plugins" (duplicate test titles); this fixes
"Enforce changed-code quality". Both had been red on this branch since
2026-08-06, and I mistook the first for the whole job.
The diagnostic logs themselves are kept — they are what a failing IME arm prints
for a reader to inspect.
* test(e2e): cover #12164 under fractional device scale factor
Fractional display scaling was #12164's last unexplored branch, and the reason
is worth recording: earlier attempts were BLOCKED, correctly, because they
proposed mutating the Windows display scale on a remote physical machine with no
console recovery. `--force-device-scale-factor` reaches the same renderer state
per process, so nothing outside the Electron instance changes and there is
nothing to restore.
The hypothesis was specific: `프프로로젝젝트트` is what a half-pixel cell boundary
could produce on a 2-column glyph, and nothing else in the suite varies dpr.
Measured at 1.25 and 1.5, both under WebGL: ink extents 25/21/16 with identical
ink groups, matching the scale-1 run. No doubling.
The arm self-certifies before asserting — if the flag does not take, the test
fails rather than silently measuring at dpr 1. That matters here: the sibling
spec's WebGL arm went two days reporting a missing GPU as a Korean rendering
defect precisely because a silent fallback looked like a result.
* fix(terminal): match Mod+letter shortcuts by physical key, not IME-rewritten key
With a CJK input source active, macOS and Windows report the physical key through
`code` but rewrite `key` to the layout's character: Korean 2-Set turns Cmd+C into
`{ key: "ㅊ", code: "KeyC", metaKey: true }`. Every `key.toLowerCase() === 'c'`
match misses it, so the shortcut is not recognised and xterm encodes the chord as
PTY input instead — issue #13033 reports `ESC[12618;9u` and a terminal that jumps
to the bottom, because user input scrolls the viewport.
This is the same key-vs-code confusion that owned #12171, where a `Shift+T`
typing ㅆ was read as Enter for want of a `code` guard, so the fix is the same
shape: trust `code` when it is present, fall back to `key` and then the legacy
`keyCode` when it is not (Chromium omits `code` on synthetic and some keypress
events, and `keyCode` keeps its US value even when `key` is rewritten).
Applied to the four terminal-side sites, including the dashboard pop-out, which
#13033 called out specifically as having its own key handler:
pty-connection.ts Cmd/Ctrl+C copy guard
keyboard-handlers.ts Cmd+G search navigation
agent-interrupt-inference.ts interrupt inference
preview-terminal-key-handler.ts pop-out paste
Nine further `key.toLowerCase()` letter matches exist outside the terminal
(TaskPage, editor, GitHub composer, browser markup). They have the same defect
and are deliberately left for a separate change rather than widening this one.
An existing case, `matchSearchNavigate > returns null for wrong key`, overrode
only `key` and left `code: 'KeyG'`, so it began passing for the wrong reason. It
now overrides both — which is what "wrong key" means once matching is physical —
and a companion case pins the Korean-rewritten chord still matching.
#13033 was closed NOT_PLANNED; the reporter's event shapes drive the new test.
* fix(renderer): match every Mod+letter shortcut by physical key, not IME-rewritten key
Completes the previous commit. A CJK input source rewrites `event.key` while
`event.code` keeps the physical key, so `key.toLowerCase() === 'z'` and friends
silently stop matching — the shortcut is not recognised and the keystroke falls
through to whatever handles unclaimed input.
The helper moves to `@/lib/ime-latin-shortcut-key` first: it now serves the
editor, GitHub composer and browser markup, and importing terminal-pane
internals into those would be the wrong direction. `lib/` already hosts
`ime-composition-keyboard-event` for the same reason.
Nine remaining sites, all previously unreachable under Korean/Japanese/Chinese/
Vietnamese input:
TaskPage, ActivityPrototypePage, ProjectViewWrapper Cmd+F search
useMarkupKeyboardShortcuts Cmd+Z undo
GitHubMarkdownComposer, RichMarkdownLinkBubble,
rich-markdown-link-shortcut Cmd+K link
native-chat-shortcut Cmd+J
rich-markdown-key-handler Cmd+Shift+X
Six of the nine test `!== 'letter'` as early-return guards and three test
`=== 'letter'`; the negation is applied per site, since a blind substitution
would have inverted six of them.
Full suite: 4249 files pass. Three files fail locally and none is caused by this
change — the branch touches no file under `src/main/` or `src/relay/`, all four
failures reproduce on an unmodified tree or pass in isolation (the worktree
poller passes 21/21 alone, so it is full-suite parallelism), and all 16 CI test
shards are green.
* docs(ime): scope the IME composition rules to the terminal-pane directory
#11893 proposed adding these to the root `AGENTS.md`, which every agent loads on
every task regardless of what it is doing. They only bind keyboard handling, the
composer and the terminal input path, so they belong next to that code —
`tests/e2e/AGENTS.md` already establishes the nested convention here.
Kept from #11893: range-derived commits, guarding above the key dispatch,
the `attachCustomKeyEventHandler` / `CompositionHelper` interaction, no
normalization at commit, and the recorded-trace evidence bar.
Added from defects found since it was written:
- match shortcuts on `event.code`, not `event.key` (#12171, #13033)
- `keyCode === 229` means an IME owns the press
- do not unmount a field mid-composition, and hiding is not a fix because
`display:none` blurs and aborts it too (#12118, STA-3219, #11332)
The evidence bar now also names the mutation check, since a test that survives
deleting the code it guards is guarding nothing — a failure this effort hit more
than once.
* fix(terminal): gate Ctrl+Enter CSI-u on a negotiated pane, porting #12462
Found while scoping the rebase onto `main`: #12462 landed on 2026-08-06 and
fixes a real defect this branch does not carry. Ctrl+Enter emitted
`\x1b[13;5u` unconditionally, so a pane that never negotiated the kitty
keyboard protocol — local Windows ConPTY, plain shell — printed the escape
verbatim into the prompt.
This branch deletes `terminal-ime-deferred-newline.ts`, which is one of the
files #12462 touched, so a rebase resolving those conflicts by taking our side
wholesale would silently reintroduce the defect. Porting it forward now means
the fix survives the rebase however the conflicts are resolved.
Mirrors the Shift+Enter guard already here: local ConPTY falls back to the
legacy CR every emulator sends, and a negotiated pane keeps the chord, so the
fallback is scoped to panes that cannot receive CSI-u rather than to Windows.
NARROWER THAN #12462 BY ONE CONDITION, deliberately. `main` also allows CSI-u
via `hasCtrlEnterCsiUAuthority()` (trusted consumer evidence, #12329); that
helper and its plumbing do not exist on this branch. Omitting it is the
conservative direction — an authorised pane gets `\r` instead of the chord,
rather than an unnegotiated pane printing an escape — but it should be restored
when the two histories are reconciled.
Test covers both directions and is mutation-checked: forcing the gate open
fails it, so it cannot pass by construction.
* fix: reconcile two more fixtures main moved while the stack waited
Both caught by CI, not locally, and the reason the local run missed one is
worth recording:
1. `browser-toolbar-profile-dialogs.ime-enter.test.tsx` did not pass
`useNativeUserAgent` / `onUseNativeUserAgentChange`, which `main` added to
`BrowserToolbarProfileDialogsProps`.
Local `pnpm typecheck` reported 0 errors on the same commit CI failed. The
cause was a stale `config/*.tsbuildinfo` — tsc reused an incremental cache
from before the merge. Deleting it reproduced CI's error exactly. Any
"typecheck clean" during this merge should be treated as unverified unless
the cache was cleared first.
2. Localization keys for `SshDisconnectedDialog` were absent from `en.json`:
the merge took this branch's component alongside `main`'s catalog.
Regenerated with `pnpm run sync:localization-catalog` rather than hand-added.
* fix(mobile): regenerate the lockfile the merge resolved by taking one side
CI's `verify` failed with `ERR_PNPM_OUTDATED_LOCKFILE` on `mermaid (lockfile:
11.16.0, manifest: 11.16.1)`. The mismatch was in `mobile/`, not the root — the
root lockfile was consistent throughout, which is why inspecting it (and even
GitHub's merge ref) found nothing wrong.
Cause: during the merge I resolved `mobile/pnpm-lock.yaml` by taking this
branch's side wholesale rather than merging it, so it kept `mermaid 11.16.0`
while `mobile/package.json` came from `main` at `11.16.1`. Taking one side of a
lockfile is only safe when the corresponding manifest comes from the same side.
Regenerated with `pnpm install --lockfile-only`; `--frozen-lockfile` now passes
in `mobile/`. Verified the xterm patch entry survives intact — same hash
`4f1b42d268f3964d…` and the parent-relative path into `config/patches/`, which
is the desktop/mobile coupling that would silently break the mobile build.
Two earlier diagnoses of this failure were wrong and are worth recording: it was
not the root lockfile, and it was not a stale merge ref (a rerun reproduced it
exactly).
---------
Co-authored-by: Orca <help@stably.ai>
* Bump mobile app.json to 0.0.42
* fix(mobile-ios): stop TestFlight CI from waiting on ASC processing
0.0.42 builds 1–2 uploaded successfully then hung for hours polling
processing with no Ready build and no Apple email. Exit after upload
and cap the job at 90m so the next cut does not repeat that hang.
* fix(mobile-ios): fully skip Pilot wait (no changelog)
Pilot only returns immediately after upload when changelog is nil;
passing notes re-enters the ASC build-list poll.
* fix(terminal): fence daemon endpoint ownership
* fix(terminal): clean failed daemon PID claims
* fix(terminal): close daemon ownership review gaps
* test(daemon): release startup IPC in boot smoke
* test(daemon): mirror production stdio in boot smoke
* fix(daemon): exit after rpc shutdown cleanup
* fix(terminal): make the socket name the daemon endpoint authority
The reported failure was a live daemon hosting PTYs that nothing could
reach: terminals acknowledged input and never ran it, listings diverged
from reality, and restarting the app never helped because the detached
helper survived. The ownership fence added for it could not fire in the
sequence that produces the split brain.
libuv unlinks the pathname a server bound to when that server closes,
with no ownership check. A daemon that lost its endpoint name therefore
deleted whichever socket then sat at that path — including a live
replacement's — stranding a daemon that still hosted every session.
Bind a private same-directory name and hard-link it into place instead:
libuv can only ever unlink our own bind name, the exclusive link is a
kernel-enforced endpoint claim, and the canonical name is removed only
under an inode ownership check. The bind name replaces the basename
rather than extending it, so it cannot overflow sun_path.
killStaleDaemon removed the PID record unconditionally immediately
before every fork, so the exclusive PID claim was always uncontested at
bind time. It also unlinked a live daemon's endpoint whenever a connect
probe merely timed out, and treated a `ps` timeout as proof of PID
recycling. Now only positive evidence of a dead endpoint authorizes
reclaiming it, SIGKILL is confirmed rather than assumed, and a daemon
that cannot be proven stopped keeps its record and endpoint while the
launcher refuses to fork beside it.
A daemon whose endpoint was taken over now retires itself, draining
rather than killing, so an unreachable orphan stops being permanent.
A repaired PID record re-derives entryPath, appVersion and the Linux
incarnation markers from the authenticated owner instead of dropping
them; without appVersion a healthy daemon read as a permanently stale
bundle and, on Windows, went unpinned against daemon-host pruning.
Repair failure now fails open — abandoning a healthy daemon over a pid
file write cost every persistent terminal on the machine.
Also: treat only ENOENT as an unclaimed record so a Windows file lock is
not reported as an ownership conflict; settle start() before close() so
an accepted connection cannot defer it forever; sweep abandoned claim
and bind names; and type the endpoint-identity seam so a rename cannot
silently disable the fence.
Adds a real-process handover smoke that reproduces the failure with two
daemons racing one endpoint, and wires it into the native-smoke job.
* fix(daemon): retire only on proven endpoint ownership loss
The ownership watchdog read a null identity for any stat failure, so a
transient EACCES or EIO on the runtime directory would retire a daemon
that was still serving every terminal on the machine. Distinguish "the
entry is gone" from "the probe failed" and act only on the former.
Also require the loss to persist across two polls: a replacement
publishes by unlink-then-link, and a single observation can land in that
gap.
* fix(daemon): source repaired ownership metadata from the authenticated hello
Adversarial review found three defects in the previous two commits.
Re-deriving entryPath from the owner's command line truncated it at the
first space. A command line is a single space-joined string, so
`C:\Program Files\Orca\...` and `/Applications/Orca 2.app/...` came back
as `"C:\Program` and `/Applications/Orca`. getDaemonLaunchIdentity treats
a present entryPath as authoritative, so a healthy daemon read as
`different_app_path` and was killed and re-forked — worse than the
missing-metadata case the derivation was added to fix. Carry entryPath
and appVersion as optional fields on the daemon hello identity instead:
the daemon already has both from its own argv, and per
docs/reference/remote-wire-compatibility.md a new optional field is safe
because every reader falls back when it is absent. This also removes a
synchronous `ps` spawn from the Electron main thread during startup.
`start()` rolled back the PID record even when it never published one.
Losing the endpoint link now runs that path, and the ownership-checked
unlink briefly renames the incumbent's record aside — enough to strand a
live daemon's ownership. Roll back only what we actually wrote.
publishDaemonSocketPath read its identity from the canonical name after
linking, so a concurrent unlink returned null: no ownership watchdog and
no endpoint cleanup on any shutdown path. Read it from the bound name
before linking, which shares the inode.
Refusing to fork beside an unconfirmed daemon left the user with no
daemon at all and no in-app recovery, since restart re-entered the same
fence. We have just proved something answers the endpoint, so adopt it
in degraded mode: live sessions keep working, fresh terminals run
locally. SIGTERM is also individually guarded now — an EPERM fell into
the blanket catch and reported "nothing alive", authorizing the very
duplicate this fence exists to prevent.
Also reset the ownership-loss streak on an inconclusive probe so the
confirmations are consecutive, and sweep scratch names before the launch
so a failed launch still reclaims them.
Mixed versions are the normal state of the remote-server feature: users update clients and servers independently. Until now nothing tested that. Every cross-version claim was made by code reading plus unit tests with hand-written old/new shapes — enough to catch design problems, not enough to catch a real skew regression.
This runs the REAL protocol implementations from two builds against each other in one process: the actual host methods and RPC dispatcher on one side, the actual renderer multiplexer on the other, with a transport that reproduces the production asymmetry — each side decodes with its OWN codec and drops frames whose opcode it does not know. A frame survives only if the RECEIVING build understands it, which is what makes this level sufficient without launching two apps. The old side is a genuine checkout extracted from the release tag; the extracted client was confirmed to lack a symbol that exists only on main.
Journey: subscribe, first snapshot, input reaching the process, live output, hide/reveal snapshot, transport drop, resubscribe, input landing again — across old->new, new->old, and a current/current control. Every step ends on an observed-state barrier; no sleeps. The oracle asserts the recorded step list, the exact 16-frame named sequence, negotiated capabilities, the exact input the host wrote to the PTY, rendered content, and zero decoder-rejected frames. A host method the stub lacks is recorded by name and asserted empty, so a harness gap cannot masquerade as a wire break.
Detection is proven per violation shape, and it attributes each to the correct side: an unnegotiated opcode goes red only where a decoder would reject it, a removed published field goes red only where an old client consumes it, and a legal additive field stays green in all three pairings so the harness will not cry wolf on safe changes.
It also documents the three compatibility rules in docs/reference/remote-wire-compatibility.md, linked from AGENTS.md, since they previously existed only as folklore — notably that "decoders reject unknown opcodes" is true for the desktop decoder but NOT for mobile, which silently drops them.
Deliberately scoped: terminal stream only. The session-tab sync channel is not covered, nor agent-session publications, file/Git RPCs, mobile E2EE framing, or the relay transport. Two version points, so a regression introduced and reverted between them is invisible.
CI selection was verified rather than assumed — `vitest list` confirms 0 matches under the shard's exclude and 4 under the dedicated job — because a lane silently running zero tests is precisely how a host-side defect escaped CI earlier in this series. Closes STA-3469.
The number answers "which build of 1.4.163 is this", so carrying it across
versions made it meaningless — 1.4.164 opened at 38 for no reason a reader
could see. It now counts titles matching the base version being built, so a
version bump restarts the series at 01.
Deriving it moves from workflow jq into the script, because the number depends
on the base version and only the script knows which base the published tags
resolved to.
Timestamps go from `07-31 13:54` to `Jul 31, 1:54PM`, still Pacific.
Co-authored-by: Orca <help@stably.ai>
The adhoc workflow checked out any requested ref and ran its scripts and
electron-builder config with MAC_CERTS, the notary password, and the adhoc
publisher token in reach — including refs/pull/* fork code a maintainer
could dispatch in one innocuous-looking click.
Vet the ref before checkout: PR refs are refused, branches/tags resolve in
a bare tree:0 scratch fetch, raw SHAs must be reachable from a repo branch
or tag (a partial clone lazily serves PR-only commits by SHA, so name
resolution alone is not a trust test), and checkout pins the vetted SHA so
a race push cannot swap the commit. Also reference an adhoc-mac-build
environment so the secrets can later be fenced off from stale workflow
copies via repo settings.
main's version only moves on `release:` commits, and stable patches are cut
from release branches that never merge back. On 2026-08-03 main read
1.4.165-rc.0 for twenty hours while 1.4.165, 1.4.166 and 1.4.167 all shipped,
so every hourly built in that window was stamped 1.4.165-hourly.* while
carrying code newer than 1.4.167 — and sorted below the stable its user was
already running.
Resolve the base from the main repo's published tags instead, taking the patch
above the highest shipped stable. package.json stays a floor for the case where
main leads the tags.
Co-authored-by: Orca <help@stably.ai>
* feat(updater): add an adhoc release channel for branch builds
Hourly covers main. This covers everything that is not main yet: a
dispatchable macOS build of an unlanded branch, published to
stablyai/orca-adhoc, so the team can run an experimental feature for a
few days instead of reasoning about it from a diff.
Adhoc sits at the bottom of the version order — 'adhoc' < 'hourly' <
'rc' < stable — so no routine check can walk anyone onto somebody's
branch; only an explicit pinned jump reaches one. It gets its own repo
rather than sharing orca-hourly's, because a branch build must not
appear in the list a developer riding main is looking at.
Signed and notarized exactly like hourly, for the same reason: macOS
anchors a notarized app's TCC grants on identifier + team, so an
unnotarized build reads as a new client and silently loses file access
under Documents/Desktop/Downloads.
Tags stamp to the second rather than the minute. Hourly runs under a
concurrency group and cannot overlap itself; adhoc builds are dispatched
on demand, so two people cutting from different branches inside one
minute is ordinary — and a minute-resolution tag would collide and fail
the second build after its whole pack-and-notarize run.
Channel-specific behaviour now derives from one DEDICATED_REPO_CHANNELS
list: repo mapping, macOS-only support, and UpdateSource. The RPC schema
that validates releaseChannelOverride was a hand-copied enum missing the
new channel, which would have rejected the override on its way to the
main process; it reads the predicate now.
* fix(updater): merge the duplicated shared/types import
Co-authored-by: Orca <help@stably.ai>
* fix(ci): default the adhoc build ref to the dispatch branch
The Actions UI puts its own "Use workflow from" branch picker directly
above the ref field, and picking a branch there is what most people read
as "build this". Making the field optional means the obvious action is
also the correct one; naming a branch explicitly still wins, so main's
copy of the workflow runs rather than a stale one on an old branch.
Co-authored-by: Orca <help@stably.ai>
---------
Co-authored-by: Orca <help@stably.ai>
macOS anchors a notarized Developer ID app's TCC grants on identifier +
team, which is cdhash-independent and so survives an in-place update.
Without a notarization ticket there is no such stable identity, so every
hourly reads as a different client: the grant row stays but stops
matching, and file access under Documents/Desktop/Downloads fails with
EPERM and no re-prompt. `tccutil reset` fixes it until the next build —
and orca-hourly has shipped as many as 14 builds in a day.
Skipping notarization was chosen because Squirrel.Mac validates the
replacement bundle's signature, not its notarization. That is true, but
it is the wrong requirement; the in-place swap was never the problem.
Budgets grow to absorb the notary round trip (publish 2x45, job 150), and
the App token is re-minted after the build so its one-hour life starts at
the first call that uses it rather than during `pnpm install`.
Remove one-off incident docs and committed test-results noise, move
dev/repro/bench tools under tests/tools, and relocate i18next config
into config/ so the GitHub root scrolls to the description faster.
* feat(updater): name hourly releases by version, build number, time, and sha
Hourly releases were titled with their raw tag
(`v1.4.163-hourly.202607312054`), which reads as one opaque digit run and
does not say which commit it came from.
Title them `1.4.163 • 01 • 07-31 13:54 • e698241` instead, and show that
same string in the in-app build picker by having the picker render the
release's stored name rather than deriving its own label. Composing it in
one place means the two surfaces cannot drift.
The build number is monotonic across the channel. It is read as the
highest number already in use rather than as a count of releases: the
prune step trims to 72, so a count would roll backwards after three days
and reissue numbers. Drafts count toward it — unlike in the freshness
check, which asks whether a commit shipped, this asks whether a number is
free, and a stranded draft still holds one.
Times are Pacific while the tag's stamp stays UTC. The stamp is a sort
key and a local one would repeat an hour at every DST fall-back, making
two distinct builds compare equal; the title is only ever read.
* fix(updater): fail the hourly build when the release name is missing
The workflow checks out `ref: main`, but a workflow_dispatch runs the
workflow file from whatever branch was dispatched. A branch that edits
this step while main still carries the old script produces an empty name
and an untitled release — silent, and only visible once someone opens the
releases page. Verified by hitting exactly that on run 30665586904.
Co-authored-by: Orca <help@stably.ai>
---------
Co-authored-by: Orca <help@stably.ai>
* fix(release): restore the SignPath composite action when cutting from an older ref
Co-authored-by: Orca <help@stably.ai>
* fix(startup): record a durable diagnostic before the bootstrap fatal-exit guard exits
Co-authored-by: Orca <help@stably.ai>
* fix(settings): make agent-skill Re-check rescan skill freshness
Co-authored-by: Orca <help@stably.ai>
* fix(startup): keep the bootstrap fatal diagnostic when the log override is unwritable
Create the parent directory an overridden ORCA_BOOTSTRAP_FATAL_LOG names and fall
back to the default location when that path still cannot be opened, so a missing
parent no longer costs the only account of the failure. Also pins the Re-check
freshness rescan to the completed install scan rather than the click.
Co-authored-by: Orca <help@stably.ai>
* refactor(settings): move the post-recheck surface sync out of the panel
Co-authored-by: Orca <help@stably.ai>
* fix(startup): retain diagnostics without node fs
* fix(skills): keep freshness scoped to the local runtime
* fix(settings): register freshness status translations
* fix(settings): scope and sequence skill freshness refreshes
* fix(settings): refresh freshness across runtime transitions
---------
Co-authored-by: Orca <help@stably.ai>
* feat(updater): add hourly dev channel and build switching
Adds an hourly macOS build channel plus a dev-only surface for switching
update channels and jumping to any published build, including older ones.
Hourly builds publish to a separate stablyai/orca-hourly repo. The routine
update path resolves tags from the main repo's releases atom feed, which
exposes only its 10 newest entries — 24 hourly tags a day would evict every
stable/RC entry there and leave real users with nothing to update to.
Hourly artifacts carry the release bundle id and Developer ID signature so
Squirrel.Mac can swap them in place; only notarization is skipped, which
in-place updates never check.
Version tails are stripped to the base (1.4.160-hourly.<stamp>, not
1.4.160-rc.3-hourly.<stamp>) so hourlies sort below both rc.N and stable and
are reachable only by an explicit pinned jump, never by an ordinary check.
The picker is revealed by Option-clicking the Updates header, matching the
Help menu's existing hidden admin affordance. Pinned jumps set allowDowngrade
and release the feed on every settle path so a jump can never leave background
checks permanently deferred.
* chore(hourly): create orca-hourly and add token provisioning script
Adds setup-hourly-release-token.sh, which provisions HOURLY_RELEASE_TOKEN
without the value ever reaching stdout, argv, or shell history: it is read
with `read -rs`, passed to gh through GH_TOKEN in the environment rather than
as an argument (argv is world-readable via ps), piped into `gh secret set` on
stdin, and scrubbed by an EXIT trap.
Verification creates and deletes a draft release in orca-hourly to prove
Contents:write for real rather than trusting the permission checkbox. Drafts
are absent from the releases atom feed, so the probe cannot disturb users.
Refuses to run without a controlling terminal instead of falling through
having set nothing, and refuses to run under xtrace, which would echo the
token on every expansion.
* fix(updater): address review feedback on the hourly channel
Renderer:
- Guard listBuilds against out-of-order responses. activeChannel flips once
getVersion resolves, and rapid channel clicks stack requests, so a slower
earlier load could land last and fill the list with builds from a channel
the picker was no longer showing.
- Selecting the running build's own channel now clears the override instead
of pinning it. There was previously no way back to "follow this build's
channel", so merely opening the panel left background checks pinned.
- Validate releaseChannelOverride on hydration, matching every other
enum-like field in that function.
Main:
- Exclude pinned jumps from recordCompletedUpdateCheck() in update-available.
A dev browsing the picker was persisting lastUpdateCheckAt and suppressing
the next real background check for a full day.
- parseHourlyVersionStamp now anchors on the whole version and round-trips
the parsed fields. It accepted garbage prefixes, and Date.UTC rolled
impossible dates forward, so ...hourly.202602300000 rendered as March 2.
Workflow:
- Publish into a draft and flip it live only after the manifest check. The
window between creating the release and verifying its assets previously
exposed a tag the picker would offer and the download would 404 on; a
draft is invisible to listReleaseBuilds, so a job that dies in that
window — including a hard kill by the job timeout, which runs no cleanup
step — leaves nothing user-visible behind.
- Add a failure handler that discards the draft, gated on the publish step
not having succeeded so a later prune failure cannot delete a live release.
- Align retry budgets with the job timeout (was 60min against a worst case
of ~185min, so a mid-retry kill skipped the cleanup that step exists for).
- Exclude drafts from the freshness and retention queries.
- persist-credentials: false; the job only reads this repo and never pushes.
* refactor(hourly): authenticate with a GitHub App instead of a PAT
A fine-grained PAT expires, and the hourly build would then fail silently on
a schedule nobody watches. A GitHub App's private key has no expiry, so this
is set up once. It is also owned by the org rather than by the person who
created it, so the credential survives that person leaving.
The workflow mints a short-lived installation token via
actions/create-github-app-token and passes it as GH_TOKEN. Installation
tokens live one hour, which is ample: this job runs no tests, no
notarization, and no Windows signing, so it is pack + upload. The retry
budgets and job timeout are re-sized to that reality rather than copied from
the release pipeline, whose 3x45 publish budget exists for notarization and
SignPath.
setup-hourly-release-token.sh now provisions HOURLY_RELEASE_APP_ID and
HOURLY_RELEASE_APP_PRIVATE_KEY. The key is redirected from a file straight
into `gh secret set` on stdin, so its contents never enter a shell variable,
argv, or the terminal.
* fix(hourly): make the xtrace guard fire and cover cancelled runs
The xtrace guard disabled tracing before testing for it, so `[[ -o xtrace ]]`
read the state the previous line had just cleared and never fired. `bash -x`
ran straight through, tracing exactly the key handling the guard exists to
prevent. Test first, then disable.
The draft cleanup only ran on failure(), but a run stopped from the Actions
UI is cancelled(), not failed — a manual cancel mid-publish stranded the
draft. Cover both.
* perf(main): add watchdog boundary memory benchmark
Add a repeatable Electron 43 RSS harness that measures the production-built watchdog entry across the child-process and worker-thread boundaries. Record per-trial samples, the median, revision, runtime, and settling procedure for reproducible PR evidence.
* perf(main): move hang watchdog into a worker thread
Keep main-thread hang detection independent of the blocked Electron event loop without paying for a second ELECTRON_RUN_AS_NODE process. Preserve the marker and telemetry contract while moving timing configuration and heartbeats onto a bundled worker entry.
* test(main): smoke packaged hang watchdog worker
* fix(main): make packaged watchdog smoke able to fail
The smoke reported failure only through process.exitCode, but its finally
block quit Electron gracefully, and Electron takes its status from the
browser exit code. Every failure mode — entry missing from app.asar, worker
error, marker timeout, non-zero worker exit — exited 0 with the diagnostic
discarded on stderr, so the required PR check could never go red.
Propagate a real status via app.exit, assert the success line in stdout, and
surface stderr. Verified against a packaged tree with the entry removed:
exit 0 before, exit 1 after.
Co-authored-by: Orca <help@stably.ai>
---------
Co-authored-by: Orca <help@stably.ai>
* perf(computer): add mac helper owner-loss benchmark
Measure the release helper's resident memory before and after its owner-session deadline. Record exact revisions, per-trial RSS, retained state, and clean-exit latency so lifecycle reclamation is reproducible.
* fix(computer): reap mac helper after client loss
Bind the detached macOS helper lifetime to authenticated socket ownership. Reap the helper after its final authenticated client disconnects, and add a startup deadline for sessions that never authenticate.
* test(computer): harden owner benchmark cleanup
* test(computer): make owner benchmark cleanup failure-safe
* test(computer): close remaining owner cleanup races
The Windows release job hard-failed in run 30125672117: every SignPath
module install attempt got 403 Forbidden from the gallery's OData API,
which is behind Azure Front Door and was also serving 502/504 at the
time. That step was the only hard-fail in an otherwise fail-open signing
chain, so a gallery incident blocked the whole release.
The gallery CDN that serves the nupkg is a separate origin and stayed
healthy throughout, so fall back to a pinned version fetched from it
after the normal install path is exhausted. The fallback verifies a
SHA-256 pin, since that route skips the gallery's own package
validation.
Extracted to a composite action so the release job and the signing
rehearsal cannot drift apart.
Co-authored-by: Jinjing <6427696+AmethystLiang@users.noreply.github.com>
Co-authored-by: Neil <4138956+nwparker@users.noreply.github.com>
Co-authored-by: JeongUk Park <jeongph.dev@gmail.com>
* fix(terminal): park SSH worktrees like local ones (C1 retention, slice A)
SSH ptys were blanket-excluded from hidden-view parking, so a hidden SSH
worktree retained every pane forever (C1: renderer heap climbs to the V8
ceiling). SSH bytes transit local main — fact-mode watchers already cover
them, and main keeps a headless model served over pty:getMainBufferSnapshot
that the SSH reattach path never consulted.
- isParkRestorableTerminalPty: snapshot-backed OR (SSH + policy); threaded
through both park verdicts, both selectors, watcher coverage, and the
watcher start guard. Remote-runtime/fail-open/foreign/null unchanged.
- Parked-SSH reveal paints from main's headless model (dimension-matched,
~5k rows) and degrades to the relay 100KiB replay unless the snapshot is a
non-empty source==='headless' payload — never a blank/stale paint.
- Kill switch: settings.terminalSshViewParking (default on).
DESIGN.md records the approved plan and the H1 magnitude non-claim.
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): bound hidden-worktree retention with a force-park budget (C1, slice B)
Un-parkable worktrees (remote-runtime ptys, uncoverable tabs, SSH with the
slice-A switch off) had unlimited retention: the parking cap/TTL only ever
saw eligibility-passing worktrees, so one bad tab pinned a whole worktree's
panes forever. Retention is now memory-bounded, not eligibility-bounded.
- terminal-hidden-worktree-retention.ts: retention budget (12 hidden / 45min
TTL, sized from the measured 2.5-19MB per-pane V8 cost, DESIGN.md §2) over
hidden worktrees ordinary parking can never evict; reuses the hot-retain
ranking so last-active exemption, deterministic ties, and deadline-driven
rechecks hold. Fail-open/foreign-pty tabs are eviction-exempt (a remount
would fresh-spawn and orphan the live shell).
- Terminal.tsx: force-parked ids join the parked set AFTER the coverage veto
(darkness for uncoverable tabs is the accepted cost); buffers captured via
the sleep-flow registry before the unmount render; retention TTL added to
the recheck deadlines for budget candidates only.
- Verdict stays out of its own effect deps; policy test asserts idempotence
and time-monotone membership (flip-loop dwell regression).
- Kill switch: settings.terminalHiddenWorktreeRetentionBudget (default on).
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): demote hidden scrollback for eviction-exempt worktrees (C1, slice C)
The retention budget (slice B) must exempt worktrees holding fail-open or
foreign-worktree ptys — a remount would fresh-spawn and orphan the live
shell — which would leave that class unbounded again. Instead, past the same
45min retention TTL their hidden panes drop to the minimum scrollback tier
(measured: ~19MB -> ~1.3MB V8 heap per 50k-row pane; trimmed history is
gone by design, reveal restores the configured cap for future output).
- terminal-hidden-scrollback-demotion.ts: module-state verdict registry
(parked-watcher pattern) with content-equality notify damping; applied in
the existing scrollback-rows effect in use-terminal-pane-lifecycle.
- selectScrollbackDemotedTerminalWorktrees: pure, TTL-gated, time-monotone.
- Retention TTL wakeups now also cover exempt worktrees so demotion fires.
- Kill switch: settings.terminalHiddenScrollbackDemotion (default on).
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): paint the SSH model snapshot inline, not via nested coordinator (C1 slice A fix)
applyMainBufferSnapshot runs its own structuralReplayCoordinator.run; calling
it from applyReattachPayload (already inside the coordinator when a relay
replay exists) deadlocks on the coordinator's tail chain. The model paint now
mirrors the daemon-snapshot branch inline (folded scrollback + rehydrate +
screen, dimension-matched, escape tail last) and arms the restored-snapshot
seq baseline so deferred/live chunks the snapshot covers dedupe instead of
double-painting. Also falls through (no early return) so reattachPayloadApplied
still latches. Adds the folder-workspace id parity unit case.
Co-authored-by: Orca <help@stably.ai>
* test(terminal): SSH park+reveal e2e round-trip + as-built design notes (C1)
Docker-gated (ORCA_E2E_SSH_DOCKER=1) spec: SSH tab parks behind a decoy and
reveal restores marker content at multi-viewport scrollback depth. DESIGN.md
records the as-built deltas (inline paint, force-park shape, last-active
floor) and the residuals so follow-ups aren't lost.
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): paint SSH reveal from main's model even when the relay replay is empty (C1 review #1)
A relay restart empties the replay buffer; the reveal previously painted
nothing even when main's headless model held the session. The reattach now
prefetches the model snapshot when no structural replay exists (SSH-shaped
ptys only) and paints it inside the coordinator; emptiness is judged on the
composed payload (scrollbackAnsi + data + pendingEscapeTailAnsi) so an
alt-screen snapshot with an empty screen frame still paints.
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): decouple scrollback demotion (slice C) from the retention-budget switch (C1 review #2)
Per the approved contract each slice reverts behind its own switch: slice C
now requires only the master terminalHiddenViewParking plus its own
terminalHiddenScrollbackDemotion flag. The TTL wakeup timer fires for
demotion candidates even with the budget switch off. No DEFAULT_SETTINGS
entries exist for sibling flags (defaults are the '!== false' optional
pattern), so no explicit defaults are added.
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): scope eviction exemption to the tab, not the worktree (C1 review #3)
One eviction-exempt tab (fail-open/foreign pty) previously vetoed force-park
for its whole worktree, pinning co-located remote-runtime tabs forever. The
worktree now force-parks while exempt tabs keep their mounted panes via a
per-tab exclusion mirroring the Activity-portal pattern (legacy watcher sync,
legacy render, and the overlay cold-parking hook). Ordinary parking is
untouched — a worktree with an exempt tab still cannot ordinary-park.
Slice C now also demotes exempt tabs' panes as soon as their worktree
force-parks under the count budget (they are the only panes left mounted).
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): demote un-parkable worktrees the force-park lever spared (C1 review #4)
The last-active exemption means a single hidden un-parkable worktree never
force-parks — and slice C previously only targeted exempt-tab worktrees, so
its panes held full scrollback forever. Demotion now also covers un-parkable
non-exempt worktrees past the retention TTL that are absent from the
force-parked set (last-active spared, or slice B switched off). Membership
stays time-monotone for fixed inputs; covered by new idempotence/monotone
selector tests.
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): keep the hidden clock running through transient background-measure windows (C1 review #5)
Whole-worktree background mounts (browser-automation bootstrap lease, mobile
mounts, agent wakes) open a ~3s self-clearing measure window that previously
deleted hiddenSince — every remount restarted the 30s hysteresis and the
45min retention TTL, so a periodically re-mounted force-parked worktree
never re-parked. The measure window still pauses parking/eviction verdicts
(all selectors skip measuring candidates); only the clock survives, so the
prior verdict resumes as soon as the window closes. Visible and
portal-holding worktrees still reset the clock.
Co-authored-by: Orca <help@stably.ai>
* test(terminal): make the SSH park+reveal depth assertion prove the model paint (C1 review #6a)
Pad the session with ~180KB of output after the numbered markers so the
earliest marker falls outside the relay's 100KiB rolling replay buffer while
staying inside main's ~5k-row headless model; asserting marker_1 after
reveal now proves the headless-model paint rather than passing under the
relay fallback.
Co-authored-by: Orca <help@stably.ai>
* docs(terminal): rewrite DESIGN.md as the single as-built C1 contract (review #7)
One contract matching the code: status IMPLEMENTED around force-park (not
the unmount proposal), real kill-switch names with coupling + revert
matrices, the true retention-floor formula with measured per-pane and
demotion numbers, an explicit when-OOM-is-still-possible paragraph naming
the H2 pendingSideEffects residual, the applyMainBufferSnapshot deadlock
constraint inside the slice-A section, stable-signal phrasing instead of a
capability latch, fail-open AND foreign-worktree exemption class, verified
cites, and a planned/landed/follow-up test matrix.
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): resolve the eviction exemption per pane, not per tab (C1 review #8)
isEvictionExemptTerminalTab read only tab.ptyId — the FIRST leaf's pty —
while the coverage veto that makes a worktree a retention candidate walks
every pane. A split tab whose second leaf held an unrestorable pty therefore
failed coverage (→ force-park target) yet looked exempt-free, so force-park
unmounted it and orphaned the live shell. The exemption now resolves panes
through the same resolveParkedTerminalPaneCandidates, keeping tab.ptyId in
the union for the no-layout/no-capture case.
Also from the same review round:
- force-park's capture passes includeLocalBuffers:false like every other
shutdownBufferCaptures caller; it was serializing up to 512KB/pane of
scrollback into the store inside a fix meant to bound renderer heap.
- Terminal.tsx unmount resets the scrollback-demotion registry — module
state with no reset path, read by a pane effect that runs before the host
effect that would clear it, so a stale verdict trimmed restore replays.
- memoize watcher coverage per tab within the parking pass; the retention
candidates re-asked it for every mounted worktree, not just the parked few.
* docs(terminal): drop DESIGN.md — the as-built C1 contract moves to the PR body
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): cap the deferred PTY side-effect queue (C1 residual H2)
pendingSideEffects grew without bound under background timer throttling
(~64 drained/s vs hundreds queued/s overnight). Cap at 512 entries with
oldest-first eviction: titles drop (last-wins), a pending bell latches
onto the next survivor, agent-status payloads collapse onto the survivor
keeping the newest 16 (last-wins store state, KB-scale strings).
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): carry command-lifecycle facts through parked watchers (C1 follow-up)
Parked fact-mode watchers omitted onCommandFinished/onCommandCode*, so
OSC 133;D and Command Code scrape signals went dark while parked. New
parked-terminal-command-status.ts ports the store-level subset: git-UI
nudge on every command finish, same-turn status-row drop for SSH PTYs
(exact mounted-path parity — the foreground tracker refuses SSH ids),
and the Command Code working seed / 1500ms done settle. Byte mode scans
the same shared parsers for authority-off parity. Local-PTY status drops
stay with the mounted pane: they need pty-connection's process-confirm
ladder to tell a leaked nested-shell 133;D from a real agent exit.
Co-authored-by: Orca <help@stably.ai>
* test(terminal): retention-budget force-park e2e with a retentionLimit override (C1 6b)
ORCA_E2E_TERMINAL_RETENTION_LIMIT flows preload → e2e-config →
getTerminalParkingPolicyOverrides (exposeStore-gated, positive-integer
only) so a spec can shrink the force-park budget to 1. The Docker-gated
spec opens two remote worktrees on one relay target (second pre-seeded
remote repo), disables terminalSshViewParking to make both un-parkable,
hides both behind the local context, and proves the older one force-parks
while the last-active exemption spares the newest; re-activating the
evicted worktree restores the marker tail via relay replay.
Co-authored-by: Orca <help@stably.ai>
* test(terminal): retention-budget e2e via same-repo remote worktrees (passes docker lane)
The first draft added a second remote repo mid-session, whose pane pty
spawn misroutes to the local daemon with the remote cwd (pre-existing
multi-repo issue, reproducible without any retention override — a seeded
local repo plus one remote repo shows the same misroute). The spec now
budgets across three worktrees of the ONE connected repo, created through
the product createWorktree path (an external git-worktree-add only lands
as a detected worktree needing adoption) and polled through the relay's
transient post-connect reconnect window. Verified green on the local
Docker lane in 20.8s.
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): prevent remount thrashing during post-measure cool-down (
Implements the C1 retention contract: preserve worktree `hiddenSinceMs` through a
background-measure window (so TTL/ranking stay honest), but re-park waits for a
full `coldParkDelayMs` cool-down after the measure ends. Without the cool-down,
every ~3s measure lease on a past-deadline worktree thrashes remount/reattach.
Core changes:
- Terminal.tsx: add measure clock (measuringTerminalWorktreeIdsRef) and post-measure
cool-down tracking (terminalWorktreeParkCooldownUntilRef); gate parking candidates
until cool-down expires.
- Extract snapshot replay choreography to shared terminal-snapshot-replay-paint.ts
(used by SSH reattach + daemon restore paths).
- Add SSH model snapshot timeout (750ms) with fallback to relay replay.
- Move cold-park recheck deadline logic to terminal-cold-park-recheck-deadlines.ts;
add cool-down deadline to scheduling.
- useTerminalTabColdParking: implement matching measure-clock contract with per-tab
cool-down gate to keep tab deadlines synced with worktree retention clock.
- Add resolveTerminalMountScrollbackRows() to demote new xterms under demoted
worktrees (pane births during demotion must take the demoted tier at create).
- Add kill switches: terminalSshViewParking, terminalHiddenWorktreeRetentionBudget,
terminalHiddenScrollbackDemotion.
* fix(terminal): detect Command Code completion in parked mid-turn panes
Seed the byte watcher with in-flight turn state from agent status: the
watcher is recreated per park cycle with no startup command to arm it,
and the banner scrolled away before parking. Also memoize
eviction-exempt checks and use SSH PTY ID builder in tests.
* fix(terminal): flush pending command-code settles on reveal remount
When a parked pane reveals mid-Command Code turn, the new detector
cannot re-observe the already-passed idle composer. Cancelling the settle
leaves the row stranded at 'working', so dispose now flushes the pending
settle instead.
Extract readInFlightCommandCodeTurn to shared space and seed detectors
with in-flight turns so remounts complete mid-flight commands. Also
memoize SSH model probes to prevent double timeouts on reattach.
* fix(terminal): remove scrollback demotion (C1 slice C)
The scrollback demotion feature for eviction-exempt hidden worktrees is no longer needed. Retention budget limits are now sufficient without this additional bound. Remove the terminal-hidden-scrollback-demotion module, the selectScrollbackDemotedTerminalWorktrees function, and related per-pane demotion logic.
* test(terminal): assert bounded probe during stalled reveal
Add assertion to verify that a stalled reveal operation makes exactly one
`getMainBufferSnapshot` call, ensuring retry logic doesn't introduce
redundant probes that would extend the timeout window before relay fallback.
* fix(terminal): implement C1 retention budget for hidden parked worktrees
Addresses OOM regressions in hidden parked terminals by force-evicting
worktrees past a retention budget: at most 12 mounted while hidden, none
past 45 minutes (absolute, not exempted by last-active). Eviction is
least-recently-hidden-first. Exempt tabs (unrestorable local PTYs) keep
their panes to avoid orphaning shells; worktrees are force-parked even
if they contain exempts, and their buffers released elsewhere. SSH/remote
worktrees serialize buffers pre-eviction for reveal; local worktrees keep
daemon snapshots. Command Code's done-settle window is transferred across
park/reveal boundaries so the row cannot strand at 'working'. Model probe
on SSH reattach is scoped to park-reveal only, not ordinary reconnects.
Includes new E2E suite proving the budget actually releases memory.
* memoize eviction-exempt terminal tabs to avoid redundant store reads
Each tab's exemption check re-reads the store and walks the layout tree.
Introduce selectEvictionExemptTerminalTabIds() to resolve all exempt tabs
for a worktree in a single pass, then memoize the result in Terminal.tsx
and useTerminalTabColdParking. This prevents O(n) store reads when checking
exemptions across multiple tabs and ensures the set remains stable across
unrelated re-renders.
* refactor: reformat hidden-worktree retention comments
Reflow to 80-character lines and remove internal ticket references
(C1, C1 slice C).
* fix(lint): split overlay slot and eviction-exempt tabs under max-lines
Static analysis failed because TerminalPaneOverlayLayer (401) and
terminal-parked-tab-watchers (304) exceeded oxlint max-lines. Extract the
slot component and eviction-exempt helpers into dedicated modules.
* test(terminal): stabilize retention budget e2e control arm
Stage un-parkable remote pty ids only after both worktrees are hidden, and
keep re-staging during the control-arm poll so a late updateTabPtyId cannot
flip the decoy back to park-restorable and ordinary-park it before budget
engages.
* test(terminal): pin retention e2e decoy to a mounted pane snapshot
Use the active pane-identity snapshot for the decoy tab instead of all
worktree tabs, and re-assert un-parkable ids after the control-arm hold so
a deferred/empty tab id cannot fail the budget-off mounted-count check.
* fix: memoize terminal eviction exemptions on layout leaf PTYs
Splits add leaf panes to the layout store without changing the tabs
array. A memo keyed only on tabs misses this change, leaving new panes
unexempted for unmount. Include layout leaf PTYs in the exemption memo
key so it recalculates when splits occur or PTYs are re-minted.
---------
Co-authored-by: Orca <help@stably.ai>
* ci(pr): run E2E when a PR touches tests/e2e paths
Regression specs under tests/e2e never ran on PR CI — only schedule and
release called e2e.yml — so a red regression test could merge green.
Path-filter and workflow_call the E2E suite when E2E-relevant files change.
Use merge-base diffs so base-branch drift does not false-trigger E2E, fail
the detector when git diff cannot compute the PR range, and pin
least-privilege contents:read on both the detector and reusable E2E workflow.
Closes#10518
Co-authored-by: Wooseong Kim <innocarpe@gmail.com>
Co-authored-by: Orca <help@stably.ai>
* ci(pr): make the E2E path gate actually block, and match the real config path
Two fixes to the new path-filtered E2E job.
The gate did not gate. pr.yml's `verify` job is the required check, and it
enumerates its dependencies explicitly — `e2e` was in neither `needs` nor the
result list, so a failing shard left `verify` green. That reproduces the exact
hole this job exists to close: a red spec merges green, just with a red box
further down the page. Add `e2e` to both.
Because the job is path-filtered, `skipped` is the normal result on a PR that
touches no E2E files and has to keep passing. That allowance is checked after
the strict loop rather than inside it, so it can never leak to the six jobs
that are always required.
The `playwright.` pattern matched nothing. The config is
tests/playwright.config.ts — beside tests/e2e/, not inside it — so no tracked
file starts with `playwright.` and editing the runner config would silently
skip E2E. Anchor it at `tests/playwright.`.
Adds a contract test alongside the existing release-e2e one. Verified it fails
when either fix is reverted, and simulated the gate across
success/skipped/failure/cancelled plus the skip-must-not-mask-a-real-failure
case.
* test(ci): close two gaps in the E2E gate contract
CodeRabbit was right on both counts — verified by reverting each and watching
the contract stay green.
The path filter was unasserted, so `e2e` could lose its `if:` and run on every
PR — the cost the filter exists to avoid — without failing anything.
The strict-loop check hardcoded four of the six required jobs, so dropping
GIT_COMPATIBILITY or SHELL_CONTRACTS left them unenforced while the contract
passed. Derive the list from verify.needs instead, so a newly added required
job that misses the loop fails here rather than silently going unchecked.
* ci(pr): land the E2E path gate advisory instead of blocking
The E2E suite is currently failing every scheduled run on main — 22 of the last
22 — so making verify depend on it would block any PR touching tests/e2e/**,
including the PRs that fix the suite. This PR's own run reproduced that: 3 of 12
shards failed on specs unrelated to it (agent-session resume, Jira linking,
plugin containment, terminal artifacts).
So the job runs and reports on E2E-path PRs but is left out of verify.needs for
now. The detector, the tests/playwright. path fix, and the contract tests are
unaffected — those stand on their own and were the substance of the review.
Flipping to blocking is a three-line change once the suite is green; the exact
wiring, including why the skipped allowance must sit outside the strict loop, is
recorded on verify's Require-successful-checks step. The contract test pins the
advisory choice so it reads as deliberate rather than as the unwired-gate bug it
originally caught, and still fails if the path filter, the strict-loop coverage,
or the config path regress.
---------
Co-authored-by: Wooseong Kim <innocarpe@gmail.com>
Co-authored-by: Orca <help@stably.ai>
Reviewed with an independent reproduction. Added the allowlist entry that unblocked verify:localization-coverage on main, the 4th drifted step, and a parity gate that fails when pnpm lint's chain contains a script absent from pr.yml.
* fix(release): restore the Windows inner-binary signature gate
electron-builder 26.9+ dropped the bundled 7zip-bin package, so the gate's
hardcoded node_modules/7zip-bin path stopped resolving in 1d2cd33c83. The
gate is fail-open, so it swallowed the error and 11 releases shipped with
no signature verification and an evidence artifact that looked clean.
Resolve 7za through app-builder-lib's toolset instead, and always record a
verdict so a degraded gate can't pass for a healthy one.
Refs #6487
* test(release): make the signing-gate structural tests assert executed code, not text
The round-2 harness matched /\bthrow\b/ and /\bcatch\b/ against raw block text, so
the word satisfied the assertion wherever it appeared. Downgrading the resolver
throw to `Write-Host "...would normally throw..."` — the exact silent fail-open
this PR exists to kill — left all 11 tests green.
Every span is now classified once (code / string / comment) by the same walk that
pairs braces, and assertions run against the string-and-comment-blanked view.
Blanking preserves length, so indices still line up across views.
Also re-anchors the catch-ordering test: `blockAfter(step, '} catch {')` picked
the first catch in the step, which stopped being the gate's own once the
persistence helpers grew theirs — moving the policy throw inside the try was
passing again.
Co-authored-by: Orca <help@stably.ai>
* test(release): pin the evidence filename the gate writes to the one the upload collects
The upload step is `if-no-files-found: ignore`, so renaming the evidence file on
one side and not the other ships a green run whose artifact silently omits the
verdict — the same silent-degradation class this PR exists to close.
Co-authored-by: Orca <help@stably.ai>
* fix(release): preserve 7za resolver failures
---------
Co-authored-by: Orca <help@stably.ai>
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
* feat(plugins): Orca plugin system — kernel, content packs, panels, workers, marketplace v0 (experimental)
Adds Orca's experimental plugin system behind a settings flag: a
supervised kernel, declarative content packs (VM recipes, commands and
keybindings, language packs), sandboxed iframe panels, forked worker
hosts, and a Git-backed marketplace v0 with consent, provenance and
kill-list enforcement.
Theme, icon-theme and terminal-theme contributions are deferred to a
follow-up pass.
* fix(plugins): make unsupported marketplace listings unreachable by key
findPlugin() backs preview/install/previewInstalledUpdate via
requireListing(), so filtering only listPlugins() hid the catalog card
while leaving the dead install path reachable one click later.
* fix(plugins): fan Pi session-only status out to plugin subscribers
The providerSessionOnly early-return in applyNormalizedStatus emitted to
onAgentStatus (main-window fanout) but skipped enrichedStatusListeners, so
plugins subscribed to agent.status.changed silently missed every Pi
session_start event. Route both emit sites through one helper so a future
early return cannot drop the plugin tap again.
Co-authored-by: Orca <help@stably.ai>
* plugins: drop dead code and hoist duplicated trust-boundary patterns
Cleanup pass over the P1 diff, no behavior change:
- Delete `readPluginTreeSnapshot`/`readSnapshotFile` and their types, plus
the now-vestigial `directories`/`signal` plumbing in `collectFiles`.
- Delete `resolveContainedPluginDirectory` (no callers).
- Delete `plugin-content-load-pool.ts`; it reimplemented the existing
`mapWithConcurrency`, whose index arg also removes the pairing wrapper
in `buildPluginList`.
- Hoist `PLUGIN_CONTENT_HASH_PATTERN` and `PLUGIN_COMMIT_PATTERN` into
the install-lockfile module; 11 sites hand-rolled these identically.
- Point the new reliability gate at the PR instead of gitignored docs
paths, matching every other gate's link form.
* fix(plugins): retry plugin state renames on Windows AV/EPERM locks
Six plugin write paths (lockfile, provenance, current pointer, kill
list, marketplace cache, staged install dir) did a plain rename, so an
antivirus or indexer holding the target open surfaced as a failed
install. The repo already retries this hazard for issue #1507, but only
through a sync helper; these paths are all async.
Adds one bounded async retry + atomic write used by all six, and trims a
consent-provenance header that restated its own JSX.
* test(plugins): cover the Windows rename retry path
The retry loop shipped untested: both existing cases hit the non-retry path,
and the temp-cleanup test passed identically with the `finally` removed.
Mock `rename` to queue errno codes so CI can exercise locks it cannot provoke.
Co-authored-by: Orca <help@stably.ai>
* fix(plugins): pin bundled plugin resources to LF
Windows CI checks out with autocrlf, so the byte-hashed launch tree arrived
as CRLF and verify-packaged-plugin-resources rejected it — the packaged build
could never pass on Windows. Reproduced locally: CRLF yields the exact CI
error, LF verifies clean. Files are already LF, so nothing renormalizes.
Co-authored-by: Orca <help@stably.ai>
* test: guard the bundled-plugin LF pin against a CRLF checkout
The byte-hash mismatch only surfaced in Windows packaging CI. Assert the
.gitattributes pin and that a CRLF tree is rejected, so a regression fails
on any platform instead of waiting for a packaged Windows build.
Co-authored-by: Orca <help@stably.ai>
* ci: trigger packaged-build check on bundled plugin resource changes
The launch tree is byte-hashed during packaging, but no trigger path covered
it — so the CRLF fix for that check would not have re-run the check. Add the
resources, verifier and .gitattributes paths that can break packaging.
Co-authored-by: Orca <help@stably.ai>
* perf(plugins): rebuild the panel frame only when its baked theme values change
The revision keys the panel iframe, so every bump destroys the sandboxed
frame and its in-panel state. It counted root attribute mutations, but
--workspace-sidebar-live-width is written every rAF of a sidebar drag, so
dragging with a panel open blanked it ~60x/sec. Compare the two values the
shell actually bakes in instead.
Co-authored-by: Orca <help@stably.ai>
* test: stop pinning a plugin name in the CRLF guard
The CRLF case rewrites every launch file, so the reported mismatch is
whichever plugin sorts first. P2 adds theme plugins that sort ahead of
orca-navigation-shortcuts, which broke the assertion there.
Co-authored-by: Orca <help@stably.ai>
* style: drop stray blank lines left by the rebase resolutions
Both sides of the agent-hooks and orca-runtime conflicts contributed a
trailing blank, which oxfmt rejects. Whitespace only.
Co-authored-by: Orca <help@stably.ai>
* test(plugins): stop the startup budget failing on machine load
P95 runs 16-34ms idle but exceeds the 50ms bound under full-suite
parallelism, so the gate flaked. Widen it to catch an order-of-magnitude
regression instead; the no-worker/no-plugin-code assertions are the real
guarantee. Verified a 400ms regression still fails.
Co-authored-by: Orca <help@stably.ai>
---------
Co-authored-by: Orca <help@stably.ai>
* fix(release-cut): gate an explicit RC against its own series
semver_gt compares through strip_pre(), so the explicit-version override
only ever checked the stable line: 1.4.156-rc.0 read as 1.4.156, cleared
a 1.4.155 stable, and republished an RC below what clients already run.
Anchor a prerelease request on highest_rc_for_base -- the same rc history
the kind path uses -- so the override can only advance the series.
Two sibling gaps in the same block:
- version_suffix was silently dropped when version was set, because the
append lives in the kind branch the override skips.
- the shape regex rejected X.Y.Z-rc.N.suffix, so a suffixed RC the rc
path can produce could never be re-cut explicitly.
* fix(release-cut): close both ends of the rc-number range the gate compares
The new explicit-rc gate compares with `[[ -le ]]`, i.e. bash machine-width
integers, and the author closed only the low end. Past INTMAX bash saturates,
so `version=1.4.156-rc.99999999999999999999` reads as "above the published
rc.3" and the gate falls open — then the tag it cuts pins
highest_rc_for_base at 1e20 for that base forever, and every later cut wraps
to a lower rc the fleet never updates to. Bound the rc number to nine digits.
Also reject leading zeros on an all-digit prerelease identifier. `npm version`
renormalizes rc.4.01 to rc.4.1 while the tag step keeps the literal input, so
the shipped package.json version and its own release tag name different
releases. The explicit path's embedded identifier now goes through the same
validator the kind path uses instead of only the shape regex.
* fix(release-cut): stop the refusal pointing minor/major RCs at the wrong series
kind=rc derives its base from bump(latest_stable, patch), so the remedy the
refusal suggested only works when the requested base *is* that next patch. A
1.5.0-rc.N series exists only because this override created it, so an operator
resuming a stuck 1.5.0-rc.2 was told to dispatch kind=rc, which would have cut
an unrelated 1.4.156-rc.4. Spell the condition out and give the fallback that
does work for a non-patch base.
Also correct the mechanism in the comment I added in 698c5beeaa: bash wraps
two's-complement, it does not saturate, which is why the hole is
value-dependent (rc.10000000000000000000 wraps negative and failed closed,
rc.99999999999999999999 wraps to 7766279631452241919 and sailed through).
And name both inputs in the suffix error, which now serves version_suffix and
the trailing identifier in version.
* fix(release-cut): count a suffixed RC from its commit subject, not just its tag
The new explicit-version gate only fails closed on a deleted tag because
highest_rc_for_base also reads `release: v<base>-rc.N` subjects. That fallback
did not parse the suffixed form: rcNumberFromTag accepts an optional
.identifier, rcNumberFromReleaseSubject did not, so `4.perf` failed its
`(\d+)(\s|$)` anchor and returned null.
So deleting a v1.4.156-rc.4.perf tag dropped the series back to rc.3, and an
explicit 1.4.156-rc.4 was waved through — below the rc.4.perf build
perf-channel clients already run. Same under-count already made kind=rc
recompute rc.4 over a deleted suffixed tag.
Mirror the tag form's optional identifier. Covered by a unit assertion and a
git-fixture test that both fail with this reverted.
* docs(release-cut): correct four operator-facing claims in the explicit path
All four are wording or consistency, no behavior change (harness: 26/26 before
and after, on bash 3.2 and bash 5.2).
- The trailing-identifier comment justified itself as preserving a shape that
"can never be re-cut through the override", but re-cutting a suffixed rc at
or below the series head is exactly what the new gate refuses. State what it
actually admits: a second spelling of version=X.Y.Z-rc.N + version_suffix.
- version_suffix's input description still said "rc kind only" after this PR
made it apply to an explicit bare X.Y.Z-rc.N.
- The suffix guard's own rc pattern was unbounded while the shape check twelve
lines up is bounded to nine digits; reuse the bounded one so a later edit to
either cannot silently drift.
- "which recovers the existing tag" was unconditional, but kind=rc recovery is
also gated on tag_matches_current_ref, so a tag cut from a ref main has moved
past advances to rc.N+1 instead.
* fix(skills): advance the release ledger at the cut so shipped revisions freeze
#10340 made the released-skill registry a function of the committed ledger
instead of a git tag walk, and #10460 reverted the cut step that advances that
ledger because it violated the #9119 contract (a version-only cut must not
regenerate or stage the content-addressed skill artifacts). Both were right;
the result is a ledger that never advances.
generate-skill-bundle-manifest.mjs:390 derives releasedCount solely from
release-mapping.json and :461 assigns a changed skill releaseRevision =
releasedCount + 1, while :518 protects only committedReleasedCounts[name] —
so index releasedCount is unprotected. A tag ships that tail revision, nothing
records it, and the next skill change rebuilds the same revision number over
different bytes. Installs carrying the shipped digest then match no snapshot
and degrade to unrecognized, which cannot be updated.
Restore the advance in a form the #9119 contract can keep enforcing:
--release now verifies that current-manifest.json and snapshot-registry.json
already match the ref being tagged, appends the mapping row, and writes only
release-mapping.json. The cut stages just that file, so it still cannot move a
content-addressed artifact — the failure #9119 guarded against — and now fails
loudly instead of recording a revision the tag does not ship.
The contract test is narrowed to match: it asserts the cut runs --release
(never --write) and stages exactly package.json and release-mapping.json.
* test(release-cut): close the staging bypasses the narrowed gate left open
The narrowed contract test anchored its `git add` scan to line start and
only inspected staged paths, so three ways to reintroduce #9119 stayed
green: a `git add` chained after `&&`, a write that never calls `git add`
at all, and `pnpm run generate:skill-bundle-manifest` — the package.json
alias for `--write`, which the hyphenated ban never matched. That last one
also passed the pre-#10460 assertions, so it was never covered.
Drop the anchor, require every `resources/skills` mention in the step to
be exactly what is staged, and ban the alias and `commit -a`. Comments are
stripped first so prose cannot trip a ban. Verified each bypass fails and
the real workflow passes.
* fix(release-cut): make the new provenance failure actionable to an operator
Verifying the content-addressed artifacts is the only new way the cut can
block, and it fails inside a step named "Bump package.json and tag" with a
lint-shaped message. That names the files and the command but not the two
things the operator needs: the regeneration has to land on main, and the
cut is safe to re-run afterwards. Say so.
Also pin down why assertReleasedHistoryPreserved takes the pre-append
mapping. It pairs with artifacts.releasedSnapshotCounts, which seeding
fixed before the row existed; handing it the post-append mapping makes
every cut throw "Released snapshot history is incomplete", which points
at tag fetching rather than the real cause. Nothing enforces the pairing.
* test(release-cut): gate the whole cut job, not just the bump step
Round-2 review defeated the previous gate twice, both proved by running
the full contract file green with #9119 reintroduced.
Every step in the cut job shares one workspace and one index, but the
contract test only inspected `Bump package.json and tag`. A step inserted
earlier could run --write and `git add resources/skills`, and the bump
step's own commit swept it into the version commit and the tag. Assert
job-wide instead: only the bump step may name the directory, and no step
may regenerate under either the flag or its package.json alias. That lives
in the generator suite because the contract file is at its max-lines cap.
Two regexes were also evadable. The mention scan required a trailing
slash, so a path held in a variable was invisible; it now matches the
directory itself. The `commit -a` ban matched nothing at all — `commit\s`
ate the only separator, so `-a`, `-am`, and `--all` all survived while
only a trailing `-a` was caught. `--allow-empty` stays allowed.
* fix(release-cut): assert the index, not the workflow text, before committing
Round-3 review defeated the job-wide grep three ways, each proved by
running both test files green with #9119 reintroduced into the tagged
commit: an `env:` block holding `--write` and `resources/skills`, a
composite action whose steps the workflow never spells out, and plain
shell concatenation (`root=resources; leaf=skills`).
Grepping shell source for path literals is inherently evadable, and the
previous fix only relocated round-2's variable-indirection hole one step
over. Move the invariant to where it cannot be dodged: immediately before
committing, the cut diffs its own index and refuses anything that is not
package.json or the release-mapping row. That does not care which step
staged what, or how the path was spelled.
The workflow grep stays as a cheap tripwire for literal spellings, now
paired with a positive assertion that the index guard exists and precedes
the commit — indirection cannot hide a missing guard. Mention matching
dedupes and trims quotes, since the guard names the row a second time.
* fix(release-cut): match the staged-path allowlist literally
`grep -vx` treats its patterns as regexes, so the `.` in `package.json`
matched any character: a staged `packageXjson` or a
`resources/skills/release-mappingXjson` was silently accepted by the
index guard. Verified both slip through `-vx` and are caught by `-vxF`.
Exercised the guard against a legitimate cut, an empty index, a staged
content-addressed artifact, paths containing a space and a non-ASCII
character (git quotes the latter, so it fails closed), and a staged
deletion. Only the two allowed paths pass.
* test(release-cut): assert the index guard aborts, not just that it exists
The positive assertion pinned the guard's shape and its position before
the commit, but not its effect: replacing `exit 1` with `:` left both
test files green while the cut logged the error and shipped the artifact
anyway. That is the same failure this whole gate keeps having — asserting
the shape of a defense rather than what it does.
Pin the abort too. Verified the neutered guard now fails the suite.
* test(release-cut): scope the abort check and catch clustered commit flags
Two holes in the guards this PR added, both in the same shape-not-effect
class the previous commit was meant to close.
The abort assertion's lazy match was not scoped to the guard's own block,
so it could borrow an `exit 1` from any later `if ... fi` in the step.
Degrading the guard to a warning while adding a plausible HEAD
precondition left every test green. Stop the match at the guard's `fi`.
The `commit -a` ban only matched when `a` led the flag cluster, so `-vam`,
`-va`, `-qam` and `-sam` all survived. That matters more than it looks:
`commit -a` stages at commit time, after the index guard has already
inspected a clean index, so it is the one way to defeat that guard. Match
`a` anywhere in a short-flag cluster; `--allow-empty` and `--amend` stay
allowed. Verified both mutants now fail.
* fix(release-cut): validate the commit, not the index, before tagging
The index guard asserted the wrong thing. `git commit` has a family of
forms that commit the working tree rather than the index — `-a`, `-i`,
`--only`, and a bare pathspec — so a rogue earlier step could leave
regenerated artifacts unstaged and any of those forms would carry them
into the tagged commit while the guard saw a clean index and passed.
Reproduced end to end: `git commit -i resources` put current-manifest.json
and snapshot-registry.json in the tag with all gates green, and
`--only resources` additionally dropped package.json from the tag.
Banning those flags one by one is the same enumeration game the earlier
rounds kept losing. Assert the outcome instead: after committing and
before tagging, diff-tree HEAD and refuse anything that is not
package.json or the release-mapping row. That is indifferent to which
step staged what and to how the commit was spelled.
Verified the whole family is now blocked (-i, --only, -a, -am, -vam,
pathspec, and an alias expanding to `commit -i`), that a stock commit and
an --allow-empty re-cut still pass, and that deleting, neutering,
un-anchoring, or relocating the guard each fails the suite.
* fix(release-cut): make the commit guard fail closed on a merge commit
Plain `git diff-tree` prints nothing for a merge commit, so the guard
would have passed silently instead of failing closed — the one direction
that matters on a release path. `-m --first-parent` reports the diff
against the first parent; verified byte-identical output for an ordinary
commit and still empty for the `--allow-empty` re-cut, so nothing else
changes. Not reachable today (nothing in the cut job creates a merge, and
npm version has no lifecycle hooks defined), but the failure mode is a
guard that looks like it ran.
Pin the flags in the assertion too, so neither dropping -m nor slipping in
a `--diff-filter` can weaken it without failing the suite.
* fix(terminal): verify Windows PTY root identity before taskkill /T /F
killWithDescendantSweep guarded its Windows tree kill with ownsRoot()
alone, which is JS state only. node-pty's ConPTY exit watcher closes the
last shell handle before it queues the JS exit callback, so Windows can
recycle the PID while the session map still looks live — force-killing an
unrelated process and its whole descendant tree.
Walk the recycled PID's ancestry back to this process before taskkill:
skip the sweep when the root is gone or resolves to a stranger, and keep
the sweep when identity is unknown so #10004 orphan cleanup still runs.
Also gate the local provider's ownsRoot on observed physical exit.
* fix(terminal): dedupe the Windows root-identity scan, drop dead exit gate
Review fixes on the PID-identity guard.
The probe read the process table through a new uncached export, bypassing
the reader that worktree teardown depends on: worktree-teardown.ts fans out
32-wide inside a 10s deadline, so a delete forked 32 powershell cold-starts
(the churn windows-foreground-process-rows.ts:25-32 warns about, #6288/#6667).
getFreshSnapshot() already guarantees a scan that starts after the request --
the exact property the bypass existed for -- and coalesces concurrent callers,
so use it. Measured on the new test: 32 scans -> 1.
The PhysicalExitTracker.hasExited gate could never fire. markExited() is only
reached at local-pty-provider.ts:985/:1431, and both are followed synchronously
by clearPtyState(), which deletes the ptyProcesses entry -- so ownsRoot's map
check is already false whenever hasExited is true. Reverting it broke no test.
Drop it and the shared getter it added; the identity probe already covers every
ownsRoot caller from inside killWithDescendantSweep.
Also point the Windows terminal-restart E2E job at the files that own this
behavior, so a change to the new Windows-only module runs the one job that
executes on a real Windows host.
* docs(terminal): state what the Windows root probe actually proves
The probe checks subtree membership, not root identity: a recycle that lands
on another Orca descendant (another pane's shell, an agent CLI, a git.exe we
spawned) still reads `own`, and that is not remote during teardown when Orca
is itself allocating pids. It bounds the blast radius rather than closing the
class. Say so at the type and at classifyWindowsTreeKillTarget, and name what
a real close would need (a CreationDate baseline -- the analogue of the POSIX
lstart check already used here -- or an inherited handle / Job Object).
Also note why our own pid must classify `foreign`.
* ci(windows): trigger the terminal-restart E2E on the shared snapshot reader
The Windows root-identity probe now reads through getFreshSnapshot, so an edit
to that module changes Windows teardown behavior without touching any path the
job already watches.
* test(terminal): guard the teardown probe against a reintroduced scan bypass
The existing volume guard covers queryWindowsProcessRowsFresh directly, but the
identity-probe cases all inject readRows, so nothing exercised the DEFAULT
reader wiring -- a bypass reintroduced inside windows-pty-root-identity would
have gone unnoticed. Drive verifyWindowsTreeKillTarget 32-wide through the real
reader and assert one scan. Verified it fails at 32 when the bypass is put back.
Jinjing
Brennan Benson
Hyeonho Lee (Ethan)
Neil
Jinwoo Hong
OrcaWin
Henry Su