Claude Code lets a subscription's one-hour prompt cache expire after an hour idle, and the first turn back re-writes the whole conversation at the cache-write rate; nearly every gap over an hour goes cold, while the saving is concentrated in gaps of one to four hours. Codeman now sends one short keep-alive turn five minutes before the expiry Claude Code reports on its statusline, so the next real prompt still reads the conversation from cache, and stops two hours (configurable) after the last real prompt.
Codeman owns the idle policy for these sessions, as agreed in Discussion #554: a launch with keep-warm on also carries idleCompaction: false in its one --settings JSON, so Claude Code's own idle compaction cannot replace the conversation with a summary a minute before the nudge. It is decided once per launch (keep-warm on, the statusLine exporter injected, Claude Code 2.1.290 or newer, an unknown version fails closed like --name), recorded on the mux record, and picked up by any relaunch; a launch with keep-warm off is byte-identical. A session switched on after launch is kept warm best-effort and the readout says so.
Every guard lives in one pure rulebook: the opt-in (an agent-spawned session does not inherit the app default), a launch that reports cache state, a warm 1h cache with enough tokens at stake and no compaction this gap, a live pane, no usage-limit pause, no respawn, Ralph, task or waiting agent, idle, no open dialog, an empty composer and no recent keystroke. The nudge is a submit and moves lastSubmitAt, which the response viewer's conversation correlation needs; the cap counts from its own persisted lastRealPromptAt instead. Read My Mind drops the exact nudge text, and the turn's stop hook and idle prompt open pre-acknowledged so nothing alerts. The last cache report and the clock survive a restart, so the timer re-arms.
GitHub Copilot CLI (`copilot`) becomes a run mode: Run menu, welcome
screen, cron and CLI management once it is installed, tmux-backed like
the other agent CLIs, working/idle and the model read off its footer,
launch options through `copilotConfig` (model, resume, continue, the
multi-user-clamped allowAll), Custom Model Endpoints via Copilot's BYOK
variables, MCP sync of its mcp-config.json, a status probe, and the
`codeman agent` verbs (spawn waits for its composer, read returns its
last answer).
The plan-usage quota row that earlier rounds carried was split out on
review and is not part of this merge.
Merge-time commits on top: the wider COPILOT_* clamp and the zh-CN
README rows (maintainer), the Docker sign-in docs, divider rows inside
an answer kept by `agent read`, route tests for the screen-read
last-response branch, and doc drift (Past Sessions, wheelForward JSDoc,
COPILOT_/OMP_ in the env prefix lists).
Verified before merge on Copilot CLI 1.0.98: injected keystrokes,
bracketed paste and Enter reach the TUI in tmux (github/copilot-cli#4180
does not reproduce), a shell tool call completes inside tmux (#4223 does
not reproduce), and a session driven through this branch answered over
send-and-wait.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Add pinCliSessionName (synced, default OFF), re-read at each spawn and respawn.
Off keeps the #457 behaviour; on pins every non-blank tab name as --name.
Update the docs, JSDoc and changeset that stated the old rule.
Copilot (`copilot`, npm @github/copilot) as a stock CLI entry with the same launch options as the other agent CLIs: model (--model), resume (--resume), continue (--continue) and allow-all (--yolo, clamped off for non-granted multi-user owners). Adds a resolver and GET /api/copilot/status probe, the agent-image package, Run-menu/mobile/logo entries, skill docs and tests.
Review round: the footer hint is drawn only at rest, so nothing arms the idle check while the Run this command list is up. The stock.ts comment, wiki, changeset and test now say a pending approval shows as working until answered, and the test pins busy during the confirmation. Adds the antigravity copy of the reattached-pane settle test, and updates the promptGlyph comment and the fixture header.
feat(codex): raise codex approval, question and trust dialogs as BLOCKED
Conflict resolved against #634 in src/session.ts _probePaneWorking():
#624's assignment (a dialog on screen is never work) is kept, and #634's
trust-dialog verdict block (log the hold once, remember the reading)
follows it, before _readWatching().
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The dialog's own prompt glyph armed idle confirmation and the quiet dialog passed it, so a fresh case reported idle about 4 s in. An agent that waited for idle and then prompted typed into the dialog, and its Enter confirmed No, exit and killed the pane.
The scan must schedule its own next read rather than wait for PTY output, but a call inside the throttle was simply dropped. The dialog's own paint is often the last output a launching pane produces and lands inside that throttle, so the session sat on the dialog with no keystroke ever sent. A stopped session no longer scans, since the kill's own output kept it answering a dialog in a dead pane.
sendInput() typed the text of a text+Enter pair with `send-keys -l` and
sent Enter 50 ms later. Codex detects an unmarked paste from a fast
character burst and folds an Enter inside that window into the paste as a
newline, so the prompt stays in the composer unsubmitted. Measured with
codex-cli 0.162.1 in a throwaway tmux pane: 1505 chars stuck 3/3 as
`[Pasted Content 1506 chars]`, a 98-char line stuck 1/4. SubmitVerifier
re-presses Enter for the short case, but it cannot match the long one: the
composer shows the placeholder instead of the prompt's head.
The text now goes in via set-buffer + `paste-buffer -p -d` (markers only
when the pane requested bracketed paste), followed by the separate Enter as
before. Same pane, same texts: 4/4 and 3/3 submitted. Each send gets its
own named buffer; a paste that fails (pane or server gone) deletes it,
since tmux never trims named buffers on its own.
Also escapes a trailing `;` in the text, which tmux treats as a command
separator (`send-keys -l 'x;'` typed `x`, a lone `;` typed nothing).
sendInputToPane() has no callers in src/ and is left as it was.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
While a codex dialog was up, the session re-read its screen every 2 s with
a synchronous `tmux capture-pane`, for as long as the dialog stayed up. That
read costs about 8 ms of blocked event loop (measured 2026-10-14), and an
approval can sit unanswered overnight: 1,800 execs an hour per waiting
session, measured as 300 in 10 quiet minutes in the unit test.
A dialog leaves the screen only through a repaint, and tmux sends every
repaint to the session's client as PTY output. So while the last reading
saw the dialog and nothing has been drawn since, it is still up and the
read is skipped; a safety-net read still runs every 30 s
(BLOCKED_QUIET_RECHECK_MS). The clear keeps its 2 to 4 s latency, because
the answer itself is output. Never skipped once a reading has missed the
dialog: tmux updates the pane before it sends the repaint, and the
confirming read must not wait for output that may already have arrived.
The session-state tests now change the fake screen the way a real pane
changes (with output), and two new tests pin the skip and the confirming
read (the second fails with the guard removed).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Codex is the one CLI Codeman launches with its approvals on by default, and
it declares `hooks: 'none'`, so nothing tells Codeman it stopped to ask. A
codex session parked on "Would you like to run the following command?" read
as plainly idle, and one parked on a question read as WORKING, because the
question's own key hints end in `esc to interrupt`, codex's working line.
This adds a generic, optional registry field,
`capabilities.workDetect.blockedLine` (plus `blockedLines`, the window), and
the session state that reads it:
- `isBlockedScreen()` (session-activity.ts, pure) joins the last
`blockedLines` non-blank rows of a capture and tests the pattern once, so a
pattern can demand several markers and anchor the dialog's footer on the
end of the window.
- `Session` reads it off the same capture the idle confirmation already
takes, emits `blockedChanged`, and exposes `screenBlocked`. It raises on
one reading, and while a dialog is up re-reads the screen every 2 s,
because answering resumes the turn and a busy pane would not reach the
settle path again until the turn ended. It clears only after two readable
captures in a row without the dialog, so a capture taken mid-repaint cannot
clear and re-raise it; an unreadable capture changes nothing. A blocked
screen wins over the working line in the same capture.
- The pattern goes through `compileVersionRegex()` in both the schema refine
and at use (lazily cached), like every other config regex.
Codex's pattern was measured on a live codex-cli 0.162.0 pane on 2026-10-14
(`-a on-request -s read-only`, widths 200 down to 36): it needs a `› N.`
option row and then the dialog's key-hint row as the last row, allowing for
that row to wrap below about 60 columns. Command, edit and network-escalation
approvals, single and multi-question dialogs and the startup trust prompt
match; the idle and working views, menus the user opened (`/model`,
`/permissions`, `/review`), the composer holding the dialogs' words and a
transcript quoting them verbatim do not. `Implement this plan?` is left out
on purpose: its footer is the same as `/model`'s and the turn has ended
there. The captures are kept as fixtures (trailing whitespace trimmed, the
scratch path replaced).
Nothing consumes `blockedChanged` yet; the next commit raises the alert.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude Code's tracker nulls recache_tokens_if_cold at a compaction or a tool-result clearing and only counts the expected rebuild at the next request, which also closes the window, so no report carried both and the old rule never fired. Stamp on the first warm null report after one that carried a value, carry it through idle re-reports and drop it at the next request; the readout now says "Context compacted at HH:MM" since a clearing opens the same window.
Also persist a changed report so a crash restores the latest one, rename the session field to statusLineExporter (statusLineTelemetry is the registry capability) and document its values, word the stake as re-cached, move the readout block above the SSE staleness heading, and drop keep-warm mentions until it lands.
Claude Code's statusline has carried a prompt_cache object since v2.1.251, and Codeman's statusline exporter already received it, but StatusTelemetrySchema stripped the whole object, so Codeman never knew whether a session's conversation was still cached. Keep it and show it.
StatusTelemetrySchema now lists every documented scalar field null-tolerant (several are documented as possibly null, and the object gains fields across CLI versions, so a strict shape would 400 the post and blank the footer), while parsePromptCache normalizes only the subset Codeman stores, so a hostile localhost payload cannot bloat state. Session.promptCache rides toState with a change-only session:promptCache event, and the route broadcasts the stored object.
The Session Options (Respawn tab) readout shows warm-until or cold with the re-read stake, the cache hit rate, the last miss's causes, and, for Claude Code's own idle compaction (recache_tokens_if_cold null while expected_rebuilds rose), a "compacted at HH:MM" note that is never raised as an alert. setPromptCache stamps that on a new object rather than mutating the parsed report, so the stamp reaches the broadcast. Sessions without a statusline of their own also get a cache:until HH:MM / cache:cold group in the terminal footer. The last report is restored on a Codeman restart (like displayModel) so the readout is not blank until the pane's next render; the readout is time-aware, so a value whose expiry has passed reads as cold, not stale-warm. A launch without the exporter records why (collection off, workspace statusline, remote) so the row says the cause instead of "not reported yet" forever.
This is the sensor a later opt-in keep-warm would read, and the TTL it reports is the source of truth rather than the plan tier.
Closing a tab took ~0.55-0.7s on an idle machine, more with child
processes or recently active subagents. Most of it was waiting:
- The web UI kept the tab until DELETE returned, then removed it on the
100ms tab-render debounce, which every session update restarts.
closeSession() is now optimistic: the tab, tile and split go and the
next session is selected before the request is sent, rendered at
once. A refused delete (checked with a GET, since a delete can land
and lose its reply) puts the row back at its old index with the
error toast. This also fixes a latent bug: _apiDelete never throws,
so an HTTP error used to report "Session closed" while the session
kept running. SSE upserts skip ids that are being closed.
- The kill path slept fixed intervals (100ms PTY grace, 200ms for the
pane's children, 100ms for the process group) and verified in 100ms
steps. waitForProcessesExit() (utils/process-exit-wait.ts) keeps
every deadline but returns once the processes are gone, counting a
zombie as exited. Signal decisions keep kill(pid, 0).
- tmux kill-session and the pane-pid lookup ran via execSync, freezing
the server for ~70ms per close. Now async.
- killSubagentsForSession() ran a full `pgrep -f claude` scan per
active/idle subagent (~85ms each with ~100 matching processes). It
now scans once for all of them.
Measured on an isolated instance: click to tab gone 540-690ms -> 58-95ms;
DELETE of a claude session ~450ms -> ~200-260ms.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Comment and doc corrections that #555 made stale, no behaviour change.
- stock.ts: the grok and omp altScreen comments compared their strip to
opencode's, which is now strip-mux-and-mouse rather than the narrow
strip. Grok now says it shares antigravity's strip until measured, and
omp drops opencode from its comparison.
- terminal-ui.js: the touch-tap comment named Claude/Codex/Gemini as the
stripped modes, but the gate is now the cliMouseTracking flag alone and
covers opencode too, so it names the two stripping flavours instead.
- src/types/session.ts: the cliMouseTracking JSDoc (the flag the browser
now gates on exclusively) listed only claude/codex/gemini; it now names
the strip-full and strip-mux-and-mouse modes, including opencode under
tmux.
- src/session.ts: the usesMux getter doc now names isMuxMouseStripMode,
since the replay strip passes usesMux to it as well.
- docs/architecture-invariants.md: the narrow-strip list gains
grok/deepseek/omp (matching the PR's own CLAUDE.md line), the
"must REMEMBER" heading covers both DECSET-stripping flavours, and the
cliMouseTracking writer is described as the full-or-mouse branch it
really is.
- docs/wiki/The-Dashboard.md: the user manual said every non-Claude CLI
scrolls locally; opencode's wheel and swipes now page its conversation.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
A gemini session stayed "working" for good after its first turn. Its braille
spinner trips the generic SPINNER_PATTERN and marks the pane working, but only
a composer glyph arms the idle confirmation and gemini declared none, so it
fell back to Claude's `❯`, which gemini never draws.
Measured on live Gemini CLI 0.63.0 panes (capture-pane every 300 ms through
real turns with a shell call at 40, 120 and 200 columns, YOLO and default
approval mode, plus the raw PTY stream). The turns ran against a local
stand-in for the Gemini API (GOOGLE_GEMINI_BASE_URL, which Codeman's custom
endpoint support already sets), since the CLI's TUI does not depend on the
backend and no account is needed for it:
- The TUI repaints its whole bottom region every frame, composer included,
and the composer sits between a `▄` bar and a `▀` bar; the submitted prompt
is echoed between the same bars. The `▀` bar arms the idle check: every
repaint carries it, tmux's reattach repaint too. The composer's prompt
character is no good: it follows the approval mode (`*` in YOLO), and its
`>` also starts the echoed prompt, which would make the submit verifier
read a submitted prompt as stranded and press Enter again.
- While a turn runs a line `⠦ Thinking... (esc to cancel, 6s)` animates about
every 80 ms (largest gap mid-turn: 214 ms). The label can be any loading
phrase, so the working line is the `(esc to cancel, <n>` suffix, or a
spinner frame opening a line for when a long phrase wraps that suffix.
Nothing at rest matches either.
- A tool confirmation (default mode) replaces the composer, stops the
spinner and the pane goes silent (3.9 s gap), so it reads as idle.
Verified on an isolated instance from this branch: a YOLO turn emitted one
session:working (+170 ms) and one session:idle (2.5 s after the last output);
a default-mode turn went working -> idle while the confirmation waited ->
working once allowed -> idle at the end; after a server restart four restored
gemini panes went busy -> idle in about 4 s; a fresh launch settled in 3 s.
The launch-settle and uncharacterised-CLI tests that used gemini as their
example of a CLI without work detection now use grok and deepseek.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
An omp session stayed "working" for good once a turn started, the same
latch pi had: omp's braille spinner trips the SPINNER_PATTERN fast path,
and only a composer glyph arms the idle confirmation. omp declared none,
so it fell back to Claude's `❯`, which omp never draws once its setup
wizard is done.
Measured on live omp 18.8.6 and 18.0.11 panes, holding a turn open
against a local endpoint that never answers: the input row is `╰─ <text>`
and is redrawn at submit, at the end of a turn, at launch and on
reattach. While a turn runs, the status bar's leading `π` becomes a
braille spinner plus the elapsed time (` ⠼ 14s > ⬢ model > ...`; 18.0.11
pads it with two spaces, past a minute it reads `1m`), with a
`⎋ Working…` row above it. The registry entry now names the input row as
the glyph and either working signal as the working line.
The glyph also switches the submit verifier on for omp, which reads the
input row the way it reads Claude's composer. A prompt sent mid-turn goes
to omp's Steering queue and clears the row, so the verifier stands down.
Text left in the row after an Enter is the one case it re-presses.
Verified end to end on a sandboxed instance (own HOME and PATH, omp
18.8.6): session:idle at launch, session:working during a turn,
session:idle about 3 s after it ended, and a restored pane settled idle
about 3 s after a server restart.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
An opencode session that ran a tool stayed "working" for good. The running
tool row draws a braille spinner (`⠋ Sleep for 12 seconds...`), which trips
the generic SPINNER_PATTERN and marks the pane working, but only a composer
glyph arms the idle confirmation and opencode declared none, so it fell back
to Claude's `❯`, which opencode never draws. Measured on an isolated
instance: a 16 s turn latched busy/isWorking for the rest of the session.
A text-only turn had the opposite problem and never showed as working.
Measured on live opencode 1.3.0 panes (capture-pane every 250-300 ms through
real turns at 40, 60, 120 and 200 columns, plus the raw PTY stream):
- Every composer row starts with a `┃` bar, and the submitted prompt lands in
the transcript with the same bar, so a turn's first repaint arms the idle
check, and tmux's reattach repaint does the same for a restored pane.
- While a turn runs the footer row starts with an 8-cell knight-rider
spinner, `⬝■■■■■■⬝ esc interrupt`, redrawn about every 40 ms (largest
gap mid-turn: 121 ms). At rest the TUI is silent and nothing on screen
draws a `⬝`/`■` run, the 200-column sidebar included.
- The working line is the spinner run, `[⬝■]{8}`, not the label: tmux ships
`esc` and `interrupt` as separate words joined by cursor moves, so the
label never reaches the stream detector, and at 40 columns the footer
wraps it to `esc` / `interr` / `upt`. All 344 spinner chunks of a turn
match the run after Codeman's ANSI strip.
- A pending permission prompt replaces the composer and stops the spinner,
so it reads as idle (waiting on the user).
- The last `┃` row on screen is the composer's agent/model row, or the
permission box's closing bar, never the prompt text, so the submit
verifier stands down and can never press Enter into a dialog.
Verified on an isolated instance from this branch: a 15 s tool turn emitted
exactly one session:working (+271 ms) and one session:idle (3 s after the
spinner stopped); a permission prompt read idle and the allowed turn went
working -> idle; after a server restart both restored opencode panes (one
at rest, one on a permission prompt) went busy -> idle in about 4 s; a
fresh launch reached an open page as idle in 3 s.
The launch-settle tests that used opencode as their example of a CLI
without work detection now use gemini and antigravity.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
40560ace made the 3 s launch settle announce its idle. For a CLI with
work detection that edge could now land in the middle of a turn: a
prompt sent ~2.9 s after launch has not been marked working yet (the
working line goes through the deferred parsers), so the settle called
the pane idle and a send-and-wait registered for that prompt resolved
before the turn even started. Before 40560ace the settle was silent, so
this edge is new.
The settle now also leaves a pane to `_confirmIdle()` when a prompt was
submitted since the timer was armed, the same way it already does for a
pane marked working; that confirmation reads the screen before it ends
the turn. A CLI without work detection still settles: nothing else ever
would.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
A pi session stayed "working" for good once a turn started. pi's braille
spinner trips the SPINNER_PATTERN fast path, which marks the pane working,
but only a composer glyph arms the idle confirmation and pi declared none,
so it fell back to Claude's `❯`, which pi never draws. Measured on beta136:
an errored turn stayed busy/isWorking for 3+ minutes after pi was back at
rest.
pi has no composer glyph. Measured on a live pi 1.1.0 pane (capture-pane
every 250 ms through a turn): its composer sits between two `─` rules, and
while a turn runs it rewrites the top rule as `── ⠏ Working ───` on every
frame. The registry entry now names the rule as the glyph that arms the
check and a spinner frame inside it as the working line.
The same glyph settles a reattached pi pane (a restored pane gets no launch
timer): tmux's reattach repaint carries `─`, which arms the confirmation.
The submit verifier reads the last rule, finds no prompt text and stands
down, so it can never press Enter on a pi pane.
Verified on an isolated instance from this branch: a real pi turn emitted
session:working then session:idle, and after a server restart the restored
pi pane went busy -> idle in about 3 s.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
A fresh codex, pi or opencode tile could spin "working" forever while the
pane sat at its composer. The external-CLI launch timer set the status to
idle WITHOUT an event, only `needsRefresh`. When the launch paint never
marked the pane working, the later idle confirmation found the status
already idle and announced nothing either, so every open browser kept the
`busy` from the spawn broadcast. A reload fixed it, which is why only open
pages were stuck. Measured on the 1.36.0 beta: w8 (codex) had
`lastPromptTime: 0`, i.e. no idle edge ever, and a page held a fresh pi
session at `busy` while GET /api/sessions said `idle`. pi and opencode hit
it on every launch (no work detection, so the timer is all they have);
codex only when its launch paint lost the race against the timer.
- `_concludeIdle()` is now the one place a pane is concluded idle: status,
working flag and prompt stamp change together and `idle` is emitted
(session:idle + state broadcast). `_confirmIdle()` uses it too.
- The launch settle (`_settlePaneStartup`) concludes a pane still in its
spawn-time `busy`. A pane already working is left to `_confirmIdle()`
only when its CLI declares `capabilities.workDetect`; for the rest the
timer is the only thing that can settle it, so it also clears a working
flag a launch spinner glyph latched.
- `_armPaneSettle()` also arms it for a RESTORED pane (Codeman restart,
auto-reattach, tile Attach) of a CLI without work detection (at this
commit shell, opencode, gemini, antigravity, pi, grok, deepseek and omp),
which used to stay `busy` server-side with nothing to clear it. Restored claude and
codex panes stay on their composer glyph, so a restart mid-turn is never
called idle. No `needsRefresh` there: an attach refetches by itself.
Pre-existing since the OpenCode integration, not a regression of this
release. Restore-path gap reported by the opencode and pi sessions working
the same symptom.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
opencode's TUI enables mouse tracking. tmux runs with `mouse off`, so it passes
the PANE's DECSETs straight through to the tmux client, and the browser's xterm
obeyed them: `mouseTrackingMode` flipped to 'any' (measured 62 none / 18 any over
16s) and xterm then reported DRAGS to the TUI instead of selecting locally.
In that state marking text produced no selection at all, so copy-on-select
silently did nothing (5/5 dead drags while `any`), and the obvious fallback —
Ctrl+C — is opencode's `app_exit`, which ended the session. Both were hit here.
opencode needs the middle strip: alt-screen toggles AND mouse DECSETs, but NOT
`3J` (a TUI is not a `clear` consumer). That is `altScreen: 'strip-mux-and-mouse'`
+ `isMuxMouseStripMode`, applied to the live stream (session.ts) and the replay
of a stored buffer, now the exported `stripReplayBuffer()` (session-routes.ts).
The browser's mouse-report gate keeps no mode list any more:
`_shouldReportMouseToCli()` reads only `cliMouseTracking`. The server sets that
flag solely in the mouse-strip branch (`_recordStrippedMouseMode`, one caller),
so it can only be true for a mode whose DECSETs are stripped, and whichever modes
the registry strips, the browser follows. Clicks still reach opencode through the
hand-encoded SGR tap it gates.
The `altScreen` JSDoc gets the decision table its three independent choices need
(alt-screen / `3J` / mouse DECSETs), written from the predicates, including that
`preserve` and `strip-mux-only` take the same runtime row. The table is pinned for
every stock CLI, with and without tmux, on both the live strip and the replay
strip, plus the published flag (test/claude-scrollback-strip.test.ts), so the two
halves cannot drift and a mis-ordered replay branch fails.
Docs and comments that still said opencode keeps its mouse reporting or gets the
narrow strip are updated (CLAUDE.md, architecture-invariants, scrollback and
copy-shortcut plans, session.ts, terminal-ui.js).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The digit rule from 21ae48a5 hid the official DeepSeek ids (`deepseek-chat`,
`deepseek-reasoner` carry no digit), so a session on the official route with
the model field on showed the logo alone (its bundle row pins a provider
alone, so the config had nothing either). It also still misread a folder name
with a digit when every field before it was off.
Now the captured field is rejected when it is what the field can be when it is
NOT the model, and read otherwise:
- capabilities.modelDetect.rejectWords (registry data, single tokens, compared
ignoring case; the schema bounds them and requires a screenLine). dsh lists
every effort id its adapters offer (pi-ai THINKING_LEVELS plus the DeepSeek
adapter's off/low/high/max) and the shipped mode ids, from dsh 0.1.1-rc.2 /
dsh-TUI 0.10.0-beta.1. A mode's drawn label (`plan mode`, `full access`,
CJK) can never be one captured field.
- In the shared screen reader, for every CLI: a field equal to the session's
own working-directory basename is the folder, never the model.
Fixtures: `deepseek-chat` and `deepseek-reasoner` with the model field on are
read; every effort id, `default`, `plan mode`, and the folder name first (with
and without a digit) are not; the live qwen footer still reads `qwen3.8-27b`.
Known gaps, all off by default, are named in stock.ts: a custom mode id drawn
raw, a git branch or a one-word session title first, and the non-compact
footer layout (nothing read there; the route config applies).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
displayModel gains a `config` source, ranked below any report from the running
CLI and above the launch model: custom endpoint, then statusline or screen,
then config, then launch, then nothing. The screen still wins whenever it
names a model, since that is what the running TUI uses.
- Registry data: capabilities.modelDetect gains `configResolver`, a NAMED
reader (src/model-config-resolvers.ts), like a launcher profile; dsh names
'deepseek-route' (the reader from the previous commit). `screenLine` becomes
optional; the schema refuses a modelDetect naming nothing, an unknown
reader, or screenLines without a screenLine.
- Session: the reader runs from _withPaneLifecycle's finally, so at every pane
start, attach and relaunch, with the session's own launch config
(legacyConfigForMode) and env (its clamped overrides, then the server's), so
a per-session DSH_HOME is the home read. Async; a read that lands after a
newer one or after the session stopped is dropped; a remote or docker
session reads nothing locally. A change emits displayModelChanged
(broadcast and persist). Not restored after a restart: the next attach
reads it again, and a restored screen value outranks it.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
A session header can only name the model a session runs if the server knows
it, so SessionState gains `displayModel: { model, source }`, resolved in a pure
module (src/session-display-model.ts), strongest first:
- custom-endpoint: a Custom Model Endpoint Profile's modelId answers the
session, whatever alias the CLI prints;
- statusline / screen: the newest report from the running CLI itself.
Claude's statusLine exporter already posts model.display_name on every
render; the status-telemetry route now records it (only for a CLI with
capabilities.statusLineTelemetry). A CLI whose registry entry declares the
new capabilities.modelDetect has its footer read off the pane capture the
idle/working probe already takes (no extra tmux call), so an in-session
/model switch is followed at the next transition;
- launch: the model the session was launched with (claude's --model or the
app-wide default, another CLI's <cli>Config.model), read where the registry
says the model param lives;
- nothing known: no field, never a placeholder.
modelDetect is registry data, measured on live panes: dsh-TUI's status line
on the row under its composer (qwen3.8-27b on the owner's route) and codex's
`<model> <effort> ·` footer on its last row. Both anchor on chrome only that
CLI draws, over the last rows of the screen only; a transcript line shaped like
the footer is never taken (fixture tests). The pattern goes through
compileVersionRegex() with exactly one capture group, checked at load time.
An unreadable or covered footer keeps the last model (unlike the watching
label: a model does not stop running when something covers its row). Model
text is untrusted: escape sequences and control characters are stripped and it
is capped at 64 characters. A change emits displayModelChanged, broadcast
(session:updated) and persisted; a restart restores a CLI-reported model until
the next report.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Maintainer merge-time fixes for the three PRs that landed together on the
session create / launch / persistence path.
#514 findings (bot verdict merge-with-fixes):
- minor, fixed: SessionState.model was published and persisted for every
mode, so a codex/opencode cron session reported the app-wide Claude
default it never ran on. toState() now emits it only where the new
cliTakesSessionModel() holds (registry capability model.source ===
'claude-settings-file', no CLI id branch). POST /api/sessions uses the
same helper for its non-claude refusal, so refusal and publication cannot
drift. Recovery then hands back undefined for other modes on its own.
- nit, fixed: the `model` schema admitted a leading dash (and '.', '[').
The first character must now be a letter or digit; still a subset of the
registry's model-claude pattern, so nothing accepted is refused at launch.
- nit, fixed (reject, the consistent choice): `model` with
attachRemoteSession was silently dropped. Now a 400 INVALID_INPUT, as
#514 does for non-claude CLIs and quick-start does for remote cases.
advisorModel (#530) gets the same refusal there. effort and envOverrides
keep their older silent ignore on that branch so no existing caller breaks.
#515 finding (bot verdict merge, one nit):
- nit, fixed: the types/session.ts @fileoverview described CodexConfig as
(model, resumeSessionId); it now lists reasoningEffort, bypass,
animations and renderMode too.
Audit of the merged combination (not reviewed before):
- The conflict resolutions in session.ts (toState), types/session.ts,
reboot-restore-routes.ts, server.ts (restoreMuxSessions), CLAUDE.md and
skills/codeman/reference/endpoints.md (+ plugin mirror) keep both sides
correctly; nothing was lost or doubled.
- A claude session with both `model` and `advisorModel` launches with
`--model <id>` and ONE merged `--settings` JSON (ultracode + advisorModel,
or advisorModel beside `--effort <level>`), on the tmux template
(including the resume || new variant and with the statusLine exporter)
and on the direct-PTY fallback. Both values (and effort) survive
restoreMuxSessions onto a dead pane, a reboot restore into a fresh pane,
and restartCli/dead-pane respawn via _buildRespawnPaneOptions.
- quick-start and ralph-loop take no per-session `model` (matching #514's
scope, POST /api/sessions only) and launch on the app-wide default, which
toState now persists for claude, so recovery stays consistent.
- No defect found in the combination beyond the findings above. Noted, not
changed: advisorModel is still published for any mode a caller sends it
with (launch-inert there; the UI and skill send it for claude only).
Tests: test/session-model-recovery.test.ts pins the pair through both
recovery shapes for effort ultracode/high/none, the recovery constructors'
fields, the tmux-manager builder hop, and the codex/opencode/shell
non-publication; test/advisor-model.test.ts pins the launch lines and a
real direct-PTY Session's pty.spawn argv; the route test covers flag-shaped
models, attach refusals and the published fields. Docs: SessionState.model
docstring, the reboot-restore-registry header, the golden test comment and
the CLAUDE.md model/advisor bullets.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude Code's advisor tool (code.claude.com/docs/en/advisor) lets the session's
main model consult a second, stronger model at decision points: before
committing to an approach, on a recurring error, and before declaring a task
done. Codeman can now start claude sessions with one.
- `advisorModel` field on POST /api/sessions, /api/quick-start and
/api/ralph-loop/start (fable, opus, sonnet or a full model id in those
families; haiku cannot advise and is refused). Stored on the session and
persisted, so respawn, boot restore and reboot restore keep it. Remote and
docker quick-starts refuse it, as they refuse effort.
- App Settings, Models, "Advisor" segment (Default / Sonnet / Opus / Fable),
synced as `claudeAdvisorModel`. Run, resume and the Ralph wizard send it.
Default sends nothing, leaving the CLI's own /advisor choice in charge.
- Carried as the `advisorModel` key in the launch's single --settings JSON,
merged with ultracode and the statusLine exporter, never the --advisor
flag: `claude --advisor haiku` exits 1 at launch, which would leave a dead
pane on every respawn, while the settings key degrades to no advisor. A
launch without an advisor is byte-identical to before.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
SessionState now carries the model a session launched with, and both
recovery constructors (mux recovery and reboot restore) pass it back, so a
recovered session relaunches on the same --model rather than the account
default. A top-level `model` sent with any other CLI is refused, since
those take their model in their own config object, and an empty string
means no per-session model, as it does for modelOverride.
CLAUDE.md now describes both routes for a Claude model. The tests pin
which of `model` and `modelOverride` reaches the launch and which the
case file, and that a model opening with a dash renders as --model's value.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
A cron job in "Paste (direct)" input mode wrote `<text>\r` into the pane
in one piece. Claude Code (measured on 2.1.283) takes a burst of about a
hundred characters as a paste, so the `\r` landed as a newline and the
prompt sat unsent on the composer while the run reported `prompt_sent`.
Delivery now lives in `deliverCronPrompt()`. Paste mode writes the text
raw, waits CRON_PASTE_ENTER_DELAY_MS (300 ms), sends `\r` as a separate
write down the same PTY (so it cannot overtake the text), and arms the
session's composer check through the new public
`Session.verifySubmitted()`, which re-presses Enter while the prompt is
still visibly unsent. A session with nothing to write to now fails the
run instead of reporting the prompt as sent. Typed mode is unchanged.
Verified on an isolated instance: a paste-mode job with a 104-character
prompt submitted on the first Enter and Claude answered.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
When Claude hands work to an ultracode workflow or background agents, it
ends its own turn and closes it with `✻ Waiting for 1 dynamic workflow to
finish` instead of `✻ Brewed for 1m 18s`, then resumes by itself when the
workers report back. The pane sits quiet with the composer up, so the idle
probe called the session idle for the whole wait. At phone width the
workflow's progress row also drops its ticking timer, so nothing on screen
changes for minutes.
A new optional registry field, `capabilities.workDetect.awaitingLine`,
names that closing row, and `_probePaneWorking()` counts it as work.
Claude renders the row once from a snapshot and never redraws it, so the
same words stay on screen after the workers finish. `isAwaitingWorkers()`
therefore tests only the newest column-0 row directly above the composer,
never the whole pane and never the PTY stream; a follow-up turn always
puts rows of its own there. The column-0 anchor also keeps an agent from
holding its own tab busy by printing the sentence.
Verified against the live Mac mini pane that reported the bug (2.1.283),
and end to end on an isolated instance: an ultracode session running a
90 s workflow at 46 columns stayed busy through the wait and the
follow-up turn, then went idle 6 s after that turn closed.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* fix(cleanup): keep .claude-images while a sibling session uses the same dir
cleanupSession() recursively removes {workingDir}/.claude-images. That
directory belongs to the working directory rather than to the session, and
several sessions routinely share one case directory, so closing one session
deleted the pasted images a live sibling still referred to.
The removal now runs only when no other live session has the same working
directory. A session that is itself being cleaned up does not count as live,
so two sessions of one case closed together still remove the dir.
Split out ahead of the exited-agent sweep for Ark0N/Codeman#446, which closes
sessions unattended and would otherwise make the loss routine.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* feat(session): close sessions whose agent exited cleanly (#446)
Part 2 of Ark0N/Codeman#446. Part 1 records an exited agent as
SessionState.paneExit. A session whose agent the user ended with /exit is
now closed through cleanupSession(), the same path the X button takes, so
finished sessions stop piling up on the board. The lifecycle log records
the reason as "agent exited cleanly (status 0)", and the conversation stays
resumable from the Resume list.
shouldCloseCleanlyExitedSession() in the new pure module pane-exit-sweep.ts
holds the rule. It closes a session only when all of these hold:
- The exit status is an explicit numeric 0 with no signal. An absent status
is how a SIGKILL presents on tmux 3.2a, so it counts as unknown and the
row stays. A non-zero status or any signal also keeps the row, with the
exit code on the tab.
- Two authoritative pane reads agreed on that exit.
TmuxManager.getPaneExitReadCount() counts them, and a failed, empty or
skipped read neither confirms nor resets the count.
- No start, attach or relaunch is running for the pane.
Session.paneLifecycleInFlight covers _setupOrAttachMuxSession(), whose
dead-pane branch revives an exited pane on purpose, and restartCli().
setPaneExit() already scopes paneExit to local mux-backed sessions, so
remote, docker and direct-PTY sessions are never closed.
planRebootRestore() now refuses a record whose persisted paneExit is a
clean exit. That covers an agent that exited just before a reboot, before
the sweep reached it. A crashed agent's record stays eligible, like its row.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* fix(web): show "exited" on the phone overview and desktop home rail (#446)
Part 1 of Ark0N/Codeman#446 taught the tab strip and the rich rail rows
to say that a session's agent has exited. The phone overview and the
desktop home rail still said "idle", beside a green or pulsing dot.
_mobileOverviewExit() in mobile-overview.js is now the one rule for all
three surfaces, and _sidebarRichRow() uses it as well. It changes what a
row shows and leaves the row's state alone, because the state still picks
the section and the sort order. An exited row gets an "exited" pill, a
neutral dot and row accent, and a duration measured from when the server
first saw the pane dead. A pending permission prompt or question still
wins, as it does on the tab.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* fix(cleanup): close the gaps review found in the #446 sweep and image guard
Four fixes from a dual review of Ark0N/Codeman#446 part 2.
- The .claude-images guard compares canonical paths, so a sibling that
reaches the same directory through a symlink keeps it. Its comment used to
say that case only missed a deletion; it caused one.
- A detached session counts as a live sibling. DELETE ?killMux=false removes
it from the server's map while its pane keeps running, so the guard now
reads persisted records too, and exempts only sessions being killed rather
than every session in cleaningUp.
- A session being closed refuses startInteractive() and startShell(). The
/interactive route awaits listener setup before the start, and a start
that raced the close could launch a CLI in a tmux session whose record was
then deleted. A failed close clears the mark again.
- The clean-exit sweep tries each exit once, keyed by session id and the
exit's at stamp, so a close that fails is not retried and logged every
two seconds.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* fix(session): keep a clean exit that lands within 10 s of a pane start (#446)
A CLI that prints a startup error ("not logged in", a bad profile, a config
error) and exits 0 used to lose its tab, and the error with it, about 4 s
after launch. The sweep now keeps any clean exit that lands within
CLEAN_EXIT_MIN_PANE_LIFETIME_MS (10 s) of the last start, attach or relaunch
finishing (Session.paneStartedAt, stamped when _withPaneLifecycle ends). The
row stays as "exited (0)" for the user to read and close.
Verified on an isolated instance: a shell that ran `exit 0` 2 s after start
kept its row, one that exited after 13 s was closed.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* fix(sessions): stop pinning the w1-myapp placeholder as Claude's /resume title
Local claude spawns passed the tab name as `--name`. That flag is not only the
cross-session peer name: it is also the prompt-box label, the `/resume` picker
entry and the terminal title, and a pinned title stops Claude generating its own
(`customTitle ?? aiTitle`). So every conversation of a case was listed in
`/resume` as the same `w1-myapp`, and none of them got a generated title. On one
workspace, 34 of 34 conversations spawned with `--name` had no ai-title, while
every conversation spawned without it had one.
Only a name the user chose is pinned now: `Session.cliPinnedName` is the name
when `nameSource === 'manual'`, carried to the builders as a separate `cliName`
so the tab/mux name is untouched. Placeholder and auto names let Claude title
the conversation again.
A rename in Codeman also reaches `/resume`: the new name is appended to the
conversation's transcript as the `custom-title` row `/rename` writes (never
creating the file, never writing an empty title). For a pane spawned without
`--name` this holds immediately; a pane spawned with one re-appends its own
title each turn, so there the new name holds from the next spawn.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(sessions): skip no-op renames and docker sessions when syncing the /resume title
A same-name PUT (the Session Options field saves on blur and recomposes the
unchanged placeholder) no longer flips nameSource to manual or appends a
custom-title row, and docker sessions skip the host transcript scan since their
transcript lives in the container. The skill pages no longer use a w<N>- name
as the peer-name example, and the changeset notes the re-append caveat.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* docs: record that nameSource decides --name and renames reach /resume
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: codeman-local <codeman@local>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
- test/setup.ts strips CLAUDE_CONFIG_DIR (pinned in test-env-isolation), so
transcript-fixture tests such as session-custom-model-restart no longer go
red on a machine that exports it for a separate Claude account (#255).
- The vanished-tmux-session branch of _setupOrAttachMuxSession() relaunches
the CLI through createSession() just like a failed respawn, so it now takes
the same resume pin. A genuinely new session is unaffected.
- After a dead-pane respawn of a fallback-chain CLI, _claudeSessionId names
the conversation the walk actually pinned instead of the chain tail, which
the walk may have passed over for lack of a transcript.
- _claudeConfigDir() trims the override like claudeProjectsDir() does.
- The remote-reattach test is labelled as documentation, since the pin
builder's own remote guard would make it pass either way.
- CLAUDE.md: the create-path pin persists through toState() as
resumeSessionId, and the end of the walk adds no pin rather than clearing
the launch seed.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
- While another device holds the pane width (_paneWidthRefused), a resize
now asks for the container's width without applying it locally
(_geometryForResizeRequest: rows follow the container, columns stay at
the PTY's). Fitting first re-wrapped the whole buffer to the container
and back on every 30s mobile retry, and throttledResize ran the
scrollback clear for a resize that brings no redraw. selectSession
clears the flag, since it belongs to the previous pane. New unit tests
run the real mixin against a fake terminal and fail without the fix.
- Session seeds _ptyCols/_ptyRows at spawn (_notePtySpawnGeometry), so a
reattached pane reports its tmux window's real size through ptyGeometry.
- Session.resize's declined-branch comment names ptyGeometry, not the
deleted ptyCols/ptyRows getters.
- Delete the dead terminalGeometryAgrees() and its window export.
- test/xterm-private-api.test.ts header: it pins the exact locked version,
so any bump fails, not only a major.
- The main-terminal fit sweep also matches fitAddon?.fit?.(), and
CLAUDE.md names the modules it actually covers.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
- session.ts: a pane capture that fails now CLEARS the watching label
(and emits watchingChanged so pages drop the badge) instead of keeping
the last one, so a failed capture degrades toward an alert rather than
pre-acknowledging the next real idle prompt. Test updated; invariant
noted in architecture-invariants.
- approvals-ui.js: the header bell counts only unacknowledged items
(pendingApprovalsCount), matching codeman tui's pendingApprovalCount();
pinned in watching-no-alert.test.ts.
- mobile-overview.js: move the orphaned "Pill copy per state" JSDoc back
onto MOBILE_OVERVIEW_PILL_LABEL.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claimed after a manual test that a session comes back from a server restart
without its badge until it next produces output. Measured instead of assumed,
and it is wrong: a codex session with a background terminal still running had
its label back within about 20 seconds of the restart, with no input from
anyone. Reconciliation re-attaches the pane, the attach repaint carries the
composer glyph, the idle confirmation arms on it, and the probe re-reads the
label — the ordinary path, doing the ordinary thing.
What produced the false claim was a session whose monitor had simply expired
while it sat there. Its footer carries no chip, so `watching: null` was the
right answer and there was nothing missing to restore.
Recorded at the field and in the invariants, because the shape of this invites
exactly one wrong fix: a polling timer to keep a value fresh that the pane
already refreshes by itself.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Reported from a manual test: a Codex session went on showing the watching badge
after its background terminal had finished. The server was right and the page
was stale — `Session.watching` changes while the session's status does not, and
nothing broadcast it.
The label is usually SET on the idle transition, which broadcasts anyway, so the
badge always appeared correctly. It CLEARS when the work ends, and a CLI can end
background work without taking a turn: codex repaints its background-terminal
row away and stays idle, so `_confirmIdle()` concludes without emitting `idle`
(that emit is guarded by `wasWorking || isInitialReady`) and no other event
fires. Every open page kept drawing a badge the server had already dropped.
`_readWatching()` now emits `watchingChanged` when, and only when, the label
really changes, and the wiring pushes the session state on it. No new SSE event:
the badge reads off the session payload every surface already has.
A/B measured on an isolated beta with the page loaded and then left untouched.
Without this commit the server dropped the label at t+50s and the page still
showed the badge at t+100s; with it, page and server cleared in the same
ten-second window.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
A dual review (Codex CLI and Claude's code-reviewer, same diff, same brief)
found the trust boundary weaker than the comments around it claimed. Eleven
findings, all applied.
The two blockers were both about who can write the row the label is read from.
Claude's window covered two rows, and the second one is the status line, whose
command a session running with permissions bypassed can write into its own
`.claude/settings.json` — so an agent could print `· 1 monitor ·` onto a row of
its own and silence its own idle alert. The default window is one row now, which
is the footer and nothing else, and the constant says why. Separately, the label
reached `data-tab-meta-sig` unescaped while the row is installed with innerHTML,
which is an injection sink for any config-supplied pattern whose capture group is
permissive; it goes through escapeHtml() like every other untrusted string in
that file.
The Codex entry could not be fixed the same way, and now says so. Its row is
third from the bottom only while a terminal runs; with none running that slot
holds the last row of the transcript, so matching the complete row (with the
`/stop to close` tail, window narrowed to three) raises the bar without closing
it. What contains it is `hooks: 'none'`: no hook event from a codex session
reaches notePrompt(), so a forged label costs a wrong badge and cannot quiet an
alert. The registry comment, `docs/cli-registry.md` and the test all state that
rather than claiming a guarantee the code does not have.
Also from the review: the TUI header badge no longer counts an acknowledged
item, which was the same gate the classifier fix already went through and was
wrong for human acknowledgement too; the TUI approval card reads the quiet
reason and drops to a new `info` tone instead of asking for a reply; the badge
carries an aria-label, because the phone it was built for has no hover target;
the schema refuses `watchingLines` without a `watchingLine`; and the pattern and
its window are resolved together rather than one memoized and one not.
Documentation moved with it. The mechanism now lives in
`docs/architecture-invariants.md` with CLAUDE.md keeping the rule and a pointer,
`docs/wiki/Notifications-And-Approvals.md` tells users why a session stopped
buzzing, and both that page and the changeset name the limitation neither did
before: a question asked in plain prose is not a dialog, so it is silenced along
with the false alarms while background work runs.
Verified live again after the narrowing, on an isolated beta: a Claude session
reported `1 monitor` and took its idle prompt acknowledged, and a Codex session
reported `1 background terminal` against the full-row anchor.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Ten findings from a two-model review of this branch. Both reviewers cleared the
detection logic itself; everything here is a gap around it.
A route that starts a command in a pane now PERSISTS as well as broadcasts.
`/interactive` and `/shell` did neither before, and the pane-exit watcher cannot
cover for them: its next tick finds `paneExit` already cleared in memory,
reports no change and writes nothing, so `state.json` kept saying the agent had
exited for as long as the session stayed quiet. Nothing reads that record for a
decision yet, which is exactly why it had to be fixed now — part 2 is designed
to read it. The `clearPaneExitForNewPane()` docstring claimed its callers
already persisted; that claim was false for these two, and now says what the
caller owes instead.
The watcher's four guards were unreachable by any test. `refreshPaneExits()`
opened with `if (IS_TEST_MODE) return;`, so the read gate, the in-flight
suppression, the generation counter and the empty-read rule could each be
deleted with the whole suite green. The tmux call moves into `readPaneRows()`,
which a test subclass overrides — the shape `runRemoteReconnectTick` already
uses in this file for the same reason — and the test-mode gate moves with it, so
what a test cannot do is spawn a process rather than exercise the bookkeeping.
Each of the four guards now has a test that fails when it is deleted.
The muted status dot turned out to be a specificity fight on three surfaces, not
two. `.tab-status.error` was not excluded, so a session whose agent exited and
whose PTY-exit breaker then tripped lost its red dot to the mute — the state the
browser answers with a "restart it?" confirm, and a needs-you colour by the same
argument that protects the two alert classes. And mobile.css gives a `busy` dot
a 9px size and a green glow with `!important`, while `status` stays `busy` for a
pane whose agent died mid-turn, so a phone rendered a grey dot still wearing the
green halo beside a badge reading "exited". Both measured against the real
stylesheets, both now excluded, and the CSS test reads mobile.css too instead of
being structurally blind to half the problem.
Six comments said things that were not true. Two named the stats collector as
what replaces a restored reading, which is the opposite of the design. The
interval constant argued that 2000 ms keeps a read inside a tick, when the
5000 ms exec timeout means it cannot — which is why the in-flight guard exists.
`MuxSession.discovered` did not say the flag is permanent, though `saveSessions()`
serializes it. The empty-read docstring claimed a distinction that `|| true`
makes impossible. The invariants doc promised more than its drift test delivers.
And CLAUDE.md had no pointer at all, leaving its two hardest prohibitions
("never set `status: 'error'`", "never null the pid") only in the file it is
meant to route people to.
Refs Ark0N/Codeman#446.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Five items, two of which he could only see by running it, plus six smaller
ones. Taking the two blockers first, because both were wrong in ways the
existing tests could not catch.
**Adopting the PTY's rows put the CLI's input line off-screen.** A phone that
took a desktop's 43 rows into a viewport with room for 18 painted an
`.xterm-screen` far taller than its container; xterm's own viewport then had
nothing to scroll, so the bottom of the frame sat below the container with no
gesture able to reach it. Output visible, typing invisible, for as long as the
desktop kept the claim hot. `reconcilePtyGeometry` adopts COLUMNS ONLY now:
width is the axis Ink's wrap and `eraseLines` arithmetic depend on, and keeping
the local row count keeps the composer at the bottom of a viewport that
scrolls. Measured at his geometry — a 360x300 container against a 198x43 pane
now keeps 13 rows, takes 198 columns, paints 202px into a 210px container, and
the input line is inside the box.
**`capture-geometry-retry.browser.test.ts` failed, and CI could not see it**
because the file is in `BROWSER_TEST_GLOBS`. Its premise WAS the clamp —
`getTerminalDimensions()` floored while `fitAddon.fit()` did not — which this
work removes at the source, so it can never hold again at any viewport. The
case survives on its own terms: a pane already drawing at the requested size
must not be replayed. Its premise is now the #464 invariant itself, that the
floored report and the terminal agree, which is a stronger guard because the
clamp coming back fails it here rather than silently restoring the replay loop.
The helper docblock that repeated the old premise is corrected too.
**A session with no pane reported 120x40 and the client adopted it.**
`resize()` writes `_ptyCols`/`_ptyRows` only when `ptyProcess` is set and
nothing seeds them from the spawn geometry, so a dead-pane session still held
the constructor defaults — clicking that tab resized the browser terminal to
120x40 and, on anything narrower, claimed another device owned the pane when
none existed. `Session.ptyGeometry` returns null without a pane, the HTTP route
answers `{}` and the socket sends no frame at all. The raw `ptyCols`/`ptyRows`
getters are deleted rather than left available to be misused again.
**The 40-column floor clipped the pane with nothing able to reach it.** The
affordance keyed on a PTY mismatch, and the floor produces no mismatch — xterm
and the PTY agree throughout, the terminal is simply wider than the box. It
keys on what does not FIT now, MEASURED (`.xterm-screen` against the container,
on the next frame, because the screen takes its width with the render) rather
than derived from cell arithmetic. Measured at 360px: font 24 applies 40
columns and paints 560px, and all 200px of the overhang is reachable.
`.pty-oversized` is renamed `.term-overflows-x`, because after this the old
name describes only one of the two causes.
**"Scroll sideways" did not work on touch for the sessions it targets.**
`touch-action: pan-x` is cancelled before it starts by the `preventDefault()`
`touchstart` calls on every 'content' tap. The terminal's own touchmove handler
pans the container now, with the axis locked once per gesture so a diagonal
cannot pan and scroll at once, and the CSS grants no `touch-action` at all —
handing the browser a pan AS WELL would move the pane twice for one finger on
the taps where that preventDefault does not run. Measured under real touch
dispatch: a 140px swipe reaches `scrollLeft` 140 where it reached 0 before, the
buffer does not move with it, and a vertical swipe still scrolls the scrollback.
Three defects in the above, found while checking it rather than by being told:
- `canPanHorizontally` first tested `scrollWidth > clientWidth` alone, which is
true of a container that is not a scroller — a sideways swipe would have
locked the axis, done nothing, AND suppressed the vertical scroll it should
have been. Gated on the class as well.
- The notice advised scrolling sideways whenever the PTY was wider, including
when it still fitted and nothing scrolled. It is gated on measured overflow,
and on a comparison against the width this container WOULD request rather
than the one it currently holds — once adopted those are equal, so the second
question answers itself false while the condition is still true.
- `_syncTerminalOverflowAffordance` could throw out of `document.getElementById`
before reaching its try block. It runs off every geometry change, so a
cosmetic affordance could have taken the resize down with it.
The smaller items:
- `docs/architecture-invariants.md` no longer explains the equality guard as a
clamp signature; it records what the clamp used to do and why it cannot any
more. Edited by hand — that file is outside the Prettier glob, and letting
Prettier near it rewrote eleven unrelated emphasis markers.
- `throttledResize`'s HTTP fallback reads the reply. It is the path where a
declined resize is least likely to be noticed, because no socket means no
`{"t":"zc"}` frame either.
- The changeset covers the whole release: the geometry work, the queued replay
clear, the renderer watchdog, the body-covering fetch deadline, the WebSocket
output-gap reconcile, the build-generated service-worker precache and
per-build cache key, and the crash-trail hygiene.
- `@xterm/headless` is declared in the root devDependencies instead of being
reached through workspace hoisting.
- The output-gap marker is cleared after any response arrives, not only when
the capture was non-empty: a server that answers with an empty capture HAS
reconciled us, and leaving the marker set refetched on every reconnect.
- `e587d845`'s message claimed a test asserted the failed-load copy against the
built asset. It did not — that assertion lived in a probe deleted with the
other scratch scripts, so the claim was false when it was written. There is a
real test now, and it reads the source rather than `dist/`, because `dist/` is
not committed and a test that skips when it is absent would pass for the wrong
reason in CI.
`Session.ptyGeometry` gets behavioural coverage against the real class in
`session-resize-arbitration.test.ts` rather than a source guard, including the
contrast — a pane that does exist still reports, and still follows a resize —
so "always null" would fail it.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Issue #464, "text gets muffled sometimes, in both TUI default and fullscreen".
The screenshot is not a dropped frame or a frozen renderer — it is arithmetic.
Claude Code's TUI wraps its frame at the width the PTY reported and erases the
previous frame by walking the cursor up the rows it believes that frame took. A
browser terminal of a different width makes each logical line occupy more
physical rows than Ink counted, so `eraseLines(n)` clears too few and the new
frame paints over rows nothing erased: doubled lines, and short tool summaries
sitting inside longer prose rows with the prose's tail still visible.
Reproduced against this repo's own xterm before changing anything — a 120-column
PTY against a 62-column terminal renders every wrapped line twice. `test/
terminal-pty-geometry.test.ts` pins that, and pins the clean render at matching
widths beside it, so the assertion cannot be satisfied by code that fixes
nothing.
Four ways the two drifted apart, none of them observable from either end:
1. `fitAddon.fit()` resizes xterm to `proposeDimensions()` RAW while every
server-facing path reported those floored at 40x10. Measured in Chrome at
430px: font size 44 proposed 13 columns, the server was told 40, and xterm
stayed at 13. Three call sites each did their own fit-then-floor, and two
re-read the proposal after the fit — `_shrinkPaddingToFit()` runs exactly
there, so the container had moved.
2. `throttledResize` (keyboard up) and `sendResize` (session detached into its
own window) reflowed locally and withheld only the SIGWINCH. That is the one
combination that cannot be right: a reflow nothing is rendering for buys
nothing and costs correctness. Both now withhold everything, and the
keyboard's settle timer still sends the one resize that stops the PTY going
stale.
3. `setFontSize`/`setFontFamily`/`setFontWeight` move the cell size — a geometry
change — and told the server nothing at all, so raising the font on a phone
left the CLI wrapping at the old column count.
4. `Session.resize` DECLINES a small-viewport request while a desktop connection
holds an active sizing claim, and said nothing, because resize was write-only.
`syncTerminalGeometry()` is now the one function that may change the terminal's
size: it fits, floors and applies as a single step, so the numbers xterm holds
are the numbers the server is told. A test sweeps every module for a bare
`fit()` on the main terminal, and finds exactly one — the owner's own.
For (4) the client cannot win, so it is told the truth instead: both transports
answer a resize with `session.ptyCols`/`ptyRows` (`{"t":"zc"}` on the socket,
the body of the resize POST) and `_onPtyGeometryReport` adopts them. A terminal
that keeps a shape the PTY refused does not render "too narrow", it renders
garbled. Adopting can leave the pane wider than the screen and the container is
`overflow: hidden`, so `.pty-oversized` grants horizontal reach for exactly as
long as the mismatch lasts: correct-and-reachable beats correct-and-clipped
beats garbled. That rule sets both overflow axes and its own `touch-action`
because mobile.css loads later and sets `.terminal-container { overflow:
visible; touch-action: none }` — a bare `overflow-x` would leave overflow-y
computing to `auto` and hand the browser a vertical scroll container the
terminal's touch handler knows nothing about.
Verified in Chrome at 430px against a live server, with a desktop client holding
the claim: the phone adopts 198x43, gets `overflow-x: auto` / `overflow-y:
hidden` / `touch-action: pan-x`, 758px of reach to the right, and keeps its own
vertical scrolling. The pre-fix build was measured in the same harness for the
control.
Two things this deliberately does not do. It does not change who owns the pane
size — the desktop still wins, and `_startMobileResizeRetry` still takes it back
once that goes idle. And `throttledResize` still holds the PTY's shape for the
whole keyboard animation rather than sending a SIGWINCH per step; that decision
predates this and was not re-tested here.
Also in this commit, Ark0N's third-pass review items on #431:
- The response viewer's byte-buffer fallback and `_onSessionClearTerminal` both
used the no-param `/terminal` form, capped only by `terminalBufferMaxBytes`
(32MB) — the largest body the frontend asks for anywhere. One carried no
deadline at all and the other got the 15s tail budget. Both now take the
full-history budget.
- A `?full=1` capture that outruns its deadline falls back to the bounded tail.
The pane is blanked before that fetch, so an abort used to leave a black
rectangle, discard the queued live output and never reach `_connectWs`. A
failed load now still opens the socket, says one dim line where the content
would have been, and clears the tab's spinner — which nothing did, so a failed
select left `aria-busy="true"` set forever.
- `_wsOutputGapSession` is cleared at the repaint that settles it, not in a
`finally` that also ran on the catch. A reconcile that threw, or hit the new
deadline — the flaky link the marker exists for — dropped the gap with nothing
to retry it. `ws.onopen` no longer clears it up front either.
- The replay-clear invariant is pinned in the gate, which is the drift this PR
exists to fix: `_resetTerminalForReplay` must be a queued write and nothing
else, and no module may blank the terminal with a `clear()+reset()` pair.
- `DIAG_ENTRY_MAX_CHARS` replaces the hardcoded 300, bound through a local
first: `CodemanDiag?.x` still throws a ReferenceError when the identifier was
never declared, and that is the one function in the app that must not throw.
- panels-ui's two kill-all clears route through the same helper, and the
xterm-version guard's comment says "resolved lockfile version" rather than
"dependency RANGE", which is what it has pinned since the last round.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
JD