* fix(terminal): keep Mac trackpad flicks continuous in fullscreen TUIs
Chromium reports wheelDelta = -3 x deltaY for continuous devices, so a fast
trackpad event crossed the 100-unit 'notched wheel' threshold and switched to
the compressed, burst-boosted wheel curve mid-flick (2 -> 6 -> 2 reports per
event). Classify the gesture instead: an off-120-grid wheelDelta marks a
continuous device and the verdict sticks until the gesture goes idle.
* fix(terminal): end a trackpad gesture when a Mac wheel notch arrives
An on-grid event now continues the continuous gesture only with the trackpad's
3x wheelDelta shape and only shortly after off-grid input, so a notched wheel
used right after the trackpad keeps its discrete feel.
* fix(terminal): match wheel notches on the zoomed grid and require the trackpad ratio
Chromium divides both deltaY and wheelDelta by page zoom, so a notch reports
120/zoom; derive the grid from the UI zoom. An on-grid event now continues a
trackpad gesture only when 3 x |deltaY| is close to |wheelDelta|, so a
Windows/Linux notch (100 px per 120) right after trackpad input stays discrete.
* refactor(terminal): classify trackpad gestures without zoom or a notch grid
Keep the base rule (a small wheelDelta is continuous) and extend it across the
gesture: a large event within 300 ms of continuous input stays continuous only
while it keeps Chromium's continuous delta shape (wheelDelta about 3x deltaY on
macOS, 1x on Windows/Linux). Both deltas scale with zoom, so the shape check
needs no zoom factor; notches never match it.
* fix(terminal): only small wheel events extend the trackpad gesture window
A notch that happens to match the continuous delta ratio (40 px per 120 on
macOS) is now captured for at most 300 ms after trackpad input instead of
renewing the window indefinitely.
* fix(terminal): keep the whole trackpad gesture continuous through long momentum
Restore the approved rule where every continuous event extends the gesture
window, so a hard flick whose momentum stays above notch size for over 300 ms
keeps its 1:1 row distance. Pins that with a long-momentum fixture.
* fix(runtime): a pane's new process does not wear its previous run's Done (STA-12043)
* fix(runtime): keep the superseded turn's segment and date process start at spawn request
* fix(runtime): take the identity row's segment from the turn clients were shown
A CSI scan now stops at ESC/CAN/SUB (as xterm's parser does), so a hidden-startup carry ending mid-CSI no longer swallows the next OSC's introducer and leaks `104;255` onto the prompt. Fixes#31410.
Co-authored-by: Neil <neil@stably.ai>
OpenCode Go credential status: a late initial read can no longer overwrite a save/clear result, and pending/failed reads show Checking…/Status unavailable instead of Not saved. Fixes#33068.
Co-authored-by: Pr1p <48266347+Pr1p@users.noreply.github.com>
A workspace created from the composer on a paired host launched its terminal
agent with the desktop's command, arguments and environment. Hosts that
negotiate agent.launch.v2 now get startupAgent (and startupPrompt) and build
the launch through their own planner and settings. Older hosts, failed probes,
hosts that predate the agent's name, linked-item drafts and VM recipes keep the
desktop-built command. When the host planned the launch, the window never runs
the desktop plan or its own agent recovery, and says it couldn't confirm the
agent started if the host reported no agent terminal.
* fix(new-workspace): read Linear status on the host the search runs on
Create worktree searched Linear on the selected repo's host but offered the
Linear tab from the focused host's status. With Linear connected only on the
desktop and a project on a paired server, the tab appeared and every search
silently came back empty. The repo-derived Linear source now reads its own
host's status, as explicit sources already did, so the tab and the selected
workspace follow the host that will answer.
* fix(new-workspace): keep Linear status per host while it re-reads, and treat loading as unknown
The host status read is keyed by host, not project scope, and a same-host
re-read (repo switch, workspace change, reconnect) keeps the last status
instead of reading as disconnected, so the Linear tab and the chosen mode no
longer reset. While the first read is pending, a chosen Linear tab stays
selected and a pasted Linear URL keeps blocking Enter. The credential-error
test now stubs the repo host's status read it depends on.
* fix(runtime): a status publication that answers a pending probe leaves the capability readable (STA-11921)
* fix(runtime): keep negative-capability re-probe across a published replacement; split the test
* refactor(runtime): one compatibility-entry factory keeps the cache under max-lines
* fix(ssh): Reconnect & Delete stays pinned to the workspace the user selected (STA-11912)
* fix(ssh): say the workspace list changed and clear the shared delete state on a send-time refusal
* fix(runtime): stop paired desktops republishing mirrored editor tabs
When two desktops pair both ways, B mirrored A's editor tabs and then
published them in its own session inventory as if B had opened them. A
then applied B's copy, which took over A's own tab.
isMobilePublishableOpenFile now skips rows marked
mirroredFromRuntimeSession, the marker is part of the open-files sync
projection so a flip republishes, and the marker is saved with the
session and restored so a relaunched client does not republish its
restored mirrors before the host's first snapshot.
Based on #25877 by BearMett.
Co-authored-by: BearMett <agomdole@gmail.com>
* fix(runtime): keep restored mirrors through unpublished frames and out of headless publication
A restored mirror now carries its marker, so an unpublished host frame
must not cull it (and its recovered draft). The windowless publication
skips persisted mirrored rows for the same reason the window does.
Co-authored-by: BearMett <agomdole@gmail.com>
---------
Co-authored-by: BearMett <agomdole@gmail.com>
* fix(github): route runtime repo paths by execution host
When a repo path is registered on both SSH and a paired runtime, path-only lookup could send the SSH-side repo ID to the runtime. Prefer the row owned by the target runtime, then retain the original first-path fallback.
AI-Assisted: OpenAI Codex
Signed-off-by: ethereal49 <hongyeluo49@gmail.com>
* fix(github): resolve the runtime repo on the target host by the caller's repo id
getRuntimeRepoTarget now takes the caller's repo identity and resolves the
row through the repo lookup index, preferring the row on the target runtime
with that id, then the row on that runtime for the path. This keeps the
refresh sweep's repo reads linear (a per-call list scan broke that ratchet).
Legacy rows with no host stamp still fall back to id-or-path lookup.
Removes the test that pinned the off-host fallback; it asserted sending
another host's repo id to the server.
---------
Signed-off-by: ethereal49 <hongyeluo49@gmail.com>
Co-authored-by: ethereal49 <hongyeluo49@gmail.com>
* test: exercise mobile controller and strengthen audit assertions
* test: use array syntax required by React Doctor
---------
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
OrcaWinandOrcaWin
Neil
Hoang Le QuocandNeil
Pr1p
Jinjing