Lifting the single-in_progress cap makes a list shape reachable that the
web surfaces never received. Two sites derived their one-line summary with
todos.find(t => t.status === 'in_progress') — the collapsed TodoPanel header
and the todo_write row — which was total under the old cap and silently
dropped every active item but the first once several could match: a plan
with three running tasks collapsed to the name of one. The expanded list
was always correct, so neither PR's tests covered it.
Both sites now take planSummary in contract/todo-plan-model.ts, the
domain-shared face both the skeleton and toolviews domains may import; the
duplicated derivation was why one find could be fixed while the other
stayed wrong. The hint names the first active item and suffixes +<n> for
the rest, so the collapsed line reports how many tasks are running.
The web fixture's todo sample now runs two items in_progress, so the
assembled web transcript replays a parallel plan: the row reads
'1/4 已完成 · 实现 fixture 样本 +1' over the built bundles.
The local PTY readiness poll held its inferred_idle fallback for exactly
one pollIntervalMs after a prompt marker, so a bash foreground handoff
that lands on the silence boundary only wins the exact stdin_read
attribution when the kernel publishes it inside that single poll. On a
slow or loaded host it does not, and the attribution flips.
handoffGraceMs replaces the hardcoded one-poll window as a validated,
deployment-owned config field defaulting to 500ms, rejected at load when
it cannot contain one readiness poll. Real-shell tests that interrupt a
send now assert the session is usable again rather than which readiness
tier observed the handoff, because no fixed grace removes the race.
`config.maxWallMs` is only checked for positivity, and it is handed to
`setTimeout`, which clamps any delay above 2^31-1 ms to 1 ms. A deployment
configuring a 25-day wall ceiling therefore gets the opposite of what it asked
for: every run times out on the first tick. The runtime now range-checks the
field at load against MAX_TIMER_DELAY_MS from dsh-timeout and throws, so the
misconfiguration fails loud where it is self-contained instead of silently
inverting the budget.
`computeMs` needs no matching bound: it is compared against measured event-loop
utilization rather than fed to a timer.
The test asserts both the rejection and that the boundary value itself loads.
Stack the parallel-in_progress change on the web todo display (#497): the
GUI is now the surface where several active items are visible, so the two
land as a chain rather than colliding on tool-todo at merge time.
Conflicts combined rather than resolved to one side: tool-todo keeps this
branch's parallel-allowing validation AND web2-todo's additionalProperties
unknown-key rejection, in src/index.ts and both README sides; the spec
drops web2-todo's 'two in_progress' rejection case and keeps its unknown-key
case; the two headless advanced-toolchain session fixtures keep this
branch's parallel transcripts.
The darwin-parity job failed intermittently on the SIGINT test with the
operation buffer holding only the echoed command line, never RAW_READY. The
harness sets idleSilenceMs to 250, so when a cold python3 start stays silent
past that bound the send settles as inferred_idle; PtySendOperation.append then
drops all later output, and the marker reaches only the scrollback.
Give the harness per-test idleSilenceMs/timeoutMs overrides and let this
scenario raise both above interpreter startup latency, so the readiness marker
lands inside the send it belongs to. waitForOutput's own deadline and the test
timeout grow to match the new bounds.
The product timings are unchanged; the pty Agent Note records why a test that
waits on an operation must outlast the child's startup.
The remote side re-merged master (i18n workflow round) and restored the
pre-architecture ModelSelector package with the Session-side state
machine. Reconciliation keeps that merge's master content and reasserts
this branch's model architecture: business state lives in ui-model's
shared per-session directory (ctx.models), not on Session/
ConversationSnapshot; the input.model seat entry and the /model popup
stay the two same-source entries; the history modelTarget piggyback stays
removed. The ui-model-selector package, its roster row, and its allowlist
entry go; the smoke-real roster names ui-model.
The required windows job moves from windows-2025 to ubuntu-latest, running
checksum-verified Windows Node under Wine at Linux-job wall clock (2m46s
warm vs 7-9min); master's serial-windows native-kernel reference is
untouched, and a new master-only wine-apt-cache job seeds the apt cache
every pull request restores. The experiment workflow folds into ci.yml,
the Agent Note moves to implemented with measured results, and the two CI
topology notes update to the shipped facts.
Review follow-up on the symmetric-cache conversion: provisioning and
caching are separable concerns, so the remaining 8 corepack enable
sites (the three enterprise restore-only PR jobs, the required Windows
job, serial-macos/serial-windows, the cache-less ci.yml jobs, and
sandbox.yml) now provision pnpm via pnpm/action-setup@v4 too — only
the provisioning line moves; every deliberate cache shape (enterprise
restore-only pairing, serial-linux producer, Windows cache-less) is
byte-preserved. corepack is gone from CI entirely.
Agent Note retitled to the decision's real shape (provisioning policy,
with caching as per-job policy on top), EN+ZH, pair re-recorded;
composite-action and stop-at-cache-workflows alternatives recorded.
The seat's dropdown follows figma 496:26454: the root pane is the
Model / Effort cell pair (14/22 label, value in the tertiary tone, right
chevron), each drilling into its own list — the provider-grouped model
list over the shared directory, and the High/Max effort levels. The
trigger (313:14108) shows both values: model name plus effort in the
caption tone. Effort is a client-local display echo on the shared
directory state for now — the design pairs the two as one selection, but
no wire carries a per-session effort override yet (the deepseek adapter's
reasoningEffort is deployment config); the directory state documents that
boundary, so wiring it later is a submit-path change, not a UI one.
Escape backs out of a drilled pane before closing.
The spare tier is retired. Steady-state pool load is one serial standby
job per master push, so six always-on instances already are the
failover capacity; pre-registered offline runners are a silently
expiring guarantee (GitHub garbage-collects them after 30 days
offline). Incident-time extra capacity is a one-minute org-token
registration, now documented in the runbook.
vi.waitFor retries every callback throw, so the mismatch branch inside
the callback waited the full 30s deadline for a fixture that writes the
marker once and cannot recover. Terminal states (complete marker, or
content that can no longer become the expected marker) now resolve out
of the retry loop and the mismatch throws after it, restoring the old
loop's immediate failure.
Two review findings on the turndown swap, both verified empirically:
- Unclosed-tag nesting makes the synchronous turndown/domino walk
superlinear (measured: depth 512 ~0.15s, 2k ~2s, 20k ~5s), during
which the cooperative fetchTimeoutMs timer cannot fire. renderBody
now preflights nesting depth with a linear tag scan and passes
bodies past 512 levels through raw; the try/catch stays for markup
the scan cannot see (comment-hidden tags), simulated in tests via a
converter throw.
- Markdown escaping can expand converted HTML ~2x (100k underscores
render as 200k chars), so provider body caps no longer bounded the
model-visible result. formatFetchOutput now caps the complete output
(header + body + footer) under new fetchMaxOutputChars config
(default 200000 = 2x the local provider's default body cap), reusing
the truncation notice.
README EN+ZH, config catalog, Agent Note EN+ZH updated; the new
web-fetch fixture is migrated to the packed layout master now
requires; tool-web coverage stays 100% per-file.
The todo row and the todo plan strip were registered without a contract in this
package's README, which still described only the bash sample. Both sides now
carry the two registrations: TodoRow's `todo_write` toolview key with its
attempted-write summary and non-ok dot semantics, and TodoDock's
`conversation.input.dock` entry at `order: -1` with the dock-owned selection,
empty-hidden panel, collapse behavior, and the composer-takeover consequence
that hides the whole dock. README.i18n.yaml re-recorded.
- ui-model README pair (short-form Model Experience + audited registry
entry + Known Limitations), recorded in the pairing manifest
- llm-deepseek README pair rebased onto master's version with this PR's
real deltas (friendly names, 256k, selector consumers) on both sides;
runtime/ui-conversation READMEs revert to master (their sections
documented the removed Session state machine and control slot)
- config catalog regenerated for the ui-model roster row
- ModelDirectory.store JSDoc; ui-model files join the TODO(gui) coverage
deferral lane beside their ui-command/ui-slash siblings
Three ds-review-bot round-8 findings on the todo display surfaces.
TodoPanel took the whole `useSession` hook and cast the snapshot to reach
`todos`, which put slot plumbing and an unchecked cast inside the presentation
component. The panel now takes `todos: readonly TodoItem[]`; TodoDock does the
selecting, matching the QueueDock posture the dock slot already establishes.
The todo row carried `onClick` with no keyboard route, so its details panel was
mouse-only. It now takes ToolRow's route verbatim: `role="button"`,
`tabIndex={0}`, and an Enter/Space handler that claims the event. The row stays
a `<div>` because a `<button>` flattens its inline spans.
`session.history`'s tail-only `todos` field was documented at the TS signature
but not in the apiproxy README pair, and the Agent Note claimed the feature
added no new wire vocabulary. Both README sides now state the tail/omission
semantics (an omitted field on a tail response is the empty plan, not unchanged
state), and the note records the one added field instead of denying it.
Tests: TodoPanel specs render the plain list; new TodoDock specs cover selection,
live follow, rollback-to-empty, and the registration shape; a new row spec pins
Enter/Space activation and non-activating keys.
- publint-all: the recursive publication view uses readdirSync
{recursive} again instead of globSync('**/*') — the glob skips
dot-prefixed segments (verified empirically), but npm pack publishes
dotfiles inside included directories, so hidden exports were
reported missing and other hidden files escaped validation
- markdown.ts/verify-type-equiv: markdownFences now reports whether a
closing delimiter terminates the block (mdast silently closes an
unterminated fence at EOF), and verify-type-equiv rejects unclosed
type-equivalence fences again — the Agent Note claimed such a block
still fails at the manifest checks, but its comparisons can succeed
- Agent Note EN+ZH: record the restored rejection; rewrite the zh
Problem section into past tense to match the English side's shipped
reality; pair re-recorded
- unused SessionId import and an unnecessary assertion in ui-model
- two over-length fake-api lines split
- trigger chrome calibrated against figma 313:14108's ToggleButton
(13/20 medium secondary at 28px; chevron in label-caption = the mock's
#ADB2B8; the mock's leading icon is invisible there, so the seat has
none, and its 'High' effort text has no wire concept to render)
An omitted `todos` on a tail request was treated as "no projection carried"
and preserved the prior value. Every installWindow caller is a tail request
(doOpen, its gap re-pull, repairGap; loadOlder prepends without it), which the
host answers with the full-log projection or omits only when the log holds no
todo/write — so the field's absence is the authoritative empty list. A live
write whose host crashed before persisting therefore left the rolled-back plan
on screen indefinitely; the assignment now clears it on the next open or
resync. Widened the parameter to an explicit `| undefined` so the two meanings
cannot be conflated again, and updated the JSDoc at both declaring seams plus
the bilingual README/note pair.
Turn 65 (todo_write) at the fx-alpha tail slides the 50-message history
window: two head-of-window messages drop out, so every trajectory cell
ordinal shifts down by two. Timing, labels, and cell content are unchanged.
The merge of origin/master at 9f218ce9d took master's re-recorded parent
session.jsonl wholesale, which reverted this branch's todo_write
description in that one file while the two child logs kept the new
parallel-in_progress text. The headless snapshot scrubs request headers
before comparison, so the three logs disagreed on the model-visible
tool contract without any test failing.
Re-record the scenario with test:snapshot:refresh, which replays the
committed scripts and rewrites all three persisted-log fixtures from the
live run. The parent regains the parallel-in_progress todo_write
description; both children pick up master's current run_code description
and its required `description` parameter, which they were stale on.
Fixture content only; no source or contract change, so the owning Agent
Note stands as written.
