The mount helper's 4th positional argument becomes an options object so a
test can set the session list row's blank flag independently of the
conversation snapshot's; the two new cases cover both branches of the
settling condition.
Address review of the ToolRow card refactor:
- ToolRow renders a visually-hidden run-state label (row.running/failed/stopped
locale keys) so a running/failed/interrupted row is not a colour-only signal;
the StateDot and the sweep are both aria-hidden. Restores the stopped-state
text assertion in diff-card.spec at the ToolRow layer.
- The file-link button stops Enter/Space from bubbling to DisclosureRow's
whole-row keydown, which would preventDefault the key and toggle expand
instead of opening the file (the keyboard analogue of the click stopPropagation).
- search-row passes output={model.output} unconditionally, matching the other
three card rows (ToolRow renders the card over the output when present).
- Assert locale: 'conversation' in the search/web/diff registration tests.
- Add the read render-intent paragraph to the ui-conversation README and a
ReadBlock atom entry + Read section to ui-primitives (both languages).
- Update the grep fixture sample lines to the post-refactor SearchRow shape.
Refresh the code-mode-round, steering, and search-card goldens for the added
status text and updated sample lines.
Resolve the agent-loop import conflict by retaining both durable request context and runtime policy context. Refresh the combined session fixtures and regenerate documentation catalogs. Mark PDF artifacts as binary so staged whitespace checks do not parse PDF bytes as text.
Rewrite the read/search/web/diff card spec suites for the collapsed-by-default
ToolRow composition: each row now needs the conversation `t`, and the card is
absent from the DOM until the row's [data-expandable] toggle is clicked. Refresh
the built-boot and search-card assembled snapshots (they expand the row before
shaping the card) and the code-mode-round golden. Fix a fixture turn-number
collision the 985/986 merge introduced (grep reused turn 67, already the
multi-hunk edit's, so both shared one callId) by renumbering search/web/todo to
68-72. Document the refactor's behavior in the ui-conversation README (both
languages) and add the Agent Note.
The user bubble's IconActions row carried an edit button with no click
handler, no client mutation, and no host operation behind it, advertising
an affordance the product cannot honor. The row now renders clock / copy /
branch only for both user and assistant chrome, and MessageIconActions
loses the `edit` prop with the button.
The package README records the missing capability under Known Limitations,
the web aria goldens pin the row without the control, and the Agent Note
states what must exist before it comes back.
Node >= 21 exposes a global `navigator` reporting the machine's own language,
so gating detection on `navigator` let a non-browser boot of the client tree
resolve to `en` instead of the documented fallback; `window` is the browser
test. `navigator.languages` is spec-required but absent on some embedders and
older WebViews, where spreading it would throw at boot, so the walk tolerates
its absence and `navigator.language` covers that host.
The per-spec pin boilerplate collapses into one suite-level
`usePinnedBrowserLanguages('zh-CN')`, which owns the rationale in
dsh-client-test-runtime, and the English-browser e2e scenario now clears the
console warnings channel too — its page has no closing inventory spec.
A first visit resolved to Chinese regardless of the browser: LocaleService
read `dsh.locale` and fell straight back to `zh` when nothing was stored,
ignoring the languages the browser already states it reads.
The initial locale now resolves through three ordered sources — the persisted
preference, then `navigator` (first entry of the ordered language list whose
primary subtag names a shipped locale, so `zh-Hans-CN` -> zh and `en-GB` ->
en), then `FALLBACK_LOCALE`. An explicit choice still wins and nothing writes
the detected locale back to storage, so "has the user chosen?" stays a
question only the stored value answers.
Specs asserting the shipped Chinese copy now state the browser they assume:
the web e2e scenarios open their page with `locale: ZH_BROWSER_LOCALE`, and
package specs pin it through the new `pinBrowserLanguages` test helper.
`settings-chrome.e2e.ts` gains an English-browser scenario as the
assembled-app proof.
The master merge added a required archivedSessionIds field to WorkspaceListState;
the read-card spec's DetailsPanel fixture state needs it to typecheck.
Public snapshot state stays in the store engine's plain-data vocabulary
(immer drafts reject Sets without the MapSet plugin, which stays off):
manager/service/contract carry readonly SessionId[] in Host order, and
the tree derivations build their own transient Set — the
expandedProjects pattern. Membership-unchanged installs still keep the
array reference for Object.is short-circuits.
WorkspaceListState gains archivedSessionIds (ReadonlySet, replaced only
on membership change), installed as full snapshots from the list
baseline, the unary echo, and the changed frame. Archiving the current
session clears the selection into the New Session view state. Test
doubles (test-runtime, fake APIs, fixture client) follow the widened
IWorkspaces/IApiClient faces.
Review caught a real divergence the earlier measurements missed: mirroring
an offset is only correct while both layers can reach it, and for a draft
ending in a newline the backdrop could not.
A textarea reserves a line box for the caret after a final newline.
`white-space: pre-wrap` collapses a text node's trailing newline and
generates none. So a draft ending in a newline made the backdrop exactly one
line shorter than the textarea — measured 628 against 652 — and the mirrored
assignment clamped, leaving the glyphs one line behind the caret at the very
bottom of the draft.
The backdrop now carries the same trailing-line sentinel the mirror div has
carried all along: its content is the decoration walk plus one newline. The
same pre-wrap collapse absorbs it when the draft does not end in a newline,
so it costs no height in the ordinary case, and it supplies the missing line
box when it does. Verified in isolation first: a bare pre-wrap div measures
180/180/198 against a textarea's 180/198/216 for zero, one and two trailing
newlines, and 180/198/216 with the sentinel.
Coverage for the shape that exposed it: the browser scenario asserts the two
extents are equal before asserting the glyphs reach the end, observing each
layer's maximum by asking for an impossible offset and reading back the
clamp rather than computing it from scrollHeight, and the golden records the
relation. The unit spec pins the backdrop's text as the draft plus exactly
one newline. Removing the sentinel fails both, the e2e with the same 628
against 652.
The scrollbar-gutter half of the same review point does not reproduce here:
both layers measure clientWidth 776 against a border box of 776 while the
draft overflows, so this engine's textarea scrollbar is an overlay and takes
no width out of the wrap.
