Serialize replay reads with event appends, acquire the live receiver under that lock, and drain the event pump during shutdown. Recovery now removes its FIFO when opening the log fails.
Own the ACP update receiver inside SessionManager, append each notification before broadcasting it, and recover structured sidecars by loading their stored ACP session id without duplicating replayed history.
Resume a stored ACP session through session/load and discard the history replayed during the load handshake so the durable event log remains the single transcript source.
<sidecar_dir>/<id>.events.jsonl holds one session/update notification
per line; seq is the line's 1-based position. open_events returns an
append handle held for the session's lifetime, and terminates a
partial last line first so a hard kill mid-write never glues the next
event onto the fragment. read_events takes a Since (forward, exclusive)
or Before (backward, exclusive, None for the tail) cursor with a limit;
unparseable lines keep their seq and are skipped. delete removes the
log with the sidecar; load_all ignores .jsonl.
Each append is a single write_all to the kernel, no fsync: a killed
daemon loses nothing already written, and an fsync per streamed chunk
is the wrong trade for a log that is not meant to survive power loss.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
SIGTERM now counts as a shutdown alongside ctrl-c: systemd stops the
unit with SIGTERM, and with KillMode=process nothing else would reach
the claude-code-acp children. After the listener drains the manager
terminates every adapter before the process exits.
A Structured spawn gets no tmux pane: the daemon starts the adapter in
the session cwd with the session env (including MAESTRO_SESSION_ID and
MAESTRO_URL), runs the handshake, and records the ACP session id on the
sidecar. An adapter that cannot come up still yields a session, Dead
with activity Error and the reason as activity_detail, so a missing
adapter reads as the install command in the sidebar rather than as a
session that never appeared.
The liveness pass checks a structured session's child instead of the
tmux list. kill sends session/cancel, waits, then SIGTERMs through the
ACP handle. SessionManager::shutdown ends every adapter concurrently
and marks those sessions Dead on disk. take_acp_updates hands #135 the
raw session/update stream once, with earlier notifications still queued.
New crate wrapping agent-client-protocol 2.1 in the client role. spawn()
resolves claude-code-acp on the PATH the child would run with, starts it
in the session cwd with the session env, runs initialize + session/new,
and returns an AcpSession holding the connection, the ACP session id,
and an unbounded channel of raw session/update notifications for #135.
initialize declares fs.readTextFile, fs.writeTextFile and terminal as
false so the adapter uses Claude Code's own tools. request_permission is
a stub that declines until #139.
terminate() sends session/cancel, waits, SIGTERMs, then SIGKILLs. The
child stays in the daemon's process group and gets SIGTERM on daemon
death via PR_SET_PDEATHSIG. CLAUDECODE is stripped from the child env:
the adapter refuses to start when it believes it is nested in another
Claude Code session, which is how a daemon launched from one behaves.
Tests run against tests/fake_acp_agent.py, a JSON-RPC-over-stdio stand-in
that logs what it receives; real_adapter_handshake is #[ignore]d.
A Structured session's adapter hands back an ACP session id from
session/new; a respawn will pass it to session/load (#135). This is a
new field rather than #161's agent_session_id: that one is the Claude
Code conversation id the activity hook reports and `claude --resume`
takes, minted by Claude Code itself. The ACP id is minted by the
adapter, only exists for Structured sessions, and is what the adapter
takes back. Overloading one field with two ids from two processes
would leave a respawn unable to tell which kind it holds.
49ab7e9 made a new-branch worktree add suffix the branch name instead of
failing, and documented it, but left two tests asserting the old hard
rejection. The rejection contract (#66) still holds for an existing
branch that is already checked out, so that test now exercises that
case.
Adds `SessionView { Terminal, Structured }` to maestro-protocol, a
defaulted `view` field on `Session` and `CreateSessionRequest`, and
sidecar round-trip tests for both a legacy payload with no `view` key
and a `Structured` session. Nothing reads the field yet; every other
struct literal across the workspace picks up `Default::default()`
(Terminal) to keep compiling.
Gitea #133.
git worktree add -b fails outright when the branch already exists, even
if the target directory is free. A random fast-forward handle commonly
collides with a branch an earlier, already-removed worktree left behind.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
The hook installer only ever wrote ~/.claude/settings.json, so a codex
session reported no activity and never posted its first prompt — leaving
it on the daemon's generated handle with no haiku name.
Adds a Target enum (claude-code, codex) over the hook file's location and
event set. One script and one marker serve both; codex swaps Notification
for PermissionRequest and has no StopFailure, so it never reports Error.
Both agents spell the prompt event UserPromptSubmit and put the text in
the payload's `prompt`, so autoname needs no change.
Codex will not run a hook it has not been told to trust; `maestro hooks
install` says so, and the approval stays the operator's.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Every attached client pushed its own dimensions straight to
`tmux resize-window`, so two clients of different sizes resized the
shared window past each other on every report — 198 resizes in 90
seconds on a four-session daemon. Each flip repaints a full-screen agent
at a new height and orphans the rows it drew at the old one, leaving a
stale input box the operator types into and reads as a stuck pane.
The window now goes to the smallest attached client per axis, tmux's own
rule for multiple clients, and only when that size actually changes.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The TUI read its leader from `[tui].leader`; the web used `DEFAULT_LEADER`,
a module constant. Rebinding one left the other on `ctrl-a`, so the two
escape hatches hanging off the leader (`leader ctrl-k`, `leader leader`)
moved on one client and not the other.
`GET /api/info` now carries `leader`. The web binds that when it arrives and
rebuilds the leader-owning `CommandSet`, so the `?` overlay and the command
bar render the configured chord; an absent or empty field — an older daemon,
or an unreachable `/api/info` — leaves it on `ctrl-a`.
Chose the existing `[tui].leader` served as-is over a new shared key: one
knob, no precedence rules between two spellings, and no migration for
operators who already set it. The name is a wart — it is the leader for every
client now, not the TUI's — and `docs/COMMAND-BAR.md` and `docs/INSTALL.md`
say so.
ctrl-a is consumed as the leader on every press, so it never reaches the
shell and readline's beginning-of-line was gone with no way back — unlike
Ctrl-K, which already has an escape hatch. Fix it the way tmux does:
pressing the leader twice sends the literal leader byte to the focused
pane and clears the pending chord, on both web and TUI, built from the
configured leader so a rebound leader carries it.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
ctrl-k now opens the command bar in both modes (#153), taking readline's
kill-to-end-of-line away from every pane. Leader then ctrl-k — ctrl-a
ctrl-k by default — sends the literal 0x0B byte to the focused pane
instead, built from the configured leader so rebinding it rebinds the
chord too.
Chording a second binding off the leader exposed a gap in ChordState:
a key bound both alone and as a chord prefix (ctrl-a alone, and ctrl-a
ctrl-k) would get stuck waiting on the next keystroke, discarding its
own meaning the moment that keystroke didn't continue the chord. Feed
now resolves the held prefix to its own binding when that happens and
hands the breaking key back via take_replay for the caller to process
fresh — otherwise, pressing leader then any other key would never
leave Insert mode again.
`tab` on a highlighted session opens its action page inside the command
bar: kill, rename, spawn similar, still fuzzy-searchable, with the
overlay titled after the session. Escape or backspace at an empty query
returns to the result list with the previous query intact.
Each action raises the flow the TUI already has — `KillConfirm`, the
rename prompt pre-filled, and the spawn form, now seeded from that
session's directory and command rather than the focused one.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
`>` narrows to commands from the live keymap for the current mode, matched
on the action's help_label only — the chord is displayed, never matched, so
results cannot depend on a user's rebinds. Rows dedupe by action, not by
chord: the default map binds both `enter` and `tab` to AttachHighlighted,
and AttachHighlighted and ToggleAttachHighlighted share a label because they
are the same match arm.
`~` narrows to recent dirs plus the project list. Enter takes the
fast-forward path pinned to that directory, with no recency fallback, so a
directory outside the project scan opens the pre-filled form rather than
spawning elsewhere.
Both fall out of parsing the query line, as in web/src/commandBar.ts: a
sigil is only a sigil in first position, and backspacing past one is just
the query becoming mixed again.
Opens from Normal and Insert mode as an ordinary keymap entry, so it is
rebindable and the help overlay picks it up from the live keymap. An
empty query lists sessions most recently focused first, matching the web
bar's `orderSessionsForBar`; typing ranks through `matcher.rs` with the
matched ranges highlighted, grouped under kind headers.
Enter attaches through the same path the sidebar's attach uses, Escape
closes and leaves focus and mode where they were.
Ports the docs/COMMAND-BAR.md scoring algorithm (dynamic program over
query/text positions, word-boundary bonuses, whole-field adjustment,
four-level tie-break) to crates/maestro-tui, matching web/src/matcher.ts.
Fixtures F1-F12 from the doc are ported as tests, asserting order,
scores, field and ranges.
The web list still said "open in the focused pane" with a 1px ring, the last
edge left in it, and the TUI said it with the ● glyph alone. Both now use a
background.
Web: `--accent-fill`, the same 30% recipe `--preview-tint` uses, so the two
tiers differ by hue rather than by strength. The row rules are also rewritten
as one ordered run of equal-specificity `:where()` selectors, which is what
`docs/SIDEBAR.md` claims the ladder is — `.is-open` moves in from the bottom
of the multi-view block, where source order had been deciding its precedence
by accident.
TUI: `palette.pane_row_bg()`, blended partway from the card surface to the
selection surface, so the sidebar reads unfilled / half filled / filled. It
borrows the selection surface's foregrounds, since those are the ones picked
for a filled row, and the cursor still wins on its own row. Blended rather
than shipped as a third `Surface` — that would be eight more colours in each
of seventeen presets — with palette tests asserting it stays a visible step
from both surfaces and stays legible in every preset.
Add Session::last_focused_at, shipped on the wire like every other
field. Both POST /api/sessions/:id/focus (focused: true) and opening
/api/attach/:id stamp it and emit SessionUpdated, so a client that
only attaches sorts the same as one that reports focus explicitly.
Losing focus does not clear it, and a never-focused session has None,
which is what lets the command bar's empty-query ordering
(docs/COMMAND-BAR.md) sort it after every session that has focused.
The pane waited 150 ms on hover and 60 ms on a keyboard step before
following the sidebar highlight, so every preview paid for the fast walk
it might have been. Replace both with a leading-edge 120 ms gap: the
first move shows at once, a burst is coalesced to its start and its end.
The row's own hover no longer eases either — `background` was on a 120 ms
transition, which is what read as lag on the row itself.
The sidebar also rebuilt its whole tree on every `SessionUpdated`, a
couple of times a minute per live session, dropping `:hover` from the row
under the pointer. Refreshes that change nothing structural now patch the
activity rail and meta line into the standing rows instead.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Insert mode bound `esc` to LeaveInsert, so full-screen apps in a session
(vim, Claude Code) never saw Escape — the only way out of Insert is now
the leader, which rebinding still follows.
A key bound to `Noop` in Insert mode is also forwarded to the pty now.
`Noop` documents itself as the way to suppress a default binding, and in
Insert mode an unbound key belongs to the session, so this is the escape
hatch for taking any other default back.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The sidebar highlight now shows its session in the pane after a 120ms
debounce, attaching if nothing already is, without the tab taking it on
and without reporting focus to the daemon. Preview-only attaches stay
warm behind the current one, bounded at three.
docs/PREVIEW.md.
Spawn was `s`/`S` in the TUI and `n`/`N` in the web client; the TUI now
uses `n`/`N` too. The web sidebar gains `c`, `r` and `x` — collapse,
rename, kill — which the TUI already bound.
docs/SIDEBAR.md now carries the shared table, so a new binding has one
place to land on both clients.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The TUI repainted on every 33ms tick whether or not anything had
changed — measured at ~725 bytes/s of escape sequences with nothing
attached and nothing happening. Over a remote terminal that is constant
traffic for no visible change.
Output was also unbuffered: `CrosstermBackend::new(std::io::stdout())`
issues one `write!` per cell run, 3551 of them for a 200x50 repaint,
each taking the global stdout lock. A BufWriter collapses a frame into
one flush (measured: 3551 writes -> 2).
Frames are now keyed on the catalog version, the active pane's screen
revision, an input epoch and the terminal size, and skipped when the key
is unchanged. Session cards carry an advancing duration, so a frame
still never stands for longer than a second.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The daemon's auto-install skipped on presence, so a hook installed
before #131 never gained UserPromptSubmit and no session was ever named
after its prompt.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The TUI arm of the umbrella binary hardcoded http://127.0.0.1:7070 instead
of going through MaestroConfigFile::resolve_daemon_url, which every CLI
subcommand uses. On a host whose [daemon].bind_addr is not the default port,
`maestro list` worked and `maestro tui` could not connect.
RunConfig.daemon_url is now the override rather than the final URL; the TUI
resolves it against the config file it already loads for its UI settings.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Keystrokes and PTY echo are tiny frames. With Nagle on, each can sit up to
40ms waiting for an ACK before the kernel sends it — invisible over
loopback, plainly felt from another device on the LAN.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The browser reached the daemon through maestro-web, a second binary that
served the SPA and reverse-proxied /api/* — two TCP connections and a
WebSocket frame re-encode on every keystroke and every byte of PTY output.
On one machine that hop buys nothing.
The daemon now embeds the bundle and serves it off its own listener, behind
a default-on `web` feature. `--no-default-features` builds without pnpm or
node; MAESTRO_WEB_DIST still takes a prebuilt bundle.
The router's fallback answers an unmatched /api/* with 404 rather than the
SPA shell, so a client expecting JSON gets a status it can read instead of
HTML.
Removes maestro-web and the [web] config section. An existing [web] block
is ignored, not rejected. The installer deletes a leftover
maestro-web.service on upgrade.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Skipping the operator's MCP servers, settings and hooks takes a measured
naming run from 12s to 7s, so the timeout is a margin rather than a race.
Refs #131
The name field starts empty behind a hint, ^R reroll is gone, and the
fetched handle seeds the worktree branch instead. Fast-forward spawns with
no name at all. A card whose name the daemon generated renders muted.
Refs #131
A session's rail glows AwaitingInput even after the operator has already
looked at it. Clients now report pane focus (POST /:id/focus); the daemon
clears the glow to Idle after 5s of continuous focus and keeps it cleared
until the session does new work, so a fresh Active cycle is required to
glow again.
Normal mode only exists while the sidebar is focused (Insert mode always
pairs with Focus::Pane), so bare `q` and `tab` there never reach a pty:
`q` quits, `tab` focuses the highlighted pane (same as Enter). `esc` is
added as a second, fixed way out of Insert alongside the leader.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
POST /api/sessions/:id/prompt claims a session's one naming attempt and runs
claude -p --model haiku in the background; the answer is reduced to a short
slug and replaces the generated handle, settling the session at Haiku. A
chosen name, a rename mid-run, a missing binary or a timeout all leave the
handle alone. The prompt goes to the child on stdin and is never logged.
Refs #131
Adds a UserPromptSubmit entry that forwards the hook payload verbatim to
/api/sessions/:id/prompt, gated by a haiku-autoname feature that the daemon,
the CLI and the umbrella binary all forward. With it off no hook points at
the route.
Refs #131
The spawn form no longer prefills a name, so create() fills one in: a
requested name is marked User, an absent one gets a generated handle marked
Generated. Rename always settles to User.
Refs #131
Miguel Palhas