Yichen Jiang
c48f8523f2
* refactor(llm-deepseek): share the /models listing machinery
Move the account route's `/models` request and decoding, its
credential-scoped listing cache, and listed-catalog resolution with the
version gate into the shared DeepSeek Messages library, so a second
route can follow the same listing. Names drop the account prefix, the
cache takes each access's endpoint settings with the call, and it gains
`invalidate()` for a forced reload. `isBuiltInCatalog()` recognizes the
catalog an uncustomized `models` setting resolves to.
The account route keeps its behavior; only its imports and the
override diagnostic wording ("model listing") change.
* feat(llm-deepseek-api-key): follow the official /models listing
While `models` equals the built-in catalog and the Messages origin is
https://api.deepseek.com, `deepseek-official` serves the official
`GET /models` listing, authenticated with the API key in `x-api-key`.
Loading, refresh, fallback, merging, and the client-version gate match
the account route: the first access waits up to `modelsTimeoutMs`, a
listing older than `modelsMaxAgeMs` refreshes in the background on the
next access, a failure keeps the last good listing or the built-in
catalog, and listed models this build cannot serve are refused before
network I/O.
A customized `models` list replaces the catalog and requests nothing.
Switching between a customized catalog and the listing discards the
cached listing, so Restore defaults reloads it on the next access.
Suites that boot the route on the official endpoint now answer the
listing themselves instead of reaching the network.
* docs: describe the official route model listing
Document when `deepseek-official` follows the official `/models`
listing, its timing fields, customization precedence, and the forced
reload after Restore defaults in the API-key README, the shared
transport README, and the provider guide. Record the decision in a new
Agent Note, keep the account note's statement of where the shared
listing code lives current, and regenerate the configuration catalog.
* test(llm-deepseek-api-key): pin the once-per-listing hidden-entry diagnostic
* docs: regenerate module and event graphs for the shared listing
* fix(llm-deepseek-api-key): report credential failures on the model request
On the official endpoint, `connect` waited for the listing view, whose
credential lookup rethrew a malformed key or a credential-store failure
before the agent loop recorded the prompt. A failed lookup now resolves
the connection from the current listing view, so `resolveAuth` reports
the failure on the model request and the turn keeps the prompt and the
failed attempt. A caller abort still rejects.
* docs(llm-deepseek-api-key): name which hidden listing entries are logged
* test: keep unit suites off the official model listing
Suites that boot the DeepSeek routes with a dummy or inherited key now
request `GET https://api.deepseek.com/models` on the official endpoint.
A unit-lane setup file rejects that request at once, so such a suite
serves the built-in catalog instead of reaching production or waiting
out the listing deadline on a network that never answers; suites that
exercise the listing answer `fetch` themselves. Other requests pass
through unchanged.
* feat(llm-deepseek): read tool_update from the model listing
Both DeepSeek routes now take a listed model's Messages tool update mode
from `api_capabilities.anthropic_messages.tool_update`, alongside the
system prompt mode. `in-history` and `addition-only` replace the
bundled mode; any other value reads as `leading-only` and removes it.
An absent or `null` mode keeps the bundled one, so a listing that does
not declare the field leaves `deepseek-flash` on `addition-only`.
`system_prompt_update: null` likewise keeps the bundled mode now; it
previously read as `leading-only`, which a serializer writing empty
values as null would have turned into a cache-unfriendly downgrade.
* fix(llm-deepseek): sync the built-in context window with the official listing
The official `/models` listing declares a 1,048,576-token context window
for `deepseek-flash` and `deepseek-v4-pro`; the built-in catalog said
1,000,000, a decimal-versus-binary mismatch. Both built-in entries now
carry the listed value, so offline starts and listing failures size the
context the same way the listing does. The Models page shows the exact
count for these rows, since 1,048,576 has no decimal K/M spelling.
`defaultContextWindow`, the fallback for model ids the catalog does not
describe, stays 1,000,000.
* test(llm-deepseek-api-key): reject the stubbed listing request with an Error
* test(cli): keep the built SDK smoke off the official model listing
The SDK profile's initialize now resolves the deepseek-official catalog,
which requests GET https://api.deepseek.com/models on the official
endpoint. With the smoke's fixture key that request answers 401 and the
route warns on stderr, so the test's empty-stderr assertion failed in the
e2e and snapshots lanes. Point the run at a loopback mock and assert it
received no model call.
* test(llm-deepseek): compose the malformed-key turn without the native executor
release/dsh-0.2.1 predates @deepseek-ai/dsh-tool-executor-native; its tool
runtime executes tools itself. Compose the official-route turn the way
account-routing.spec.ts does on this branch.
* fix(llm-deepseek-api-key): gate nothing when the credential lookup fails
The listing cache switches scope only after it reads a credential, so a
failed lookup left `connect` serving the previous credential's listing.
After a key whose listing refuses a model was replaced by a malformed one,
`admit` refused that model with CLIENT_UPGRADE_REQUIRED before the turn
recorded the prompt, instead of `resolveAuth` reporting the credential
error on the request. Serve the built-in catalog without listing gates
when the lookup fails.
* test(llm-deepseek-api-key): pin the once-per-listing diagnostic across request preparation
The hidden-entry diagnostic test repeated the same discovery assertion to
prove a second pass reports nothing new, which read as a duplicate. Assert
the single warning after the request preparations instead; they resolve
the same listing, so dropping the per-listing deduplication still fails
the test.
* docs(llm-deepseek-api-key): note that a list matching an earlier built-in catalog counts as customized
Customization is judged against the running version's built-in catalog,
so a saved `models` list equal to an earlier version's catalog stops
following the listing once the built-in catalog changes, until Restore
defaults. Record the consequence in the Agent Note and the README's known
limitations.
apps/web browser e2e
English | 中文
These tests boot the real web composition in-process and drive it with real browsers over real HTTP. Chromium runs the full lane; the model and reasoning picker scenario also runs in WebKit to cover native mouse focus behavior. Session replay recovery also runs in WebKit, sharing the fresh-round-trip recording and expected output without rewriting them. It covers active-stream reload, completed-turn reload, and reopening a Session. The lane's mechanics — modes, fixtures, goldens, and the deliberate composition divergences from dsh web — are documented in scaffold.ts and the browser e2e Agent Note.
After installing workspace dependencies, install the browsers and their system dependencies from the repository root:
pnpm --filter @deepseek-ai/dsh-web-frontend exec playwright install --with-deps chromium webkit
On Linux, --with-deps installs dependencies through the system package manager. The persistent CI VM must provide these dependencies through image maintenance; CI installs only the browser binaries, as required by the failover runbook.
Ordinary scenarios begin with no registered Workspace or Session and a durable marker recording a removed default Workspace, so explicit folder-selection scenarios retain control of their cwd. launchWebScaffold({ firstUse: true }) leaves initialization eligible for startup scenarios.
Completion observations
State-sensitive cases use Workspace, admission, attachment, and model-stream barriers to separate visible intermediate states from completed operations. Model-picker persistence assertions wait for the saved default, independently of menu closure. Details close waits for frame transitions; archive verification assigns an explicit title to the seeded Session and follows that identity across reload. The CI test reliability workflow owns these synchronization rules.
Idle-submission handoff waits up to the scaffold's 30-second turn-completion budget for Host frames and browser placement. Vitest owns scaffold teardown so a replay-consumption error does not replace the failed scenario assertion.
Plugin-manager switches wait up to PLUGIN_TOGGLE_SETTLE_MS (10 seconds) for the Host's profile write and the store's follow-up inventory read, the two steps that republish aria-checked; that constant's JSDoc records the measured chain.
Explicit scrolling uses scrollIntoView from support.ts: it resolves the locator again when its old element detaches and checks connection in the same browser task as native scrolling. Scenarios retain their visibility and geometry assertions after scrolling.
These are Host-face tests
They type-check in the root tsconfig.host.json, not in the Client aggregate, because they read Host services directly: ctx.connection, the Host SessionStore, and ctx.sessionProjectionCache. Driving a browser at runtime does not make a file part of the Client program — the two faces merge Cordis Context under the same keys with different services, so one program cannot see both. Moving these files into the Client aggregate makes every Host-service access fail to compile.
Do not import @deepseek-ai/dsh-client-* here
Importing a Client package — a value or a type — pulls its whole TypeScript project, and every project it references, into the Host build graph. That has bitten this lane once already: four Client consumer packages reference api/remotes' Client face, which cannot compile until Host tsdown has generated @deepseek-ai/dsh-goal/remote, so the Host build phase ended up waiting on an artifact it produces itself.
When a scenario needs a Client-owned constant or pure function, mirror it here instead, next to the commented-out import that names the source module. A drift then surfaces as a missed selector or a stale mirrored value — a loud failure, never a silent pass. scaffold.ts follows this rule for the welcome-notice namespace, acknowledgement field, version, and asserted Chinese copy.
The built-client harness is the exception. assembled-boot.ts imports AppWebEntry, the boot-manifest type, and RemoteMock; assembled-remote.ts imports the Client test runtime's default responses and RemoteMock. These packages are explicit project references for booting the real shell against a test-owned carrier. The chat scenarios mirror conversationContextKey in support.ts instead of importing its Client owner.
Nothing mechanically enforces this rule; keep it in review.