Address the three open review threads on the dictionary parity gate.
Regex/TEXT: localeOf now requires an uppercase ASCII [A-Z] flat-letter at
the third position of a name-prefix shape, so zh2Foo/zh_probe are no longer
treated as dictionaries in localeOf while the admission pre-filter skips
them. The two now agree exactly.
register detection now also admits a bare register identifier callee in
addition to a property access, covering a future destructured
register(NS, 'zh'|'en', dict) call instead of silently dropping it.
A 3-arg register whose dictionary argument is a local variable is resolved
through module-scope const initializers; one that cannot be resolved to an
object literal makes the gate refuse with a named error rather than skipping
the registration and narrowing the sweep.
Also restate the FALLBACK_LOCALE rationale: the residual case points at
English because a browser naming neither shipped language is the reader
least likely to read Chinese, not because English is the copy's source
language (Chinese is; packages/client/AGENTS.md). Sync the identical claim
in the bilingual Agent Note and re-record its .i18n.yaml pairing.
The parity gate recognized only a `[['zh',{...}],['en',{...}]]` array, so the
two separate ctx.locale.register(NS, 'zh'|'en', {...}) calls in
ui-permission-presets were unchecked: deleting a key from one side left the
gate green. Pair those calls by their namespace argument. Widen the pre-filter
to admit zhSettings/accessZh spellings, which a bare \b(zh|en)\b misses and
would have skipped before parsing.
Assert document.documentElement.lang in the assembled app. The served markup
already ships lang="en", so the fr-FR scenario passes whether or not the sync
runs; the zh scenario is the discriminating half and now asserts zh-CN before
the switch and en after it.
Drop the dead vi.unstubAllGlobals() from the document-language spec, which
manages navigator with defineProperty and never calls vi.stubGlobal.
Restore @deepseek-ai/dsh-client-test-runtime in ui-settings-general: the
package still imports bindSnapshotSelector from it in
tests/components.client.spec.tsx, so removing it was manifest drift. The
earlier knip report predated that file arriving on this branch.
Swallow only ENOENT when reading a directory in the parity gate. A broad catch
treated EACCES or an I/O failure as "absent", which would narrow the sweep and
let the gate pass while checking less.
setLocale returned early when the id already matched the active locale, so
choosing the language already on screen wrote nothing. That value may be a
provisional browser-derived or fallback resolution nothing has stored, so a
different browser sharing the DSH home still resolved on its own. Write
unconditionally; keep the render publish conditional.
Broaden the dictionary parity gate to every workspace package, pair zh/en
across sibling files and inline registrations, and fail when a dictionary has
no counterpart. It previously scanned only packages/client and packages/
extensions, compared within a single module, and silently skipped unpaired
dictionaries -- so the split locales/zh.ts + en.ts common pair, the inline
directory-picker-browse dictionary, and session-log-export were unchecked.
Normalize paths at ingestion so the sweep does not narrow on Windows.
Regenerate the client API catalog and update the locale README pair: both
described the old zh fallback direction.
Add the English fallback dialog golden, and drop a dead afterEach plus the
blank lines left where the dead browser-language pins were removed.
An image with a side above the deployed routes' 2000px many-image bound
could be durably committed by read_image, ride every later request, and
permanently fail the session with provider 400s. Admission now enforces a
configurable maxImageDimension (default 2000) during the full decode, so
read_image surfaces a recoverable tool error naming the limit instead of
poisoning durable history; the Web composer gets dedicated copy for the
new IMAGE_DIMENSION_TOO_LARGE reason.
Fixes#2626
master introduced AttachmentStore.saveImages as the batch admission
(count/aggregate-byte/media-type limits, validate-all-before-save,
ordered commit). admitEncodedImages narrows to the shared wire entry:
canonical-base64 enforcement plus delegation to saveImages, keeping one
home for batch policy while both wire endpoints (prompt RPC and the
command executor) still call one function. Test doubles gain saveImages;
batch-limit error texts follow saveImages' wording.
A claimed slash command consumed only the text half of the composer
submission: /goal with reference images executed, cleared the draft, and
silently stranded the images in the rail. Model-visible attachment intent
had no route through the command plane.
The submission envelope is now modeled end to end. CommandDefinition
input.images declares acceptance; the declaration rides the descriptor to
every client, onto the minted CommandClaim, and into the input machine's
claim snapshot. commands.execute carries the submission's base64 images
and enforces the declaration in the executor: non-declaring commands, a
missing attachment store, and exceeded batch limits settle as logged
error results before the handler runs. Admission reuses the attachment
package's new admitEncodedImages, extracted from api-proxy's prompt path
so both wire endpoints share one limits/validation/commit sequence.
Producers own model visibility: /goal submits one user followup (image
blocks + a fixed reference line) after a successful create/edit so goal
rounds read the images from session history; /plan folds them into its
steered message. Grammar misfits (/goal pause, bare /plan, /plan off)
return direct errors and the composer keeps the images.
On the client, enter adjudication carries a SubmitEnvelope and every
command route that cannot consume images throws a localized refusal that
renders as one composer notice with draft and images retained; the
claimed pre-gate applies the same copy. An accepting claim serializes the
draft images, forwards them to commands.execute, and clears plus releases
them only on a success outcome.
The assembled web test roster gains the ui-input-trigger and ui-commands
plugins, mirroring the shipped composition, so slash submissions exercise
the command plane; a new keyless snapshot pins the refusal banner and the
accepting /goal flow over the built client graph.
The provisional locale fell back to zh, so a browser asking for neither
zh nor en (fr, de) opened the product in Chinese. Resolve to en instead,
and use en as the dictionary fallback: the shipped zh/en dictionaries
declare identical key sets, so one constant serves both roles.
Add scripts/locale-dictionary-parity.spec.ts to gate that symmetry, and
set the asserted locale explicitly in specs that had relied on the old
zh fallback through a dead usePinnedBrowserLanguages call (those files
declare no jsdom environment, so browser detection never ran there).
Resolutions:
- docs/module-graph.md, docs/config-catalog.md: generated files. Regenerated
with gen-module-graph and gen-config-catalog on the merged tree, then carried
the new package's entries into the Chinese sides and re-recorded both
pairings. Each side now differs from master by exactly the
code-runtime-python rows.
- scripts/verify-package-readme-model-experience.ts, tsconfig.host.json: master
renamed packages/bash -> packages/shell, packages/pty -> packages/terminal,
code-runtime-worker -> code-runtime-worker-thread and agent-tool-mode ->
agent-tool-presentation. Kept master's names and re-added this branch's
code-runtime-python entry.
Adapted the package to conventions master introduced while the branch was open:
version 0.1.0-rc.6 with publishConfig.access "public" (the release-member rule
check-workspace-constraints now enforces), the invariants project reference
moved to packages/runtime-diagnostics/invariants, and the README companion link
retargeted to code-runtime-worker-thread.
The Python lane never compared what the minimal composition shows the
model: the mock model only asserted system-role messages, and the
advanced snapshot tokenizes the assembled system prompt and tool
schemas.
The sdk-minimal scenario now records model-visible.json — every model
request's advertised tool schemas verbatim and its message list, with
system and user text kept and assistant/tool payloads reduced to call
identity so the expected output replays on macOS and Linux. It excludes
the dynamic runtime-context snapshot, which the same composition emits
on macOS and not on Linux (#2488).
AGENTS.md and the testing policy name both SDKs as independent
projections of the agent loop, session lifecycle, and SessionEventMap.
A max-tokens response that included a tool call persisted assembler-transformed
content next to replay metadata projected from the untransformed native message,
so the next request died in history reconstruction with INVALID_REPLAY_STATE and
the session stayed permanently stuck.
Write side: the finish chunk's replayState becomes a typed ReplayEnvelope —
opaque response-level metadata plus optional per-block entries aligned with the
emitted block sequence. BlockAssembler computes one keep/drop decision for
blocks and entries together, so stored metadata always describes stored content
and retained blocks keep their signatures. pi-ai splits its state into a
version-2 response half and per-block signature entries.
Read side: durable content is authoritative. toPiAssistant degrades any
unusable state — foreign kind, other versions (including the flat v1 form
already on disk), malformed metadata, or content/block mismatches — to the
existing provider-neutral conversion with an onReplayDegrade diagnostic instead
of failing the request, which un-bricks sessions poisoned before this change.
Covered by assembler and replay unit tests, an agent-loop continuation
regression, keyless real-composition continuation tests (native pruned-envelope
replay and legacy flat-state degrade), and the authored keyless snapshot
scenario max-tokens-continue through the assembled ACP app.
creatixchu