Return CLI-style browser connect status messages from the gateway and render them in the TUI so local Chrome launch attempts are visible instead of ending in a silent delayed failure.
Share Chrome CDP launch helpers between the classic CLI and TUI so default /browser connect uses loopback consistently, retries local Chrome launch, and reports a copyable manual-start command instead of claiming a dead connection.
Keep the latest prompt sticky while the viewport is in live assistant output beyond history, and clear stale sticky state at the real bottom using fresh scroll height.
* fix(tui): honor documented mouse_tracking config key
The TUI runtime was reading display.tui_mouse while docs and user-facing
examples pointed users at display.mouse_tracking. That made persistent
mouse-disable config look like a no-op for users trying to restore native
terminal selection/copy behavior on Linux/SSH/tmux terminals.
Use display.mouse_tracking as the canonical key, keep display.tui_mouse as
a legacy fallback, and have /mouse write the documented key. Both gateway
config.get and client-side config sync now share the same precedence: the
canonical key wins, then the legacy key, then default on.
* review(copilot): align mouse tracking config coercion
- Load gateway config once before deriving display.mouse_tracking state.
- Use key-presence precedence on the TUI client too, so canonical
mouse_tracking wins over legacy tui_mouse even when the value is null.
- Treat numeric 0 as disabled on both gateway and client, matching the
existing string "0" handling.
- Widen ConfigDisplayConfig mouse fields because config.get full returns raw
YAML, not normalized booleans.
This PR groups the TUI fixes that restore macOS Terminal usability and clean up the theme/composer regressions:
- copy transcript selections on macOS drag-release so Terminal.app users can copy while mouse tracking is enabled
- copy composer selections on macOS drag-release; composer selection is internal to TextInput and does not use the global Ink selection bus
- keep IDE Cmd+C forwarding setup macOS-only, and make keybinding conflict checks respect simple when-clause overlap/negation
- force truecolor before chalk initializes (unless NO_COLOR / FORCE_COLOR / HERMES_TUI_TRUECOLOR opt-outs apply) so the default banner keeps its gold/amber/bronze gradient in Terminal.app
- move TUI surfaces onto semantic theme tokens and preserve skin prompt symbols as bare tokens with renderer-owned spacing
- render focused placeholders as dim hint text in TTY mode instead of inverse/selected-looking synthetic cursor text
* feat(tui): pluggable busy-indicator styles (kaomoji/emoji/unicode/ascii)
The status-bar `FaceTicker` rotated through wide-and-variable kaomoji
glyphs (`(｡•́︿•̀｡)`, `( ͡° ͜ʖ ͡°)`, …) every 2.5s. Real display widths range
from ~5 to ~16 columns, so the rest of the bar (cwd, ctx %, voice,
bg counter) shifted on every cycle. Padding the verb alone (#17116)
helped but didn't address the dominant jitter source — the glyph
itself.
Add four indicator styles, configurable + hot-swappable:
* `kaomoji` (default — preserves the existing vibe; verb is now
pad-stable so the only width churn left is the kaomoji itself).
* `emoji` — single 2-col emoji frame (`⚕ 🌀🤔✨🍵🔮`).
* `unicode` — `unicode-animations` braille spinner (1-col, smooth).
* `ascii` — `| / - \` (1-col, max compat).
Wires:
* `display.tui_status_indicator` in `DEFAULT_CONFIG` (default
`kaomoji`).
* New JSON-RPC `config.set/get indicator` keys, narrow allow-list.
* `applyDisplay` reads the field and patches `UiState.indicatorStyle`,
so the existing `mtime` poll picks up `~/.hermes/config.yaml` edits
within ~5s without a TUI restart.
* `/indicator [style]` slash command (alias `/indicator-style`,
subcommand completion `kaomoji|emoji|unicode|ascii`). Bare form
shows the current style; setter fires `config.set` and
optimistically `patchUiState({ indicatorStyle })` so the live TUI
swaps immediately, matching the `/skin` UX.
* `CommandDef("indicator", ..., subcommands=...)` so classic CLI
autocomplete + TUI `complete.slash` both surface it.
* `FaceTicker` decouples spinner cadence from verb cadence — the
glyph runs at the spinner's authored interval (or `FACE_TICK_MS`
for kaomoji), the verb stays on the original 2.5s cycle, and both
re-arm cleanly when style changes.
Tests:
* `normalizeIndicatorStyle` rejects unknown / non-string input.
* `applyDisplay → tui_status_indicator` covers fan-out + fallback.
* `/indicator <style>` hot-swaps `UiState.indicatorStyle` after a
successful `config.set`.
* `/indicator sparkle` rejects with the usage hint and never hits
the gateway.
* Slash-parity matrix gets `'/indicator'` → `config.get`.
Validation:
cd ui-tui && npm run type-check — clean; npm test --run — 398/398.
scripts/run_tests.sh tests/test_tui_gateway_server.py
tests/hermes_cli/test_commands.py — 220/220.
* chore(tui): drop /indicator-style alias to declutter autocomplete
* fix(tui): drop verb-width pad — /indicator handles glyph jitter directly
* fix(tui): unicode indicator style hides the verb (cleanest option)
* refactor(tui): single source of truth for INDICATOR_STYLES; cleaner error format
Round 1 Copilot review on PR #17150:
- Exported `INDICATOR_STYLES` const tuple from `interfaces.ts`;
`IndicatorStyle` union type is derived from it. `useConfigSync`
builds its validation Set from the tuple, and `session.ts` uses it
for both the usage hint and the runtime allow-list — adding/removing
a style now touches one line.
- Backend `config.set indicator` error message: switched
`sorted(allowed)` list repr to `pick one of ascii|emoji|kaomoji|unicode`
(matches the TUI usage hint), and reports the normalized `raw`
instead of the original `value`. Backend allowed tuple now has a
comment pointing back at `INDICATOR_STYLES` so the two stay aligned.
Note: kept the verb portion unpadded per design intent — fixed-width
padding was the exact UX the `/indicator` command was added to remove.
Stable width comes from the glyph; verbs cycling is part of the kawaii
aesthetic. Reply on the verb thread will explain.
* fix(tui): drop type collapse + gate verb timer + DEFAULT_INDICATOR_STYLE
Round 2 Copilot review on PR #17150:
- `tui_status_indicator?: 'ascii' | ... | string` collapses to `string`
in TS — consumers got no narrowing. Documented as plain `string` with
a comment about runtime validation via `normalizeIndicatorStyle`.
- `FaceTicker` always started a 2.5s verb interval, even for the
`unicode` style which hides the verb entirely. Now gated on
`showVerb` from `renderIndicator` — `unicode` stays calm.
Pre-emptive self-review (avoid round 3):
- Three call sites duplicated the literal `'kaomoji'` default
(uiStore, normalizeIndicatorStyle, slash command). Added
`DEFAULT_INDICATOR_STYLE` to interfaces.ts and threaded it through
so changing the default touches one line.
* fix(tui-gateway): normalize config.get indicator output to match TUI render
Round 4 Copilot review on PR #17150: `config.get` for `indicator`
returned the raw `display.tui_status_indicator` value without
validation, so a hand-edited config.yaml with stray casing or an
unknown style would leave `/indicator` printing one thing while
the TUI rendered the kaomoji default (frontend's
`normalizeIndicatorStyle` does this normalization on receive).
Lifted the allow-list to module scope as `_INDICATOR_STYLES` /
`_INDICATOR_DEFAULT`, reused by both `config.set` and `config.get`.
Comment notes the alignment with `INDICATOR_STYLES` /
`DEFAULT_INDICATOR_STYLE` in interfaces.ts so adding/removing a
style is a one-line change on each end.
Tests cover: known value verbatim, casing/whitespace normalize,
unknown→default, unset→default.
* fix(tui-gateway): preserve falsy-input diagnostics in config.set indicator error
Round 5 Copilot review on PR #17150: `raw = str(value or "").strip().lower()`
collapsed any falsy non-string (`0`, `False`, `[]`) to empty string,
so the error message read `unknown indicator: ` with nothing after —
losing the original input.
Switched to `("" if value is None else str(value)).strip().lower()`
so only `None` (the genuine 'no value' case) becomes blank. Used
`{raw!r}` in the error so the diagnostic is unambiguous (`'0'` vs `0`).
Tests:
- known-value happy path (`'EMOJI'` → `'emoji'`)
- falsy non-string inputs (`0` / `False` / `[]`) surface meaningfully
- `None` keeps the blank-repr error
* feat(tui): expand light-terminal auto-detection (HERMES_TUI_THEME, BG hex)
Modern terminals (Ghostty, Warp, iTerm2) don't set COLORFGBG, so the
auto-light path was effectively COLORFGBG-only and silently broken for
many users. Two pragmatic additions, both opt-in, plus a clearer
priority chain:
1. **`HERMES_TUI_THEME=light|dark`** as a symmetric explicit override.
The existing `HERMES_TUI_LIGHT` is fine but reads as boolean noise;
a named theme env var matches `display.skin` muscle memory.
2. **`HERMES_TUI_BACKGROUND` hex/rgb hint.** Lets advanced users
(or a future OSC11 query helper that caches the answer) state a
ground-truth background colour. Decoded to Rec. 709 luma; ≥ 0.6
counts as light.
Priority order is now fully ordered and explainable:
1. `HERMES_TUI_LIGHT` (1/0/true/false/on/off).
2. `HERMES_TUI_THEME=light|dark`.
3. `HERMES_TUI_BACKGROUND` luminance.
4. `COLORFGBG` last field — light slots 7/15 → light, 0–15 → dark
(authoritative when set, so the new TERM_PROGRAM path can never
stomp on a terminal that already volunteered a dark answer).
5. `TERM_PROGRAM` allow-list — empty by default. The slot is left
in place because folks asked for it but populating it risks
wrongly flipping users on Apple_Terminal / iTerm2 dark profiles
to light. Easy to add per terminal once we have signal.
Tests: 5 new cases in `theme.test.ts` covering theme env, background
hex (3- and 6-char), invalid hex falling through, and COLORFGBG taking
precedence over the future allow-list.
Validation: `npm run type-check` clean, `npm test --run` 392/392.
* review(copilot): tighten theme detection comments + drop unnecessary cast
* review(copilot): strict hex regex so partial garbage doesn't slip into luminance
* test(tui): make TERM_PROGRAM allow-list injectable so precedence is provable
Copilot review on PR #17113: `LIGHT_DEFAULT_TERM_PROGRAMS` is empty
in production, so the prior assertion would have passed even if
`detectLightMode` ignored `COLORFGBG` entirely. That defeats the
test's purpose.
`detectLightMode` now takes the allow-list as an optional second
argument (defaults to the production set). The test injects a set
containing `Apple_Terminal`, asserts the allow-list alone WOULD
return light, then asserts `COLORFGBG: '15;0'` overrides it — the
precedence rule is now exercised, not assumed.
* fix(tui): COLORFGBG empty-trailing-field falls through; isolate DEFAULT_THEME tests
Round 2 Copilot review on PR #17113:
1. `Number(colorfgbg.split(';').at(-1))` returns 0 for an empty trailing
field (e.g. `COLORFGBG='15;'` → bg===0), which would have looked
like an authoritative dark slot and incorrectly blocked the
TERM_PROGRAM allow-list. Added a `/^\d+$/` guard before coercion;
non-numeric trailing fields now fall through.
2. Fixed the misleading '0–6 / 8–15 ranges are dark' comment — the
block returns true for bg===15, so the range is actually 0–6 / 8–14.
3. `DEFAULT_THEME` is computed from `process.env` at module-load.
A developer shell with `HERMES_TUI_THEME=light` (or a bright
`HERMES_TUI_BACKGROUND`) would flip it and break local tests.
The DEFAULT_THEME describe blocks now sterilize the relevant env
vars + dynamically import theme.ts (vi.resetModules pattern from
platform.test.ts). fromSkin tests compare against DARK_THEME
directly to decouple them from ambient env.
* test(tui): isolate ALL env-coupled theme symbols, not just DEFAULT_THEME
Round 3 Copilot review on PR #17113: the static top-level imports of
`fromSkin`, `DARK_THEME`, `LIGHT_THEME` evaluated theme.ts before
`importThemeWithCleanEnv` had a chance to clean the env. Because
`fromSkin` closes over `DEFAULT_THEME`, an ambient `HERMES_TUI_THEME=light`
or bright `HERMES_TUI_BACKGROUND` would still flip the base palette
and cause local-only failures.
Removed the static import entirely. Every test now obtains its theme
symbols via `importThemeWithCleanEnv`, including `detectLightMode`
(for consistency, even though it takes env as a parameter).
`fromSkin` tests assert against the cleaned `DEFAULT_THEME` from the
same dynamic import — preserves the actual contract (skins extend the
ambient base palette) without coupling the test to dev-shell state.
Verified by running with HERMES_TUI_THEME=light + HERMES_TUI_BACKGROUND=#ffffff:
all 20 theme tests still pass.
Self-review (avoid round 4):
- Audited other test files importing DEFAULT_THEME (syntax.test.ts,
streamingMarkdown.test.ts, constants.test.ts) — all just pass it as
a parameter or assert palette property existence (works on both
light + dark), so no env coupling there.
* fix(tui): honor display.busy_input_mode in TUI v2
The TUI v2 frontend hard-coded `composerActions.enqueue(full)` whenever
`ui.busy` was true. The classic CLI and gateway adapters honor the
`display.busy_input_mode` config key (`interrupt` | `queue` | `steer`),
but Ink ignored it — sending a message during a long-running turn always
landed in the queue regardless of config. The config default is already
`interrupt` (hermes_cli/config.py), so users who explicitly opted into
that experience were silently stuck on the legacy queue path.
This wires the value through the existing config-sync surface:
* `applyDisplay` now reads `display.busy_input_mode`, defaults to
`interrupt` (matching `_load_busy_input_mode` in tui_gateway), and
drops it into a new `UiState.busyInputMode` field.
* `dispatchSubmission` and the queue-edit fall-through call a shared
`handleBusyInput` helper that branches on the mode:
* `queue` — legacy behavior, append to the queue.
* `steer` — call `session.steer`; on rejection, fall back to
queue with a sys note.
* `interrupt` — `turnController.interruptTurn(...)` then `send()`,
so the new prompt actually moves.
* Mtime polling in `useConfigSync` already re-applies `config.full`, so
flipping `display.busy_input_mode` in `~/.hermes/config.yaml` takes
effect on the next 5s tick without restarting the TUI.
Tests:
* `applyDisplay → busy_input_mode` covers normalization + UiState fan-out.
* `normalizeBusyInputMode` mirrors the Python side's allow-list.
Validation:
* `npm run type-check` (in `ui-tui/`) — clean.
* `npm test --run` (in `ui-tui/`) — 394/394.
* review(copilot): narrow busy_input_mode type, preserve queue order on steer fallback
* review(copilot): clarify handleBusyInput comment (option, not return value)
* fix(tui): default busy_input_mode to queue in TUI (CLI keeps interrupt)
In a full-screen TUI users typically author the next prompt while the
agent is still streaming, so an unintended interrupt loses in-flight
typing. TUI fallback now defaults to `queue`; CLI / messaging
adapters keep `interrupt` as the framework default.
Override per-config via `display.busy_input_mode: interrupt` (or
`steer`) — the normalize/wire path is unchanged, only the missing-
value branch differs from the Python default.
uiStore initial value also flipped to `queue` so first-frame render
before `config.full` lands matches the eventual normalized value.
`turnController.recordMessageComplete` and `recordMessageDelta` both
prioritised `payload.rendered` over `payload.text`. `payload.rendered`
is the Rich-Console output `tui_gateway` builds for terminals that
can't render markdown themselves; the TUI already renders markdown via
`<Md>`. Two real bugs follow:
1. **Final answer garbled when `display.final_response_markdown: render`
is set** (#16391). Raw ANSI escape sequences pass through into the
React tree and the user sees overlapping coloured text instead of
their answer.
2. **Streaming silently drops content.** Per-delta `rendered` is an
*incremental* Rich fragment. The previous code did
`this.bufRef = rendered ?? this.bufRef + text`, which on every tick
replaced the whole accumulated buffer with the latest mid-sequence
ANSI fragment. Long replies arrived truncated and looked
half-painted — easy to miss as "model is being terse" instead of a
client bug.
Fix:
* `recordMessageComplete` now prefers `payload.text`, falling back to
`payload.rendered` only when the gateway elected not to send any.
* `recordMessageDelta` always accumulates `text`; `rendered` is ignored
on the streaming path entirely (Ink does its own markdown render via
`<Md>` / `streamingMarkdown.tsx`).
Tests:
* `prefers raw text over Rich-rendered ANSI on message.complete` —
the assistant message reflects raw markdown, not ANSI.
* `falls back to payload.rendered when text is missing` — preserves
the legacy "no `text`, only ANSI" path used by some adapters.
* `always accumulates raw text in message.delta and ignores rendered` —
pre-fix code would have made this assertion fail because each delta
overwrote the buffer.
Validation: `npm run type-check` clean, `npm test --run` 392/392 pass.
* feat(tui): opt-in auto-resume of the most recent session
`hermes --tui` always forges a fresh session at startup unless the user
sets `HERMES_TUI_RESUME=<id>`. Disconnects, terminal-window crashes,
and accidental Ctrl+D therefore lose every piece of in-flight context
even though `state.db` still has the full history a `/resume` away.
Add an opt-in path that mirrors classic CLI's `hermes -c` muscle
memory: when `display.tui_auto_resume_recent: true` is set in
`~/.hermes/config.yaml`, the TUI looks up the most recent human-facing
session and resumes it instead of starting fresh. Default off so
existing users aren't surprised; explicit `HERMES_TUI_RESUME` always
wins.
Wires:
* New `session.most_recent` JSON-RPC in `tui_gateway/server.py` that
returns the first non-`tool` row from `list_sessions_rich`, or
`{"session_id": null}` when none. Uses the same deny-list as
`session.list` so sub-agent rows can't sneak in.
* `createGatewayEventHandler.handleReady` re-ordered: explicit
`STARTUP_RESUME_ID` first (unchanged), then conditional auto-resume
via `config.get full → display.tui_auto_resume_recent`, then the
legacy `newSession()` fallback. Failures of either RPC fall back
to `newSession()` so the path is always finite.
* Default `display.tui_auto_resume_recent: False` added to
`DEFAULT_CONFIG` in `hermes_cli/config.py` (no `_config_version`
bump per AGENTS.md — deep-merge handles the additive key).
Tests:
* 4 new vitest cases in `createGatewayEventHandler.test.ts` cover
every gate-and-fallback combination (env wins, config off, config
on with hit, config on with miss).
* 3 new pytest cases for `session.most_recent` (denied row skip,
tool-only → null, db-unavailable → null).
Validation:
scripts/run_tests.sh tests/test_tui_gateway_server.py — 93/93.
cd ui-tui && npm run type-check — clean; npm test --run — 393/393.
* review(copilot): fold session.most_recent errors into null + extend ConfigDisplayConfig
* review(copilot): cover RPC-rejection fallbacks in auto-resume tests
* fix(tui): drop stale stream events after ctrl-c interrupt
Once interruptTurn() flips this.interrupted, only recordMessageDelta
short-circuited. recordReasoningDelta/Available, recordToolStart/
Progress/Complete, and recordInlineDiffToolComplete kept populating
turnState until the python loop reached its next _interrupt_requested
check (~1s on busy turns), making it look like ctrl-c was ignored
while late "thinking" + tool calls kept landing in the UI.
Add the same interrupted guard to every stream-side recorder, and
clear the flag at startMessage() so the next turn isn't suppressed
if the previous turn never delivered message.complete.
* fix(tui): guard recordTodos against post-interrupt mutation; fake-timers in test
Copilot review on PR #16706:
1. `recordToolStart` is interruption-guarded, but `tool.start`
handler also calls `recordTodos(payload.todos)` first — so a
late tool.start carrying todos could still mutate `turnState.todos`
after Ctrl-C, leaving ghost rows in the panel. Adds the same
`if (this.interrupted) return` early-exit to `recordTodos` so
*all* tool.start side-effects are dropped post-interrupt.
2. The interrupt test was leaking a real `setTimeout` (interrupt
cooldown) across test files, which could fire later and mutate
uiStore from the wrong test context. Wraps the test in
`vi.useFakeTimers()` + `vi.runAllTimers()` and restores real
timers in finally.
3. Extends the same test with a todos payload on the post-interrupt
tool.start so we have explicit regression coverage for #1.
* fix(tui): guard pushTrail post-interrupt; harden interrupt-test cleanup
Round 2 Copilot review on PR #16706:
1. `tool.generating` events route through `pushTrail`, which was not
interruption-guarded — late events could still write 'drafting …'
into `turnTrail` after Ctrl-C, leaving a stale shimmer in the UI.
Adds the same `if (this.interrupted) return` early-exit.
2. Test cleanup moved `vi.runAllTimers()` into `finally` (before
`vi.useRealTimers()`) so a mid-test assertion failure can't leak
the interrupt-cooldown setTimeout across other test files.
3. Replaced the misleading 'pre-interrupt todos … expected to be
cleared by the interrupt cycle' comment with an accurate one
reflecting current behaviour (interrupt does NOT clear todos).
4. Added an explicit assertion that a post-interrupt `tool.generating`
event does not extend `turnTrail` — regression coverage for #1.
* fix(tui): append gateway stderr tail to start_timeout activity
`gateway.start_timeout` previously published only `cwd` + `python`,
which made TUI startup failures hard to disambiguate. The user saw
`gateway startup timed out · /path/to/python /repo · /logs to inspect`
with no signal whether the actual cause was a wrong python interpreter,
a missing dependency, or a config parse failure.
Plumb a 20-line stderr tail through the event so the most useful lines
land directly in the TUI activity feed, capped to the last 8 non-empty
lines for readability:
* `gatewayClient.ts` — collect `getLogTail(20)` when the readyTimer
fires and attach it as `payload.stderr_tail`.
* `gatewayTypes.ts` — extend the `gateway.start_timeout` event union
with the new optional field.
* `createGatewayEventHandler.ts` — emit the trimmed lines after the
existing `gateway startup timed out` activity entry, classified
`error`.
Tests: regression test in `createGatewayEventHandler.test.ts` checks
that `ModuleNotFoundError` / `FileNotFoundError` lines from the tail
land in `getTurnState().activity` so they show up in the UI immediately.
Validation: `npm run type-check` clean, `npm test --run` 390/390.
* review(copilot): filter blanks before slice and cap stderr tail at 120 chars
- Rename `removeAt` → `removeAtInPlace` and document the mutation
contract; the old name read like a non-mutating helper.
- Hotkey table + queue header: use `Ctrl+X` / `Esc` to match the
rest of the UI (was `⌃X` / `esc`).
- Render the queued header as a single template literal so JSX
text-node whitespace can't sneak into the rendered line.
- Make `Esc` while editing beat the `terminal.hasSelection` clear:
the header promises 'Esc cancel', so an active selection
shouldn't silently consume the keystroke.
Today there's no way to remove a queued message — ↑ loads it for edit,
ctrl-K dispatches the head, but a draft you no longer want stays put
forever. ctrl-C just clears the composer and exits edit mode without
touching the queue.
Two new bindings, both gated on queueEditIdx !== null so they're
inert when the user isn't pointing at a queue item:
- ctrl-X — delete the queue item being edited, clear composer, exit
edit mode. "cut" matches the mental model and doesn't collide with
any existing binding.
- esc — cancel the edit (composer clears, item stays in queue).
Mirrors ctrl-C's existing behavior so muscle memory has two paths.
Header line now reads `queued (3) · editing 2 · ⌃X delete · esc cancel`
when in edit mode, so the affordance is discoverable without /help.
The /help hotkey table also gets a Ctrl+X entry.
ctrl-C is intentionally unchanged: it should never destroy queued
content. Cancel is non-destructive (esc / ctrl-C); only ctrl-X
removes the item.
Keep the parity test backed by the real Python command registry while avoiding hard failures in Node-only Vitest environments that cannot import hermes_cli.commands.
Use explicit repaint patch semantics for Ctrl/Cmd+L and narrow the hotkey assertion to the actual +L entry so unrelated descriptions do not cause false failures.
Make Ctrl+L non-destructive by redrawing the current screen state instead of starting a new session, and stop auto-appending --global for typed /model commands so session scope remains the default unless explicitly requested.
Route /browser, /reload-mcp, /rollback, /stop, /fast, and /busy through direct TUI RPC handlers so state changes hit the live gateway session instead of slash-worker fallback. Add TUI session finalize/reset parity hooks (memory commit + plugin boundaries) and parity matrix tests to keep mutating commands off fallback.
Route TUI /title through session.title RPC and queue titles when the session DB row is still initializing, so renamed sessions reliably appear in /resume and browse flows.
- drop unused TUI helpers, test-only layout scaffolding, and stale public debug exports
- remove an unused profiler import and trim test-only coverage for deleted helpers
CPU profile (Apr 2026, real-user scroll on 11k-line session) showed three
hot loops in the per-frame render path:
Output.get() per-frame walk: 24% total
└─ sliceAnsi(line, from, to) per write: 18% total
stringWidth(line) chain (cached + JS): 14% total
All three were re-doing identical work every frame: same string → same
clipped slice → same width.
Fixes:
1. Memoize stringWidth (8k-entry LRU) for non-ASCII strings; ASCII fast-path
skips the cache (inline scan beats Map.get for short ASCII, the >90%
case). String.charCodeAt scan up to 64 chars is cheaper than the regex
fallback.
2. Memoize wrapText (4k-entry LRU keyed by maxWidth|wrapType|text) — wrapAnsi
is pure and the same content reflows identically every frame.
3. Memoize sliceAnsi (4k-entry LRU keyed by start|end|str) for the
end-defined hot path used by Output.get().
4. Skip the slice entirely in Output.get() when the line already fits the
clip box (startsBefore=false && endsAfter=false). Most transcript lines
never exceed their container width, and tokenizing them just to slice
(line, 0, width) was pure overhead. This single fast-path drops
sliceAnsi from 18% → ~0% in the profile.
Also tighten virtualization constants (MAX_MOUNTED 260→120, OVERSCAN 40→20,
SLIDE_STEP 25→12) and cap historical-message render at 800 chars / 16
lines via HISTORY_RENDER_MAX_*; messages inside the FULL_RENDER_TAIL_ITEMS
window still render in full so reading-zone behavior is unchanged.
Validation, real-user CPU profile, page-up scroll on 11k-line session:
Output.get() self-time: 24% → 0.3%
sliceAnsi total: 18% → not in top 25
stringWidth family: 14% → ~3%
idle: 60.7% → 77.3%
Frame timings (synthetic page-up profile harness):
dur p95: ~10ms → 4.87ms
dur p99: 25ms+ → 12.80ms
yoga p99: ~20ms → 1.87ms
The remaining CPU in the profile is Yoga layoutNode + React commit,
which is the irreducible work for this UI tree size.
Replaces the static WHEEL_SCROLL_STEP=1 multiplier on wheel events
with an adaptive accel state machine that infers user intent from
inter-event timing.
Algorithm ported straight from claude-code's
src/components/ScrollKeybindingHandler.tsx. All tuning constants,
the native/xterm.js path split, the encoder-bounce detection, the
trackpad-burst signature → all theirs. This file is a mechanical
port into our module structure.
What it does:
precision click (>500ms gap) 1 row/event (deliberate scan)
sustained mouse (40-200ms) 2-6 rows (decay curve)
detected wheel bounce ramps to 15 (sticky wheel-mode)
trackpad flick (5+ <5ms) 1 row/event (burst detect)
direction reversal reset to base
Two implementation paths:
* native terminals (ghostty, iTerm2, Kitty, WezTerm) — linear
window-ramp + optional wheel-mode curve triggered by detected
encoder bounce. SGR proportional reporting handled via the
burst-count guard.
* xterm.js (VS Code / Cursor / browser terminals) — pure
exponential-decay curve with fractional carry. Events arrive
1-per-notch with no pre-amplification, so the curve is more
aggressive.
Selected at construction via isXtermJs() from @hermes/ink (now
exported). Per-user tune via HERMES_TUI_SCROLL_SPEED (alias
CLAUDE_CODE_SCROLL_SPEED for portability).
13 unit tests covering direction flip/bounce/reversal, idle
disengage, trackpad-burst disengage, frac invariants, and the
native vs xterm.js branches.
Profiled under --rate 30 (stress test) and --rate 10 (realistic
sustained scroll): accel ramps to cap=6 at 30Hz burst, decays to
1-3 rows at sparse 10Hz clicks. Perf is comparable to baseline
because accel IS multiplying step — the win is perceptual (fast
flicks cover distance, slow clicks keep precision), not raw fps.
Companion to the earlier WHEEL_SCROLL_STEP=1 change: that set the
base; this modulates around it.
Split in-flight assistant text at the last stable block boundary so only
the unclosed tail re-tokenizes per stream delta. Previously the full
text was rendered as plain <Text> during streaming and only flipped to
<Md> at message.complete — cheap per delta but loses live markdown
formatting.
New StreamingMd component holds a monotonically-growing stablePrefix
in a ref (idempotent under StrictMode double-render), renders it as
one <Md> that memoizes across deltas, and renders the unstable suffix
as a second <Md> that re-parses on each delta. Cost per delta drops
from O(total length) to O(unstable length).
findStableBoundary walks back to the last "\n\n" outside an open
fenced code block — splitting inside an open fence would orphan the
opener and break highlighting in the prefix.
Adapted from claude-code's src/components/Markdown.tsx:186 but built
on our line-based tokenizer instead of marked.lexer. 9 new tests cover
fence balance, boundary walk, and empty input.
Part of the --tui perf audit (see audit #7).
