The TTFT and tok/s readings divide by measured wall time, so they are not
reproducible: the same replayed scenario yielded 69 and 70 tok/s on
consecutive local runs, and a 3 ms replayed stream reads 26333 tok/s. Baking
those into committed goldens made the Web lane flaky by construction, and the
goldens for the readings themselves were never refreshed.
Three fixes, then a refresh:
The footer's decorative dots are `aria-hidden`, so the readings concatenated
into one accessible string — `Ran for 13sTTFT 0.2s12 tok/s`. That is a real
defect on its own (a reader hears one run-on instead of three facts) and it
also denied `{{duration}}` the word boundary it matches on, so even the
previously-stable `Ran for` duration started leaking raw. The separators now
carry flanking spaces.
`normalizeAria` gains `{{throughput}}` beside `{{duration}}`, and its duration
alternation accepts the stats line's compact `2m42s` as well as the
message-chrome template's `2m 42s` — the compact form had no pattern at all,
which is why `LLM 382m39s` survived the first refresh.
Refreshed 17 goldens. They also record that the stats line's `LLM` group now
renders at all: it folds assistant `timing`, which the live transcript adapter
only began attaching in this branch, so the group was previously dead in Chat.
Verified by running the lane in replay three times after the refresh: 41/41
files green each time, goldens untouched. Before this change two consecutive
runs disagreed on both the values and the failure count.
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.
Settled history now exposes user/assistant chrome (including date-aware
clocks) in the accessibility tree; collapse clocks via scaffold and update
keyless scenario goldens.
The assembled fixture app now takes over the composer while fx-alpha's
resident approval is pending — the feature under test, not chrome noise —
so the session-title snapshot answers it (允许一次) before asserting the
model seat, mirroring what a user sees. The eleven keyless ui.expected.md
goldens re-record (DSH_SNAPSHOT=refresh, no key) for the Access placeholder
→ projection chip swap (Danger Full Access under the roster default) plus
the master-side chrome drift the goldens had accumulated (copy/branch/edit
buttons, tab rows) since their last refresh.
The startup-selection flow leaves a fresh world (no Workspace) in the
locked view state, so every e2e scenario that types into the composer
now connects one first via the shared connectFreshWorkspace helper
(hero picker create-by-name dialog; the default 'workspace' name keeps
the session-header cwd assertions intact).
Golden refreshes carry the current composer chrome: the plan/model
control seats are empty until their owning plugins register (the seats
shipped without occupants on this branch), the sidebar shows the
connected workspace group pre-send, and the bash details material
renders Input/code/Output as separate nodes. The cancel scenario polls
the frozen-partial swap instead of counting synchronously — the abort
frame reaches the browser over SSE after the host settles.
Every spec now commits at least one golden and the interactive ones one
per distinct end-state (nine new .expected.md):
- live-interactions: cancel.expected.md (frozen partial + 已停止 marker),
error-auth.expected.md (the prompt bubble alone — the committed artifact
of the web-error-surface gap, the diff that flips when error rendering
lands), retry.expected.md (indistinguishable from a clean completion —
retries are deliberately invisible in the transcript).
- question-composer: answered.expected.md (the question resolved into its
tool round trip plus the final reply, takeover gone) beside the existing
waiting-state golden.
- steering: mid-steer.expected.md pins the accepted-but-INVISIBLE state
(the loop drains steering only at the step boundary, so no interjection
bubble exists while the question still blocks — if the client ever
renders pending steers eagerly, this golden flips first) and
settled.expected.md the badged bubble plus obeying reply.
- navigation-panes: waterfall.expected.md and details-open.expected.md
(tool-name header, Input args, Output result) beside the trajectory one.
- lifecycle-chrome: reloaded.expected.md — rendering the same settled
transcript from persistence alone IS the recovery claim.
Fixture inventories extended to the new closed sets; the Agent Note's
expected-outputs policy updated in both languages (per-end-state goldens
for interactive scenarios), pairing re-recorded.
Yichen Jiang
Yif
kingwl
imccyu
Tianyi Cui
creatixchu
imccyu
NI0317
ZiyaZhang
_Kerman
07akioni