Dismissed shell runs leave the transcript and fold into a "N commands"
chip in the footer of the first answer at or after each run (the last
answer for newer runs, or a row of its own without any answer). The chip
opens the turn sheet (wide) or a full-screen page (narrow) listing the
commands; each opens its output read-only with Copy output, Send to agent
and Restore.
On a wide conversation the answer's Copy button moves out of the footer
into a 48 px gutter left of the reading column (chatCopyGutter), shown on
hover or focus on desktops and always on touch; narrow widths keep it in
the footer.
Finished command cards get a Dismiss button and the chat's More menu a
Dismiss all finished commands entry. Dismissal applies to every client of
the chat; a SnackBar (Material) or transient notice (Cupertino) offers Undo
for 5 s. Running commands cannot be dismissed until they end.
Finished terminals are sized synchronously from their rows at the view's
width (the post-frame fit made the transcript re-anchor forever); blank rows
xterm.dart keeps when rewrapping are dropped; desktop typing uses key events;
the composer leaves Cmd/Ctrl-V to a focused terminal; Ctrl-A reaches the
process. The probe fixture gains coloured terminal runs and an interactive
login. Notes record behaviour, protocol, security and verification.
On macOS, Linux, Windows and desktop browsers the "Thought · tool calls ·
background tasks" row is as wide as its label, so its hover and click area
no longer span the transcript. Phones keep the full-width touch row. The
check uses the device platform because web themes are pinned to Android.
The scroll fixture gains ?turnactivity=1 for captures.
Show Expand only when output can overflow and drop the final newline from
the drawn output. The full-app probe lists shell runs with ?shell=1 so the
cards, Shell mode and Run command dialog can be captured in Chrome.
The answer footer shows '3 files changed +9 -2'; tapping opens the turn's
files in the sheet (wide) or a full-screen page (narrow), and a file opens
the run Changes diff viewer. Patches load through the chat's own gateway.
The list applies chat_activity pushes and each row's activity by seq, so
every listed chat shows a running turn, not only the chat open here. An
idle chat with running background tasks wears a quiet badge with tooltip and
semantics "Background tasks running (N)". Older servers keep the old
behaviour. List rows that never changed carry the server's current seq so
they order after snapshots kept from an earlier server process.
Markdown links, autolinks and bare http(s)/www. URLs open externally through
url_launcher (http, https and mailto only) in assistant markdown and operator
messages; code is never linkified. Links use the host primary colour with an
underline that thickens on hover, and offer Copy link on long press and
right-click without breaking text selection.
The just-sent (or queued) message slides up 20px, fades and grows from
0.96 over 280ms on an emphasized decelerate curve. Only a send marks it:
history, scroll restore, other devices, replies and the server copy
replacing the local one appear in place. Row keys follow the operation
id so the server copy keeps the row. Reduced motion shows it at once.
The capture fixture now accepts chat_prompt.
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.
A search with kind chat matches the operator-session registry's titles and
directories and the newest messages of recent transcripts through the chat
broker (bounded count, concurrency and one-second budget), answered off the
connection's read loop. The all-kind search is unchanged for older clients.
The app sends the chat search after its run search and shows a Chats group
with title, harness, directory and snippet; a chat hit opens that chat in the
Chat tab. The probe fixture answers chat searches over ?chats=N.
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.
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).
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.
FadeThroughStack(lazy: true) builds only the shown page first and the rest
one per frame once no animation runs. Opening a run detail no longer lays out
Timeline, Artifacts and Changes offstage during the route transition
(transition build 62-103 ms -> 38-59 ms); first tab switches are unchanged.
Adds tool/perf/interaction_bench.mjs (FrameTiming per interaction over the
full-app fixture, CPU profiles mapped to Dart files), fixture options for
hits/chats/graph, and notes/perf-round3.md with the survey.
Blur sigma 16 -> 4 keeps CanvasKit on its direct blur path; under the 84%
fill the look is unchanged (max 13/255 inside the pieces). SegmentedBar's
pieces share a BackdropGroup so Impeller reads the content behind them once.
Adds a scroll frame-time bench, a backdrop census and filler rows to the
full-app probe. CPU CanvasKit, 560x900: run timeline p50 31.8 -> 22.1 ms,
runs list 25.3 -> 19.9 ms.