The composer's /new used to create a separate chat. It now sends
chat_new_session. The server starts a fresh native session on the chat's
own harness (OpenCode, Codex or Claude Code) and re-keys the
operator_sessions row in one transaction. The chat keeps its title,
directory, backend, colour, star, model and mode. The old row is kept
unlisted with replaced_by, and its native session stays resumable.
The reply and the chat_session_changed push carry previous_id, so every
client swaps the row in place and an open conversation follows it to the
fresh, empty transcript with a notice.
Two or more consecutive Claude Code task notices fold into one
task_notices row: a chip 'N background tasks' (tinted when one failed)
that opens a sheet with one compact row per task (outcome icon, summary,
outcome word), like the tools and permissions chips. A lone notice stays
one line. The server also treats Claude Code's [SYSTEM NOTIFICATION]
banner as harness text, so a bannered notification without an origin is
still a notice, never the operator's turn. The chat fixture gains
?tasks=N.
A chat opened from the keyboard asks its composer for focus; the composer
takes it after the frame it is built in (new detail pane or phone page), or
at once when that chat is already open. Taps and other opens cancel the
request, so they keep focus where it was. The capture kit's press step
accepts modifier chords such as Control+2.
A chat's colour is picked from its row menu (Colour...), a long-press or
the chat information sheet. The list shows it as a leading bar, a faint
row tint and a coloured avatar; the open chat shows a faint wash under its
top bar. Shades are Cupertino's dynamic system colours and Material
swatches tuned per brightness. Changes from other devices apply live.
Ctrl+1...Ctrl+9 open the Nth item of the open tab's list (chats, runs,
pipelines) in on-screen order, also while typing. Starred chats are sorted
by name, case-insensitively, as one alphabetical group.
After a send the view jumps to the end of the conversation and resumes
following, even when the operator had scrolled up or the message is taller
than the screen, so the message's end and the reply stay in view.
The chat information sheet gains a Session section: the harness's native
session id (copy), harness, directory, a verified resume command (copy)
and Export conversation (Markdown), Export JSON and, for Claude Code,
Export native session (the JSONL transcript).
Server: chat_sessions rows carry resume_command and native_export; new
chat_export command builds Markdown/JSON from the same paged history
chat_messages serves (bounded at 8 MiB), refuses non-operator sessions.
App: exports go through the artifact file bridge (web download, share
sheet on phones).
At the field's line limit, a newline at the end left the caret on an empty
line the field only scrolled to after the next character. The field now
scrolls to the end right after such an edit.
The message field grows up to 12 lines (was 6) before it scrolls, capped at
about 40% of the visible height with the keyboard excluded, so the
conversation stays in view on a phone.
Root causes found in release web captures (1280x800 and 560x900):
- Follow mode undid small wheel/trackpad ticks at the end: each tick is its
own scroll, re-judged as following within 32 px and jumped back. A user
scroll toward older rows now releases the end however small; the jump
control appears once streamed content moves the end beyond the threshold.
- The end-of-scroll jump ran inside the position's activity switch, so a
touch or scrollbar drag released near the end recursed without end (stack
overflow during the gesture). The jump now runs after the notification.
- Horizontally scrolling code blocks and tables claimed diagonal trackpad
gestures whole. HorizontalScrollView gives mostly vertical gestures to the
enclosing list; sideways and Shift + wheel still scroll the block.
- The wheel over the floating chat bar, composer and jump control moved
nothing. VerticalWheelForwarder scrolls the transcript from them unless
something under the pointer claims the event.
Adds widget tests per content kind and scroll mode, follow/touch/Load older
tests, a capture-kit check (tool/capture/examples/chat-wheel-scroll.json)
with eval/expect steps, wheel dx/repeat, --params and eval: targets, a
transcript scroll probe and fixture options, and fixes the kit ignoring a
script's size. Notes in notes/chat-wheel-scroll.md.
Usage frames carry optional usage.account {harness, source, plan, as_of,
windows[{id,label,used_percent,resets_at}], unsupported}. Claude Code quota
comes from its rate_limit_event stream lines and the initialize account
(plan), shared across chats on the same account. Codex reads
account/rateLimits/read once per connection (re-read when a chat opens after
5 min) and merges account/rateLimits/updated. OpenCode exposes no quota API
and says so. The app's chat info sheet shows an Account usage section with a
progress bar and reset countdown per window, Material and Cupertino.
Replace the single glass pill (host NavigationBar/CupertinoTabBar with one
gliding highlight) with a row of separate circular GlassPill buttons, 52/50 px
with 12/10 px gaps, at the same bottom-right place. The selected button fills
with the host selection tint and cross-fades; each button pops in on its own,
staggered, only when the bar appears or its tabs change; presses sink like the
top bar's. Tab semantics, keys, tooltips and badges are kept; the row shares
one BackdropGroup for its blurs.
Claude Code records subagent/peer reports, task notifications, reminders and
skill bodies as user-role entries. Classify them server-side (history and live
stream) by origin/isMeta/queued_command mode: reports become kind agent_report,
task notifications kind notice, harness meta is hidden, and the mid-turn
wrapper is unwrapped to the operator's text. The app draws a collapsed Agent
report card and a one-line system notice on the agent's side.
A created chat's reply carried no mode, and the chat destination kept the
first row it opened for the whole visit, ignoring later list re-reads, so the
mode control only appeared after closing and reopening the chat. The server
now names the mode on a created row, and the destination adopts each listed
row that a re-read changed (excluding the live running flag), so the open
conversation follows new, refreshed and other-device mode changes in place.
Chat sections are slivers: a lazy, keyed SliverList under the same heading,
with SliverAnimatedExtent reproducing the clipped AnimatedSize for open,
close and arriving rows. The first Chat tab visit no longer builds every row
(150 chats: worst build frame 116 ms to 19 ms).
Reuse unchanged transcript rows, the composer bar and the header across
streamed deltas, and share parsed markdown by text so rows scrolled back
into view are not parsed again. Adds a rebuild/parse harness and a
release-web frame-timing probe with a long-transcript fixture.
Server: list_directory returns a directory_listing (path, home, parent,
entries with is_dir/is_git_repo/hidden, truncated) confined to the server
user's home subtree after cleaning and resolving symlinks, refusing pseudo
filesystems and capping entries at 2000. Documented in docs/cli.md.
App: "Choose folder…" in the create dialog opens a picker (full-screen page
on phones, large dialog on wide windows, Material and Cupertino) with
breadcrumbs, filter, Show hidden, git badges, dimmed files, recent folders,
inline errors and "Start chat here". The probe app gains a folder fixture
for captures.
The thread reads the session's queued prompts from the chat_subscribe reply
and chat_queue frames: a reload draws them as Queued after the transcript,
a delivered one enters the transcript as Sent, a refused one becomes Not
sent, and Remove sends chat_unqueue. A prompt the server reports queued is
dropped from the device unsent store; the any-frame queued heuristic is
used only with a server that does not report its queue.
On web the engine closes the text connection on the editing element's blur,
which the browser fires just before the window blur, so the field is already
unfocused when lifecycle inactive arrives and FocusManager's own suspend and
resume restores only the route scope. Remember whether the composer held (or
had just lost) focus when the window left, and refocus it on resume, keeping
the caret and selection, unless another control took focus or a dialog,
sheet or route covers the chat. Desktop hosts only, so phones never get an
unasked on-screen keyboard.
The room that records a server-run turn is torn down when its last client
leaves, so a page reload re-subscribed to a new room that no longer knew the
turn was running and never replayed the step-started marker. Claude Code does
not re-announce a running turn to a new stream either, so the activity row and
Stop control vanished until the next streamed delta.
chat_subscribe now also replays the marker when the room's stream announced a
turn (CLI-originated or re-announced on attach) or when the broker reports its
own turn in flight through the optional chatTurnReporter (Claude Code reads its
in-process running flag; the chat manager routes it without I/O).
A sent message is stored on the device until the server accepts it, so a
reload mid-request no longer loses it: on reopening, a stored intent the
history records is cleared as sent, otherwise it shows as Not sent with
Retry (same operation id) and Edit. The operator's bubble shows Sending,
Sent (newest only), Queued or Not sent. A live refusal keeps the old
contract (exact draft and chips back in the composer, unchanged re-send
reuses the id) and marks the bubble Not sent with Retry. On web, leaving
or reloading the page while a prompt is in flight asks first.
The transcript listener read the lazily built _controller while _attach was
still building it (load() notifies synchronously), so it built another
controller, recursively, until the test process ran out of memory.
Consecutive text parts of an agent turn (including folded messages) and
markdown blocks within a part are now separated by ~0.9x the scaled body
font size; a heading stays 8px from what it introduces. Adds a ?paras=1
chunked reply to the composer probe and gap tests at 1x/2x.
Settings leaves the tab bar (Pipelines, Runs, Chat, Search) and opens as a
page from a gear in the Pipelines bar, keeping /settings, its keys and a Back
to Pipelines (also the system Back). Every floating tab bar shows icons only,
named for assistive technology. On wide windows the shell's tab bar stays at
the bottom right; a split detail learns its height plus a 12px gap, so its
tabs or the chat composer float above it, and pages the detail opens (the
stage page) stay in its pane through its own navigator.
Review r2 findings P2-4, P2-5, P3-13: a shared relative-time helper for
every list row's time, formatDuration switches to days past 24h ('9d
16h'), the stage page uses one visit numbering, drops a repeated stage
heading and fixes '1 files', and search run hits read 'Run 1'.
Implemented by pipeline run #205/#207 (Terra); its UIReview blocked on a
fixture without timestamps, verified by the manager.
AppListRow (Material + Cupertino) with avatar, bold title and trailing
time, one muted line, one overflow menu style, inset dividers, selected
and emphasised states and a compact mode. Pipelines get a legible initial
avatar, schedule rows shrink to two lines with Edit/Pause/Delete in an
overflow menu, and the shell chat list scrolls behind the bar.
One floating frosted tab bar (host NavigationBar / CupertinoTabBar in the
shared GlassPill) replaces the NavigationRail, the Cupertino sidebar and the
run detail, stage page and pipeline detail tab strips. Below 600px it spans
the width minus 16px margins; wider it is sized to its tabs at the right edge.
Content keeps the full width, scrolls behind the bar and learns its height as
bottom media padding; lists pad their end clear of it and FABs sit above it.
The bar now sizes itself from its own constraints instead of the screen size,
which fixes the widget_test failure where a MediaQuery without a size gave it
negative width constraints.