From 83ae7046f2fc9e8ea953f392feffeb5fefa0391e Mon Sep 17 00:00:00 2001 From: Turtle Date: Thu, 30 Jul 2026 17:54:48 +0800 Subject: [PATCH 01/13] fix(install): preserve pnpm version errors --- scripts/install.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/install.sh b/scripts/install.sh index 41d5c749c1..e4e8c175da 100755 --- a/scripts/install.sh +++ b/scripts/install.sh @@ -186,7 +186,7 @@ fi # pnpm is the only dependency we offer to install for you. if command -v pnpm >/dev/null 2>&1; then - info "pnpm $(pnpm --version 2>/dev/null) ... ok" + info "pnpm $(pnpm --version) ... ok" else warn "pnpm is not installed." if confirm "Install pnpm now?" Y; then From 7aa9e5108480eb2aa6cf2f5145d41e9eaf12b3f6 Mon Sep 17 00:00:00 2001 From: ZiyaZhang <199893125+ZiyaZhang@users.noreply.github.com> Date: Sat, 1 Aug 2026 06:15:16 -0700 Subject: [PATCH 02/13] fix(cli): give shipped web search a 60s budget --- .../feature/2026-07-31-web-default-search.i18n.yaml | 4 ++-- .../implemented/feature/2026-07-31-web-default-search.md | 6 ++++-- .../feature/2026-07-31-web-default-search.zh.md | 6 ++++-- apps/cli/config/base.cordis.yml | 5 ++++- apps/cli/tests/built-bin.e2e.ts | 7 +++++++ 5 files changed, 21 insertions(+), 7 deletions(-) diff --git a/.agents/notes/implemented/feature/2026-07-31-web-default-search.i18n.yaml b/.agents/notes/implemented/feature/2026-07-31-web-default-search.i18n.yaml index 2c8f3eea48..1c04bc1d55 100644 --- a/.agents/notes/implemented/feature/2026-07-31-web-default-search.i18n.yaml +++ b/.agents/notes/implemented/feature/2026-07-31-web-default-search.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-31-web-default-search.md -2026-07-31-web-default-search.md: 121a1dff5fffd4223eefcc7475fff874276658aa -2026-07-31-web-default-search.zh.md: ac98f4806413cb6050a08354a553942437866fe1 +2026-07-31-web-default-search.md: 8f2cc6f5d1a9ec9f33279ae8e5c992bdc660e7fa +2026-07-31-web-default-search.zh.md: 7d528617143cb9756d0a4532bb1ed6a3a3e70ca8 diff --git a/.agents/notes/implemented/feature/2026-07-31-web-default-search.md b/.agents/notes/implemented/feature/2026-07-31-web-default-search.md index 121a1dff5f..8f2cc6f5d1 100644 --- a/.agents/notes/implemented/feature/2026-07-31-web-default-search.md +++ b/.agents/notes/implemented/feature/2026-07-31-web-default-search.md @@ -10,7 +10,7 @@ The harness had a complete Web capability family—provider registry, DeepSeek/E ## Decision -`apps/cli/config/base.cordis.yml` explicitly mounts `dsh-web` with `searchProvider: deepseek-official`, `dsh-web-search-deepseek`, and `dsh-tool-web` with `fetch: false`. It does not mount `dsh-web-fetch-local` or select a fetch provider. The shared base makes only `web_search` a default for TUI, browser, and headless sessions. The explicit search provider id keeps selection independent of registration order and leaves personal or `--config` overlays able to replace or disable the rows. +`apps/cli/config/base.cordis.yml` explicitly mounts `dsh-web` with `searchProvider: deepseek-official`, `dsh-web-search-deepseek`, and `dsh-tool-web` with `fetch: false` and `searchTimeoutMs: 60000`. It does not mount `dsh-web-fetch-local` or select a fetch provider. The shared base makes only `web_search` a default for TUI, browser, and headless sessions. The explicit search provider id keeps selection independent of registration order and leaves personal or `--config` overlays able to replace or disable the rows. The one-minute shipped budget covers an auxiliary DeepSeek Messages request plus server-side retrieval while leaving `dsh-tool-web`'s provider-neutral 30-second default unchanged for custom compositions. DeepSeek search uses the same `DEEPSEEK_API_KEY` credential reference as the official conversation adapter. The provider resolves that reference inside every search through the optional `ctx.credentials` service; only a composition without the seam falls back to the launching process environment, and a non-empty literal `apiKey` remains the programmatic last resort. A stored or rotated Web Models key therefore reaches the next search without restarting or retaining the value on the provider. Because `WebSearchProvider.available()` is synchronous, it treats an installed resolver as locally usable and missing dynamic credentials fail the operation with the provider-specific `WEB_PROVIDER_CREDENTIAL_MISSING` code while the stable tool schema stays registered. @@ -28,8 +28,10 @@ The default mount does not create a Web-specific permission policy. `web_search` **Keep Web tools in `web.cordis.yml`.** Rejected because it preserves an unexplained tool-roster difference between TUI and Web/headless. The rows are not surface-specific, so `base.cordis.yml` is their one home; the [tool-roster decision](2026-07-31-even-out-shipped-tool-rosters.md) records the shared composition. +**Raise `dsh-tool-web`'s provider-neutral timeout.** Rejected because custom providers and deployments own different latency expectations; the shipped DeepSeek composition owns this deployment budget. + **Enable search and fetch together.** Rejected because default `web_fetch` would allow model-selected anonymous outbound HTTP(S) retrieval to arbitrary URLs. Search covers discovery; deployments that accept broader retrieval can opt into `dsh-web-fetch-local` and set `dsh-tool-web`'s `fetch` option to `true` in their overlay. ## Consequences -Native model requests on every shipped surface carry only the `web_search` schema and search-only prompt guidance; Web/headless Code Mode exposes the same search capability beneath `run_code`. The prompt tells the model to use returned snippets and never advertises the disabled `web_fetch` tool. Search adds a complete auxiliary model call and may use the native server tool multiple times; its exact secret-free request remains reconstructable from the initiating session log. The default offers search-result snippets and source metadata but no arbitrary page retrieval; deployments that need full-page fetch must opt in. The Web snapshot lane boots the shipped tree, drives a replayed `web_search` call through the real DeepSeek provider against a local Messages fixture, asserts the durable auxiliary request and structured result, and pins the settled browser presentation. The TUI/Web composition smokes pin the shared `web_search` roster and absence of `web_fetch`; provider tests pin missing, stored, and rotated credential behavior plus literal and ambient compatibility. +Native model requests on every shipped surface carry only the `web_search` schema and search-only prompt guidance; Web/headless Code Mode exposes the same search capability beneath `run_code`. The prompt tells the model to use returned snippets and never advertises the disabled `web_fetch` tool. Search adds a complete auxiliary model call and may use the native server tool multiple times; its exact secret-free request remains reconstructable from the initiating session log. The default offers search-result snippets and source metadata but no arbitrary page retrieval; deployments that need full-page fetch must opt in. The Web snapshot lane boots the shipped tree, drives a replayed `web_search` call through the real DeepSeek provider against a local Messages fixture, asserts the durable auxiliary request and structured result, and pins the settled browser presentation. The TUI/Web composition smokes pin the shared `web_search` roster and absence of `web_fetch`; the built composition dump pins the one-minute shipped search budget; provider tests pin missing, stored, and rotated credential behavior plus literal and ambient compatibility. diff --git a/.agents/notes/implemented/feature/2026-07-31-web-default-search.zh.md b/.agents/notes/implemented/feature/2026-07-31-web-default-search.zh.md index ac98f48064..7d52861714 100644 --- a/.agents/notes/implemented/feature/2026-07-31-web-default-search.zh.md +++ b/.agents/notes/implemented/feature/2026-07-31-web-default-search.zh.md @@ -10,7 +10,7 @@ Status: implemented ## 决策 -`apps/cli/config/base.cordis.yml` 明确挂载 `dsh-web`,配置 `searchProvider: deepseek-official`,同时挂载 `dsh-web-search-deepseek`,并以 `fetch: false` 挂载 `dsh-tool-web`。它不挂载 `dsh-web-fetch-local`,也不选择抓取提供方。共享 base 只将 `web_search` 设为 TUI、浏览器与无头会话的默认工具。显式搜索提供方 id 使选择不受注册顺序影响,同时个人覆盖层或 `--config` 覆盖层仍可替换或禁用这些配置项。 +`apps/cli/config/base.cordis.yml` 明确挂载 `dsh-web`,配置 `searchProvider: deepseek-official`,同时挂载 `dsh-web-search-deepseek`,并以 `fetch: false` 和 `searchTimeoutMs: 60000` 挂载 `dsh-tool-web`。它不挂载 `dsh-web-fetch-local`,也不选择抓取提供方。共享 base 只将 `web_search` 设为 TUI、浏览器与无头会话的默认工具。显式搜索提供方 id 使选择不受注册顺序影响,同时个人覆盖层或 `--config` 覆盖层仍可替换或禁用这些配置项。已交付的一分钟预算用于覆盖一次辅助 DeepSeek Messages 请求及服务端检索,同时保持 `dsh-tool-web` 提供方无关的 30 秒默认值不变,以供自定义组合使用。 DeepSeek 搜索使用与官方会话适配器相同的 `DEEPSEEK_API_KEY` 凭据引用。提供方在每次搜索内部通过可选的 `ctx.credentials` 服务解析该引用;只有未挂载该 seam 的组合才会回退到启动进程的环境变量,非空的 `apiKey` 字面值仍作为程序化配置的最后兜底。因此,由 Web 的 Models 页存储或轮换的密钥无需重启即可用于下一次搜索,提供方也无需保留该值。由于 `WebSearchProvider.available()` 是同步方法,它会将已安装解析器视为本地可用;若动态凭据缺失,操作会以提供方专属错误码 `WEB_PROVIDER_CREDENTIAL_MISSING` 失败,而稳定的工具 schema 仍保持注册。 @@ -28,8 +28,10 @@ DeepSeek 搜索使用与官方会话适配器相同的 `DEEPSEEK_API_KEY` 凭据 **将 Web 工具保留在 `web.cordis.yml` 中。** 不予采纳:这会保留 TUI 与 Web/无头界面之间无法解释的工具清单差异。这些配置行并非界面特有,因此其唯一归属是 `base.cordis.yml`;[工具清单决策](2026-07-31-even-out-shipped-tool-rosters.md)记录了这一共享组合。 +**提高 `dsh-tool-web` 的提供方无关超时。** 不予采纳:自定义提供方和部署有各自不同的延迟预期;这一部署预算应归已交付的 DeepSeek 组合所有。 + **同时启用搜索和抓取。** 不予采纳:默认启用 `web_fetch` 会允许模型自行选择任意 URL,执行匿名出站 HTTP(S) 抓取。搜索负责发现信息;接受更广泛抓取范围的部署可以在覆盖层中选择启用 `dsh-web-fetch-local`,并将 `dsh-tool-web` 的 `fetch` 选项设为 `true`。 ## 后果 -每个已交付界面的原生模型请求都只会携带 `web_search` schema,以及仅用于搜索的提示词指引;Web/无头 Code Mode 通过 `run_code` 公开相同的搜索能力。该提示词要求模型使用返回的 snippet,且绝不会向模型提及已禁用的 `web_fetch` 工具。搜索会增加一次完整的辅助模型调用,并可能多次使用原生服务器工具;发起会话的日志仍可精确重建其不含密钥的请求。默认配置会提供搜索结果 snippet 与来源元数据,但不支持任意页面抓取;需要抓取完整页面的部署必须自行选择启用抓取。Web 快照通道会启动已交付配置树,使用本地 Messages fixture(测试前置数据),经由真实 DeepSeek 提供方驱动一次回放的 `web_search` 调用,断言持久化的辅助请求与结构化结果,并固定最终浏览器呈现。TUI/Web 组合冒烟测试固定了共享的 `web_search` 清单及不提供 `web_fetch` 这一事实;提供方测试固定缺失、已存储及已轮换凭据的行为,以及字面值与环境变量的兼容性。 +每个已交付界面的原生模型请求都只会携带 `web_search` schema,以及仅用于搜索的提示词指引;Web/无头 Code Mode 通过 `run_code` 公开相同的搜索能力。该提示词要求模型使用返回的 snippet,且绝不会向模型提及已禁用的 `web_fetch` 工具。搜索会增加一次完整的辅助模型调用,并可能多次使用原生服务器工具;发起会话的日志仍可精确重建其不含密钥的请求。默认配置会提供搜索结果 snippet 与来源元数据,但不支持任意页面抓取;需要抓取完整页面的部署必须自行选择启用抓取。Web 快照通道会启动已交付配置树,使用本地 Messages fixture(测试前置数据),经由真实 DeepSeek 提供方驱动一次回放的 `web_search` 调用,断言持久化的辅助请求与结构化结果,并固定最终浏览器呈现。TUI/Web 组合冒烟测试固定了共享的 `web_search` 清单及不提供 `web_fetch` 这一事实;构建后组合配置的转储固定了已交付的一分钟搜索预算;提供方测试固定缺失、已存储及已轮换凭据的行为,以及字面值与环境变量的兼容性。 diff --git a/apps/cli/config/base.cordis.yml b/apps/cli/config/base.cordis.yml index e88cbdb39d..b1810215e4 100644 --- a/apps/cli/config/base.cordis.yml +++ b/apps/cli/config/base.cordis.yml @@ -332,7 +332,9 @@ # chat, at each search; its Messages endpoint is separate from the # chat-completions endpoint, so it takes its own base-URL override. Fetch stays # disabled and no fetch provider is mounted: that provider defers SSRF -# protection and the model would choose the request target. +# protection and the model would choose the request target. Search is a full +# auxiliary model request with server-side retrieval, so this shipped DeepSeek +# route gets 60s while the provider-neutral tool default remains 30s. - id: web name: '@deepseek-ai/dsh-web' config: @@ -348,6 +350,7 @@ name: '@deepseek-ai/dsh-tool-web' config: fetch: false + searchTimeoutMs: 60000 # ── rows every surface mounts, whose values each overlay states ────────────── diff --git a/apps/cli/tests/built-bin.e2e.ts b/apps/cli/tests/built-bin.e2e.ts index 9591ca3105..3592d438dd 100644 --- a/apps/cli/tests/built-bin.e2e.ts +++ b/apps/cli/tests/built-bin.e2e.ts @@ -94,6 +94,13 @@ describe.skipIf(!existsSync(dshBin))('dsh BUILT bin (node lib/bin.js, no tsx)', expect(stdout).toContain('model: deepseek-v4-pro') expect(stdout).toContain('cwd: !!js process.cwd()') expect(stdout).toContain("name: '@deepseek-ai/dsh-tui'") + expect(stdout).toContain([ + '- id: tool-web', + " name: '@deepseek-ai/dsh-tool-web'", + ' config:', + ' fetch: false', + ' searchTimeoutMs: 60000', + ].join('\n')) // Provenance comment separators name each section's source file. expect(stdout).toContain('# == base.cordis.yml') expect(stdout).toContain('# == base.cordis.yml, patched by tui.cordis.yml') From 955a12cca4921a36bacf856bcccc240073c13667 Mon Sep 17 00:00:00 2001 From: kingwl Date: Thu, 30 Jul 2026 03:17:54 +0800 Subject: [PATCH 03/13] feat(web): steer queued messages into active turns --- ...-29-addressable-queue-operations.i18n.yaml | 4 +- ...2026-07-29-addressable-queue-operations.md | 10 +- ...6-07-29-addressable-queue-operations.zh.md | 10 +- ...026-07-30-web-queue-steer-action.i18n.yaml | 6 ++ .../2026-07-30-web-queue-steer-action.md | 67 +++++++++++++ .../2026-07-30-web-queue-steer-action.zh.md | 67 +++++++++++++ .../queue-actions/editing.expected.md | 2 + .../snapshots/queue-actions/ui.expected.md | 2 + apps/web/tests/steering.e2e.ts | 69 +++++-------- docs/architecture.i18n.yaml | 4 +- docs/architecture.md | 2 + docs/architecture.zh.md | 2 + docs/cordis-catalog/events.md | 32 +++---- docs/core-data-structures/core.i18n.yaml | 4 +- docs/core-data-structures/core.md | 14 ++- docs/core-data-structures/core.zh.md | 14 ++- docs/event-producer-consumer.md | 32 +++---- packages/client/runtime/README.i18n.yaml | 4 +- packages/client/runtime/README.md | 2 +- packages/client/runtime/README.zh.md | 2 +- .../runtime/src/client/contract/session.ts | 4 +- .../client/runtime/tests/queue-store.spec.ts | 19 +++- .../client/ui-conversation/README.i18n.yaml | 4 +- packages/client/ui-conversation/README.md | 4 +- packages/client/ui-conversation/README.zh.md | 4 +- .../ui-conversation/src/client/locales.ts | 6 ++ .../src/client/queue/QueueDock.tsx | 19 +++- .../ui-conversation/src/client/service.ts | 4 +- .../ui-conversation/tests/queue-dock.spec.tsx | 45 ++++++++- .../cordis/tool-cordis/src/api-catalog.ts | 4 +- packages/core/agent-loop/README.i18n.yaml | 4 +- packages/core/agent-loop/README.md | 2 +- packages/core/agent-loop/README.zh.md | 2 +- packages/core/agent-loop/src/agent.ts | 16 ++++ .../tests/contract-regressions.spec.ts | 96 +++++++++++++++++++ packages/core/agent/README.i18n.yaml | 4 +- packages/core/agent/README.md | 2 +- packages/core/agent/README.zh.md | 2 +- packages/core/agent/src/types.ts | 12 ++- packages/host/apiproxy/README.i18n.yaml | 4 +- packages/host/apiproxy/README.md | 2 +- packages/host/apiproxy/README.zh.md | 2 +- packages/host/apiproxy/src/api-proxy.ts | 17 +++- packages/host/apiproxy/src/api/rpc.schema.ts | 1 + packages/host/apiproxy/src/api/rpc.ts | 1 + .../host/apiproxy/src/api/sessions.schema.ts | 1 + packages/host/apiproxy/src/api/sessions.ts | 3 +- .../apiproxy/tests/api-proxy-commands.spec.ts | 18 +++- .../host/apiproxy/tests/rpc-schemas.spec.ts | 4 + 49 files changed, 508 insertions(+), 147 deletions(-) create mode 100644 .agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml create mode 100644 .agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md create mode 100644 .agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md diff --git a/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.i18n.yaml b/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.i18n.yaml index 99b5efd70f..217bcd32db 100644 --- a/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.i18n.yaml +++ b/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.md -2026-07-29-addressable-queue-operations.md: 02519f8fe8be60823ac773ac4cceecb9f12f88b2 -2026-07-29-addressable-queue-operations.zh.md: 998fc49ae7289b07c97312d4814d85b5cc84387b +2026-07-29-addressable-queue-operations.md: e5fbacfa7de244daa1c5f8e504ee98a4a7730393 +2026-07-29-addressable-queue-operations.zh.md: a532d18d1b29fcebc7e9d8222da9eb5a162b8ec6 diff --git a/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.md b/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.md index 02519f8fe8..e5fbacfa7d 100644 --- a/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.md +++ b/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.md @@ -12,13 +12,13 @@ The Web queue rendered pending messages but could not edit or delete one row. `M **Each accepted FIFO occurrence has its own identity.** AgentLoop mints an opaque `InboxItemId` and publishes an `InboxItem` containing that id, the identified `UserMessage`, and its acceptance-time `queued | steering` placement. Reusing one `MessageId` creates distinct inbox identities. Injection bypasses the FIFOs and receives no inbox identity. -**Mutation ends at driver claim.** `Agent.updateInbox(id, action)` synchronously searches the pending queued FIFO. Edit replaces frozen content while preserving `InboxItemId`, `MessageId`, source, wake policy, and position. Remove emits the occurrence’s terminal discard. Steering and driver-claimed occurrences return `not-found`, so queue operations never rewrite active-turn input or durable history. +**Mutation ends at driver claim.** `Agent.updateInbox(id, action)` synchronously searches the pending queued FIFO. Edit replaces frozen content while preserving `InboxItemId`, `MessageId`, source, wake policy, and position. Remove emits the occurrence’s terminal discard. Strict steer transfers the message into an open next-step window as a new steering occurrence; a closed window returns `steer-unavailable` without changing the queued item. Pending steering and driver-claimed occurrences return `not-found`, so later mutations never rewrite active-turn input or durable history. **The live ledger is authoritative.** `agent/inbox/enqueue`, `update`, `dequeue`, and `discard` maintain a Host mirror of queued occurrences. A synchronously re-entrant update or terminal event may reach the mirror before its outer enqueue listener; the mirror retains that unseen outcome for the current dispatch and folds it into the enqueue, so listener registration order cannot publish stale content or a ghost row. The wire sends complete `session/queue` snapshots rather than incremental guesses. Reconnect sends the current baseline, and every queued mutation or terminal event replaces it. The client applies no optimistic edit and never retires a row from durable turn events or status changes. **Queue addresses require a live ordinary-session Agent.** `session.updateQueue` queries only the mounted Agent registry and never resumes a cold session: an `InboxItemId` is process-local and cannot name work after restart or disposal. A session-backed subagent returns `agent-busy` before inbox access and retains its continuation owner; for ordinary sessions, a missing Agent and a driver-claimed occurrence both return `queue-item-not-found`. -**Web actions address Queue only.** The Host excludes pending steering from `session/queue`; steering retains its existing durable transcript path after consumption. QueueDock hides while empty, renders one pending occurrence directly, and defaults two or more occurrences to a collapsed `" 条排队消息"` header that expands or collapses the complete list. The header exposes `aria-expanded` and `aria-controls`; the expanded list scrolls within a 180px height bound. An active edit or mutation keeps its rows visible, and emptying the queue restores the collapsed default for the next queue. Visible rows expose edit and delete, but no send-now control. The UI derives queue row and mutation types from the runtime `SessionFace` contract rather than importing the connection plugin, so plugin cooperation continues through services and snapshots. Edit is available only when all content blocks are text; the editor cannot silently drop non-text blocks. An editing row exposes only save and cancel, with Enter and Escape as their keyboard equivalents. Delete removes the exact occurrence. The Web stop action preserves pending Queue work; AgentLoop claims the next waking occurrence only after the interrupted turn reaches quiescence, and its dequeue event retires that row without a browser resend. +**Web actions address Queue only.** The Host excludes pending steering from `session/queue`; steering retains its existing durable transcript path after consumption. QueueDock hides while empty, renders one pending occurrence directly, and defaults two or more occurrences to a collapsed `" 条排队消息"` header that expands or collapses the complete list. The header exposes `aria-expanded` and `aria-controls`; the expanded list scrolls within a 180px height bound. An active edit or mutation keeps its rows visible, and emptying the queue restores the collapsed default for the next queue. Visible rows expose edit, delete, and a running-only strict-steer action. The UI derives queue row and mutation types from the runtime `SessionFace` contract rather than importing the connection plugin, so plugin cooperation continues through services and snapshots. Edit is available only when all content blocks are text; the editor cannot silently drop non-text blocks. An editing row exposes only save and cancel, with Enter and Escape as their keyboard equivalents. Delete removes the exact occurrence, while strict steer preserves every content block and retires the row only through the authoritative snapshot. The Web stop action preserves pending Queue work; AgentLoop claims the next waking occurrence only after the interrupted turn reaches quiescence, and its dequeue event retires that row without a browser resend. The [Web Queue steer action](2026-07-30-web-queue-steer-action.md) owns the strict transfer contract. ## Alternatives considered @@ -26,7 +26,7 @@ The Web queue rendered pending messages but could not edit or delete one row. `M **Apply optimistic browser mutations.** Rejected because driver claim and another client can win before the Host action. Waiting for the authoritative snapshot makes the ownership boundary visible and lets `queue-item-not-found` report a real race. -**Include pending steering in the queue mutation protocol.** Rejected because QueueDock has no steering interaction, and editing or deleting active-turn input would widen this feature beyond its current consumer. A dedicated steering interaction owns that delivery contract. +**Allow editing or removal of pending steering.** Rejected because QueueDock only addresses independent queued turns. Once strict steer succeeds, the new steering occurrence belongs to the active turn and remains outside this mutation surface. **Expose a protocol-only promotion operation.** Rejected because no product interaction reorders Queue. A public operation without a current consumer would add ordering semantics and tests for speculative use. @@ -34,10 +34,10 @@ The Web queue rendered pending messages but could not edit or delete one row. `M ## Verification -AgentLoop contract tests hold prompt admission while editing and removing exact queued occurrences, reject mutations of steering occurrences, and verify the resulting independent turn and terminal lifecycle events. Host schema and proxy tests cover queued-only authoritative snapshots, synchronous re-entrant mutation order, reconnect, cold-Agent rejection, typed not-found errors, and the RPC transport. Client runtime and QueueDock tests cover non-optimistic projection, single-row presentation, default multi-row collapse, interaction-forced visibility, reset after emptying, expansion, text-only editing, save and cancel affordances, removal, retirement races, and disabled mixed-content editing. A keyless browser scenario captures the default collapsed header, drives edit and delete through the built Web composition and real HTTP/SSE wire, then stops consecutive active turns to prove the preserved FIFO advances without clearing its tail. +AgentLoop contract tests hold prompt admission while editing, removing, and strictly steering exact queued occurrences; they reject mutations of steering occurrences and verify the resulting independent turn and terminal lifecycle events. Host schema and proxy tests cover queued-only authoritative snapshots, synchronous re-entrant mutation order, reconnect, cold-Agent rejection, typed race errors, and the RPC transport. Client runtime and QueueDock tests cover non-optimistic projection, single-row presentation, default multi-row collapse, interaction-forced visibility, reset after emptying, expansion, text-only editing, save and cancel affordances, removal, strict steer, retirement races, and disabled mixed-content editing. Keyless browser scenarios drive all three exposed actions through the built Web composition and real HTTP/SSE wire, then stop consecutive active turns to prove the preserved FIFO advances without clearing its tail. ## Consequences -Queued work gains precise row operations without becoming durable session history. Occurrence identity is a live process-local capability and disappears at claim, broad cancellation, disposal, or restart; the Web stop action preserves it until a later claim, while reconnect recovers only queued items still held by the live Agent. Editing excludes mixed content until an editor can preserve every block, while pending steering remains outside this operation surface. +Queued work gains precise row operations without becoming durable session history. Occurrence identity is a live process-local capability and disappears at claim, strict transfer, broad cancellation, disposal, or restart; the Web stop action preserves queued occurrences until a later claim, while reconnect recovers only queued items still held by the live Agent. Editing excludes mixed content until an editor can preserve every block, while pending steering remains outside the projection and operation surface. The protocol now carries full queue snapshots on each change. Queues are expected to remain short, so deterministic recovery and multi-client convergence are preferred over an incremental mutation protocol. diff --git a/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.zh.md b/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.zh.md index 998fc49ae7..a532d18d1b 100644 --- a/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.zh.md +++ b/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.zh.md @@ -12,13 +12,13 @@ Web 队列能够渲染待处理消息,但无法编辑或删除其中某一行 **每次获准进入 FIFO 的项都有独立标识。** AgentLoop 会铸造不透明的 `InboxItemId`,并发布一个 `InboxItem`,其中包含该 id、已有标识的 `UserMessage`,以及接受时确定的 `queued | steering` 放置方式。复用同一个 `MessageId` 会创建不同的 inbox 标识。注入绕过 FIFO,因此不会获得 inbox 标识。 -**变更边界止于驱动器认领。** `Agent.updateInbox(id, action)` 会同步搜索待处理的 queued FIFO。编辑会替换已冻结的内容,同时保留 `InboxItemId`、`MessageId`、来源、唤醒策略和位置。移除会发出该次入队项的终态 discard。steering(中途引导)项和已被驱动器认领的项会返回 `not-found`,因此队列操作绝不会改写活动轮次输入或持久历史。 +**变更边界止于驱动器认领。** `Agent.updateInbox(id, action)` 会同步搜索待处理的 queued FIFO。编辑会替换已冻结的内容,同时保留 `InboxItemId`、`MessageId`、来源、唤醒策略和位置。移除会发出该次入队项的终态 discard。严格 steering(中途引导)会把消息作为新的 steering 单次入队项转移到开放的 next-step 窗口;窗口关闭时返回 `steer-unavailable`,且不改变 queued 项。待处理 steering 和已被驱动器认领的项会返回 `not-found`,因此后续变更绝不会改写活动轮次输入或持久历史。 **实时账本是权威状态。** `agent/inbox/enqueue`、`update`、`dequeue` 和 `discard` 共同维护 queued 入队项的 Host 镜像。同步可重入的 update 或终态事件可能先于外层 enqueue 监听器到达镜像;镜像会在当前分发期间保留这一尚不可见的结果,并在处理 enqueue 时把它合并进去,因此监听器注册顺序不会导致系统发布陈旧内容或不存在的行。协议发送完整的 `session/queue` 快照,而非增量猜测。重连会发送当前基线,每次 queued 变更或终态事件都会整体替换它。客户端不会进行乐观编辑,也绝不根据持久轮次事件或状态变化退役队列行。 **Queue 寻址要求普通会话的 Agent 存活。** `session.updateQueue` 只查询已挂载的 Agent 注册表,绝不恢复冷会话:`InboxItemId` 属于进程本地标识,无法在重启或资源释放后继续指向工作。由会话支撑的 subagent 会在访问 inbox 前返回 `agent-busy`,并保留其继续执行 owner;对于普通会话,Agent 缺失和单次入队项已被驱动器认领这两种情况都返回 `queue-item-not-found`。 -**Web 操作只面向 Queue。** Host 从 `session/queue` 中排除待处理 steering;steering 消费后仍沿用既有的持久 transcript(文本记录)路径。QueueDock 在队列为空时隐藏,只有一个待处理项时直接渲染该行,存在两个或更多待处理项时则默认收起为可展开或收起完整列表的 `" 条排队消息"` 表头。表头暴露 `aria-expanded` 和 `aria-controls`;展开后的列表以 180px 为高度上限,并可滚动。存在进行中的编辑或变更时,列表行会保持可见;队列清空后,下一次出现队列时会恢复默认收起状态。可见行暴露编辑和删除操作,不提供立即发送控件。UI 从运行时 `SessionFace` 契约派生队列行与变更类型,而不是导入连接插件,因此插件仍通过服务和快照协作。仅当所有内容块都是文本时才提供编辑功能;编辑器不能静默丢弃非文本块。编辑中的行只展示保存和取消操作,对应的键盘操作分别是 Enter 和 Escape。删除会移除对应的精确入队项。Web 停止操作会保留待处理 Queue 工作;只有在被中断轮次达到完全停稳后,AgentLoop 才认领下一个可唤醒入队项,其出队事件会退役该行,无需浏览器重发。 +**Web 操作只面向 Queue。** Host 从 `session/queue` 中排除待处理 steering;steering 消费后仍沿用既有的持久 transcript(文本记录)路径。QueueDock 在队列为空时隐藏,只有一个待处理项时直接渲染该行,存在两个或更多待处理项时则默认收起为可展开或收起完整列表的 `" 条排队消息"` 表头。表头暴露 `aria-expanded` 和 `aria-controls`;展开后的列表以 180px 为高度上限,并可滚动。存在进行中的编辑或变更时,列表行会保持可见;队列清空后,下一次出现队列时会恢复默认收起状态。可见行暴露编辑、删除以及仅在运行期间可用的严格 steering 操作。UI 从运行时 `SessionFace` 契约派生队列行与变更类型,而不是导入连接插件,因此插件仍通过服务和快照协作。仅当所有内容块都是文本时才提供编辑功能;编辑器不能静默丢弃非文本块。编辑中的行只展示保存和取消操作,对应的键盘操作分别是 Enter 和 Escape。删除会移除对应的精确入队项;严格 steering 会保留每个内容块,并且只通过权威快照退役该行。Web 停止操作会保留待处理 Queue 工作;只有在被中断轮次达到完全停稳后,AgentLoop 才认领下一个可唤醒入队项,其出队事件会退役该行,无需浏览器重发。[Web Queue steering 操作](2026-07-30-web-queue-steer-action.md)负责定义严格转移契约。 ## 考虑过的替代方案 @@ -26,7 +26,7 @@ Web 队列能够渲染待处理消息,但无法编辑或删除其中某一行 **在浏览器中进行乐观变更。** 不予采纳,因为驱动器认领或另一个客户端可能先于 Host 操作完成。等待权威快照可以显式呈现所有权边界,并让 `queue-item-not-found` 报告真实竞态。 -**将待处理 steering 纳入队列变更协议。** 不予采纳,因为 QueueDock 没有 steering 交互,而编辑或删除活动轮次输入会把此功能扩展到当前消费方之外。应由专用 steering 交互负责该投递契约。 +**允许编辑或移除待处理 steering。** 不予采纳,因为 QueueDock 只寻址独立的 queued 轮次。严格 steering 一旦成功,新的 steering 单次入队项就属于活动轮次,并且不再位于此变更接口内。 **暴露仅协议层的前移操作。** 不予采纳,因为当前没有产品交互会重新排序 Queue。公开一个没有当前消费方的操作,会为了推测性用途引入排序语义和测试。 @@ -34,10 +34,10 @@ Web 队列能够渲染待处理消息,但无法编辑或删除其中某一行 ## 验证 -AgentLoop 契约测试会在编辑和移除精确 queued 入队项时阻塞提示词接纳,拒绝对 steering 入队项的变更,并验证所得独立轮次及终态生命周期事件。Host schema 与代理测试覆盖仅含 queued 项的权威快照、同步可重入变更顺序、重连、拒绝冷 Agent、类型化 not-found 错误和 RPC 传输。客户端运行时和 QueueDock 测试覆盖非乐观投影、单行呈现、多行默认收起、交互期间强制保持可见、清空后重置、展开、仅文本编辑、保存与取消入口、移除、退役竞态,以及禁用混合内容编辑。一个无密钥浏览器场景会捕获默认收起的表头,通过构建后的 Web 组合和真实 HTTP/SSE 协议执行编辑和删除,随后连续停止活动轮次,证明保留的 FIFO 会继续推进且不清空队尾。 +AgentLoop 契约测试会在编辑、移除和严格 steering 精确 queued 入队项时保持提示词接纳窗口打开,拒绝对 steering 入队项的变更,并验证所得独立轮次及终态生命周期事件。Host schema 与代理测试覆盖仅含 queued 项的权威快照、同步可重入变更顺序、重连、拒绝冷 Agent、类型化竞态错误和 RPC 传输。客户端运行时和 QueueDock 测试覆盖非乐观投影、单行呈现、多行默认收起、交互期间强制保持可见、清空后重置、展开、仅文本编辑、保存与取消入口、移除、严格 steering、退役竞态,以及禁用混合内容编辑。无密钥浏览器场景会通过构建后的 Web 组合和真实 HTTP/SSE 协议执行公开的全部三项操作,随后连续停止活动轮次,证明保留的 FIFO 会继续推进且不清空队尾。 ## 后果 -queued 工作获得精确的行操作,但不会因此成为持久会话历史。单次入队标识是进程本地的实时寻址凭据,会在认领、广义取消、dispose 或重启时消失;Web 停止操作会将其保留到后续认领,而重连只能恢复仍由活跃 Agent 持有的 queued 项。编辑会排除混合内容,直至编辑器能够保留每个块;待处理 steering 则不属于此操作接口。 +queued 工作获得精确的行操作,但不会因此成为持久会话历史。单次入队标识是进程本地的实时寻址凭据,会在认领、严格转移、广义取消、dispose 或重启时消失;Web 停止操作会将 queued 入队项保留到后续认领,而重连只能恢复仍由活跃 Agent 持有的 queued 项。编辑会排除混合内容,直至编辑器能够保留每个块;待处理 steering 则不属于此投影和操作接口。 现在,协议会在每次变更时携带完整队列快照。队列预期保持较短,因此系统优先选择确定性恢复和多客户端收敛,而非增量变更协议。 diff --git a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml new file mode 100644 index 0000000000..780858e1d8 --- /dev/null +++ b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml @@ -0,0 +1,6 @@ +# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each +# side as of the last confirmed-consistent state. Both languages carry equal authority; +# after editing either side, bring the other along and re-record with: +# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md +2026-07-30-web-queue-steer-action.md: 4a8a04c056748d5c607c5ec868de4f2dd8342f2c +2026-07-30-web-queue-steer-action.zh.md: 92225f3a61e24b9eb687c50400aa6fe5e6ff5623 diff --git a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md new file mode 100644 index 0000000000..4a8a04c056 --- /dev/null +++ b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md @@ -0,0 +1,67 @@ +# Agent Note: Steer a queued Web message into the active turn + +Status: implemented + +English | [中文](2026-07-30-web-queue-steer-action.zh.md) + +## Problem + +The Web composer deliberately queues Enter submissions while an agent runs. QueueDock already gives each pending message an addressable row, and the durable transcript already renders consumed `steering/message` events with an interjection badge, but Web has no action connecting those two surfaces. + +Implementing the row action as a client-side delete followed by `session.prompt(mode: 'steer')` would split one user intent across two RPCs. Driver claim could win between them, the steer could fail after deletion, or the existing best-effort `agent.steer()` fallback could silently append a new Queue item after the original occurrence was removed. A send-now action must therefore distinguish current-turn steering from Queue promotion and preserve the original row when steering is no longer possible. + +## Decision + +### Product contract + +Each non-editing QueueDock row exposes the upward-arrow action as “插话发送”. The action is enabled only while the session reports a running agent; mixed-content messages remain eligible because steering forwards the complete immutable `UserMessage` rather than the row's text projection. Edit and delete keep their existing behavior, and the composer continues to submit Enter as Queue. + +Activating the action requests strict current-turn steering for that exact `InboxItemId`. Success removes the Queue row through the authoritative Host snapshot. When AgentLoop drains it, the existing durable `steering/message` event and transcript badge render the message without a new chat presentation path. + +The running bit is only an interaction hint. AgentLoop's `acceptsNextStep` value is authoritative at the synchronous mutation boundary. If that window has closed, the operation leaves the Queue occurrence unchanged and returns a typed `steer-unavailable` error; if the driver already claimed the occurrence, it returns the existing `queue-item-not-found` error. The UI reports either race without optimistically removing the row. + +### Agent and lifecycle boundary + +`InboxAction` gains a consumer-backed `{ kind: 'steer' }` operation alongside edit and remove. `Agent.updateInbox()` handles it only after locating the queued occurrence and proving `acceptsNextStep`; it never delegates to the best-effort `agent.steer()` alias. + +An applied action ends the queued occurrence and accepts the same immutable `UserMessage` as a new steering occurrence. The steering occurrence receives a new `InboxItemId` and truthful `placement: 'steering'`, while the message retains its `MessageId`, content, and source. AgentLoop installs the new outbox entry before publishing lifecycle events, then emits its enqueue before the old occurrence's discard so re-entrant cancellation cannot observe or retire an unannounced item. The existing inbox conservation invariant therefore continues to require one enqueue and one terminal dequeue or discard for each occurrence. + +The action does not run `agent/prompt-submit`: choosing steering intentionally changes delivery from an independently admitted turn to current-turn next-step input. It neither cancels current work nor reorders the remaining Queue. + +### Host and client boundary + +`session.updateQueue` carries the `steer` action and maps the two negative outcomes to typed RPC errors. The conversion is one synchronous Agent operation; the Host never reconstructs it by combining remove and prompt calls. + +The Host's transient `session/queue` projection remains Queue-only. It ignores the new pending steering occurrence and removes the old row when its discard arrives. Pending steering does not gain edit, delete, or reconnect presentation in this cut. A later dedicated pending-steering projection may add that observability without widening Queue mutation semantics. + +The existing `session.prompt(mode: 'steer')` contract remains best-effort for new input: outside the next-step window it may become a waking follow-up. Only the Queue row action is strict, because failure can safely leave its already-pending message untouched. + +### Verification + +AgentLoop contract coverage holds prompt admission open, converts one exact queued occurrence, and proves the replacement steering occurrence keeps the message value, drains as `steering/message`, and never starts its former independent turn. It also pins unavailable-window retention, claimed-address rejection, and re-entrant cancellation lifecycle conservation. + +Host schema and proxy tests cover the new action, both typed errors, authoritative Queue snapshots, and the absence of pending steering from reconnect snapshots. QueueDock tests cover running-state enablement, complete-content eligibility, failure retention, and authoritative success retirement. + +The keyless Web steering scenario queues a message through the real composer while the first response streams, activates the row arrow, then uses `ask_user_question` as a stable pending-steering barrier. After the answer, it proves one badged interjection becomes durable and the next model request obeys it. Queue edit/delete scenarios continue to prove those actions are unchanged. + +## Alternatives considered + +**Delete the row, then call `session.prompt(mode: 'steer')` from Web.** Rejected because two RPCs cannot make deletion and steering atomic; failure and driver-claim races can lose or duplicate the user's message. + +**Restore Queue promotion under the upward arrow.** Rejected because moving an item to the front still creates an independent admitted turn. The control promises current-turn steering, not priority within Queue. + +**Use the existing best-effort `agent.steer()` behavior.** Rejected for this action because a closed next-step window would silently turn the selected row back into queued work, possibly at a different position and identity. Strict failure preserves the original occurrence and makes the semantic race visible. + +**Change `agent.steer()` to be strict for every caller.** Rejected because TUI and plugin callers use its safe follow-up fallback for newly submitted input. A queued row has recoverable state that those callers do not. + +**Preserve the same `InboxItemId` while changing placement.** Rejected because `InboxItemId` identifies one FIFO acceptance and `placement` records that acceptance's resolved delivery. Ending one queued occurrence and accepting one steering occurrence keeps lifecycle facts truthful and leaves the conservation invariant unchanged. + +**Expose pending steering in `session/queue`.** Deferred because the existing product design provides no pending-steering row state or operations. Authoritative Queue retirement plus the durable consumed bubble is sufficient for the first interaction cut; reconnect visibility can be added through a dedicated projection if product testing shows the gap matters. + +**Cancel the active turn and run the selected Queue item.** Rejected because it destroys unrelated in-flight work and starts a new turn rather than steering the current one. + +## Consequences + +A successful action can be pending but absent from the Web after its Queue row retires and before `steering/message` commits; a refresh during that interval has no pending-steering indication. The running bit can also remain true briefly after the strict next-step window closes, so the button may be enabled for an operation that correctly returns `steer-unavailable`. + +The explicit action changes delivery from an independently admitted turn to current-turn steering, so prompt-admission plugins do not process the converted message. Enqueue-before-discard lifecycle publication remains required for re-entrant cancellation safety; focused regression coverage protects that ordering. diff --git a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md new file mode 100644 index 0000000000..92225f3a61 --- /dev/null +++ b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md @@ -0,0 +1,67 @@ +# Agent Note(agent 决策记录):将 Web 已排队消息转为活动轮次的 steering(中途引导) + +Status: implemented + +[English](2026-07-30-web-queue-steer-action.md) | 中文 + +## 问题 + +Web composer 会在 agent 运行期间有意把 Enter 提交作为 Queue 入队。QueueDock 已经为每条待处理消息提供可寻址的行,持久 transcript(文本记录)也已能把消费后的 `steering/message` 事件渲染为带插话徽标的消息,但 Web 没有连接这两个界面的操作。 + +如果 Web 先在客户端删除该行,再调用 `session.prompt(mode: 'steer')`,就会把用户的一次意图拆分到两个 RPC 中。驱动器可能在两次调用之间先认领该项,steering 投递也可能在删除后失败;现有尽力而为的 `agent.steer()` 回退还可能在原单次入队项被移除后,静默追加一个新的 Queue 项。因此,立即发送操作必须区分当前轮次 steering 与 Queue 前移,并在 steering 已不可用时保留原行。 + +## 决策 + +### 产品契约 + +每个非编辑态的 QueueDock 行都会提供名为“插话发送”的向上箭头操作。仅当会话报告 agent 正在运行时,该操作才会启用;包含混合内容的消息仍可使用,因为 steering 会转发完整且不可变的 `UserMessage`,而非该行的文本投影。编辑和删除保持现有行为,composer 也继续把 Enter 提交为 Queue。 + +触发该操作会针对对应的 `InboxItemId` 请求严格的当前轮次 steering。操作成功后,权威 Host 快照会移除 Queue 行。AgentLoop 排空该项时,现有持久 `steering/message` 事件与 transcript 插话徽标会渲染这条消息,无需新增聊天展示路径。 + +running 标志位只用于提示交互状态。在同步变更边界上,AgentLoop 的 `acceptsNextStep` 值才是权威依据。如果该窗口已经关闭,操作会保持 Queue 单次入队项不变,并返回类型化的 `steer-unavailable` 错误;如果驱动器已经认领该项,则返回现有的 `queue-item-not-found` 错误。UI 会报告任一竞态,不会乐观地移除该行。 + +### Agent 与生命周期边界 + +`InboxAction` 会在编辑和移除之外,新增由实际消费方支撑的 `{ kind: 'steer' }` 操作。`Agent.updateInbox()` 只有在找到 queued 单次入队项并确认 `acceptsNextStep` 后才会处理该操作,绝不会委托给尽力而为的 `agent.steer()` 别名。 + +操作成功应用后,系统会结束 queued 单次入队项,并把同一个不可变 `UserMessage` 接受为新的 steering 单次入队项。steering 单次入队项会获得新的 `InboxItemId` 和如实反映投递方式的 `placement: 'steering'`,消息则保留其 `MessageId`、内容和来源。AgentLoop 会先安装新的 outbox 项,再发布生命周期事件;随后先发出新单次入队项的 enqueue,再发出旧单次入队项的 discard,确保可重入取消无法观察或退役一个尚未公布的项。因此,现有 inbox 守恒不变量仍然要求每个单次入队项恰好对应一个 enqueue,以及一个终态 dequeue 或 discard。 + +该操作不会运行 `agent/prompt-submit`:选择 steering 会有意把投递方式从经独立接纳的轮次改为当前轮次的 next-step 输入。它既不会取消当前工作,也不会重新排序 Queue 中的剩余项。 + +### Host 与客户端边界 + +`session.updateQueue` 会携带 `steer` 操作,并把两种负面结果映射为类型化 RPC 错误。这项转换是一次同步 Agent 操作;Host 绝不会通过组合移除和提示词调用来重建它。 + +Host 的瞬态 `session/queue` 投影仍然只包含 Queue。它会忽略新的待处理 steering 单次入队项,并在收到旧项的 discard 时移除原行。本阶段不会为待处理 steering 增加编辑、删除或重连展示。未来可以用专用的待处理 steering 投影补充这种可观测性,而无需扩大 Queue 变更语义。 + +现有 `session.prompt(mode: 'steer')` 对新输入仍采用尽力而为的契约:在 next-step 窗口之外,它可能变为会唤醒 agent 的后续轮次。只有 Queue 行操作采用严格语义,因为失败时可以安全地保留其已经待处理的消息。 + +### 验证 + +AgentLoop 契约覆盖保持提示词接纳窗口打开,转换一个精确的 queued 单次入队项,并证明替代它的 steering 单次入队项保留消息值、以 `steering/message` 的形式排空,且绝不启动原本的独立轮次。该覆盖还钉住窗口不可用时保留原项、拒绝已被认领的地址,以及可重入取消下的生命周期守恒。 + +Host schema 和代理测试覆盖新操作、两种类型化错误、权威 Queue 快照,以及重连快照不包含待处理 steering。QueueDock 测试覆盖按运行状态启用、混合内容消息仍可完整投递、失败时保留原行,以及成功后由权威快照退役。 + +无密钥 Web steering 场景在第一次响应流式输出期间,通过真实 composer 排队一条消息并触发行上的箭头,再用 `ask_user_question` 作为稳定的待处理 steering 屏障。回答问题后,该场景证明一条带徽标的插话成为持久记录,并且下一次模型请求遵循它。Queue 编辑/删除场景继续证明这些操作没有变化。 + +## 考虑过的替代方案 + +**在 Web 中删除该行,再调用 `session.prompt(mode: 'steer')`。** 不予采纳,因为两个 RPC 无法让删除和 steering 成为原子操作;失败和驱动器认领竞态可能丢失或重复用户消息。 + +**恢复向上箭头对应的 Queue 前移操作。** 不予采纳,因为把某个项移到队首仍然会创建一个独立接纳的轮次。该控件承诺的是当前轮次 steering,而不是 Queue 内的优先级。 + +**使用现有尽力而为的 `agent.steer()` 行为。** 不予采纳,因为关闭的 next-step 窗口会静默地把选中行重新变成 queued 工作,而且位置和标识可能不同。严格失败会保留原单次入队项,并让这项语义竞态明确可见。 + +**让每个调用方使用的 `agent.steer()` 都采用严格语义。** 不予采纳,因为 TUI 和插件调用方会针对新提交的输入使用其安全的后续轮次回退。queued 行具有这些调用方不具备的可恢复状态。 + +**改变投递方式时保留同一个 `InboxItemId`。** 不予采纳,因为 `InboxItemId` 标识一次 FIFO 接受,而 `placement` 记录该次接受解析出的投递方式。结束一个 queued 单次入队项并接受一个 steering 单次入队项,能够使生命周期事实保持如实,并让守恒不变量保持不变。 + +**在 `session/queue` 中暴露待处理 steering。** 暂缓,因为现有产品设计没有为待处理 steering 提供行状态或操作。权威的 Queue 退役加上持久的已消费气泡,足以支撑首个交互阶段;如果产品测试表明这一缺口影响显著,可以通过专用投影增加重连可见性。 + +**取消活动轮次并运行选中的 Queue 项。** 不予采纳,因为这会破坏无关的进行中工作,并且会启动新轮次,而不是 steering 当前轮次。 + +## 后果 + +操作成功后,从 Queue 行退役到 `steering/message` 提交之间,对应消息可能仍处于待处理状态,却不会出现在 Web 中;如果在此期间刷新,界面不会显示待处理 steering。严格 next-step 窗口关闭后,running 标志位仍可能短暂保持为 true,因此按钮可能会为一个最终正确返回 `steer-unavailable` 的操作保持启用。 + +这项显式操作会把投递方式从经独立接纳的轮次改为当前轮次 steering,因此提示词接纳插件不会处理转换后的消息。为保证可重入取消安全,生命周期事件仍必须先发布 enqueue 再发布 discard;有针对性的回归覆盖会保护这一顺序。 diff --git a/apps/web/tests/snapshots/queue-actions/editing.expected.md b/apps/web/tests/snapshots/queue-actions/editing.expected.md index 3a70840713..3f6ebab8a2 100644 --- a/apps/web/tests/snapshots/queue-actions/editing.expected.md +++ b/apps/web/tests/snapshots/queue-actions/editing.expected.md @@ -23,6 +23,8 @@ - img - button "Remove queued message": - img + - button "插话发送": + - img - listitem: - textbox "Edit queued message": Edited queue item - button "Save queued message": diff --git a/apps/web/tests/snapshots/queue-actions/ui.expected.md b/apps/web/tests/snapshots/queue-actions/ui.expected.md index 24edb57417..db860c30d1 100644 --- a/apps/web/tests/snapshots/queue-actions/ui.expected.md +++ b/apps/web/tests/snapshots/queue-actions/ui.expected.md @@ -22,6 +22,8 @@ - img - button "Remove queued message": - img + - button "Steer queued message": + - img - textbox "Message the agent" - button "Commands": - img diff --git a/apps/web/tests/steering.e2e.ts b/apps/web/tests/steering.e2e.ts index a329f2c02e..bfa8aae162 100644 --- a/apps/web/tests/steering.e2e.ts +++ b/apps/web/tests/steering.e2e.ts @@ -1,15 +1,7 @@ -// Web e2e scenario: mid-turn steering over the host wire. The Web UI has no -// steer entry, so the steer is POSTed from the page over the same -// same-origin /api transport the client uses. Everything downstream is -// product: the gateway routes mode:'steer' to Agent.steer, the loop drains -// it at the step boundary into a durable steering/message event, the SSE mux -// pushes it, and the transcript shows the text as a plain bubble (no -// interjection chrome). The question composer supplies the deterministic -// mid-turn window: while ask_user_question blocks, the turn is provably -// running, so record and replay perform the identical steer-then-answer -// sequence with zero timing dependence — and the recorded final reply proves -// the steer reached the MODEL (it obeys an instruction that only the -// steering message carries). +// Web e2e scenario: queue a message while the first response streams, strictly +// transfer that exact occurrence to steering through QueueDock, then prove it +// is logged, rendered, and obeyed. The following question tool supplies a +// deterministic pending-steering snapshot before the step can drain. import { readFile } from 'node:fs/promises' import { fileURLToPath } from 'node:url' import { join } from 'node:path' @@ -56,15 +48,12 @@ describe('web e2e: mid-turn steering lands durably and visibly', () => { let browser: Browser let page: Page let tripwire: ReturnType - let liveSessionId: string | undefined const sessionEvents: SessionEvent[] = [] beforeAll(async () => { - scaffold = await launchWebScaffold(MODE === 'record' ? {} : { replayFixture: FIXTURE, paceMs: 15 }) - scaffold.ctx.on('session/event', (session, event) => { - liveSessionId ??= session.id - sessionEvents.push(event) - }) + // The slower replay keeps the Queue action available until the recorded question barrier arrives. + scaffold = await launchWebScaffold(MODE === 'record' ? {} : { replayFixture: FIXTURE, paceMs: 100 }) + scaffold.ctx.on('session/event', (_session, event) => { sessionEvents.push(event) }) browser = await chromium.launch() page = await newEnglishPage(browser) tripwire = watchConsole(page) @@ -79,7 +68,7 @@ describe('web e2e: mid-turn steering lands durably and visibly', () => { await scaffold?.close() }) - it('steers during the blocked step; the message is logged, rendered, and obeyed', async () => { + it('strictly steers one queued row; the interjection is logged, rendered, and obeyed', async () => { onTestFailed(() => saveFailureShot(page, 'web-e2e-steering')) if (MODE !== 'record') { // The steer must NOT be a user/message — it lands as steering/message. @@ -91,35 +80,27 @@ describe('web e2e: mid-turn steering lands durably and visibly', () => { await input.fill(PROMPT) await input.press('Enter') - // The blocked composer is the mid-turn barrier: its presence proves the - // ask_user_question step is executing, i.e. the turn is running NOW. + // Enter remains the Queue gesture. The row action then atomically moves + // this exact occurrence into the current turn's steering outbox. + await input.fill(STEER) + await input.press('Enter') + const queued = page.getByText(STEER, { exact: true }) + await queued.waitFor({ timeout: 10_000 }) + const queuedRow = page.getByRole('listitem').filter({ hasText: STEER }) + const steerButton = queuedRow.getByRole('button', { name: 'Steer queued message' }) + await expect.poll(() => steerButton.isEnabled(), { timeout: 10_000 }).toBe(true) + await steerButton.click({ timeout: 10_000 }) + await expect.poll(() => page.getByText(STEER, { exact: true }).count(), { timeout: 10_000 }).toBe(0) + + // The blocked composer is the mid-turn barrier: the tool cannot finish + // this step, so the accepted steering remains pending and invisible. const composer = page.locator('[data-question-key]') await composer.waitFor({ timeout: MODE === 'record' ? 120_000 : 30_000 }) - // Steer through the real wire from the page (same envelope + endpoint the - // web client's session.prompt uses). accepted:true is the transport proof. - expect(liveSessionId).toBeDefined() - const reply = await page.evaluate(async ({ sessionId, text }) => { - const response = await fetch('/api/session.prompt', { - method: 'POST', - headers: { 'content-type': 'application/json' }, - body: JSON.stringify({ - type: 'client-request', - rpcId: crypto.randomUUID(), - method: 'session.prompt', - payload: { sessionId, mode: 'steer', content: [{ type: 'text', text }] }, - }), - }) - return await response.json() as { result?: { ok?: boolean } } - }, { sessionId: liveSessionId!, text: STEER }) - expect(reply.result?.ok).toBe(true) - if (MODE !== 'record') { - // Mid-turn golden: the ACCEPTED steer is durable in the inbox but the - // loop drains steering only at the step boundary, so no steering/message - // exists yet and no steer text renders — the composer still blocks, - // alone. The DOM is stable here (no further SSE frames can arrive until - // the question is answered), making this state capturable. + // Mid-turn golden: the converted steer is pending in the loop but the + // loop drains steering only at the step boundary, so no Queue row or + // steering/message bubble renders while the question still blocks. expect(await page.getByText(STEER, { exact: true }).count()).toBe(0) expect(await page.getByRole('button', { name: 'Edit queued message' }).count()).toBe(0) const snapshot = await captureStableAria(page, '[class*="centerCol"]', scaffold.workspaceCwd) diff --git a/docs/architecture.i18n.yaml b/docs/architecture.i18n.yaml index 576ef05d32..f0e2a9892c 100644 --- a/docs/architecture.i18n.yaml +++ b/docs/architecture.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write docs/architecture.md -architecture.md: 6aa942ba2702d8d30ae94d9968f07abb5e1fe88d -architecture.zh.md: c8aaa68527f34f4879f882a08260a4e0bd4f4c5f +architecture.md: 4bb0cb1f29bb48adf89c97af7c85c90219d0558a +architecture.zh.md: 4c6ff721d894aaee4cc14ff321967580997aba56 diff --git a/docs/architecture.md b/docs/architecture.md index 6aa942ba27..4bb0cb1f29 100644 --- a/docs/architecture.md +++ b/docs/architecture.md @@ -123,6 +123,8 @@ Each step assembles ordered stable system sections, cache-safe dynamic contexts, Admission-time and active-turn `inject()` stage for the next step; tool-time injection and post-tool `additionalContexts` settle after results. Steering shares the outbox but remains provisional until a request admits it. `steer()` returns a message-owned receipt: after `agent/step` and asynchronous prompt assembly succeed, the loop commits the stable batch, snapshots request history, opens `step/start`, then resolves its receipts as admitted with the turn and step; later arrivals wait. A turn-concluding tool result, broad cancellation, disposal, or a claimed idle-steering turn that never opens a step rejects affected receipts, while `cancel(..., { keepInbox: true })` and non-terminal routing preserve pending delivery. Idle `inject()` appends immediately without changing turn numbers; persistence drains eagerly. +Before driver claim, `updateInbox()` may edit or remove a queued occurrence, or strictly transfer its immutable message into an open next-step window. That transfer ends the queued occurrence and accepts a new steering occurrence; a closed window leaves Queue unchanged. Direct `steer()` remains best-effort for newly submitted input and falls back to a waking follow-up outside the window ([decision](../.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md)). + Pruning precedes summaries; overflow retries require durable progress. `agent/request-error` may authorize one retry turn between failed-step and turn close; cancellation wins. Adapter-owned `retryPolicy` makes normal mode bounded; always mode delegates specialized recovery before retrying until success or cancellation ([compaction](../.agents/notes/implemented/architecture/2026-07-10-after-call-compaction-pressure-and-overflow-recovery.md), [retry foundation](../.agents/notes/implemented/architecture/2026-06-21-bounded-llm-request-recovery.md), [provider policy](../.agents/notes/implemented/feature/2026-07-24-provider-retry-policies.md)). ### Failure Boundaries diff --git a/docs/architecture.zh.md b/docs/architecture.zh.md index c8aaa68527..4c6ff721d8 100644 --- a/docs/architecture.zh.md +++ b/docs/architecture.zh.md @@ -123,6 +123,8 @@ idle inject: 接纳期间和活跃轮次内的 `inject()` 会为下一步骤暂存;工具执行期间的注入和工具执行后的 `additionalContexts` 会在结果记录完毕后落定。steering 与其共用 outbox,但在请求接纳前始终处于待准入状态。`steer()` 会返回归属于该消息的回执:`agent/step` 和异步提示词组装成功后,循环提交稳定批次、捕获请求历史并开启 `step/start`,再将其回执解析为已准入并附带轮次与步骤;后续消息继续等待。结束轮次的工具结果、广义取消、dispose(资源释放),以及已领取 idle-steering 消息却从未开启步骤的轮次,都会拒绝受影响的回执;`cancel(..., { keepInbox: true })` 和非终止型路由则保留待处理投递。空闲状态下的 `inject()` 会立即追加,且不改变轮次编号;持久化层会尽快排空。 +驱动器认领之前,`updateInbox()` 可以编辑或移除 queued 单次入队项,也可以严格地把其不可变消息转移到开放的 next-step 窗口。该转移会结束 queued 单次入队项,并接受一个新的 steering 单次入队项;窗口关闭时 Queue 保持不变。直接调用 `steer()` 时,对新提交的输入仍采用尽力而为的语义,并在窗口之外回退为会唤醒 agent 的后续轮次([决策](../.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md))。 + 裁剪先于摘要;溢出重试必须取得持久进展。`agent/request-error` 可以在失败步骤与轮次关闭之间授权一个重试轮次;取消优先。适配器拥有的 `retryPolicy` 使 normal mode 保持有界;always mode 先委托专门恢复,再持续重试直至成功或取消([压缩](../.agents/notes/implemented/architecture/2026-07-10-after-call-compaction-pressure-and-overflow-recovery.md)、[重试基础](../.agents/notes/implemented/architecture/2026-06-21-bounded-llm-request-recovery.md)、[提供方策略](../.agents/notes/implemented/feature/2026-07-24-provider-retry-policies.md))。 ### 失败边界 diff --git a/docs/cordis-catalog/events.md b/docs/cordis-catalog/events.md index 55785ad536..521a2fa735 100644 --- a/docs/cordis-catalog/events.md +++ b/docs/cordis-catalog/events.md @@ -32,7 +32,7 @@ Effective broad cancellation was requested, before queued/outbox work is cleared Types: [Agent](../core-data-structures/core.md) · [AgentCancelCause](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) -Source: [`packages/core/agent/src/types.ts:349`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:353`](../../packages/core/agent/src/types.ts) ### `agent/created` — emit @@ -54,7 +54,7 @@ A fully configured agent and live session were published. Setup is composition-o Types: [Agent](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) -Source: [`packages/core/agent/src/types.ts:280`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:284`](../../packages/core/agent/src/types.ts) ### `agent/disposed` — emit @@ -74,7 +74,7 @@ An agent left the registry; AgentLoop emits this after driver quiescence and sco Types: [Agent](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) -Source: [`packages/core/agent/src/types.ts:289`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:293`](../../packages/core/agent/src/types.ts) ### `agent/error` — emit @@ -96,7 +96,7 @@ A step or turn errored. The machine reports a failure here (plus the logger) eve Types: [Agent](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) -Source: [`packages/core/agent/src/types.ts:463`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:467`](../../packages/core/agent/src/types.ts) ### `agent/inbox/dequeue` — emit @@ -117,7 +117,7 @@ The driver claimed one item out of the inbox: a queued item at a turn boundary, Types: [Agent](../core-data-structures/core.md) · [InboxItem](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) -Source: [`packages/core/agent/src/types.ts:327`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:331`](../../packages/core/agent/src/types.ts) ### `agent/inbox/discard` — emit @@ -140,7 +140,7 @@ Pending inbox items were dropped without delivering them, so every enqueue occur Types: [Agent](../core-data-structures/core.md) · [InboxItem](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) -Source: [`packages/core/agent/src/types.ts:339`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:343`](../../packages/core/agent/src/types.ts) ### `agent/inbox/enqueue` — emit @@ -161,7 +161,7 @@ An item entered the queued or steering inbox. `placement` is the acceptance-time Types: [Agent](../core-data-structures/core.md) · [InboxItem](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) -Source: [`packages/core/agent/src/types.ts:308`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:312`](../../packages/core/agent/src/types.ts) ### `agent/inbox/update` — emit @@ -181,7 +181,7 @@ A still-pending queued item changed content. The item id, placement, and positio Types: [Agent](../core-data-structures/core.md) · [InboxItem](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) -Source: [`packages/core/agent/src/types.ts:317`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:321`](../../packages/core/agent/src/types.ts) ### `agent/prompt-submit` — waterfall @@ -204,7 +204,7 @@ Allow, rewrite, or block one claimed prompt before it becomes a user message or Types: [Agent](../core-data-structures/core.md) · [PromptDecision](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) · [UserMessage](../core-data-structures/session.md) -Source: [`packages/core/agent/src/types.ts:376`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:380`](../../packages/core/agent/src/types.ts) ### `agent/request` — waterfall @@ -228,7 +228,7 @@ Replace the frozen call configuration. `await next()` yields the config the mach Types: [Agent](../core-data-structures/core.md) · [LlmCallConfig](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) -Source: [`packages/core/agent/src/types.ts:402`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:406`](../../packages/core/agent/src/types.ts) ### `agent/request-error` — waterfall @@ -258,7 +258,7 @@ Handle a model-request failure after its failed step has closed but before the f Types: [Agent](../core-data-structures/core.md) · [LlmFailure](../core-data-structures/llm-streaming.md) · [RequestError](../core-data-structures/core.md) · [RequestErrorAction](../core-data-structures/core.md) · [ResolvedRetryPolicy](../core-data-structures/llm-streaming.md) · [Scoped](../core-data-structures/scope.md) -Source: [`packages/core/agent/src/types.ts:421`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:425`](../../packages/core/agent/src/types.ts) ### `agent/session-start` — emit @@ -280,7 +280,7 @@ The session lifecycle began, once before the first turn. Use `agent.inject()` to Types: [Agent](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) · [SessionStartSource](../core-data-structures/core.md) -Source: [`packages/core/agent/src/types.ts:362`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:366`](../../packages/core/agent/src/types.ts) ### `agent/settled` — emit @@ -305,7 +305,7 @@ One drain chain reached its terminal turn: that turn's `turn/end` is already com Types: [Agent](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) · [SettleReason](../core-data-structures/core.md) -Source: [`packages/core/agent/src/types.ts:450`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:454`](../../packages/core/agent/src/types.ts) ### `agent/status` — emit @@ -325,7 +325,7 @@ Agent status changed (`idle` ⇄ `running`). `send()` does not enter `running` s Types: [Agent](../core-data-structures/core.md) · [AgentStatus](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) -Source: [`packages/core/agent/src/types.ts:298`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:302`](../../packages/core/agent/src/types.ts) ### `agent/step` — serial @@ -349,7 +349,7 @@ Awaited serial checkpoint before EVERY request of a turn is built (the first as Types: [Agent](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) -Source: [`packages/core/agent/src/types.ts:389`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:393`](../../packages/core/agent/src/types.ts) ### `agent/turn-stopping` — serial @@ -375,7 +375,7 @@ The turn is about to close: the model owes no response (no live tool calls, no f Types: [Agent](../core-data-structures/core.md) · [Scoped](../core-data-structures/scope.md) -Source: [`packages/core/agent/src/types.ts:436`](../../packages/core/agent/src/types.ts) +Source: [`packages/core/agent/src/types.ts:440`](../../packages/core/agent/src/types.ts) ## `agent-loop/*` diff --git a/docs/core-data-structures/core.i18n.yaml b/docs/core-data-structures/core.i18n.yaml index a156394d97..415d74e6f6 100644 --- a/docs/core-data-structures/core.i18n.yaml +++ b/docs/core-data-structures/core.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write docs/core-data-structures/core.md -core.md: 795256a2b30e44771baf8bcb7c1d541134692a02 -core.zh.md: 5cebff049aae53df9f0494da1fadc8dfa5d9ad09 +core.md: 1b5704384157688b45ae0900bf2d9924426bbd6b +core.zh.md: 05802039920163ac8185703ab483c5603087ed96 diff --git a/docs/core-data-structures/core.md b/docs/core-data-structures/core.md index 795256a2b3..1b57043841 100644 --- a/docs/core-data-structures/core.md +++ b/docs/core-data-structures/core.md @@ -516,11 +516,12 @@ interface InboxItem { type InboxAction = | { readonly kind: 'edit'; readonly content: ContentBlock[] } | { readonly kind: 'remove' } + | { readonly kind: 'steer' } ``` ```ts type-equiv /** Result of applying an inbox action at the synchronous ownership boundary. */ -type InboxActionResult = 'applied' | 'not-found' +type InboxActionResult = 'applied' | 'not-found' | 'steer-unavailable' ``` ```ts type-equiv @@ -546,7 +547,7 @@ interface SendOptions { } ``` -The fixed-preset aliases own `target` and `wakeup`; their already identified `UserMessage` carries role, content, and provenance. Its `MessageId` remains stable when an edit replaces the message content, while the enclosing `InboxItemId` identifies one accepted occurrence across `agent/inbox/enqueue`, `agent/inbox/update`, and its terminal dequeue or discard. Injection bypasses the FIFOs and never appears on those events. +The fixed-preset aliases own `target` and `wakeup`; their already identified `UserMessage` carries role, content, and provenance. Its `MessageId` remains stable when an edit replaces content or strict steer transfers the immutable message. The original queued occurrence ends and strict steer accepts a new steering occurrence with a distinct `InboxItemId`. Injection bypasses the FIFOs and never appears on inbox lifecycle events. ```ts type-equiv /** Options for {@link Agent.cancel}. */ @@ -633,10 +634,13 @@ interface Agent { /** * Mutate one still-pending queued occurrence synchronously. Editing preserves * the message identity and queue position; removal publishes its terminal - * discard. Steering occurrences and driver-claimed items return `not-found`. + * discard. Steer strictly transfers the message into the current next-step + * window, or returns `steer-unavailable` without changing the queued + * occurrence. Steering occurrences and driver-claimed items return + * `not-found`. * @param id - independently addressable queued occurrence. - * @param action - edit or remove operation. - * @returns whether the pending occurrence was found and updated. + * @param action - edit, remove, or strict steer operation. + * @returns the applied outcome or the reason no mutation occurred. */ updateInbox(id: InboxItemId, action: InboxAction): InboxActionResult diff --git a/docs/core-data-structures/core.zh.md b/docs/core-data-structures/core.zh.md index 5cebff049a..0580203992 100644 --- a/docs/core-data-structures/core.zh.md +++ b/docs/core-data-structures/core.zh.md @@ -524,11 +524,12 @@ interface InboxItem { type InboxAction = | { readonly kind: 'edit'; readonly content: ContentBlock[] } | { readonly kind: 'remove' } + | { readonly kind: 'steer' } ``` ```ts type-equiv /** Result of applying an inbox action at the synchronous ownership boundary. */ -type InboxActionResult = 'applied' | 'not-found' +type InboxActionResult = 'applied' | 'not-found' | 'steer-unavailable' ``` ```ts type-equiv @@ -554,7 +555,7 @@ interface SendOptions { } ``` -固定预设的别名方法自带 `target` 与 `wakeup`;其已有标识的 `UserMessage` 会携带角色、内容与 provenance。编辑替换消息内容时,其 `MessageId` 保持稳定;外层 `InboxItemId` 则在 `agent/inbox/enqueue`、`agent/inbox/update` 及终态 dequeue 或 discard 之间标识同一次入队。注入绕过两个 FIFO,从不出现在这些事件中。 +固定预设的别名方法自带 `target` 与 `wakeup`;其已有标识的 `UserMessage` 会携带角色、内容与 provenance。编辑替换内容或严格 steering(中途引导)转移不可变消息时,其 `MessageId` 都保持稳定。原 queued 单次入队项会结束,严格 steering 则接受一个具有不同 `InboxItemId` 的新 steering 单次入队项。注入绕过两个 FIFO,从不出现在 inbox 生命周期事件中。 ```ts type-equiv /** Options for {@link Agent.cancel}. */ @@ -641,10 +642,13 @@ interface Agent { /** * Mutate one still-pending queued occurrence synchronously. Editing preserves * the message identity and queue position; removal publishes its terminal - * discard. Steering occurrences and driver-claimed items return `not-found`. + * discard. Steer strictly transfers the message into the current next-step + * window, or returns `steer-unavailable` without changing the queued + * occurrence. Steering occurrences and driver-claimed items return + * `not-found`. * @param id - independently addressable queued occurrence. - * @param action - edit or remove operation. - * @returns whether the pending occurrence was found and updated. + * @param action - edit, remove, or strict steer operation. + * @returns the applied outcome or the reason no mutation occurred. */ updateInbox(id: InboxItemId, action: InboxAction): InboxActionResult diff --git a/docs/event-producer-consumer.md b/docs/event-producer-consumer.md index 1c4309004b..d1457a75af 100644 --- a/docs/event-producer-consumer.md +++ b/docs/event-producer-consumer.md @@ -8,22 +8,22 @@ This matrix shows which packages dispatch each harness-owned event and which pac | Event | Mode | Declared in | Dispatchers | Listeners | | --- | --- | --- | --- | --- | | `agent-loop/config-start-failed` | `emit` | [`packages/core/agent-loop/src/index.ts:157`](../packages/core/agent-loop/src/index.ts) | [`agent-loop`](../packages/core/agent-loop) (`events.dispatch`) | [`tui`](../packages/ui/tui) | -| `agent/cancel-requested` | `emit` | [`packages/core/agent/src/types.ts:349`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | [`goal-session`](../packages/goal/goal-session) | -| `agent/created` | `emit` | [`packages/core/agent/src/types.ts:280`](../packages/core/agent/src/types.ts) | [`agent`](../packages/core/agent) (`events.dispatch`) | [`goal-session`](../packages/goal/goal-session), [`tui`](../packages/ui/tui) | -| `agent/disposed` | `emit` | [`packages/core/agent/src/types.ts:289`](../packages/core/agent/src/types.ts) | [`agent`](../packages/core/agent) (`events.dispatch`) | [`agent-loop`](../packages/core/agent-loop), [`goal-session`](../packages/goal/goal-session), [`subagent`](../packages/subagent/subagent), [`tui`](../packages/ui/tui) | -| `agent/error` | `emit` | [`packages/core/agent/src/types.ts:463`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | `apiproxy`, [`goal-session`](../packages/goal/goal-session), [`session-telemetry`](../packages/telemetry/session-telemetry), [`tui`](../packages/ui/tui) | -| `agent/inbox/dequeue` | `emit` | [`packages/core/agent/src/types.ts:327`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | [`agent`](../packages/core/agent), `apiproxy`, [`subagent`](../packages/subagent/subagent), [`tui`](../packages/ui/tui) | -| `agent/inbox/discard` | `emit` | [`packages/core/agent/src/types.ts:339`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | [`agent`](../packages/core/agent), `apiproxy`, [`subagent`](../packages/subagent/subagent), [`tui`](../packages/ui/tui) | -| `agent/inbox/enqueue` | `emit` | [`packages/core/agent/src/types.ts:308`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | [`agent`](../packages/core/agent), `apiproxy`, [`goal-session`](../packages/goal/goal-session) | -| `agent/inbox/update` | `emit` | [`packages/core/agent/src/types.ts:317`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | `apiproxy` | -| `agent/prompt-submit` | `waterfall` | [`packages/core/agent/src/types.ts:376`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`goal-session`](../packages/goal/goal-session), [`hooks-claude`](../packages/hooks/hooks-claude), [`hooks-codex`](../packages/hooks/hooks-codex), [`repeat-tool-guard`](../packages/guard/repeat-tool-guard), [`tui`](../packages/ui/tui) | -| `agent/request` | `waterfall` | [`packages/core/agent/src/types.ts:402`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`agent`](../packages/core/agent) | -| `agent/request-error` | `waterfall` | [`packages/core/agent/src/types.ts:421`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`compact-basic`](../packages/compact/compact-basic), [`llm-retry`](../packages/llm/llm-retry) | -| `agent/session-start` | `emit` | [`packages/core/agent/src/types.ts:362`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | [`goal`](../packages/goal/goal), [`goal-session`](../packages/goal/goal-session), [`hooks-claude`](../packages/hooks/hooks-claude), [`hooks-codex`](../packages/hooks/hooks-codex), [`workspace-context`](../packages/context/workspace-context) | -| `agent/settled` | `emit` | [`packages/core/agent/src/types.ts:450`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | [`compact-basic`](../packages/compact/compact-basic) | -| `agent/status` | `emit` | [`packages/core/agent/src/types.ts:298`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | [`agent`](../packages/core/agent), `apiproxy`, [`goal-session`](../packages/goal/goal-session), [`tui`](../packages/ui/tui) | -| `agent/step` | `serial` | [`packages/core/agent/src/types.ts:389`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`serial`) | [`compact-basic`](../packages/compact/compact-basic), [`plan-mode`](../packages/plan/plan-mode), [`session-checkpoint-policy`](../packages/session-persistence/session-checkpoint-policy), [`subagent-inprocess`](../packages/subagent/subagent-inprocess), [`time-context`](../packages/context/time-context), [`tmux-context`](../packages/context/tmux-context), [`tool-skill`](../packages/skill/tool-skill), [`workspace-context`](../packages/context/workspace-context) | -| `agent/turn-stopping` | `serial` | [`packages/core/agent/src/types.ts:436`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`serial`) | [`hooks-claude`](../packages/hooks/hooks-claude), [`hooks-codex`](../packages/hooks/hooks-codex) | +| `agent/cancel-requested` | `emit` | [`packages/core/agent/src/types.ts:353`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | [`goal-session`](../packages/goal/goal-session) | +| `agent/created` | `emit` | [`packages/core/agent/src/types.ts:284`](../packages/core/agent/src/types.ts) | [`agent`](../packages/core/agent) (`events.dispatch`) | [`goal-session`](../packages/goal/goal-session), [`tui`](../packages/ui/tui) | +| `agent/disposed` | `emit` | [`packages/core/agent/src/types.ts:293`](../packages/core/agent/src/types.ts) | [`agent`](../packages/core/agent) (`events.dispatch`) | [`agent-loop`](../packages/core/agent-loop), [`goal-session`](../packages/goal/goal-session), [`subagent`](../packages/subagent/subagent), [`tui`](../packages/ui/tui) | +| `agent/error` | `emit` | [`packages/core/agent/src/types.ts:467`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | `apiproxy`, [`goal-session`](../packages/goal/goal-session), [`session-telemetry`](../packages/telemetry/session-telemetry), [`tui`](../packages/ui/tui) | +| `agent/inbox/dequeue` | `emit` | [`packages/core/agent/src/types.ts:331`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | [`agent`](../packages/core/agent), `apiproxy`, [`subagent`](../packages/subagent/subagent), [`tui`](../packages/ui/tui) | +| `agent/inbox/discard` | `emit` | [`packages/core/agent/src/types.ts:343`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | [`agent`](../packages/core/agent), `apiproxy`, [`subagent`](../packages/subagent/subagent), [`tui`](../packages/ui/tui) | +| `agent/inbox/enqueue` | `emit` | [`packages/core/agent/src/types.ts:312`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | [`agent`](../packages/core/agent), `apiproxy`, [`goal-session`](../packages/goal/goal-session) | +| `agent/inbox/update` | `emit` | [`packages/core/agent/src/types.ts:321`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | `apiproxy` | +| `agent/prompt-submit` | `waterfall` | [`packages/core/agent/src/types.ts:380`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`goal-session`](../packages/goal/goal-session), [`hooks-claude`](../packages/hooks/hooks-claude), [`hooks-codex`](../packages/hooks/hooks-codex), [`repeat-tool-guard`](../packages/guard/repeat-tool-guard), [`tui`](../packages/ui/tui) | +| `agent/request` | `waterfall` | [`packages/core/agent/src/types.ts:406`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`agent`](../packages/core/agent) | +| `agent/request-error` | `waterfall` | [`packages/core/agent/src/types.ts:425`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`compact-basic`](../packages/compact/compact-basic), [`llm-retry`](../packages/llm/llm-retry) | +| `agent/session-start` | `emit` | [`packages/core/agent/src/types.ts:366`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | [`goal`](../packages/goal/goal), [`goal-session`](../packages/goal/goal-session), [`hooks-claude`](../packages/hooks/hooks-claude), [`hooks-codex`](../packages/hooks/hooks-codex), [`workspace-context`](../packages/context/workspace-context) | +| `agent/settled` | `emit` | [`packages/core/agent/src/types.ts:454`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | [`compact-basic`](../packages/compact/compact-basic) | +| `agent/status` | `emit` | [`packages/core/agent/src/types.ts:302`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | [`agent`](../packages/core/agent), `apiproxy`, [`goal-session`](../packages/goal/goal-session), [`tui`](../packages/ui/tui) | +| `agent/step` | `serial` | [`packages/core/agent/src/types.ts:393`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`serial`) | [`compact-basic`](../packages/compact/compact-basic), [`plan-mode`](../packages/plan/plan-mode), [`session-checkpoint-policy`](../packages/session-persistence/session-checkpoint-policy), [`subagent-inprocess`](../packages/subagent/subagent-inprocess), [`time-context`](../packages/context/time-context), [`tmux-context`](../packages/context/tmux-context), [`tool-skill`](../packages/skill/tool-skill), [`workspace-context`](../packages/context/workspace-context) | +| `agent/turn-stopping` | `serial` | [`packages/core/agent/src/types.ts:440`](../packages/core/agent/src/types.ts) | [`agent-loop`](../packages/core/agent-loop) (`serial`) | [`hooks-claude`](../packages/hooks/hooks-claude), [`hooks-codex`](../packages/hooks/hooks-codex) | | `approval/request` | `waterfall` | [`packages/ui/user-approval/src/index.ts:30`](../packages/ui/user-approval/src/index.ts) | [`user-approval`](../packages/ui/user-approval) (`waterfall`) | [`acp`](../packages/acp/acp), `apiproxy` | | `commands/change` | `emit` | [`packages/ui/commands/src/index.ts:154`](../packages/ui/commands/src/index.ts) | [`commands`](../packages/ui/commands) (`events.dispatch`) | `apiproxy`, [`tui`](../packages/ui/tui) | | `credentials/updated` | `emit` | [`packages/credentials/credentials/src/index.ts:67`](../packages/credentials/credentials/src/index.ts) | [`credentials`](../packages/credentials/credentials) (`events.dispatch`) | `apiproxy`, [`credentials`](../packages/credentials/credentials) | diff --git a/packages/client/runtime/README.i18n.yaml b/packages/client/runtime/README.i18n.yaml index f20967ee2f..0bbb962352 100644 --- a/packages/client/runtime/README.i18n.yaml +++ b/packages/client/runtime/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/client/runtime/README.md -README.md: f956be22384a42e9ed30e8aa5f25fe8173cc9f9c -README.zh.md: 49449c51d89d957b5bd39798c9167607f72a5c3a +README.md: a98e97d796ca4e4be07d8d8d25ebc0a24d066a8b +README.zh.md: a973c4fbbf16633fed11d137ae548604516d8a51 diff --git a/packages/client/runtime/README.md b/packages/client/runtime/README.md index f956be2238..a98e97d796 100644 --- a/packages/client/runtime/README.md +++ b/packages/client/runtime/README.md @@ -22,7 +22,7 @@ SlotsService gives the renderer separate bare observables for `useSessions` and ## Pending queue projection -`ConversationSnapshot.queue` is the Host's authoritative transient Queue snapshot; pending steering stays outside this projection. Each row carries its `InboxItemId`, complete editable text when every content block is text, and a flattened preview. `session/queue` replaces the whole projection; reconnect buffering retains only the latest snapshot, and neither durable turn events nor running-status changes guess that an item was claimed. `Session.updateQueue()` sends edit/remove operations without optimistic mutation, so the next Host snapshot is the sole visible commit and a claim race can surface `queue-item-not-found`. +`ConversationSnapshot.queue` is the Host's authoritative transient Queue snapshot; pending steering stays outside this projection. Each row carries its `InboxItemId`, complete editable text when every content block is text, and a flattened preview. `session/queue` replaces the whole projection; reconnect buffering retains only the latest snapshot, and neither durable turn events nor running-status changes guess that an item was claimed. `Session.updateQueue()` sends edit, remove, and strict-steer operations without optimistic mutation, so the next Host snapshot is the sole visible commit; claim and closed-window races surface `queue-item-not-found` and `steer-unavailable`. ## The human transcript diff --git a/packages/client/runtime/README.zh.md b/packages/client/runtime/README.zh.md index 49449c51d8..a973c4fbbf 100644 --- a/packages/client/runtime/README.zh.md +++ b/packages/client/runtime/README.zh.md @@ -22,7 +22,7 @@ SlotsService 分别为 renderer 提供 `useSessions` 与 `useWorkspaces` 的裸 ## 待处理队列投影 -`ConversationSnapshot.queue` 是 Host 提供的权威瞬态 Queue 快照;待处理 steering(中途引导)不进入此投影。每行都携带其 `InboxItemId`、所有内容块均为文本时的完整可编辑文本,以及扁平化预览。`session/queue` 会整体替换该投影;重连缓冲只保留最新快照,持久轮次事件和 running 状态变化都不会猜测某个项已被认领。`Session.updateQueue()` 发送编辑/移除操作,不进行乐观更新,因此下一份 Host 快照是唯一可见的提交结果,认领竞态则会返回 `queue-item-not-found`。 +`ConversationSnapshot.queue` 是 Host 提供的权威瞬态 Queue 快照;待处理 steering(中途引导)不进入此投影。每行都携带其 `InboxItemId`、所有内容块均为文本时的完整可编辑文本,以及扁平化预览。`session/queue` 会整体替换该投影;重连缓冲只保留最新快照,持久轮次事件和 running 状态变化都不会猜测某个项已被认领。`Session.updateQueue()` 发送编辑、移除和严格 steering 操作,不进行乐观更新,因此下一份 Host 快照是唯一可见的提交结果;认领与窗口关闭竞态分别会返回 `queue-item-not-found` 和 `steer-unavailable`。 ## 面向人的 transcript(文本记录) diff --git a/packages/client/runtime/src/client/contract/session.ts b/packages/client/runtime/src/client/contract/session.ts index f2c7d4183d..a68fececa3 100644 --- a/packages/client/runtime/src/client/contract/session.ts +++ b/packages/client/runtime/src/client/contract/session.ts @@ -39,9 +39,9 @@ export interface ISession { */ prompt(content: ContentBlock[], mode: 'queue' | 'steer'): Promise> /** - * Apply one mutation to a still-pending queue occurrence. + * Apply one edit, remove, or strict steer action to a still-pending queue occurrence. * @param itemId - agent-owned inbox occurrence identity. - * @param action - edit or remove operation. + * @param action - requested queue operation. * @returns acceptance, or a business/transport error. */ updateQueue(itemId: InboxItemId, action: QueueAction): Promise> diff --git a/packages/client/runtime/tests/queue-store.spec.ts b/packages/client/runtime/tests/queue-store.spec.ts index 192885b66b..b594e9a9df 100644 --- a/packages/client/runtime/tests/queue-store.spec.ts +++ b/packages/client/runtime/tests/queue-store.spec.ts @@ -110,11 +110,20 @@ describe('queue operation transport', () => { await expect(session.updateQueue(iid('q-op'), { kind: 'edit', content: text('next') })) .resolves.toEqual({ ok: true, value: { accepted: true } }) - expect(api.callsOf('session.updateQueue')).toEqual([{ - sessionId: SID, - itemId: 'q-op', - action: { kind: 'edit', content: text('next') }, - }]) + await expect(session.updateQueue(iid('q-op'), { kind: 'steer' })) + .resolves.toEqual({ ok: true, value: { accepted: true } }) + expect(api.callsOf('session.updateQueue')).toEqual([ + { + sessionId: SID, + itemId: 'q-op', + action: { kind: 'edit', content: text('next') }, + }, + { + sessionId: SID, + itemId: 'q-op', + action: { kind: 'steer' }, + }, + ]) expect(session.getSnapshot().queue).toBe(before) }) }) diff --git a/packages/client/ui-conversation/README.i18n.yaml b/packages/client/ui-conversation/README.i18n.yaml index 316b685585..1a8fc3a22a 100644 --- a/packages/client/ui-conversation/README.i18n.yaml +++ b/packages/client/ui-conversation/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/client/ui-conversation/README.md -README.md: d92202e6dc2b1003db0cdf97910746a0de133357 -README.zh.md: a75ee75a5fb2b3d8e5278283b6b6da15bf9c02cd +README.md: 2bf74454fab14303f305b9822e1032f67de0d3ae +README.zh.md: 2677ab5cc85c16fffeca688a4648caf2410ba895 diff --git a/packages/client/ui-conversation/README.md b/packages/client/ui-conversation/README.md index d92202e6dc..2bf74454fa 100644 --- a/packages/client/ui-conversation/README.md +++ b/packages/client/ui-conversation/README.md @@ -62,5 +62,5 @@ None; this package neither assembles nor sends a provider request. - **The sparkle icon for the others tool row is a hand-drawn approximation** — the design glyph's vector geometry is not exportable locally; promotion into ui-primitives waits on an exact export. - **The approval panel's "Always allow this type" is deferred** — durable grants need a grant-storage design; only allow-once/reject answer today. - **TodoPanel truncates long item text to one ellipsized line** — the figma strip has no wrap or expand affordance; full text is not readable inline. -- **Queue edit is text-only** — rows containing non-text blocks still show a flattened preview, but their edit control is disabled because the inline editor cannot preserve those blocks. A text row's edit mode replaces delete with save and cancel; Enter saves and Escape cancels. QueueDock exposes no send-now control. -- **Web exposes pending Queue only** — the composer and `conversation.send` never submit `mode:'steer'`. The Host omits pending steering from the Queue snapshot. A consumed `steering/message` still folds into the durable transcript as a plain bubble (no interjection chrome) so external/host steering remains truthful on replay. +- **Queue edit is text-only** — rows containing non-text blocks still show a flattened preview, but their edit control is disabled because the inline editor cannot preserve those blocks. A text row's edit mode replaces delete and strict steer with save and cancel; Enter saves and Escape cancels. +- **Queue strict steer preserves complete messages** — while the Agent is running, the steer action atomically transfers the addressed Queue occurrence into the current next-step window. Mixed-content rows remain eligible because the action forwards the immutable message instead of the text projection. The Host omits pending steering from the Queue snapshot; a consumed `steering/message` still folds into the durable transcript as a plain bubble so replay remains truthful. diff --git a/packages/client/ui-conversation/README.zh.md b/packages/client/ui-conversation/README.zh.md index a75ee75a5f..2677ab5cc8 100644 --- a/packages/client/ui-conversation/README.zh.md +++ b/packages/client/ui-conversation/README.zh.md @@ -62,5 +62,5 @@ todo 两个面就是在该形状上的两个注册项,都是普通注册方插 - **others 工具行的闪光图标是手绘近似版本**:无法在本地导出设计字形的矢量几何;等到存在精确导出后再将其提升到 ui-primitives。 - **审批面板的「始终允许此类」暂缓**:持久授权需要授权存储设计;今天只能回答允许一次/拒绝。 - **TodoPanel 将过长条目截成单行省略号**:figma 条没有换行或展开入口,完整文本无法在行内读完。 -- **Queue 编辑仅支持文本**:包含非文本块的行仍显示扁平化预览,但由于内联编辑器无法保留这些块,其编辑控件会被禁用。文本行进入编辑模式后,删除会替换为保存和取消;Enter 保存,Escape 取消。QueueDock 不提供立即发送控件。 -- **Web 仅暴露待处理 Queue**:composer 与 `conversation.send` 从不提交 `mode:'steer'`。Host 不会把待处理 steering(中途引导)纳入 Queue 快照。已消费的 `steering/message` 仍会折叠进持久 transcript(文本记录),并以无「插话」徽章的普通气泡呈现,因此从外部/Host 提交的 steering 在回放时仍能如实呈现。 +- **Queue 编辑仅支持文本**:包含非文本块的行仍显示扁平化预览,但由于内联编辑器无法保留这些块,其编辑控件会被禁用。文本行进入编辑模式后,删除和严格 steering(中途引导)操作会被保存和取消取代;Enter 保存,Escape 取消。 +- **Queue 严格 steering 会保留完整消息**:Agent 运行期间,steering 操作会以原子方式把所寻址的 Queue 单次入队项转移到当前 next-step 窗口。包含混合内容的行仍可使用此操作,因为它会转发不可变消息,而非文本投影。Host 不会把待处理 steering 纳入 Queue 快照;已消费的 `steering/message` 仍会折叠进持久 transcript(文本记录),并以普通气泡呈现,因此回放仍然如实。 diff --git a/packages/client/ui-conversation/src/client/locales.ts b/packages/client/ui-conversation/src/client/locales.ts index 76560ca9e4..21897a529b 100644 --- a/packages/client/ui-conversation/src/client/locales.ts +++ b/packages/client/ui-conversation/src/client/locales.ts @@ -89,8 +89,11 @@ export const zh = { 'queue.save': '保存排队消息', 'queue.cancelEdit': '取消编辑', 'queue.remove': '删除排队消息', + 'queue.steer': '插话发送', + 'queue.steer.unavailable': '仅运行中可插话发送', 'queue.editFailed': '编辑失败:这条消息可能已经开始发送。', 'queue.removeFailed': '删除失败:这条消息可能已经开始发送。', + 'queue.steerFailed': '插话失败:当前回复已结束,或这条消息已经开始发送。', 'terminal.signal': '信号 {signal}', 'terminal.exitCode': '退出码 {code}', 'terminal.running': '运行中', @@ -189,8 +192,11 @@ export const en = { 'queue.save': 'Save queued message', 'queue.cancelEdit': 'Cancel editing', 'queue.remove': 'Remove queued message', + 'queue.steer': 'Steer queued message', + 'queue.steer.unavailable': 'Steering is available only while the agent is running', 'queue.editFailed': 'Edit failed: this message may have already started sending.', 'queue.removeFailed': 'Removal failed: this message may have already started sending.', + 'queue.steerFailed': 'Steering failed: the current response ended or this message already started sending.', 'terminal.signal': 'signal {signal}', 'terminal.exitCode': 'exit code {code}', 'terminal.running': 'Running', diff --git a/packages/client/ui-conversation/src/client/queue/QueueDock.tsx b/packages/client/ui-conversation/src/client/queue/QueueDock.tsx index e62e5ab389..b374fa36d6 100644 --- a/packages/client/ui-conversation/src/client/queue/QueueDock.tsx +++ b/packages/client/ui-conversation/src/client/queue/QueueDock.tsx @@ -9,7 +9,7 @@ import type { PropsLocale, PropsRuntime } from '@deepseek-ai/dsh-client-ui-slots import type { SessionId } from '@deepseek-ai/dsh-client-runtime/client' import { IconCheckOutline16, IconChevronDownOutline14, IconChevronUpOutline14, - IconCloseOutline16, IconEditOutline16, IconTrashOutline16, + IconCloseOutline16, IconEditOutline16, IconRightUpOutline16, IconTrashOutline16, } from '@deepseek-ai/dsh-client-ui-primitives' import type { QueueAction, QueueItemId } from '../contract/queue.ts' import { NS } from '../locales.ts' @@ -30,6 +30,7 @@ export type QueueDockProps = PropsRuntime<'conversation.input.dock'> & QueueDock */ export function QueueDock({ useSession, updateQueue, notify, t }: QueueDockProps) { const queue = useSession(s => s.queue) + const running = useSession(s => s.running) const [editing, setEditing] = useState<{ id: QueueItemId; text: string } | null>(null) const [busy, setBusy] = useState(null) const [collapsed, setCollapsed] = useState(true) @@ -170,6 +171,22 @@ export function QueueDock({ useSession, updateQueue, notify, t }: QueueDockProps > + )} diff --git a/packages/client/ui-conversation/src/client/service.ts b/packages/client/ui-conversation/src/client/service.ts index 9a7d8fd5bb..4ecf41fd10 100644 --- a/packages/client/ui-conversation/src/client/service.ts +++ b/packages/client/ui-conversation/src/client/service.ts @@ -31,9 +31,9 @@ export interface IConversation { */ send(text: string): Promise /** - * Apply one operation to a pending queue occurrence. + * Apply one edit, remove, or strict steer operation to a pending queue occurrence. * @param itemId - agent-owned inbox occurrence identity. - * @param action - edit or remove operation. + * @param action - requested queue operation. * @returns completion; business failures reject. */ updateQueue(itemId: QueueItemId, action: QueueAction): Promise diff --git a/packages/client/ui-conversation/tests/queue-dock.spec.tsx b/packages/client/ui-conversation/tests/queue-dock.spec.tsx index 6c11119c0a..eb5c293501 100644 --- a/packages/client/ui-conversation/tests/queue-dock.spec.tsx +++ b/packages/client/ui-conversation/tests/queue-dock.spec.tsx @@ -1,7 +1,7 @@ // @vitest-environment jsdom /** * QueueDock rendering and operations: authoritative rows, inline editing, - * collapse state, removal, failure notices, and live retirement. + * collapse state, removal, strict steering, failure notices, and live retirement. */ import { afterEach, describe, expect, it, vi } from 'vitest' import { act, cleanup, fireEvent, render, waitFor } from '@testing-library/react' @@ -187,10 +187,10 @@ describe('QueueDock', () => { fireEvent.click(getByRole('button', { name: '2 条排队消息' })) expect([...container.querySelectorAll('li')].map(item => item.textContent)) .toEqual(['第一条排队消息', 'image [image]']) - expect(container.querySelectorAll('button')).toHaveLength(5) + expect(container.querySelectorAll('button')).toHaveLength(7) expect(container.querySelectorAll('[aria-label="编辑排队消息"]')).toHaveLength(2) expect(container.querySelectorAll('[aria-label="删除排队消息"]')).toHaveLength(2) - expect(container.querySelectorAll('[aria-label="立即发送排队消息"]')).toHaveLength(0) + expect(container.querySelectorAll('[aria-label="插话发送"]')).toHaveLength(2) expect((container.querySelectorAll('[aria-label="编辑排队消息"]')[0] as HTMLButtonElement).disabled).toBe(false) expect((container.querySelectorAll('[aria-label="编辑排队消息"]')[1] as HTMLButtonElement).disabled).toBe(true) expect(container.querySelectorAll('[aria-label="编辑排队消息"]')[1]?.getAttribute('title')) @@ -273,6 +273,45 @@ describe('QueueDock', () => { }) }) + it('strictly steers complete row content only while the agent is running', async () => { + const running = snapshotWith([row('i-steer', null, 'image [image]')]) + const source = liveSession(running) + const updateQueue = vi.fn(() => Promise.resolve()) + const rendered = render( + , + ) + + const button = rendered.getByLabelText('插话发送') + expect(button).toHaveProperty('disabled', false) + fireEvent.click(button) + await waitFor(() => { + expect(updateQueue).toHaveBeenCalledWith(iid('i-steer'), { kind: 'steer' }) + }) + + act(() => { source.push({ ...running, running: false }) }) + expect(rendered.getByLabelText('插话发送')).toHaveProperty('disabled', true) + expect(rendered.getByLabelText('插话发送').getAttribute('title')).toBe('仅运行中可插话发送') + }) + + it('keeps the row and reports a strict steer race', async () => { + const snap = snapshotWith([row('i-steer-race', 'pending steer')]) + const source = liveSession(snap) + const notify = vi.fn() + const updateQueue = vi.fn(() => Promise.reject(new Error('steer unavailable'))) + const { getByLabelText, getByText } = render( + , + ) + + fireEvent.click(getByLabelText('插话发送')) + await waitFor(() => { + expect(notify).toHaveBeenCalledWith( + 'error', + '插话失败:当前回复已结束,或这条消息已经开始发送。', + ) + }) + expect(getByText('pending steer')).toBeTruthy() + }) + it('keeps the row and surfaces a notice when an operation loses the claim race', async () => { const snap = snapshotWith([row('i-race', 'pending')]) const source = liveSession(snap) diff --git a/packages/cordis/tool-cordis/src/api-catalog.ts b/packages/cordis/tool-cordis/src/api-catalog.ts index a2a1bf3c5d..786d4431de 100644 --- a/packages/cordis/tool-cordis/src/api-catalog.ts +++ b/packages/cordis/tool-cordis/src/api-catalog.ts @@ -2025,11 +2025,11 @@ export const TYPE_API: readonly TypeApiEntry[] = [ }, { name: 'InboxAction', - declaration: 'export type InboxAction = {\n readonly kind: \'edit\';\n readonly content: ContentBlock[];\n} | {\n readonly kind: \'remove\';\n};', + declaration: 'export type InboxAction = {\n readonly kind: \'edit\';\n readonly content: ContentBlock[];\n} | {\n readonly kind: \'remove\';\n} | {\n readonly kind: \'steer\';\n};', }, { name: 'InboxActionResult', - declaration: 'export type InboxActionResult = \'applied\' | \'not-found\';', + declaration: 'export type InboxActionResult = \'applied\' | \'not-found\' | \'steer-unavailable\';', }, { name: 'InboxItemId', diff --git a/packages/core/agent-loop/README.i18n.yaml b/packages/core/agent-loop/README.i18n.yaml index 6e9e63a22f..691115619b 100644 --- a/packages/core/agent-loop/README.i18n.yaml +++ b/packages/core/agent-loop/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/core/agent-loop/README.md -README.md: 1662b1076cc116888d048cb6af1be1c7ab8196f6 -README.zh.md: 2fca32a02fdd73961c912c988933e1cd1a1a5817 +README.md: c71b350adfe06a19d4c24cb7e67de895a662bd87 +README.zh.md: d30dfc85e1597a8e193019cc23f4c7c39c991776 diff --git a/packages/core/agent-loop/README.md b/packages/core/agent-loop/README.md index 1662b1076c..c71b350adf 100644 --- a/packages/core/agent-loop/README.md +++ b/packages/core/agent-loop/README.md @@ -59,7 +59,7 @@ The unified `send()` primitive routes content and source by (`target` × `wakeup `steer()` attaches a one-shot admission receipt to its exact accepted message. After `agent/step` and asynchronous prompt assembly succeed, the loop commits a stable pending batch as `steering/message`, snapshots derived history, and opens `step/start`; only then does each receipt resolve `admitted` with that turn and step. Later arrivals remain pending. Idle steering enters the ordinary FIFO and uses the first request of its eventual turn as the same admission boundary. A turn-concluding tool result, broad cancellation, disposal, or a claimed idle-steering turn that never reaches a request resolves affected receipts `rejected`; `cancel(..., { keepInbox: true })` and non-terminal routing preserve pending delivery. Open-turn `inject()` still commits after all tool results, including accepted context finalized during an interrupted batch, while steering remains provisional until a request admits it. -Every FIFO acceptance mints an `InboxItemId` and publishes `agent/inbox/enqueue` with the complete occurrence. `updateInbox()` owns the synchronous queued-item boundary: edit freezes replacement content without changing message identity or position, while remove publishes discard. Edit publishes `agent/inbox/update`; steering and claimed occurrences return `not-found`. Claim publishes `agent/inbox/dequeue` and irrevocably removes the live address before prompt admission, so a racing update cannot rewrite durable history; `cancel()` without `keepInbox` publishes `agent/inbox/discard`. +Every FIFO acceptance mints an `InboxItemId` and publishes `agent/inbox/enqueue` with the complete occurrence. `updateInbox()` owns the synchronous queued-item boundary: edit freezes replacement content without changing message identity or position, remove publishes discard, and strict steer transfers the immutable message into an open next-step window as a new steering occurrence. A closed window returns `steer-unavailable` without mutation; pending steering and claimed occurrences return `not-found`. Claim publishes `agent/inbox/dequeue` and irrevocably removes the live address before prompt admission, so a racing update cannot rewrite durable history; `cancel()` without `keepInbox` publishes `agent/inbox/discard`. ### Loop lifecycle (`agent.ts`) diff --git a/packages/core/agent-loop/README.zh.md b/packages/core/agent-loop/README.zh.md index 2fca32a02f..d30dfc85e1 100644 --- a/packages/core/agent-loop/README.zh.md +++ b/packages/core/agent-loop/README.zh.md @@ -59,7 +59,7 @@ interface Config { `steer()` 会把一次性准入回执附着到其准确的已接收消息。`agent/step` 和异步提示词组装成功后,循环把稳定的待处理批次提交为 `steering/message`、捕获派生历史并开启 `step/start`;只有此时,每个回执才会解析为 `admitted`,并附带轮次与步骤。之后到达的消息继续待处理。空闲 steering 会进入普通 FIFO,并以其最终轮次的首次请求作为相同准入边界。结束轮次的工具结果、广义取消、dispose(资源释放),或已领取 idle-steering 消息却从未到达请求的轮次,会把受影响回执解析为 `rejected`;`cancel(..., { keepInbox: true })` 和非终止型路由会保留待处理投递。活跃轮次内的 `inject()` 仍会在所有工具结果后提交,包括被中断批次中已最终确认的上下文;steering 则保持待准入,直到请求接纳它。 -每次 FIFO 接受项时都会铸造一个 `InboxItemId`,并通过 `agent/inbox/enqueue` 发布完整的单次入队项。`updateInbox()` 持有同步 queued 项边界:编辑会冻结替换内容,但不改变消息标识或位置;移除会发布 discard。编辑会发布 `agent/inbox/update`;steering 项和已被认领的项会返回 `not-found`。认领操作会发布 `agent/inbox/dequeue`,并在提示词接纳前不可逆地移除实时寻址标识,因此竞态中的更新无法改写持久历史;`cancel()` 在不带 `keepInbox` 时会发布 `agent/inbox/discard`。 +每次 FIFO 接受项时都会铸造一个 `InboxItemId`,并通过 `agent/inbox/enqueue` 发布完整的单次入队项。`updateInbox()` 持有同步 queued 项边界:编辑会冻结替换内容,但不改变消息标识或位置;移除会发布 discard;严格 steering 会把不可变消息作为新的 steering 单次入队项转移到开放的 next-step 窗口。窗口关闭时返回 `steer-unavailable`,且不做任何变更;待处理 steering 和已被认领的项会返回 `not-found`。认领操作会发布 `agent/inbox/dequeue`,并在提示词接纳前不可逆地移除实时寻址标识,因此竞态中的更新无法改写持久历史;`cancel()` 在不带 `keepInbox` 时会发布 `agent/inbox/discard`。 ### 循环生命周期(`agent.ts`) diff --git a/packages/core/agent-loop/src/agent.ts b/packages/core/agent-loop/src/agent.ts index 90354f0cc2..88fcd64ead 100644 --- a/packages/core/agent-loop/src/agent.ts +++ b/packages/core/agent-loop/src/agent.ts @@ -256,6 +256,22 @@ export class ReactLoopAgent implements Agent { emitAgentEvent(this.loopCtx, this, 'agent/inbox/discard', [pending.item]) return 'applied' } + case 'steer': { + if (!this.acceptsNextStep) return 'steer-unavailable' + this.queued.splice(queuedIndex, 1) + const item: InboxItem = Object.freeze({ + id: InboxItemId(randomUUID()), + message: pending.item.message, + placement: 'steering', + }) + this.outbox.push({ message: item.message, steering: true, item }) + // Publish the replacement only after it is owned by the outbox. Its + // enqueue precedes the old occurrence's discard so reentrant + // cancellation can terminally account for both occurrences. + emitAgentEvent(this.loopCtx, this, 'agent/inbox/enqueue', item) + emitAgentEvent(this.loopCtx, this, 'agent/inbox/discard', [pending.item]) + return 'applied' + } default: /* v8 ignore next -- InboxAction is a closed discriminated union. */ return assertNever(action) diff --git a/packages/core/agent-loop/tests/contract-regressions.spec.ts b/packages/core/agent-loop/tests/contract-regressions.spec.ts index 31e6d7c2b1..f61af0833c 100644 --- a/packages/core/agent-loop/tests/contract-regressions.spec.ts +++ b/packages/core/agent-loop/tests/contract-regressions.spec.ts @@ -156,6 +156,102 @@ describe('addressable inbox operations', () => { : '')) .toEqual(['keep me']) }) + + it('strictly transfers a queued occurrence into the open turn', async () => { + const adapter = new MockAdapter([textResponse('done')]) + const ctx = await harness(adapter) + const agent = ctx.agentLoop.create(SessionId('queue-to-steer'), { provider: 'mock', model: 'mock' }) + const entered = Promise.withResolvers() + const decision = Promise.withResolvers<{ kind: 'allow' }>() + ctx.on('agent/prompt-submit', async () => { + entered.resolve(undefined) + return decision.promise + }) + + const enqueued: InboxItem[] = [] + const discarded: InboxItem[] = [] + ctx.on('agent/inbox/enqueue', (subject, item) => { + if (subject === agent) enqueued.push(item) + }) + ctx.on('agent/inbox/discard', (subject, items) => { + if (subject === agent) discarded.push(...items) + }) + + const idle = waitForIdle(ctx, agent) + send(agent, 'open the turn') + await entered.promise + send(agent, 'steer this message') + const queued = enqueued.find(item => inboxText(item) === 'steer this message')! + + expect(agent.updateInbox(queued.id, { kind: 'steer' })).toBe('applied') + const steering = enqueued.find(item => item.placement === 'steering')! + expect(steering.id).not.toBe(queued.id) + expect(steering.message).toBe(queued.message) + expect(discarded).toEqual([queued]) + + decision.resolve({ kind: 'allow' }) + await idle + expect(agent.session.events.flatMap(event => + event.type === 'steering/message' ? [event.data.message] : [], + )).toEqual([queued.message]) + expect(agent.updateInbox(queued.id, { kind: 'steer' })).toBe('not-found') + }) + + it('keeps a queued occurrence when the next-step window is closed', () => { + const ctx = new Context() + const session = new Session(SessionId('queue-to-steer-closed')) + const agent = new ReactLoopAgent(ctx, session.id, {}, session) + const enqueued: InboxItem[] = [] + const discarded: InboxItem[] = [] + ctx.on('agent/inbox/enqueue', (_subject, item) => { enqueued.push(item) }) + ctx.on('agent/inbox/discard', (_subject, items) => { discarded.push(...items) }) + + agent.send( + createUserMessage({ content: [{ type: 'text', text: 'stay queued' }], source: { kind: 'user' } }), + { target: 'next-turn', wakeup: false }, + ) + const queued = enqueued[0]! + expect(agent.updateInbox(queued.id, { kind: 'steer' })).toBe('steer-unavailable') + expect(discarded).toEqual([]) + expect(agent.updateInbox(queued.id, { kind: 'remove' })).toBe('applied') + }) + + it('accounts for both occurrences when steering enqueue cancels reentrantly', async () => { + const adapter = new MockAdapter([textResponse('unused')]) + const ctx = await harness(adapter) + const agent = ctx.agentLoop.create(SessionId('queue-to-steer-cancel'), { provider: 'mock', model: 'mock' }) + const entered = Promise.withResolvers() + const decision = Promise.withResolvers<{ kind: 'allow' }>() + ctx.on('agent/prompt-submit', async () => { + entered.resolve(undefined) + return decision.promise + }) + + const enqueued: InboxItem[] = [] + const discarded: InboxItem[] = [] + ctx.on('agent/inbox/enqueue', (subject, item) => { + if (subject !== agent) return + enqueued.push(item) + if (item.placement === 'steering') agent.cancel({ kind: 'user' }) + }) + ctx.on('agent/inbox/discard', (subject, items) => { + if (subject === agent) discarded.push(...items) + }) + + const idle = waitForIdle(ctx, agent) + send(agent, 'open the turn') + await entered.promise + send(agent, 'cancel during conversion') + const queued = enqueued.find(item => inboxText(item) === 'cancel during conversion')! + + expect(agent.updateInbox(queued.id, { kind: 'steer' })).toBe('applied') + const steering = enqueued.find(item => item.placement === 'steering')! + expect(discarded).toEqual([steering, queued]) + + decision.resolve({ kind: 'allow' }) + await idle + expect(agent.session.events.some(event => event.type === 'steering/message')).toBe(false) + }) }) describe('assistant replay provenance', () => { diff --git a/packages/core/agent/README.i18n.yaml b/packages/core/agent/README.i18n.yaml index 78473df4a7..f8431cf022 100644 --- a/packages/core/agent/README.i18n.yaml +++ b/packages/core/agent/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/core/agent/README.md -README.md: 8a6028352127c4638c0b5e0e3ee85964d1d7d734 -README.zh.md: ffa71ea987ab355ff2f30b6164376199cd5d0170 +README.md: 4c6a6dd95541cfa559e95858fede01d7cd76637f +README.zh.md: 07bf887c557b410005bbe0fa1a988e63a765ad99 diff --git a/packages/core/agent/README.md b/packages/core/agent/README.md index 8a60283521..4c6a6dd955 100644 --- a/packages/core/agent/README.md +++ b/packages/core/agent/README.md @@ -62,7 +62,7 @@ The handle every plugin programs against: - `agent.send(message, options)` — the one delivery primitive over the (`target` × `wakeup`) matrix. `message` is an already identified, frozen `UserMessage`; callers normally create it with `createUserMessage()` before routing begins. `SendOptions` owns only the `target` and `wakeup` policy. Each accepted FIFO occurrence receives its own `InboxItemId`, even when callers reuse a `MessageId`; `agent/inbox/enqueue`/`update` and the terminal `dequeue` or `discard` carry that complete `InboxItem`. `target: 'next-turn'` queues one independent FIFO item that, if admitted, becomes the sole ordinary prompt in its turn. `target: 'next-step'` with `wakeup: true` submits steering, while `target: 'next-step'` with `wakeup: false` injects durable context without running the model. The [one-send-one-turn Agent Note](../../../.agents/notes/implemented/simplification/2026-07-17-one-send-one-turn.md) owns the turn rationale. - `agent.reserveTurnAdmission()` — synchronously reserve the idle boundary before any queued waking prompt can claim its turn. An accepted prompt, including a same-tick pending wake, has right of way and makes reservation return `undefined`. Later sends keep their ordinary IDs, FIFO placement, and wakeup facts while held; `acceptsNextStep` remains false, `inject()` is not withheld, `whenIdle()` counts the reservation as activity, and the returned release is idempotent. This narrow coordination capability lets standalone durable operations such as manual compaction finish and flush before queued prompts derive from the session. -- `agent.updateInbox(itemId, action)` — synchronously edits or removes one still-pending queued occurrence. Edit keeps its `MessageId`, `InboxItemId`, source, and FIFO position while replacing frozen content; remove emits the occurrence's terminal discard. Steering and claimed occurrences return `not-found`. +- `agent.updateInbox(itemId, action)` — synchronously edits, removes, or strictly steers one still-pending queued occurrence. Edit keeps its `MessageId`, `InboxItemId`, source, and FIFO position while replacing frozen content; remove emits the occurrence's terminal discard. Strict steer requires `acceptsNextStep`, ends the queued occurrence, and accepts the same immutable message as a new steering occurrence with a new `InboxItemId`; a closed window returns `steer-unavailable` without mutation. Pending steering and claimed occurrences return `not-found`. - `agent.followup(input)` — the `next-turn`/wakeup preset of `send()`: queue an ordinary follow-up turn and wake the driver. - `agent.steer(input)` — the `next-step`/wakeup preset: submit one identified message and receive its `SteeringReceipt`. During prompt admission or an open turn, the message stages for the next safe request boundary without dispatching `agent/prompt-submit`; outside that acceptance window, it becomes a woken queued prompt. `receipt.outcome` resolves `admitted` with the turn and step only after the loop logs the message, captures it in immutable request history, and commits `step/start`. A turn-concluding tool result, broad cancellation, disposal, or pre-admission failure resolves it `rejected`; `cancel(..., { keepInbox: true })` and non-terminal routing preserve pending delivery. Reliable callers await the receipt, while best-effort UI steering may ignore it. - `agent.inject(input)` — the `next-step`/no-wakeup preset: append model-facing context without running the model; the next request sees a verbatim user-role message whose provenance is carried by the required `input.source`. During prompt admission or an open turn, injection waits in the outbox for the next safe boundary. Outside that acceptance window, it appends immediately without opening a turn; a context-only admission batch takes this fallback if admission closes without a turn, while context staged beside steering remains pending with it. Persistence reacts to `session/event` independently. Injection emits no `agent/inbox/*` event. diff --git a/packages/core/agent/README.zh.md b/packages/core/agent/README.zh.md index ffa71ea987..07bf887c55 100644 --- a/packages/core/agent/README.zh.md +++ b/packages/core/agent/README.zh.md @@ -62,7 +62,7 @@ Agent *创建* 由实现 `AgentFactory` 的插件(`dsh-agent-loop`)提供, - `agent.send(message, options)`:覆盖(`target` × `wakeup`)矩阵的唯一投递原语。`message` 是已有标识且已冻结的 `UserMessage`;调用方通常会在开始路由前使用 `createUserMessage()` 创建它。`SendOptions` 只持有 `target` 与 `wakeup` 策略。每次获准进入 FIFO 的项都会获得独立的 `InboxItemId`,即使调用方复用了同一个 `MessageId`;`agent/inbox/enqueue`/`update` 及终态 `dequeue` 或 `discard` 都会携带这一完整 `InboxItem`。`target: 'next-turn'` 排队一条独立 FIFO 项,获准后成为其轮次中唯一的普通提示词。`target: 'next-step'` 且 `wakeup: true` 提交 steering(中途引导),而 `target: 'next-step'` 且 `wakeup: false` 注入持久上下文,不运行模型。轮次原理由 [one-send-one-turn Agent Note](../../../.agents/notes/implemented/simplification/2026-07-17-one-send-one-turn.md)拥有。 - `agent.reserveTurnAdmission()`:在任何已排队唤醒提示词认领其轮次之前,同步预留空闲边界。已获接纳的提示词拥有优先权,包括同一 tick 内仍在等待唤醒的项,此时预留返回 `undefined`。预留期间,之后发送的项保留其普通 ID、FIFO 位置与唤醒信息;`acceptsNextStep` 保持 false,`inject()` 不受阻塞,`whenIdle()` 将该预留计为活动,返回的释放函数可幂等调用。这项范围有限的协调能力使手动压缩(compaction)等独立持久操作能够在排队提示词从会话派生内容前完成并 flush。 -- `agent.updateInbox(itemId, action)`:同步编辑或移除一个仍处于待处理状态的 queued 入队项。编辑会替换已冻结的内容,同时保留其 `MessageId`、`InboxItemId`、来源与 FIFO 位置;移除会发出该项的终态 discard。steering 项和已被认领的项会返回 `not-found`。 +- `agent.updateInbox(itemId, action)`:同步编辑、移除一个仍处于待处理状态的 queued 入队项,或对其执行严格 steering。编辑会替换已冻结的内容,同时保留其 `MessageId`、`InboxItemId`、来源与 FIFO 位置;移除会发出该项的终态 discard。严格 steering 要求 `acceptsNextStep` 为 true;它会结束 queued 单次入队项,并把同一条不可变消息接受为新的 steering 单次入队项,后者使用新的 `InboxItemId`。窗口关闭时返回 `steer-unavailable`,且不做任何变更。待处理 steering 和已被认领的项会返回 `not-found`。 - `agent.followup(input)`:`send()` 的 `next-turn`/wakeup 预设:排队一个普通后续轮次并唤醒驱动器。 - `agent.steer(input)`:`next-step`/wakeup 预设:提交一条已有标识的消息,并取得其 `SteeringReceipt`。提示词接纳期间或轮次打开时,消息会为下一个安全请求边界暂存,且不分发 `agent/prompt-submit`;该接收窗口之外则成为会唤醒驱动器的排队提示词。只有循环记录消息、将其捕获到不可变请求历史并提交 `step/start` 后,`receipt.outcome` 才会解析为 `admitted`,并附带轮次与步骤。结束轮次的工具结果、广义取消、dispose(资源释放)或准入前故障会使其解析为 `rejected`;`cancel(..., { keepInbox: true })` 和非终止型路由会保留待处理投递。需要可靠投递的调用方应等待回执;尽力执行的 UI steering 可以忽略它。 - `agent.inject(input)`:`next-step`/不唤醒预设:追加面向模型的上下文而不运行模型;下一次请求会看到一条逐字的 user role 消息,其来源由必填的 `input.source` 携带。提示词接纳期间或轮次打开时,注入会在 outbox 中等待下一个安全边界。该接收窗口之外,它会立即追加而不开启轮次;如果接纳结束却未开启轮次,仅含上下文的接纳批次会采用这一回退,而与 steering 一同暂存的上下文则会随其继续待处理。持久化独立地响应 `session/event`。注入不发出 `agent/inbox/*` 事件。 diff --git a/packages/core/agent/src/types.ts b/packages/core/agent/src/types.ts index 1a1010e90e..80d289c9f0 100644 --- a/packages/core/agent/src/types.ts +++ b/packages/core/agent/src/types.ts @@ -54,9 +54,10 @@ export interface InboxItem { export type InboxAction = | { readonly kind: 'edit'; readonly content: ContentBlock[] } | { readonly kind: 'remove' } + | { readonly kind: 'steer' } /** Result of applying an inbox action at the synchronous ownership boundary. */ -export type InboxActionResult = 'applied' | 'not-found' +export type InboxActionResult = 'applied' | 'not-found' | 'steer-unavailable' /** Final admission outcome for one call to {@link Agent.steer}. */ export type SteeringOutcome = @@ -209,10 +210,13 @@ export interface Agent { /** * Mutate one still-pending queued occurrence synchronously. Editing preserves * the message identity and queue position; removal publishes its terminal - * discard. Steering occurrences and driver-claimed items return `not-found`. + * discard. Steer strictly transfers the message into the current next-step + * window, or returns `steer-unavailable` without changing the queued + * occurrence. Steering occurrences and driver-claimed items return + * `not-found`. * @param id - independently addressable queued occurrence. - * @param action - edit or remove operation. - * @returns whether the pending occurrence was found and updated. + * @param action - edit, remove, or strict steer operation. + * @returns the applied outcome or the reason no mutation occurred. */ updateInbox(id: InboxItemId, action: InboxAction): InboxActionResult diff --git a/packages/host/apiproxy/README.i18n.yaml b/packages/host/apiproxy/README.i18n.yaml index 69093a7692..5f3bec73b0 100644 --- a/packages/host/apiproxy/README.i18n.yaml +++ b/packages/host/apiproxy/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/host/apiproxy/README.md -README.md: 933b5f6167263545b3bef5ca9fb8f8b945e86ef9 -README.zh.md: f1f1106dbd0c50889eca6f6ae52fbb29d1c4c03f +README.md: 87c477362dc159621a38a1ff923bebf07cbd3c6f +README.zh.md: 7316f3a6d93d30e101537c40608037fedfa77321 diff --git a/packages/host/apiproxy/README.md b/packages/host/apiproxy/README.md index 933b5f6167..87c477362d 100644 --- a/packages/host/apiproxy/README.md +++ b/packages/host/apiproxy/README.md @@ -22,7 +22,7 @@ Session model routing is a session-domain contract. `session.models` returns the Generic Agent-bound session, command, and goal operations serve ordinary sessions only. They return `agent-busy` for a session-backed subagent instead of resuming or driving it; explicit-id `session.create` adoption and the attached-only queue controls enforce the same ownership boundary. Subagent conversation reads and continuation use the dedicated `subagent.*` domain, which retains catalog-mode and direct-parent authorization. -Pending queued input is a live control-plane contract, not session history. The gateway mirrors queued `InboxItem` occurrences from `agent/inbox/*` and broadcasts authoritative `session/queue` snapshots on every queued change and reconnect; pending steering stays outside this Web projection. `session.updateQueue` addresses one `InboxItemId`: edit replaces pending content and remove discards it. `session.cancel` aborts only the active turn and preserves pending inbox work; after cancellation reaches quiescence and the closing turn flushes, AgentLoop claims the next waking occurrence in FIFO order. The browser never resends or promotes that occurrence. A driver claim wins races by retiring the address before admission; a later operation returns `queue-item-not-found`. Queue operations query only an attached Agent and never resume a cold session because process-local inbox identities do not survive restart or disposal. The client never infers retirement from turn or status events. +Pending queued input is a live control-plane contract, not session history. The gateway mirrors queued `InboxItem` occurrences from `agent/inbox/*` and broadcasts authoritative `session/queue` snapshots on every queued change and reconnect; pending steering stays outside this Web projection. `session.updateQueue` addresses one `InboxItemId`: edit replaces pending content, remove discards it, and strict steer transfers its complete message into the current next-step window. A closed window returns `steer-unavailable` without changing the row. `session.cancel` aborts only the active turn and preserves pending inbox work; after cancellation reaches quiescence and the closing turn flushes, AgentLoop claims the next waking occurrence in FIFO order. The browser never resends or promotes that occurrence. A driver claim wins races by retiring the address before admission; a later operation returns `queue-item-not-found`. Queue operations query only an attached ordinary-session Agent and never resume a cold session because process-local inbox identities do not survive restart or disposal. The client never infers retirement from turn or status events. Workspace and Session lists are separate reconnect baselines. `workspace.create({ name })` creates a uniquely titled directory under the configured root, while `workspace.create({ path })` adopts an existing canonical directory and permits basename-derived titles to repeat. `workspace.delete` removes only the Workspace registration, `session.create` accepts an optional preallocated Session id, and `host/workspace-changed`, `host/workspace-removed`, plus `host/session-added` carry committed increments in either arrival order. `workspace.archiveSession` adds one session to the registry-global archive set and answers the full updated set; `workspace.list` carries that set as the reconnect baseline and `host/archived-sessions-changed` pushes the full snapshot after every durable change. Archiving hides the session from grouping surfaces without touching its log or its workspace account; a session neither live nor persisted fails with `session-not-found`. Registration deletion preserves the directory and session logs; its Sessions remain in `session.list` and become Ungrouped. `SessionSummary.blank` and the `host/session-added` frame carry the derived zero-events bit: clients hide blank sessions and reuse them per workspace, flip blank on the first `host/session-status(running:true)`, and treat `session.list` as the reconnect authority; cold summaries are never blank because lazy persistence keeps never-appended sessions out of `list()`. The `session.list` summaries and `host/session-added` frames also carry the optional durable `origin: 'subagent'` classification so navigation can suppress duplicate child rows immediately and after reconnect; that bit is never continuation authority. diff --git a/packages/host/apiproxy/README.zh.md b/packages/host/apiproxy/README.zh.md index f1f1106dbd..7316f3a6d9 100644 --- a/packages/host/apiproxy/README.zh.md +++ b/packages/host/apiproxy/README.zh.md @@ -22,7 +22,7 @@ 绑定到 Agent 的通用会话、命令与目标操作只服务普通会话。对于由会话支撑的 subagent,它们会返回 `agent-busy`,而不是恢复或驱动它;显式 id 的 `session.create` 接纳与仅针对已附加会话的队列控件也会执行同一所有权边界。subagent 对话读取与继续执行使用专用的 `subagent.*` 领域,该领域保留目录 mode 与直接 parent 授权。 -待处理的 queued 输入属于实时控制平面契约,而非会话历史。网关镜像来自 `agent/inbox/*` 的 queued `InboxItem` 入队项,并在每次 queued 变更和重连时广播权威的 `session/queue` 快照;待处理 steering(中途引导)不进入此 Web 投影。`session.updateQueue` 通过 `InboxItemId` 寻址单个项:编辑会替换待处理内容,移除会将其丢弃。`session.cancel` 仅中止活动轮次,并保留待处理 inbox 工作;取消达到完全停稳且结束中的轮次完成 flush 后,AgentLoop 按 FIFO 顺序认领下一个可唤醒入队项。浏览器绝不重发或提升该入队项。驱动器在接纳前退役寻址标识,因此认领会赢得竞态;之后的操作返回 `queue-item-not-found`。队列操作只查询当前已挂载的 Agent,绝不恢复冷会话,因为进程本地 inbox 标识无法在重启或资源释放后存活。客户端绝不根据轮次或状态事件推断项已退役。 +待处理的 queued 输入属于实时控制平面契约,而非会话历史。网关镜像来自 `agent/inbox/*` 的 queued `InboxItem` 入队项,并在每次 queued 变更和重连时广播权威的 `session/queue` 快照;待处理 steering(中途引导)不进入此 Web 投影。`session.updateQueue` 通过 `InboxItemId` 寻址单个项:编辑会替换待处理内容,移除会将其丢弃,严格 steering 会把其完整消息转移到当前 next-step 窗口。窗口关闭时返回 `steer-unavailable`,且不改变该行。`session.cancel` 仅中止活动轮次,并保留待处理 inbox 工作;取消达到完全停稳且结束中的轮次完成 flush 后,AgentLoop 按 FIFO 顺序认领下一个可唤醒入队项。浏览器绝不重发或提升该入队项。驱动器在接纳前退役寻址标识,因此认领会赢得竞态;之后的操作返回 `queue-item-not-found`。队列操作只查询当前已挂载的普通会话 Agent,绝不恢复冷会话,因为进程本地 inbox 标识无法在重启或资源释放后存活。客户端绝不根据轮次或状态事件推断项已退役。 Workspace 列表与 Session 列表是相互独立的重连基线。`workspace.create({ name })` 会在配置根目录下创建显示标题唯一的目录,而 `workspace.create({ path })` 会接纳已有的规范目录,并允许由 basename 派生的标题重复。`workspace.delete` 只移除 Workspace 注册记录,`session.create` 接受可选的预分配 Session id,`host/workspace-changed`、`host/workspace-removed` 与 `host/session-added` 则以任意到达顺序携带已提交的增量。`workspace.archiveSession` 向注册表级全局归档集合添加一个会话,并应答完整的更新后集合;`workspace.list` 携带该集合作为重连基线,`host/archived-sessions-changed` 在每次持久变更后推送完整快照。归档只把会话从各分组视图中隐藏,不触碰其日志和 workspace 记账;既非实时也未持久化的会话以 `session-not-found` 失败。删除注册记录会保留目录和会话日志;相关 Session 仍留在 `session.list` 中,并进入 Ungrouped。`SessionSummary.blank` 与 `host/session-added` 帧携带派生的零事件位:客户端隐藏空白会话并按 workspace 复用它们,在首个 `host/session-status(running:true)` 时翻转 blank,并以 `session.list` 作为重连权威;冷会话摘要永远不是空白:惰性持久化让从未追加过事件的会话根本不出现在 `list()` 中。`session.list` 摘要与 `host/session-added` 帧还会携带可选的持久化分类 `origin: 'subagent'`,使导航在实时创建与重连后都能隐藏重复的 child 行;该标记绝不是继续执行的权威依据。 diff --git a/packages/host/apiproxy/src/api-proxy.ts b/packages/host/apiproxy/src/api-proxy.ts index fd4626e573..7d1c033393 100644 --- a/packages/host/apiproxy/src/api-proxy.ts +++ b/packages/host/apiproxy/src/api-proxy.ts @@ -1847,13 +1847,28 @@ export function createApiProxy(ctx: Context, defaults: ApiProxyDefaults): ApiPro if (agent !== undefined && hasSubagentOwner(agent.session, agent)) { return Promise.resolve(err(request, subagentOwnershipError(sessionId))) } - if (agent === undefined || agent.updateInbox(itemId, action) === 'not-found') { + if (agent === undefined) { return Promise.resolve(err(request, { code: 'queue-item-not-found', message: 'queued item is no longer pending', details: { itemId }, })) } + const result = agent.updateInbox(itemId, action) + if (result === 'not-found') { + return Promise.resolve(err(request, { + code: 'queue-item-not-found', + message: 'queued item is no longer pending', + details: { itemId }, + })) + } + if (result === 'steer-unavailable') { + return Promise.resolve(err(request, { + code: 'steer-unavailable', + message: 'current turn no longer accepts steering', + details: { itemId }, + })) + } return Promise.resolve(ok(request, { accepted: true as const })) }, diff --git a/packages/host/apiproxy/src/api/rpc.schema.ts b/packages/host/apiproxy/src/api/rpc.schema.ts index 40289b91ef..4424b5527c 100644 --- a/packages/host/apiproxy/src/api/rpc.schema.ts +++ b/packages/host/apiproxy/src/api/rpc.schema.ts @@ -48,6 +48,7 @@ export const rpcErrorSchema: z.ZodType = z.discriminatedUnion('code', z.object({ code: z.literal('directory-picker-unavailable'), message: z.string(), details: z.object({ capability: z.string() }) }), z.object({ code: z.literal('agent-busy'), message: z.string(), details: z.object({ reason: z.string() }) }), z.object({ code: z.literal('queue-item-not-found'), message: z.string(), details: z.object({ itemId: z.string() }) }), + z.object({ code: z.literal('steer-unavailable'), message: z.string(), details: z.object({ itemId: z.string() }) }), z.object({ code: z.literal('command-error'), message: z.string(), details: z.object({}) }), z.object({ code: z.literal('unknown-command'), message: z.string(), details: z.object({}) }), z.object({ code: z.literal('settings-rejected'), message: z.string(), details: z.object({ ns: z.string() }) }), diff --git a/packages/host/apiproxy/src/api/rpc.ts b/packages/host/apiproxy/src/api/rpc.ts index 09cd1ce3f1..f48a2ca562 100644 --- a/packages/host/apiproxy/src/api/rpc.ts +++ b/packages/host/apiproxy/src/api/rpc.ts @@ -46,6 +46,7 @@ export interface RpcErrorDetailsMap { 'directory-picker-unavailable': { capability: string } 'agent-busy': { reason: string } 'queue-item-not-found': { itemId: InboxItemId } + 'steer-unavailable': { itemId: InboxItemId } /** A known slash command reported a usage/state error; the message is the command's own text. */ 'command-error': {} /** A leading-/ prompt named no registered command; the message names the token. */ diff --git a/packages/host/apiproxy/src/api/sessions.schema.ts b/packages/host/apiproxy/src/api/sessions.schema.ts index 8fcce1e648..1675e396d6 100644 --- a/packages/host/apiproxy/src/api/sessions.schema.ts +++ b/packages/host/apiproxy/src/api/sessions.schema.ts @@ -269,6 +269,7 @@ export const sessionUpdateQueueRequestSchema = z.object({ action: z.discriminatedUnion('kind', [ z.object({ kind: z.literal('edit'), content: z.array(contentBlockSchema) }), z.object({ kind: z.literal('remove') }), + z.object({ kind: z.literal('steer') }), ]), }) as unknown as z.ZodType> diff --git a/packages/host/apiproxy/src/api/sessions.ts b/packages/host/apiproxy/src/api/sessions.ts index e19ab08ef5..f55cb87001 100644 --- a/packages/host/apiproxy/src/api/sessions.ts +++ b/packages/host/apiproxy/src/api/sessions.ts @@ -129,6 +129,7 @@ export interface SessionModels { export type QueueAction = | { kind: 'edit'; content: ContentBlock[] } | { kind: 'remove' } + | { kind: 'steer' } /** Session list entry (v1 builds no index: list does readdir+stat). */ export interface SessionSummary { @@ -285,7 +286,7 @@ export interface SessionsApi { Promise> /** - * Edits or removes one pending queued occurrence on an ordinary session. + * Edits, removes, or strictly steers one pending queued occurrence on an ordinary session. * Session-backed subagents reject with `agent-busy`. */ updateQueue(request: RpcRequest<{ sessionId: SessionId; itemId: InboxItemId; action: QueueAction }>): diff --git a/packages/host/apiproxy/tests/api-proxy-commands.spec.ts b/packages/host/apiproxy/tests/api-proxy-commands.spec.ts index f2e9e352d7..61e4c6f4d3 100644 --- a/packages/host/apiproxy/tests/api-proxy-commands.spec.ts +++ b/packages/host/apiproxy/tests/api-proxy-commands.spec.ts @@ -291,13 +291,14 @@ function inboxItem(id: string, message: UserMessage, placement: InboxPlacement): } describe('session.updateQueue', () => { - it('routes an addressable action and reports a lost claim race', async () => { + it('routes addressable actions and reports strict steer races', async () => { const ctx = await harness() const agent = stubAgent(ctx) const seen: unknown[] = [] agent.updateInbox = (id, action) => { seen.push({ id, action }) - return id === InboxItemId('present') ? 'applied' : 'not-found' + if (id === InboxItemId('present')) return 'applied' + return id === InboxItemId('closed') ? 'steer-unavailable' : 'not-found' } const api = createApiProxy(ctx, DEFAULTS) @@ -319,9 +320,22 @@ describe('session.updateQueue', () => { }, }) expect(expectErr(missing)).toMatchObject({ code: 'queue-item-not-found' }) + const closed = await api.sessions.updateQueue({ + rpcId: RpcId('q-closed'), + payload: { + sessionId: agent.id, + itemId: InboxItemId('closed'), + action: { kind: 'steer' }, + }, + }) + expect(expectErr(closed)).toMatchObject({ + code: 'steer-unavailable', + details: { itemId: 'closed' }, + }) expect(seen).toEqual([ { id: 'present', action: { kind: 'edit', content: [{ type: 'text', text: 'edited' }] } }, { id: 'claimed', action: { kind: 'remove' } }, + { id: 'closed', action: { kind: 'steer' } }, ]) }) diff --git a/packages/host/apiproxy/tests/rpc-schemas.spec.ts b/packages/host/apiproxy/tests/rpc-schemas.spec.ts index 34910acb9f..541ddab1c8 100644 --- a/packages/host/apiproxy/tests/rpc-schemas.spec.ts +++ b/packages/host/apiproxy/tests/rpc-schemas.spec.ts @@ -77,6 +77,7 @@ describe('rpcErrorSchema', () => { }).code).toBe('model-unavailable') expect(rpcErrorSchema.parse({ code: 'agent-busy', message: 'm', details: { reason: 'r' } }).code).toBe('agent-busy') expect(rpcErrorSchema.parse({ code: 'queue-item-not-found', message: 'm', details: { itemId: 'i' } }).code).toBe('queue-item-not-found') + expect(rpcErrorSchema.parse({ code: 'steer-unavailable', message: 'm', details: { itemId: 'i' } }).code).toBe('steer-unavailable') expect(rpcErrorSchema.parse({ code: 'command-error', message: 'm', details: {} }).code).toBe('command-error') expect(rpcErrorSchema.parse({ code: 'unknown-command', message: 'm', details: {} }).code).toBe('unknown-command') expect(rpcErrorSchema.parse({ code: 'title-invalid', message: 'm', details: { sessionId: 's' } }).code).toBe('title-invalid') @@ -280,6 +281,9 @@ describe('sessions domain schemas', () => { expect(sessionUpdateQueueRequestSchema.parse({ sessionId: 's1', itemId: 'i1', action: { kind: 'remove' }, }).action.kind).toBe('remove') + expect(sessionUpdateQueueRequestSchema.parse({ + sessionId: 's1', itemId: 'i1', action: { kind: 'steer' }, + }).action.kind).toBe('steer') expect(() => sessionUpdateQueueRequestSchema.parse({ sessionId: 's1', itemId: 'i1', action: { kind: 'promote' }, })).toThrow() From 5b17d3bab5fc09eb7557554a9c9f6b3ebf513d03 Mon Sep 17 00:00:00 2001 From: kingwl Date: Thu, 30 Jul 2026 03:29:12 +0800 Subject: [PATCH 04/13] fix(web): match queue steer icon to design --- .../client/ui-conversation/src/client/queue/QueueDock.tsx | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/packages/client/ui-conversation/src/client/queue/QueueDock.tsx b/packages/client/ui-conversation/src/client/queue/QueueDock.tsx index b374fa36d6..10c16e6dc5 100644 --- a/packages/client/ui-conversation/src/client/queue/QueueDock.tsx +++ b/packages/client/ui-conversation/src/client/queue/QueueDock.tsx @@ -9,7 +9,7 @@ import type { PropsLocale, PropsRuntime } from '@deepseek-ai/dsh-client-ui-slots import type { SessionId } from '@deepseek-ai/dsh-client-runtime/client' import { IconCheckOutline16, IconChevronDownOutline14, IconChevronUpOutline14, - IconCloseOutline16, IconEditOutline16, IconRightUpOutline16, IconTrashOutline16, + IconCloseOutline16, IconEditOutline16, IconSendOutline16, IconTrashOutline16, } from '@deepseek-ai/dsh-client-ui-primitives' import type { QueueAction, QueueItemId } from '../contract/queue.ts' import { NS } from '../locales.ts' @@ -185,7 +185,7 @@ export function QueueDock({ useSession, updateQueue, notify, t }: QueueDockProps ) }} > - + )} From dffe955ed203b85c324fbf3ad77d1996b01e27b0 Mon Sep 17 00:00:00 2001 From: imccyu Date: Sun, 2 Aug 2026 15:36:17 +0800 Subject: [PATCH 05/13] feat(web): surface and configure composer steering --- ...026-07-30-web-queue-steer-action.i18n.yaml | 4 +- .../2026-07-30-web-queue-steer-action.md | 24 +-- .../2026-07-30-web-queue-steer-action.zh.md | 24 +-- apps/web/tests/settings-chrome.e2e.ts | 31 +++- .../queue-actions/editing.expected.md | 2 +- .../queue-actions/preserved.expected.md | 2 + .../settings-chrome/dialog.expected.md | 4 + .../snapshots/steering/mid-steer.expected.md | 1 + apps/web/tests/steering.e2e.ts | 150 +++++++++++++++--- .../client/session-history/history-fold.ts | 3 +- .../src/client/sessions/conversation.ts | 13 +- .../runtime/src/client/sessions/session.ts | 3 + .../src/client/sessions/transcript-adapter.ts | 3 +- .../client/runtime/tests/queue-store.spec.ts | 47 +++++- .../client/ui-conversation/README.i18n.yaml | 4 +- packages/client/ui-conversation/README.md | 8 +- packages/client/ui-conversation/README.zh.md | 8 +- .../ui-conversation/src/client/apply.ts | 23 ++- .../src/client/chat/ChatView.tsx | 20 ++- .../src/client/chat/MessageItem.tsx | 19 ++- .../client/contract/composer-submission.ts | 10 ++ .../src/client/contract/slots.ts | 3 + .../src/client/input/contract.ts | 22 +-- .../src/client/input/facade.ts | 26 ++- .../ui-conversation/src/client/input/hub.ts | 7 +- .../src/client/input/machine.ts | 31 ++-- .../src/client/input/submission-policy.ts | 72 +++++++++ .../ui-conversation/src/client/locales.ts | 8 + .../src/client/queue/QueueDock.tsx | 5 +- .../ui-conversation/src/client/queue/store.ts | 4 +- .../settings/EnterBehaviorRow.module.css | 56 +++++++ .../src/client/settings/EnterBehaviorRow.tsx | 76 +++++++++ .../src/client/skeleton/InputBar.tsx | 17 +- .../ui-conversation/tests/chat-apply.spec.tsx | 3 + .../ui-conversation/tests/chat-view.spec.tsx | 45 ++++++ .../tests/enter-behavior-row.spec.tsx | 67 ++++++++ .../ui-conversation/tests/input-bar.spec.tsx | 50 ++++-- .../tests/input-machine.spec.ts | 70 +++----- .../tests/input-matrix.spec.tsx | 5 +- .../tests/input-scenarios.spec.tsx | 5 +- .../ui-conversation/tests/queue-dock.spec.tsx | 14 +- .../ui-conversation/tests/skeleton.spec.tsx | 3 +- .../tests/submission-policy.spec.ts | 65 ++++++++ packages/host/apiproxy/src/api-proxy.ts | 45 ++++-- .../host/apiproxy/src/api/events.schema.ts | 1 + packages/host/apiproxy/src/api/events.ts | 13 +- .../apiproxy/tests/api-proxy-commands.spec.ts | 113 +++++++++++-- .../host/apiproxy/tests/rpc-schemas.spec.ts | 2 +- 48 files changed, 1006 insertions(+), 225 deletions(-) create mode 100644 packages/client/ui-conversation/src/client/contract/composer-submission.ts create mode 100644 packages/client/ui-conversation/src/client/input/submission-policy.ts create mode 100644 packages/client/ui-conversation/src/client/settings/EnterBehaviorRow.module.css create mode 100644 packages/client/ui-conversation/src/client/settings/EnterBehaviorRow.tsx create mode 100644 packages/client/ui-conversation/tests/enter-behavior-row.spec.tsx create mode 100644 packages/client/ui-conversation/tests/submission-policy.spec.ts diff --git a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml index 780858e1d8..1220cc39fa 100644 --- a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml +++ b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md -2026-07-30-web-queue-steer-action.md: 4a8a04c056748d5c607c5ec868de4f2dd8342f2c -2026-07-30-web-queue-steer-action.zh.md: 92225f3a61e24b9eb687c50400aa6fe5e6ff5623 +2026-07-30-web-queue-steer-action.md: b84ee5acf472ba8482f6dfb7f466a905ddfd8f4f +2026-07-30-web-queue-steer-action.zh.md: fd4b4f8f8ab18db8a5372ff1a419fa186e29f399 diff --git a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md index 4a8a04c056..b84ee5acf4 100644 --- a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md +++ b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md @@ -6,7 +6,7 @@ English | [中文](2026-07-30-web-queue-steer-action.zh.md) ## Problem -The Web composer deliberately queues Enter submissions while an agent runs. QueueDock already gives each pending message an addressable row, and the durable transcript already renders consumed `steering/message` events with an interjection badge, but Web has no action connecting those two surfaces. +The Web composer originally queued every Enter submission while an agent ran. QueueDock already gives each pending message an addressable row, and the durable transcript already renders consumed `steering/message` events as user-style bubbles without message actions, but Web had neither an action connecting those two surfaces nor a direct composer gesture for choosing current-turn steering. Implementing the row action as a client-side delete followed by `session.prompt(mode: 'steer')` would split one user intent across two RPCs. Driver claim could win between them, the steer could fail after deletion, or the existing best-effort `agent.steer()` fallback could silently append a new Queue item after the original occurrence was removed. A send-now action must therefore distinguish current-turn steering from Queue promotion and preserve the original row when steering is no longer possible. @@ -14,12 +14,14 @@ Implementing the row action as a client-side delete followed by `session.prompt( ### Product contract -Each non-editing QueueDock row exposes the upward-arrow action as “插话发送”. The action is enabled only while the session reports a running agent; mixed-content messages remain eligible because steering forwards the complete immutable `UserMessage` rather than the row's text projection. Edit and delete keep their existing behavior, and the composer continues to submit Enter as Queue. +Each non-editing QueueDock row exposes the upward-arrow action as “插话发送”. The action is enabled only while the session reports a running agent; mixed-content messages remain eligible because steering forwards the complete immutable `UserMessage` rather than the row's text projection. Edit and delete keep their existing behavior. -Activating the action requests strict current-turn steering for that exact `InboxItemId`. Success removes the Queue row through the authoritative Host snapshot. When AgentLoop drains it, the existing durable `steering/message` event and transcript badge render the message without a new chat presentation path. +Activating the action requests strict current-turn steering for that exact `InboxItemId`. Success removes the Queue row through the authoritative Host snapshot. When AgentLoop drains it, the existing durable `steering/message` event renders the same user-style bubble without a separate durable presentation path. The running bit is only an interaction hint. AgentLoop's `acceptsNextStep` value is authoritative at the synchronous mutation boundary. If that window has closed, the operation leaves the Queue occurrence unchanged and returns a typed `steer-unavailable` error; if the driver already claimed the occurrence, it returns the existing `queue-item-not-found` error. The UI reports either race without optimistically removing the row. +The composer uses a separate best-effort contract for newly typed input. While the addressed session is idle, Enter and Cmd/Ctrl+Enter both perform an ordinary Queue send. While it is running, a General Settings preference assigns plain Enter to Queue (the default) or Steer, and Cmd/Ctrl+Enter performs the other behavior; Shift+Enter inserts a newline. The browser persists that preference, and it affects only the busy-state gesture pair. If a direct composer Steer misses the current next-step window, AgentLoop automatically admits it as the next waking Queue turn and the Web does not report a failure. + ### Agent and lifecycle boundary `InboxAction` gains a consumer-backed `{ kind: 'steer' }` operation alongside edit and remove. `Agent.updateInbox()` handles it only after locating the queued occurrence and proving `acceptsNextStep`; it never delegates to the best-effort `agent.steer()` alias. @@ -32,17 +34,19 @@ The action does not run `agent/prompt-submit`: choosing steering intentionally c `session.updateQueue` carries the `steer` action and maps the two negative outcomes to typed RPC errors. The conversion is one synchronous Agent operation; the Host never reconstructs it by combining remove and prompt calls. -The Host's transient `session/queue` projection remains Queue-only. It ignores the new pending steering occurrence and removes the old row when its discard arrives. Pending steering does not gain edit, delete, or reconnect presentation in this cut. A later dedicated pending-steering projection may add that observability without widening Queue mutation semantics. +The Host's existing `queuedMirror` remains the sole transient inbox authority. Its `session/queue` snapshot carries every live occurrence with `placement: 'queued' | 'steering'`: QueueDock renders only queued rows, while ChatView renders pending steering at the conversation tail without edit or delete actions. Reconnect replays the same snapshot, so this visibility does not require client optimism or a second registry. -The existing `session.prompt(mode: 'steer')` contract remains best-effort for new input: outside the next-step window it may become a waking follow-up. Only the Queue row action is strict, because failure can safely leave its already-pending message untouched. +When AgentLoop claims pending steering, it emits `agent/inbox/dequeue` immediately before synchronously appending `steering/message`. The Host retires that steering row on the following microtask, allowing the durable session event to enter the linear mux stream first. ChatView matches the shared `MessageId` and suppresses the transient projection as soon as the durable node exists, so one bubble changes authority without a visible gap or duplicate; an append failure still retires the claimed row. + +The existing `session.prompt(mode: 'steer')` contract remains best-effort for new input: outside the next-step window it becomes a waking follow-up. The composer carries an explicit `queue | steer` mode through slash adjudication and reference serialization before calling that contract. A browser-local submission policy owns the persisted busy-Enter preference and resolves plain versus accelerated Enter as complementary gestures; the Settings row and InputBar share that policy without duplicating storage or delivery-window authority. Only the Queue row action is strict, because failure can safely leave its already-pending message untouched. ### Verification AgentLoop contract coverage holds prompt admission open, converts one exact queued occurrence, and proves the replacement steering occurrence keeps the message value, drains as `steering/message`, and never starts its former independent turn. It also pins unavailable-window retention, claimed-address rejection, and re-entrant cancellation lifecycle conservation. -Host schema and proxy tests cover the new action, both typed errors, authoritative Queue snapshots, and the absence of pending steering from reconnect snapshots. QueueDock tests cover running-state enablement, complete-content eligibility, failure retention, and authoritative success retirement. +Host schema and proxy tests cover the new action, both typed errors, placement-aware snapshots and reconnect replay, plus durable-before-retirement ordering. QueueDock tests cover running-state enablement, complete-content eligibility, failure retention, authoritative success retirement, and filtering of steering occurrences. ChatView tests cover the transient bubble and its single-copy handoff to the durable node. -The keyless Web steering scenario queues a message through the real composer while the first response streams, activates the row arrow, then uses `ask_user_question` as a stable pending-steering barrier. After the answer, it proves one badged interjection becomes durable and the next model request obeys it. Queue edit/delete scenarios continue to prove those actions are unchanged. +The keyless Web steering scenario queues a message through the real composer while the first response streams, activates the row arrow, then uses `ask_user_question` as a stable pending-steering barrier. It proves the Host-backed pending bubble appears before admission, hands off to one durable interjection after the answer, and affects the next model request. Assembled composer scenarios prove default-mode Cmd+Enter reaches the same pending and durable path without creating a Queue row, while Steer-mode Cmd+Enter creates a Queue row instead. Settings and submission-policy coverage pin the default, persistence, busy-only scope, and complementary gesture mapping; Queue edit/delete scenarios continue to prove those actions are unchanged. ## Alternatives considered @@ -50,18 +54,18 @@ The keyless Web steering scenario queues a message through the real composer whi **Restore Queue promotion under the upward arrow.** Rejected because moving an item to the front still creates an independent admitted turn. The control promises current-turn steering, not priority within Queue. -**Use the existing best-effort `agent.steer()` behavior.** Rejected for this action because a closed next-step window would silently turn the selected row back into queued work, possibly at a different position and identity. Strict failure preserves the original occurrence and makes the semantic race visible. +**Use the existing best-effort `agent.steer()` behavior for the Queue row.** Rejected for that action because a closed next-step window would silently turn the selected row back into queued work, possibly at a different position and identity. Strict failure preserves the original occurrence and makes the semantic race visible. Newly typed composer input has no existing Queue occurrence to preserve, so it intentionally uses the best-effort behavior. **Change `agent.steer()` to be strict for every caller.** Rejected because TUI and plugin callers use its safe follow-up fallback for newly submitted input. A queued row has recoverable state that those callers do not. **Preserve the same `InboxItemId` while changing placement.** Rejected because `InboxItemId` identifies one FIFO acceptance and `placement` records that acceptance's resolved delivery. Ending one queued occurrence and accepting one steering occurrence keeps lifecycle facts truthful and leaves the conservation invariant unchanged. -**Expose pending steering in `session/queue`.** Deferred because the existing product design provides no pending-steering row state or operations. Authoritative Queue retirement plus the durable consumed bubble is sufficient for the first interaction cut; reconnect visibility can be added through a dedicated projection if product testing shows the gap matters. +**Add a dedicated pending-steering projection and client store.** Rejected because queued and steering occurrences already share one Agent inbox lifecycle and one Host mirror. A second projection would duplicate reconnect state and ordering authority; a placement tag lets each client surface select its rows without widening Queue mutation semantics. **Cancel the active turn and run the selected Queue item.** Rejected because it destroys unrelated in-flight work and starts a new turn rather than steering the current one. ## Consequences -A successful action can be pending but absent from the Web after its Queue row retires and before `steering/message` commits; a refresh during that interval has no pending-steering indication. The running bit can also remain true briefly after the strict next-step window closes, so the button may be enabled for an operation that correctly returns `steer-unavailable`. +`session/queue` describes a placement-aware transient inbox snapshot rather than a Queue-only list, so every consumer must filter by placement. Pending steering survives reconnect and appears immediately, but remains non-durable until `steering/message` commits. The running bit can also remain true briefly after the strict next-step window closes, so the button may be enabled for an operation that correctly returns `steer-unavailable`. The explicit action changes delivery from an independently admitted turn to current-turn steering, so prompt-admission plugins do not process the converted message. Enqueue-before-discard lifecycle publication remains required for re-entrant cancellation safety; focused regression coverage protects that ordering. diff --git a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md index 92225f3a61..fd4b4f8f8a 100644 --- a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md +++ b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md @@ -6,7 +6,7 @@ Status: implemented ## 问题 -Web composer 会在 agent 运行期间有意把 Enter 提交作为 Queue 入队。QueueDock 已经为每条待处理消息提供可寻址的行,持久 transcript(文本记录)也已能把消费后的 `steering/message` 事件渲染为带插话徽标的消息,但 Web 没有连接这两个界面的操作。 +Web composer 原本会在 agent 运行期间把所有 Enter 提交作为 Queue 入队。QueueDock 已经为每条待处理消息提供可寻址的行,持久 transcript(文本记录)也已能把消费后的 `steering/message` 事件渲染为不带消息操作的用户样式气泡,但 Web 既没有连接这两个界面的操作,也没有让用户从 composer 直接选择当前轮次 steering 的手势。 如果 Web 先在客户端删除该行,再调用 `session.prompt(mode: 'steer')`,就会把用户的一次意图拆分到两个 RPC 中。驱动器可能在两次调用之间先认领该项,steering 投递也可能在删除后失败;现有尽力而为的 `agent.steer()` 回退还可能在原单次入队项被移除后,静默追加一个新的 Queue 项。因此,立即发送操作必须区分当前轮次 steering 与 Queue 前移,并在 steering 已不可用时保留原行。 @@ -14,12 +14,14 @@ Web composer 会在 agent 运行期间有意把 Enter 提交作为 Queue 入队 ### 产品契约 -每个非编辑态的 QueueDock 行都会提供名为“插话发送”的向上箭头操作。仅当会话报告 agent 正在运行时,该操作才会启用;包含混合内容的消息仍可使用,因为 steering 会转发完整且不可变的 `UserMessage`,而非该行的文本投影。编辑和删除保持现有行为,composer 也继续把 Enter 提交为 Queue。 +每个非编辑态的 QueueDock 行都会提供名为“插话发送”的向上箭头操作。仅当会话报告 agent 正在运行时,该操作才会启用;包含混合内容的消息仍可使用,因为 steering 会转发完整且不可变的 `UserMessage`,而非该行的文本投影。编辑和删除保持现有行为。 -触发该操作会针对对应的 `InboxItemId` 请求严格的当前轮次 steering。操作成功后,权威 Host 快照会移除 Queue 行。AgentLoop 排空该项时,现有持久 `steering/message` 事件与 transcript 插话徽标会渲染这条消息,无需新增聊天展示路径。 +触发该操作会针对对应的 `InboxItemId` 请求严格的当前轮次 steering。操作成功后,权威 Host 快照会移除 Queue 行。AgentLoop 排空该项时,现有持久 `steering/message` 事件会渲染相同的用户样式气泡,无需另建持久展示路径。 running 标志位只用于提示交互状态。在同步变更边界上,AgentLoop 的 `acceptsNextStep` 值才是权威依据。如果该窗口已经关闭,操作会保持 Queue 单次入队项不变,并返回类型化的 `steer-unavailable` 错误;如果驱动器已经认领该项,则返回现有的 `queue-item-not-found` 错误。UI 会报告任一竞态,不会乐观地移除该行。 +Composer 对新输入采用另一套尽力而为契约。所寻址会话空闲时,Enter 和 Cmd/Ctrl+Enter 都执行普通 Queue 发送。会话运行期间,General Settings 偏好会把普通 Enter 分配为 Queue(默认值)或 Steer,Cmd/Ctrl+Enter 则执行另一种行为;Shift+Enter 用于换行。浏览器会持久化该偏好,并且它只影响繁忙态下这对手势。如果 composer 直接发出的 Steer 错过当前 next-step 窗口,AgentLoop 会自动将其接纳为下一条唤醒 Queue 轮次,Web 不显示失败。 + ### Agent 与生命周期边界 `InboxAction` 会在编辑和移除之外,新增由实际消费方支撑的 `{ kind: 'steer' }` 操作。`Agent.updateInbox()` 只有在找到 queued 单次入队项并确认 `acceptsNextStep` 后才会处理该操作,绝不会委托给尽力而为的 `agent.steer()` 别名。 @@ -32,17 +34,19 @@ running 标志位只用于提示交互状态。在同步变更边界上,AgentL `session.updateQueue` 会携带 `steer` 操作,并把两种负面结果映射为类型化 RPC 错误。这项转换是一次同步 Agent 操作;Host 绝不会通过组合移除和提示词调用来重建它。 -Host 的瞬态 `session/queue` 投影仍然只包含 Queue。它会忽略新的待处理 steering 单次入队项,并在收到旧项的 discard 时移除原行。本阶段不会为待处理 steering 增加编辑、删除或重连展示。未来可以用专用的待处理 steering 投影补充这种可观测性,而无需扩大 Queue 变更语义。 +Host 仍以现有 `queuedMirror` 作为唯一的瞬态 inbox 权威。`session/queue` 快照会携带所有存活单次入队项及其 `placement: 'queued' | 'steering'`:QueueDock 只渲染 queued 行,ChatView 则在会话流末尾渲染待处理 steering,且不提供编辑或删除操作。重连会重放同一份快照,因此这项可见性既不依赖客户端乐观展示,也不需要第二个 registry。 -现有 `session.prompt(mode: 'steer')` 对新输入仍采用尽力而为的契约:在 next-step 窗口之外,它可能变为会唤醒 agent 的后续轮次。只有 Queue 行操作采用严格语义,因为失败时可以安全地保留其已经待处理的消息。 +AgentLoop 认领待处理 steering 时,会在同步追加 `steering/message` 之前立即发出 `agent/inbox/dequeue`。Host 会等到下一个微任务才退役该 steering 行,让持久 session 事件先进入线性 mux 流。ChatView 会匹配两边共享的 `MessageId`,并在持久节点出现时立即抑制瞬态投影,因此同一个气泡切换权威时不会产生可见空档或重复;如果追加失败,已认领行仍会退役。 + +现有 `session.prompt(mode: 'steer')` 对新输入仍采用尽力而为的契约:在 next-step 窗口之外,它会变为唤醒 agent 的后续轮次。Composer 会让显式 `queue | steer` 模式经过 slash 裁决与引用序列化,再调用该契约。浏览器本地的提交策略拥有持久化的繁忙态 Enter 偏好,并把普通 Enter 与加速 Enter 解析为互补手势;Settings 行和 InputBar 共享该策略,不重复实现存储或投递窗口权威。只有 Queue 行操作采用严格语义,因为失败时可以安全地保留其已经待处理的消息。 ### 验证 AgentLoop 契约覆盖保持提示词接纳窗口打开,转换一个精确的 queued 单次入队项,并证明替代它的 steering 单次入队项保留消息值、以 `steering/message` 的形式排空,且绝不启动原本的独立轮次。该覆盖还钉住窗口不可用时保留原项、拒绝已被认领的地址,以及可重入取消下的生命周期守恒。 -Host schema 和代理测试覆盖新操作、两种类型化错误、权威 Queue 快照,以及重连快照不包含待处理 steering。QueueDock 测试覆盖按运行状态启用、混合内容消息仍可完整投递、失败时保留原行,以及成功后由权威快照退役。 +Host schema 和代理测试覆盖新操作、两种类型化错误、带 placement 的快照与重连重放,以及先持久化再退役的顺序。QueueDock 测试覆盖按运行状态启用、混合内容消息仍可完整投递、失败时保留原行、成功后由权威快照退役,以及过滤 steering 单次入队项。ChatView 测试覆盖瞬态气泡及其只保留一份的持久节点交接。 -无密钥 Web steering 场景在第一次响应流式输出期间,通过真实 composer 排队一条消息并触发行上的箭头,再用 `ask_user_question` 作为稳定的待处理 steering 屏障。回答问题后,该场景证明一条带徽标的插话成为持久记录,并且下一次模型请求遵循它。Queue 编辑/删除场景继续证明这些操作没有变化。 +无密钥 Web steering 场景在第一次响应流式输出期间,通过真实 composer 排队一条消息并触发行上的箭头,再用 `ask_user_question` 作为稳定的待处理 steering 屏障。该场景证明 Host 支撑的待处理气泡会在准入前出现,在回答后交接为唯一一条持久插话,并影响下一次模型请求。组装后的 composer 场景证明默认模式下的 Cmd+Enter 无需创建 Queue 行,也会进入同一条待处理与持久路径;Steer 模式下的 Cmd+Enter 则会创建 Queue 行。Settings 与提交策略覆盖会固定默认值、持久化、仅限繁忙态的作用域和互补手势映射;Queue 编辑/删除场景继续证明这些操作没有变化。 ## 考虑过的替代方案 @@ -50,18 +54,18 @@ Host schema 和代理测试覆盖新操作、两种类型化错误、权威 Queu **恢复向上箭头对应的 Queue 前移操作。** 不予采纳,因为把某个项移到队首仍然会创建一个独立接纳的轮次。该控件承诺的是当前轮次 steering,而不是 Queue 内的优先级。 -**使用现有尽力而为的 `agent.steer()` 行为。** 不予采纳,因为关闭的 next-step 窗口会静默地把选中行重新变成 queued 工作,而且位置和标识可能不同。严格失败会保留原单次入队项,并让这项语义竞态明确可见。 +**为 Queue 行使用现有尽力而为的 `agent.steer()` 行为。** 不予采纳,因为关闭的 next-step 窗口会静默地把选中行重新变成 queued 工作,而且位置和标识可能不同。严格失败会保留原单次入队项,并让这项语义竞态明确可见。新输入的 composer 消息没有需要保留的现有 Queue 单次入队项,因此有意采用尽力而为行为。 **让每个调用方使用的 `agent.steer()` 都采用严格语义。** 不予采纳,因为 TUI 和插件调用方会针对新提交的输入使用其安全的后续轮次回退。queued 行具有这些调用方不具备的可恢复状态。 **改变投递方式时保留同一个 `InboxItemId`。** 不予采纳,因为 `InboxItemId` 标识一次 FIFO 接受,而 `placement` 记录该次接受解析出的投递方式。结束一个 queued 单次入队项并接受一个 steering 单次入队项,能够使生命周期事实保持如实,并让守恒不变量保持不变。 -**在 `session/queue` 中暴露待处理 steering。** 暂缓,因为现有产品设计没有为待处理 steering 提供行状态或操作。权威的 Queue 退役加上持久的已消费气泡,足以支撑首个交互阶段;如果产品测试表明这一缺口影响显著,可以通过专用投影增加重连可见性。 +**增加专用的待处理 steering 投影和客户端 store。** 不予采纳,因为 queued 与 steering 单次入队项已经共享同一套 Agent inbox 生命周期和 Host mirror。第二份投影会重复保存重连状态与顺序权威;placement 标签能让各客户端界面选取自己的行,而不扩大 Queue 变更语义。 **取消活动轮次并运行选中的 Queue 项。** 不予采纳,因为这会破坏无关的进行中工作,并且会启动新轮次,而不是 steering 当前轮次。 ## 后果 -操作成功后,从 Queue 行退役到 `steering/message` 提交之间,对应消息可能仍处于待处理状态,却不会出现在 Web 中;如果在此期间刷新,界面不会显示待处理 steering。严格 next-step 窗口关闭后,running 标志位仍可能短暂保持为 true,因此按钮可能会为一个最终正确返回 `steer-unavailable` 的操作保持启用。 +`session/queue` 表示带 placement 的瞬态 inbox 快照,而不只是 Queue 列表,因此每个消费方都必须按 placement 过滤。待处理 steering 会在界面中立即出现并能在重连后恢复,但在 `steering/message` 提交前仍不持久。严格 next-step 窗口关闭后,running 标志位仍可能短暂保持为 true,因此按钮可能会为一个最终正确返回 `steer-unavailable` 的操作保持启用。 这项显式操作会把投递方式从经独立接纳的轮次改为当前轮次 steering,因此提示词接纳插件不会处理转换后的消息。为保证可重入取消安全,生命周期事件仍必须先发布 enqueue 再发布 discard;有针对性的回归覆盖会保护这一顺序。 diff --git a/apps/web/tests/settings-chrome.e2e.ts b/apps/web/tests/settings-chrome.e2e.ts index 7bdfcc1d59..19e1b33538 100644 --- a/apps/web/tests/settings-chrome.e2e.ts +++ b/apps/web/tests/settings-chrome.e2e.ts @@ -2,8 +2,9 @@ // section switching, both close paths), the Appearance preference row (the // real theme gesture — click 深色 and the whole cascade runs: ThemeService preference -> localStorage dsh.theme // -> theme/change -> ui-layout's presenter -> body attribute -> alias token) -// and the Language row (settings-scoped localization + persisted dsh.locale), -// plus Permission as the persisted default for subsequently created sessions. +// the Language row (settings-scoped localization + persisted dsh.locale), +// the busy-state Enter preference, plus Permission as the persisted default +// for subsequently created sessions. // Zero model calls: everything is pure client + persistence state on a blank // frame, so there is no fixture and a stray stream would fail loud on the // open llm seam. @@ -182,6 +183,32 @@ describe('web e2e: settings modal and General preferences', () => { expect(tripwire.pageErrors).toEqual([]) }, 90_000) + it('persists the busy-state Enter behavior across reload and restores Queue', async () => { + onTestFailed(() => saveFailureShot(page, 'web-e2e-settings-enter-behavior')) + await page.getByRole('button', { name: '设置', exact: true }).click() + const dialog = page.getByRole('dialog', { name: '设置' }) + await dialog.waitFor({ timeout: 10_000 }) + await dialog.getByRole('button', { name: '排队发送' }).click() + await page.getByRole('menuitem', { name: '插话发送' }).click() + await dialog.getByRole('button', { name: '插话发送' }).waitFor({ timeout: 10_000 }) + expect(await page.evaluate(() => localStorage.getItem('dsh.conversation.busyEnter'))).toBe('steer') + await page.keyboard.press('Escape') + + const warningStart = tripwire.warnings.length + await page.reload({ waitUntil: 'load' }) + await page.waitForSelector('[class*="frame"]', { timeout: 30_000 }) + acknowledgeReloadConnectionLoss(tripwire, warningStart) + await page.getByRole('button', { name: '设置', exact: true }).click() + const reloaded = page.getByRole('dialog', { name: '设置' }) + await reloaded.getByRole('button', { name: '插话发送' }).waitFor({ timeout: 10_000 }) + await reloaded.getByRole('button', { name: '插话发送' }).click() + await page.getByRole('menuitem', { name: '排队发送' }).click() + await reloaded.getByRole('button', { name: '排队发送' }).waitFor({ timeout: 10_000 }) + expect(await page.evaluate(() => localStorage.getItem('dsh.conversation.busyEnter'))).toBe('queue') + await page.keyboard.press('Escape') + expect(tripwire.pageErrors).toEqual([]) + }, 90_000) + it('switches the settings surface language and persists dsh.locale', async () => { onTestFailed(() => saveFailureShot(page, 'web-e2e-settings-language')) await page.getByRole('button', { name: '设置', exact: true }).click() diff --git a/apps/web/tests/snapshots/queue-actions/editing.expected.md b/apps/web/tests/snapshots/queue-actions/editing.expected.md index 3f6ebab8a2..bfff0cd9f0 100644 --- a/apps/web/tests/snapshots/queue-actions/editing.expected.md +++ b/apps/web/tests/snapshots/queue-actions/editing.expected.md @@ -23,7 +23,7 @@ - img - button "Remove queued message": - img - - button "插话发送": + - button "Steer queued message": - img - listitem: - textbox "Edit queued message": Edited queue item diff --git a/apps/web/tests/snapshots/queue-actions/preserved.expected.md b/apps/web/tests/snapshots/queue-actions/preserved.expected.md index 0a43cee68b..fe6f5b6087 100644 --- a/apps/web/tests/snapshots/queue-actions/preserved.expected.md +++ b/apps/web/tests/snapshots/queue-actions/preserved.expected.md @@ -33,6 +33,8 @@ - img - button "Remove queued message": - img + - button "Steer queued message": + - img - textbox "Message the agent" - button "Commands": - img diff --git a/apps/web/tests/snapshots/settings-chrome/dialog.expected.md b/apps/web/tests/snapshots/settings-chrome/dialog.expected.md index 118e4ff3e1..cb693b62a8 100644 --- a/apps/web/tests/snapshots/settings-chrome/dialog.expected.md +++ b/apps/web/tests/snapshots/settings-chrome/dialog.expected.md @@ -28,3 +28,7 @@ - button "跟随系统" [pressed]: - img - text: 跟随系统 + - text: 繁忙时 Enter 键行为 仅在智能体运行时生效;Cmd/Ctrl+Enter 使用另一行为 + - button "排队发送": + - text: 排队发送 + - img diff --git a/apps/web/tests/snapshots/steering/mid-steer.expected.md b/apps/web/tests/snapshots/steering/mid-steer.expected.md index 27b40ef442..9f077b21a5 100644 --- a/apps/web/tests/snapshots/steering/mid-steer.expected.md +++ b/apps/web/tests/snapshots/steering/mid-steer.expected.md @@ -23,6 +23,7 @@ - img - text: Ask question waiting - status: Deep diving... +- text: "Interjection: include the word BANANA in your final reply." - region "Ready to continue?": - text: Checkpoint - heading "Ready to continue?" [level=2] diff --git a/apps/web/tests/steering.e2e.ts b/apps/web/tests/steering.e2e.ts index bfa8aae162..794a92b74b 100644 --- a/apps/web/tests/steering.e2e.ts +++ b/apps/web/tests/steering.e2e.ts @@ -1,7 +1,7 @@ -// Web e2e scenario: queue a message while the first response streams, strictly -// transfer that exact occurrence to steering through QueueDock, then prove it -// is logged, rendered, and obeyed. The following question tool supplies a -// deterministic pending-steering snapshot before the step can drain. +// Web e2e scenarios for both steering entry points: QueueDock strictly +// transfers one queued occurrence, while the complementary composer gestures +// choose Queue or Steer. The question tool supplies a deterministic pending- +// steering snapshot before the step can drain. import { readFile } from 'node:fs/promises' import { fileURLToPath } from 'node:url' import { join } from 'node:path' @@ -18,13 +18,10 @@ import { connectFreshWorkspace, newEnglishPage, saveFailureShot } from './suppor const SNAPSHOT_DIR = fileURLToPath(new URL('./snapshots/steering', import.meta.url)) const FIXTURE = join(SNAPSHOT_DIR, 'session.jsonl') -// Two goldens for the two distinct states this interaction produces: the -// mid-turn moment (steer ACCEPTED but deliberately invisible — the loop -// drains steering at the step boundary, so no steering text exists while -// the question still blocks the step) and the settled transcript (plain -// bubble in place, final reply obeying it). The pair pins the timing -// semantics visually: if the client ever starts rendering pending steers -// eagerly, the mid-steer golden flips first. +// Two goldens pin the transient Host projection and its durable handoff: the +// mid-turn state renders accepted steering from session/queue while the +// question blocks admission, then the settled state renders the same message +// from steering/message beside the reply that obeys it. const MID_EXPECTED = join(SNAPSHOT_DIR, 'mid-steer.expected.md') const SETTLED_EXPECTED = join(SNAPSHOT_DIR, 'settled.expected.md') const MODE = webSnapshotMode() @@ -90,18 +87,17 @@ describe('web e2e: mid-turn steering lands durably and visibly', () => { const steerButton = queuedRow.getByRole('button', { name: 'Steer queued message' }) await expect.poll(() => steerButton.isEnabled(), { timeout: 10_000 }).toBe(true) await steerButton.click({ timeout: 10_000 }) - await expect.poll(() => page.getByText(STEER, { exact: true }).count(), { timeout: 10_000 }).toBe(0) + const pendingSteering = page.locator('[data-pending-steering]').filter({ hasText: STEER }) + await pendingSteering.waitFor({ timeout: 10_000 }) - // The blocked composer is the mid-turn barrier: the tool cannot finish - // this step, so the accepted steering remains pending and invisible. + // The blocked composer keeps steering pending long enough to observe the + // Host-authoritative mirror before the loop admits it durably. const composer = page.locator('[data-question-key]') await composer.waitFor({ timeout: MODE === 'record' ? 120_000 : 30_000 }) if (MODE !== 'record') { - // Mid-turn golden: the converted steer is pending in the loop but the - // loop drains steering only at the step boundary, so no Queue row or - // steering/message bubble renders while the question still blocks. - expect(await page.getByText(STEER, { exact: true }).count()).toBe(0) + expect(await page.getByText(STEER, { exact: true }).count()).toBe(1) + expect(await pendingSteering.count()).toBe(1) expect(await page.getByRole('button', { name: 'Edit queued message' }).count()).toBe(0) const snapshot = await captureStableAria(page, '[class*="centerCol"]', scaffold.workspaceCwd) await compareOrRefreshGolden(MID_EXPECTED, snapshot, MODE) @@ -137,6 +133,7 @@ describe('web e2e: mid-turn steering lands durably and visibly', () => { // Visible: the plain steering bubble plus the reply that obeys it // (steer text + final reply each contain the marker word). await expect.poll(() => page.getByText(STEER, { exact: true }).count(), { timeout: 15_000 }).toBe(1) + expect(await pendingSteering.count()).toBe(0) await expect.poll(() => page.getByText('BANANA', { exact: false }).count(), { timeout: 10_000 }).toBeGreaterThanOrEqual(2) expect(await page.locator('[data-question-key]').count()).toBe(0) // Settled golden: steer text between the question round trip and the @@ -151,3 +148,120 @@ describe('web e2e: mid-turn steering lands durably and visibly', () => { await assertFixtureInventory(SNAPSHOT_DIR, ['session.jsonl', 'mid-steer.expected.md', 'settled.expected.md']) }) }) + +describe('web e2e: composer shortcut steers directly', () => { + let scaffold: WebScaffold + let browser: Browser + let page: Page + let tripwire: ReturnType + const sessionEvents: SessionEvent[] = [] + + beforeAll(async () => { + scaffold = await launchWebScaffold({ replayFixture: FIXTURE, paceMs: 100 }) + scaffold.ctx.on('session/event', (_session, event) => { sessionEvents.push(event) }) + browser = await chromium.launch() + page = await newEnglishPage(browser) + tripwire = watchConsole(page) + await page.goto(scaffold.baseUrl, { waitUntil: 'load' }) + await page.waitForSelector('[class*="frame"]', { timeout: 30_000 }) + await connectFreshWorkspace(page, scaffold.workspaceCwd) + }, 120_000) + + afterAll(async () => { + await browser?.close() + await scaffold?.close() + }) + + it.skipIf(MODE === 'record')('uses Cmd+Enter without creating a Queue row', async () => { + onTestFailed(() => saveFailureShot(page, 'web-e2e-composer-steering')) + expect(fixtureUserPrompts(await readFile(FIXTURE, 'utf8'))).toEqual([PROMPT]) + const input = page.locator('textarea').first() + await input.waitFor({ timeout: 10_000 }) + const settled = scaffold.whenTurnSettled(30_000) + await input.fill(PROMPT) + await input.press('Enter') + await page.getByRole('button', { name: 'Stop generating' }).waitFor({ timeout: 10_000 }) + + await input.fill(STEER) + await input.press('Meta+Enter') + await expect.poll(() => input.inputValue(), { timeout: 5_000 }).toBe('') + expect(await page.locator('[data-queue-dock]').count()).toBe(0) + + const composer = page.locator('[data-question-key]') + await composer.waitFor({ timeout: 30_000 }) + const pendingSteering = page.locator('[data-pending-steering]').filter({ hasText: STEER }) + await pendingSteering.waitFor({ timeout: 10_000 }) + await composer.getByRole('radio', { name: 'Yes' }).click() + await composer.getByRole('radio', { name: 'Yes' }).press('Enter') + await settled + + const steerEvents = sessionEvents.filter(event => event.type === 'steering/message') + expect(steerEvents).toHaveLength(1) + expect((steerEvents[0] as SessionEvent & { data: { turn: number } }).data.turn).toBe(1) + await expect.poll(() => page.getByText(STEER, { exact: true }).count(), { timeout: 15_000 }).toBe(1) + expect(await pendingSteering.count()).toBe(0) + await expect.poll(() => page.getByText('BANANA', { exact: false }).count(), { timeout: 10_000 }) + .toBeGreaterThanOrEqual(2) + expect(tripwire.pageErrors).toEqual([]) + expect(tripwire.warnings).toEqual([]) + }, 90_000) +}) + +describe('web e2e: composer shortcut follows the swapped busy behavior', () => { + let scaffold: WebScaffold + let browser: Browser + let page: Page + let tripwire: ReturnType + const sessionEvents: SessionEvent[] = [] + + beforeAll(async () => { + scaffold = await launchWebScaffold({ replayFixture: FIXTURE, paceMs: 100 }) + scaffold.ctx.on('session/event', (_session, event) => { sessionEvents.push(event) }) + browser = await chromium.launch() + page = await newEnglishPage(browser) + tripwire = watchConsole(page) + await page.goto(scaffold.baseUrl, { waitUntil: 'load' }) + await page.waitForSelector('[class*="frame"]', { timeout: 30_000 }) + await connectFreshWorkspace(page, scaffold.workspaceCwd) + }, 120_000) + + afterAll(async () => { + await browser?.close() + await scaffold?.close() + }) + + it.skipIf(MODE === 'record')('queues Cmd+Enter when plain Enter is configured to Steer', async () => { + onTestFailed(() => saveFailureShot(page, 'web-e2e-composer-swapped-shortcut')) + await page.getByRole('button', { name: 'Settings', exact: true }).click() + const dialog = page.getByRole('dialog', { name: 'Settings' }) + await dialog.getByRole('button', { name: 'Queue' }).click() + await page.getByRole('menuitem', { name: 'Steer' }).click() + await dialog.getByRole('button', { name: 'Steer' }).waitFor({ timeout: 10_000 }) + await page.keyboard.press('Escape') + + const input = page.locator('textarea').first() + const settled = scaffold.whenTurnSettled(30_000) + await input.fill(PROMPT) + await input.press('Enter') + await page.getByRole('button', { name: 'Stop generating' }).waitFor({ timeout: 10_000 }) + + const queuedText = 'Queued by the complementary Cmd+Enter shortcut.' + await input.fill(queuedText) + await input.press('Meta+Enter') + const queuedRow = page.locator('[data-queue-dock]').getByRole('listitem').filter({ hasText: queuedText }) + await queuedRow.getByText(queuedText, { exact: true }).waitFor({ timeout: 10_000 }) + expect(await page.locator('[data-pending-steering]').filter({ hasText: queuedText }).count()).toBe(0) + expect(sessionEvents.filter(event => event.type === 'steering/message')).toHaveLength(0) + + // Remove the asserted Queue row, then finish the recorded question turn + // so replay teardown still proves that every fixture call was consumed. + await queuedRow.getByRole('button', { name: 'Remove queued message' }).click() + const composer = page.locator('[data-question-key]') + await composer.waitFor({ timeout: 30_000 }) + await composer.getByRole('radio', { name: 'Yes' }).click() + await composer.getByRole('radio', { name: 'Yes' }).press('Enter') + await settled + expect(tripwire.pageErrors).toEqual([]) + expect(tripwire.warnings).toEqual([]) + }, 90_000) +}) diff --git a/packages/client/runtime/src/client/session-history/history-fold.ts b/packages/client/runtime/src/client/session-history/history-fold.ts index e7f9922425..c4bc6ed9b5 100644 --- a/packages/client/runtime/src/client/session-history/history-fold.ts +++ b/packages/client/runtime/src/client/session-history/history-fold.ts @@ -146,7 +146,8 @@ function materializeNode( } case 'steering/message': return { - kind: 'steering', seq: event.seq, time: event.time, turn: event.data.turn, + kind: 'steering', messageId: event.data.message.id, + seq: event.seq, time: event.time, turn: event.data.turn, content: event.data.message.content, source: event.data.message.source, } case 'tool/result': { diff --git a/packages/client/runtime/src/client/sessions/conversation.ts b/packages/client/runtime/src/client/sessions/conversation.ts index a4c260a068..31872356c8 100644 --- a/packages/client/runtime/src/client/sessions/conversation.ts +++ b/packages/client/runtime/src/client/sessions/conversation.ts @@ -4,6 +4,7 @@ // string here (narrow to real brands when convenient). import type { CommandId } from '@deepseek-ai/dsh-commands/brand' +import type { MessageId } from '@deepseek-ai/dsh-llm/brand' import type { ContentBlock } from '@deepseek-ai/dsh-llm/types' import type { LlmRetryEventData } from '@deepseek-ai/dsh-llm-retry/types' import type { TodoItem } from '@deepseek-ai/dsh-session/types' @@ -104,6 +105,8 @@ export interface AssistantMessageNode { /** A steering message injected mid-turn. */ export interface SteeringMessageNode { kind: 'steering' + /** Stable identity shared with its pre-admission inbox occurrence. */ + messageId: MessageId seq: number /** Unix epoch ms from the source session event. */ time: number @@ -271,9 +274,15 @@ export interface RunningToolCall { } -/** One independently addressable row from the transient queue snapshot. */ +/** One transient inbox occurrence from the authoritative `session/queue` snapshot. */ export interface QueuedMessage { readonly id: InboxItemId + /** Stable message identity used for transient-to-durable steering handoff. */ + readonly messageId: MessageId + /** Agent-resolved placement; only queued rows accept queue mutations. */ + readonly placement: 'queued' | 'steering' + /** Complete content used to render pending steering before it becomes durable. */ + readonly content: readonly ContentBlock[] readonly preview: string /** Complete editable text; null when the message contains non-text blocks. */ readonly text: string | null @@ -332,7 +341,7 @@ export interface ConversationSnapshot { */ codeDispatches: ReadonlyMap pending: readonly PendingInteraction[] - /** Authoritative transient inbox snapshot, replaced after every host-side change. */ + /** Authoritative transient inbox snapshot, including queued and steering placements. */ queue: readonly QueuedMessage[] running: boolean /** diff --git a/packages/client/runtime/src/client/sessions/session.ts b/packages/client/runtime/src/client/sessions/session.ts index 2bcb2ac5fc..490693a6a4 100644 --- a/packages/client/runtime/src/client/sessions/session.ts +++ b/packages/client/runtime/src/client/sessions/session.ts @@ -446,6 +446,9 @@ export class Session implements SessionFace { case 'session/queue': { this.queued = frame.items.map(item => ({ id: item.id, + messageId: item.message.id, + placement: item.placement, + content: item.message.content, preview: queuePreviewOf(item.message.content), text: queueTextOf(item.message.content), })) diff --git a/packages/client/runtime/src/client/sessions/transcript-adapter.ts b/packages/client/runtime/src/client/sessions/transcript-adapter.ts index 6fae44ede4..05ac3cb067 100644 --- a/packages/client/runtime/src/client/sessions/transcript-adapter.ts +++ b/packages/client/runtime/src/client/sessions/transcript-adapter.ts @@ -74,7 +74,8 @@ function materializeNode( } case 'steering/message': return { - kind: 'steering', seq: event.seq, time: event.time, turn: event.data.turn, + kind: 'steering', messageId: event.data.message.id, + seq: event.seq, time: event.time, turn: event.data.turn, content: event.data.message.content, source: event.data.message.source, } case 'tool/result': { diff --git a/packages/client/runtime/tests/queue-store.spec.ts b/packages/client/runtime/tests/queue-store.spec.ts index b594e9a9df..cc2f3d0e53 100644 --- a/packages/client/runtime/tests/queue-store.spec.ts +++ b/packages/client/runtime/tests/queue-store.spec.ts @@ -22,6 +22,7 @@ interface QueueFixture { id: string body: string content?: ContentBlock[] + placement?: 'queued' | 'steering' } /** Build one authoritative queue snapshot. */ @@ -31,6 +32,7 @@ function queueFrame(items: QueueFixture[]): MuxFrame { sessionId: SID, items: items.map(item => ({ id: iid(item.id), + placement: item.placement ?? 'queued', message: createUserMessage({ content: item.content ?? text(item.body), source: { kind: 'user', rpcId: rid(`rpc-${item.id}`) } as never, @@ -49,8 +51,14 @@ describe('queue snapshot intake', () => { session.handleMuxEnvelope(rid('env-1'), queueFrame([ { id: 'q-1', body: '第一条 排队\n消息' }, ])) - expect(session.getSnapshot().queue).toEqual([ - { id: 'q-1', preview: '第一条 排队 消息', text: '第一条 排队\n消息' }, + const queue = session.getSnapshot().queue + expect(typeof queue[0]?.messageId).toBe('string') + expect(queue).toMatchObject([ + { + id: 'q-1', placement: 'queued', + content: [{ type: 'text', text: '第一条 排队\n消息' }], + preview: '第一条 排队 消息', text: '第一条 排队\n消息', + }, ]) }) @@ -61,8 +69,14 @@ describe('queue snapshot intake', () => { body: '', content: [{ type: 'text', text: 'hi' }, { type: 'image', data: 'x' } as never], }])) - expect(session.getSnapshot().queue).toEqual([ - { id: 'q-image', preview: 'hi [image]', text: null }, + const queue = session.getSnapshot().queue + expect(typeof queue[0]?.messageId).toBe('string') + expect(queue).toMatchObject([ + { + id: 'q-image', placement: 'queued', + content: [{ type: 'text', text: 'hi' }, { type: 'image', data: 'x' }], + preview: 'hi [image]', text: null, + }, ]) }) @@ -85,8 +99,14 @@ describe('queue snapshot intake', () => { session.handleMuxEnvelope(rid('env-5'), queueFrame([ { id: 'q-2', body: 'two edited' }, ])) - expect(session.getSnapshot().queue).toEqual([ - { id: 'q-2', preview: 'two edited', text: 'two edited' }, + const queue = session.getSnapshot().queue + expect(typeof queue[0]?.messageId).toBe('string') + expect(queue).toMatchObject([ + { + id: 'q-2', placement: 'queued', + content: [{ type: 'text', text: 'two edited' }], + preview: 'two edited', text: 'two edited', + }, ]) session.handleMuxEnvelope(rid('env-6'), queueFrame([])) expect(session.getSnapshot().queue).toEqual([]) @@ -99,6 +119,21 @@ describe('queue snapshot intake', () => { session.handleAgentError('unrelated') expect(session.getSnapshot().queue).toBe(before) }) + + it('retains steering placement and complete content in the same authoritative snapshot', () => { + const session = makeSession() + session.handleMuxEnvelope(rid('env-steering'), queueFrame([ + { id: 'q-next', body: 'later' }, + { id: 's-now', body: 'interrupt now', placement: 'steering' }, + ])) + + expect(session.getSnapshot().queue.map(item => ({ + id: item.id, placement: item.placement, content: item.content, + }))).toEqual([ + { id: 'q-next', placement: 'queued', content: text('later') }, + { id: 's-now', placement: 'steering', content: text('interrupt now') }, + ]) + }) }) describe('queue operation transport', () => { diff --git a/packages/client/ui-conversation/README.i18n.yaml b/packages/client/ui-conversation/README.i18n.yaml index 1a8fc3a22a..f3f3c43124 100644 --- a/packages/client/ui-conversation/README.i18n.yaml +++ b/packages/client/ui-conversation/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/client/ui-conversation/README.md -README.md: 2bf74454fab14303f305b9822e1032f67de0d3ae -README.zh.md: 2677ab5cc85c16fffeca688a4648caf2410ba895 +README.md: c80ca786fb289055fcb3ad2b5a7aab64127ca485 +README.zh.md: 601139a00d2d40fb34e0cdcb8554984f7db9a724 diff --git a/packages/client/ui-conversation/README.md b/packages/client/ui-conversation/README.md index 2bf74454fa..c80ca786fb 100644 --- a/packages/client/ui-conversation/README.md +++ b/packages/client/ui-conversation/README.md @@ -34,7 +34,11 @@ Tool rows are slots too — the standalone tool ring (`ToolViewRegistry`/`ctx.to The todo surfaces are two registrations over that shape, both plain registrant plugins with `inject: ['slots', 'conversation']`. `TodoRow` takes the `'conversation.chat.toolview'` key `todo_write` and summarizes what the call attempted (`/ 已完成 · ` parsed from its args, falling back to the generic summary on malformed or wrongly-shaped model JSON, and keeping the generic dot for non-ok execution states so a cancelled call never reads as a completed update). `TodoDock` takes the `'conversation.input.dock'` list slot at `order: 0` — before Goal and Queue — and is the plan strip: it reads the host-computed `todos` projection via `useProjection` (standing plan: latest `todo/write` with no later `turn/start`) and renders `TodoPanel`, which takes the plain list, hides itself while the list is empty, and starts collapsed as a header of title plus `"/ tasks · in progress"` (status glyphs are the figma check / progress / dashed-pending set). The dock adapter owns the selection so the panel stays a pure function of its props; the standing list lives here rather than in the row so the row stays one line. Anything the input-zone composer chain hides (a `conversation.composer` takeover such as ui-question's) hides the whole dock, this strip included. -`QueueDock` is the terminal input-dock entry at `order: 20`. It hides while empty, renders one pending row directly, and defaults two or more rows to a collapsed `" 条排队消息"` header whose button expands or collapses the complete list. The header exposes `aria-expanded` and `aria-controls`; the expanded list scrolls within a 180px height bound. An active edit or mutation keeps its rows visible, and emptying the queue restores the collapsed default for the next queue. Each visible row remains a single-line preview with its exact-occurrence edit and delete actions. +`QueueDock` is the terminal input-dock entry at `order: 20`. It hides while empty, renders one pending row directly, and defaults two or more rows to a collapsed `" 条排队消息"` header whose button expands or collapses the complete list. The header exposes `aria-expanded` and `aria-controls`; the expanded list scrolls within a 180px height bound. An active edit or mutation keeps its rows visible, and emptying the queue restores the collapsed default for the next queue. Each visible row remains a single-line preview with its exact-occurrence edit, delete, and strict-steer actions; an unavailable steering window leaves the Queue occurrence in place and reports the failure. + +The Host's placement-aware `session/queue` snapshot also carries pending steering. QueueDock filters it out, while ChatView projects it as a user-style bubble at the conversation tail. The Host delays steering retirement until the durable `steering/message` has entered the mux stream, and ChatView deduplicates the two projections by their shared `MessageId`; the bubble therefore hands off without a gap or duplicate, while reconnect restores pending state from the same authority. + +Keyboard message submission resolves delivery from the addressed session's running state. While idle, Enter and Cmd/Ctrl+Enter both perform an ordinary Queue send. While running, the browser-persisted General Settings preference assigns plain Enter to `Queue` (the default) or `Steer`, and Cmd/Ctrl+Enter performs the other behavior; Shift+Enter remains a newline. The preference affects only the busy-state gesture pair, and the send button and non-keyboard submit actions remain Queue. Composer Steer uses the existing best-effort `session.prompt(mode: 'steer')` contract: if the current next-step window closes before acceptance, AgentLoop admits the message as the next waking Queue turn without surfacing a failure or losing the draft transaction. Per-session UI state for selection and the active view lives in the declared chat store (`stores.ts` `createChatStore`); the InputHub owns the composer state machine and mirrors its draft into that store for persistence. Apply passes one store handle to the strict session subtree, chat view, and details registrations, so each session shares one instance and the framework owns its lifecycle. Components are pure: the framework standard kit supplies `useSession`/`sessionId`, global `useSessions`/`useWorkspaces`, and the input machine's `useInput`/`inputActions`; store faces and inject factories supply the remaining state and callbacks. @@ -63,4 +67,4 @@ None; this package neither assembles nor sends a provider request. - **The approval panel's "Always allow this type" is deferred** — durable grants need a grant-storage design; only allow-once/reject answer today. - **TodoPanel truncates long item text to one ellipsized line** — the figma strip has no wrap or expand affordance; full text is not readable inline. - **Queue edit is text-only** — rows containing non-text blocks still show a flattened preview, but their edit control is disabled because the inline editor cannot preserve those blocks. A text row's edit mode replaces delete and strict steer with save and cancel; Enter saves and Escape cancels. -- **Queue strict steer preserves complete messages** — while the Agent is running, the steer action atomically transfers the addressed Queue occurrence into the current next-step window. Mixed-content rows remain eligible because the action forwards the immutable message instead of the text projection. The Host omits pending steering from the Queue snapshot; a consumed `steering/message` still folds into the durable transcript as a plain bubble so replay remains truthful. +- **Queue strict steer preserves complete messages** — while the Agent is running, the steer action atomically transfers the addressed Queue occurrence into the current next-step window. Mixed-content rows remain eligible because the action forwards the immutable message instead of the text projection. The placement-aware Host snapshot renders pending steering at the conversation tail until the consumed `steering/message` folds into the durable transcript, so immediate display, reconnect, and replay share one linear authority. diff --git a/packages/client/ui-conversation/README.zh.md b/packages/client/ui-conversation/README.zh.md index 2677ab5cc8..601139a00d 100644 --- a/packages/client/ui-conversation/README.zh.md +++ b/packages/client/ui-conversation/README.zh.md @@ -34,7 +34,11 @@ todo 两个面就是在该形状上的两个注册项,都是普通注册方插件,`inject: ['slots', 'conversation']`。`TodoRow` 占用 `'conversation.chat.toolview'` 的 `todo_write` key,摘要该次调用「试图写入」的内容(从其 args 解析出 `<已完成>/<总数> 已完成 · <进行中条目>`;模型 JSON 残缺或形状不对时回落到通用摘要;非 ok 执行状态保留通用状态点,使被取消的调用绝不读成一次已完成的更新)。`TodoDock` 以 `order: 0` 占用 `'conversation.input.dock'` 列表 slot(位于 Goal 与 Queue 之前),是计划条:它经 `useProjection` 读取 host 计算的 `todos` 投影(站立计划:其后没有更晚 `turn/start` 的最近一次 `todo/write`)并渲染 `TodoPanel`,后者接收纯列表,在列表为空时自我隐藏;列表非空时面板初始折叠,表头显示标题加 `"<已完成>/<总数> tasks · in progress"`(状态图标为 figma 的勾选/进行中/虚线未开始一组)。选取由 dock 适配器负责,因此面板保持为其 props 的纯函数;站立列表放在此处而非行内,行才能保持单行。输入区 composer 链隐藏的一切(例如 ui-question 对 `conversation.composer` 的接管)也会隐藏整个 dock,包括这条计划条。 -`QueueDock` 是 `order: 20` 的末端 input-dock 条目。队列为空时隐藏;只有一个待处理项时直接渲染该行;存在两个或更多待处理项时,默认收起为 `" 条排队消息"` 表头,其按钮可展开或收起完整列表。表头暴露 `aria-expanded` 和 `aria-controls`;展开后的列表以 180px 为高度上限,并可滚动。存在进行中的编辑或变更时,列表行会保持可见;队列清空后,下一次出现队列时会恢复默认收起状态。每条可见行仍是单行预览,并提供针对精确单次入队项的编辑和删除操作。 +`QueueDock` 是 `order: 20` 的末端 input-dock 条目。队列为空时隐藏;只有一个待处理项时直接渲染该行;存在两个或更多待处理项时,默认收起为 `" 条排队消息"` 表头,其按钮可展开或收起完整列表。表头暴露 `aria-expanded` 和 `aria-controls`;展开后的列表以 180px 为高度上限,并可滚动。存在进行中的编辑或变更时,列表行会保持可见;队列清空后,下一次出现队列时会恢复默认收起状态。每条可见行仍是单行预览,并提供针对精确单次入队项的编辑、删除和严格 steering(中途引导)操作;steering 窗口不可用时,Queue 单次入队项会留在原处并显示失败。 + +Host 带 placement 的 `session/queue` 快照也会携带待处理 steering。QueueDock 会将其过滤掉,ChatView 则把它投影为会话流末尾的用户样式气泡。Host 会等持久 `steering/message` 进入 mux 流之后再退役 steering,ChatView 则按两份投影共享的 `MessageId` 去重;气泡交接时因而不会产生空档或重复,重连也能从同一权威恢复待处理状态。 + +键盘消息提交会根据所寻址会话的运行状态解析投递方式。空闲时,Enter 和 Cmd/Ctrl+Enter 都执行普通 Queue 发送。运行期间,浏览器持久化的 General Settings 偏好会把普通 Enter 分配为 `Queue`(默认值)或 `Steer`,Cmd/Ctrl+Enter 则执行另一种行为;Shift+Enter 仍然换行。该偏好只影响繁忙态下这对手势,发送按钮与非键盘提交操作仍使用 Queue。Composer Steer 复用现有尽力而为的 `session.prompt(mode: 'steer')` 契约:如果当前 next-step 窗口在接纳前关闭,AgentLoop 会把消息接纳为下一条唤醒 Queue 轮次,不显示失败,也不会丢失草稿事务。 逐 Session UI 状态中的选择与活跃视图位于已声明的聊天 store(`stores.ts` `createChatStore`)中;InputHub 拥有输入区状态机,并将草稿镜像到该 store 以便持久化。apply 将同一个 store handle 传给严格限定于会话的子树、聊天视图和详情注册,因此每个会话内共享一个实例,框架拥有其生命周期。组件保持纯粹:框架标准工具包提供 `useSession`/`sessionId`、全局 `useSessions`/`useWorkspaces`,以及输入状态机的 `useInput`/`inputActions`;store 表层与 inject factory 提供其余状态和回调。 @@ -63,4 +67,4 @@ todo 两个面就是在该形状上的两个注册项,都是普通注册方插 - **审批面板的「始终允许此类」暂缓**:持久授权需要授权存储设计;今天只能回答允许一次/拒绝。 - **TodoPanel 将过长条目截成单行省略号**:figma 条没有换行或展开入口,完整文本无法在行内读完。 - **Queue 编辑仅支持文本**:包含非文本块的行仍显示扁平化预览,但由于内联编辑器无法保留这些块,其编辑控件会被禁用。文本行进入编辑模式后,删除和严格 steering(中途引导)操作会被保存和取消取代;Enter 保存,Escape 取消。 -- **Queue 严格 steering 会保留完整消息**:Agent 运行期间,steering 操作会以原子方式把所寻址的 Queue 单次入队项转移到当前 next-step 窗口。包含混合内容的行仍可使用此操作,因为它会转发不可变消息,而非文本投影。Host 不会把待处理 steering 纳入 Queue 快照;已消费的 `steering/message` 仍会折叠进持久 transcript(文本记录),并以普通气泡呈现,因此回放仍然如实。 +- **Queue 严格 steering 会保留完整消息**:Agent 运行期间,steering 操作会以原子方式把所寻址的 Queue 单次入队项转移到当前 next-step 窗口。包含混合内容的行仍可使用此操作,因为它会转发不可变消息,而非文本投影。带 placement 的 Host 快照会在会话流末尾渲染待处理 steering,直到已消费的 `steering/message` 折叠进持久 transcript(文本记录),因此立即展示、重连和回放共享同一个线性权威。 diff --git a/packages/client/ui-conversation/src/client/apply.ts b/packages/client/ui-conversation/src/client/apply.ts index 625728d4cf..bb34164a3f 100644 --- a/packages/client/ui-conversation/src/client/apply.ts +++ b/packages/client/ui-conversation/src/client/apply.ts @@ -1,6 +1,6 @@ /** Registers the conversation components, shared store, and service callbacks. */ import type { Context } from 'cordis' -import { resolveSlotLabel, type BoundActions } from '@deepseek-ai/dsh-client-ui-slots' +import { deferRegistration, resolveSlotLabel, type BoundActions } from '@deepseek-ai/dsh-client-ui-slots' import type { ISessions, SessionId } from '@deepseek-ai/dsh-client-runtime/client' import type {} from '@deepseek-ai/dsh-client-ui-layout/client' // Type-only: pulls the locale plugin's Context merge (ctx.locale). @@ -16,7 +16,10 @@ import { createChatStore } from './stores.ts' import { ConversationService } from './service.ts' import type { IConversation } from './service.ts' import { InputHub } from './input/hub.ts' +import { ComposerSubmissionPolicy } from './input/submission-policy.ts' import { InputBar } from './skeleton/InputBar.tsx' +import { EnterBehaviorRow } from './settings/EnterBehaviorRow.tsx' +import type { EnterBehaviorRowInjected } from './settings/EnterBehaviorRow.tsx' import { ChatView } from './chat/ChatView.tsx' import { StatsLine } from './chat/StatsLine.tsx' import { bashToolviewSample } from './toolviews/bash-sample.tsx' @@ -93,6 +96,22 @@ export function apply(ctx: Context): void { // Apply-time construction keeps store identity bound to this fiber. const chatStore = createChatStore() + const submissionPolicy = new ComposerSubmissionPolicy() + + ctx.effect(() => { + const row = deferRegistration(ctx.slots, 'settings.general.item', EnterBehaviorRow, () => + ctx.slots.register({ + name: 'settings.general.item', + id: 'composer-enter', + order: 20, + locale: NS, + inject: (): EnterBehaviorRowInjected => ({ + hooks: { busyEnter: submissionPolicy.busyEnter }, + setBusyEnter: (behavior) => { submissionPolicy.setBusyEnter(behavior) }, + }), + }, EnterBehaviorRow)) + return () => { row.dispose() } + }, 'ui-conversation: Enter behavior settings row') // Chat scroll offsets by session, surviving view switches (the chat view // unmounts under the tab ring). Deliberately not persisted: a fresh page @@ -203,6 +222,7 @@ export function apply(ctx: Context): void { if (sessionId === undefined) { return { keyboard: undefined, + resolveSubmitMode: (running, gesture) => submissionPolicy.resolve(running, gesture), toggleCommandMenu: undefined, stop: undefined, command: undefined, @@ -213,6 +233,7 @@ export function apply(ctx: Context): void { const slash = inputHub.slash(sessionId) return { keyboard: shell, + resolveSubmitMode: (running, gesture) => submissionPolicy.resolve(running, gesture), toggleCommandMenu: slash === undefined ? undefined : (selection) => { diff --git a/packages/client/ui-conversation/src/client/chat/ChatView.tsx b/packages/client/ui-conversation/src/client/chat/ChatView.tsx index 4a8e715bcf..4f237c2d0c 100644 --- a/packages/client/ui-conversation/src/client/chat/ChatView.tsx +++ b/packages/client/ui-conversation/src/client/chat/ChatView.tsx @@ -34,7 +34,7 @@ import { assistantActionsSeqs, deriveChatFlow, type ChatFlowItem } from './chat- import { AssistantMarkdown } from './AssistantMarkdown.tsx' import { GenericCommandCard } from './GenericCommandCard.tsx' import { GenericToolCard } from './GenericToolCard.tsx' -import { MessageItem } from './MessageItem.tsx' +import { MessageItem, PendingSteeringBubble } from './MessageItem.tsx' import css from './ChatView.module.css' const FOLLOW_THRESHOLD = 24 @@ -236,6 +236,7 @@ export function ChatView({ useSession, useSessions, useStore, renderSlot, sessionId, openFile, loadOlder, inspectCall, chatScroll, forkAt, t, }: ChatViewSlotProps) { const nodes = useSession(s => s.nodes) + const inbox = useSession(s => s.queue) // Workspace root off the session list row: path summaries display relative to it. const cwd = useSessions(s => s.byId[sessionId]?.cwd) const running = useSession(s => s.running) @@ -248,6 +249,10 @@ export function ChatView({ const selectedCallId = useStore(s => s.selection?.callId) const items = useMemo(() => deriveChatFlow(nodes), [nodes]) + const pendingSteering = useMemo(() => { + const durable = new Set(nodes.flatMap(node => node.kind === 'steering' ? [node.messageId] : [])) + return inbox.filter(item => item.placement === 'steering' && !durable.has(item.messageId)) + }, [inbox, nodes]) const activeRetry = useMemo(() => activeRetrySeq(nodes, running), [nodes, running]) // Only the last content assistant of each turn owns IconActions; mid-turn // text (before tools) omits `time` so AssistantMarkdown stays chrome-free. @@ -261,6 +266,7 @@ export function ChatView({ const firstSeqRef = useRef(null) const openedRef = useRef(false) const lastKeyRef = useRef(null) + const lastSteeringIdRef = useRef(null) /** Flow tip signature — follow-scroll only when this moves, never on a * scroll-driven at-bottom chrome re-render (that was snapping inertial * scrolls the rest of the way to the floor). */ @@ -269,7 +275,8 @@ export function ChatView({ const firstSeq = nodes[0]?.seq ?? null const lastItem = items[items.length - 1] const lastKey = lastItem?.key ?? null - const followSig = `${openState}:${firstSeq}:${lastKey}:${nodes.length}:${running ? 1 : 0}:${runningCalls.length}` + const lastSteeringId = pendingSteering[pendingSteering.length - 1]?.id ?? null + const followSig = `${openState}:${firstSeq}:${lastKey}:${nodes.length}:${running ? 1 : 0}:${runningCalls.length}:${lastSteeringId ?? ''}` const toBottom = (el: HTMLElement): void => { el.scrollTop = el.scrollHeight @@ -298,6 +305,7 @@ export function ChatView({ } firstSeqRef.current = firstSeq lastKeyRef.current = lastKey + lastSteeringIdRef.current = lastSteeringId followSigRef.current = followSig return } @@ -308,6 +316,7 @@ export function ChatView({ firstSeqRef.current = firstSeq /* v8 ignore next -- ?? arm: a prepend adds nodes, so the flow list here is never empty. */ lastKeyRef.current = lastKey + lastSteeringIdRef.current = lastSteeringId followSigRef.current = followSig return } @@ -316,12 +325,14 @@ export function ChatView({ // (send lives in the composer, so arrival is detected here, not armed there). const appendedUser = lastKey !== lastKeyRef.current && lastItem !== undefined && lastItem.kind === 'node' && lastItem.node.kind === 'user' + const appendedSteering = lastSteeringId !== null && lastSteeringId !== lastSteeringIdRef.current const tipMoved = followSigRef.current !== followSig lastKeyRef.current = lastKey + lastSteeringIdRef.current = lastSteeringId followSigRef.current = followSig // Follow new flow content while pinned; do NOT re-pin on every render // merely because atBottomRef is true (scroll threshold → setState → snap). - if (appendedUser || (tipMoved && atBottomRef.current)) toBottom(el) + if (appendedUser || appendedSteering || (tipMoved && atBottomRef.current)) toBottom(el) }) const onScrollRef = useRef(() => {}) @@ -467,6 +478,9 @@ export function ChatView({ {/* Turn-level loading signal: rides the whole running turn (first-token wait, tool execution, streaming) so it never flickers per step. */} {running && } + {pendingSteering.map(item => ( + + ))} {!atBottom && (
diff --git a/packages/client/ui-conversation/src/client/chat/MessageItem.tsx b/packages/client/ui-conversation/src/client/chat/MessageItem.tsx index b54604060a..5783d0211d 100644 --- a/packages/client/ui-conversation/src/client/chat/MessageItem.tsx +++ b/packages/client/ui-conversation/src/client/chat/MessageItem.tsx @@ -169,17 +169,19 @@ function projectUserText(text: string): ReactNode { /** Right-aligned bubble shared by user and steering rows (steering has no actions). */ function UserStyleBubble({ - content, actions, t, + content, actions, pending = false, t, }: { content: readonly unknown[] /** Optional IconActions (or similar) below the bubble; receives the joined text. */ actions?: (text: string) => ReactNode + /** Whether this is the Host-authoritative pre-admission steering projection. */ + pending?: boolean t: ChatViewSlotProps['t'] }): ReactNode { const { text, rest } = contentText(content) const truncated = (total: number): string => t('json.truncated', { total }) return ( -
+
{projectUserText(text)} {rest.map((block, i) => )} @@ -189,6 +191,19 @@ function UserStyleBubble({ ) } +/** + * Render one Host-authoritative pending steering item with the same visual + * language as its eventual durable transcript node. + * @param props - Pending message content and conversation translator. + * @returns the pending steering bubble. + */ +export function PendingSteeringBubble({ content, t }: { + content: readonly unknown[] + t: ChatViewSlotProps['t'] +}): ReactNode { + return +} + export const MessageItem = memo(function MessageItem({ node, retryActive = false, onFork, t, }: MessageItemProps) { diff --git a/packages/client/ui-conversation/src/client/contract/composer-submission.ts b/packages/client/ui-conversation/src/client/contract/composer-submission.ts new file mode 100644 index 0000000000..c5bcdc7826 --- /dev/null +++ b/packages/client/ui-conversation/src/client/contract/composer-submission.ts @@ -0,0 +1,10 @@ +/** Composer submission vocabulary shared by the input and settings domains. */ + +/** Delivery mode requested for one ordinary composer message. */ +export type InputSubmitMode = 'queue' | 'steer' + +/** Configurable meaning of plain Enter while the addressed agent is busy. */ +export type BusyEnterBehavior = InputSubmitMode + +/** Keyboard gesture whose delivery mode the submission policy resolves. */ +export type ComposerSubmitGesture = 'enter' | 'accelerated' diff --git a/packages/client/ui-conversation/src/client/contract/slots.ts b/packages/client/ui-conversation/src/client/contract/slots.ts index a6c9d16e61..16f789f70c 100644 --- a/packages/client/ui-conversation/src/client/contract/slots.ts +++ b/packages/client/ui-conversation/src/client/contract/slots.ts @@ -7,6 +7,7 @@ import type { CommandNode, ConversationSnapshot, ObservableSnapshot, PendingInte import type {} from '@deepseek-ai/dsh-client-ui-layout/client' import type { ComposerKeyboard, EditSelection, InputActions, InputNotice, InputState } from '../input/contract.ts' import type { createChatStore } from '../stores.ts' +import type { ComposerSubmitGesture, InputSubmitMode } from './composer-submission.ts' import type { CallId, SelectionTarget, ViewTab } from './views.ts' declare module '@deepseek-ai/dsh-client-ui-slots' { @@ -285,6 +286,8 @@ export interface ComposerBarOwnerProps { export interface ComposerBarInjected { /** The InputBar-exclusive keyboard/DOM command face (decision 20 private plane); absent with the session. */ keyboard: ComposerKeyboard | undefined + /** Resolve one keyboard submission gesture against the current running state and persisted preference. */ + resolveSubmitMode: (running: boolean, gesture: ComposerSubmitGesture) => InputSubmitMode /** Toggle the shared slash menu with only its command source; absent without ui-slash or a session. */ toggleCommandMenu: ((selection: EditSelection) => void) | undefined /** Cancel the in-flight turn; absent with the session. */ diff --git a/packages/client/ui-conversation/src/client/input/contract.ts b/packages/client/ui-conversation/src/client/input/contract.ts index f503381bbb..6af78bdb72 100644 --- a/packages/client/ui-conversation/src/client/input/contract.ts +++ b/packages/client/ui-conversation/src/client/input/contract.ts @@ -11,6 +11,7 @@ import type { ReferenceInsert, SubmitOutcome, TokenSpan, } from '@deepseek-ai/dsh-client-ui-slash/client' import type { QueueRow } from '../contract/queue.ts' +import type { InputSubmitMode } from '../contract/composer-submission.ts' /** * The scoped-event application verbs: the hub's bail listeners call these, @@ -28,8 +29,11 @@ export interface InputTarget { export interface SessionInput extends InputTarget { /** Single write path for draft text (all mutation rides machine events). */ setDraft(text: string): void - /** THE complexity sink: enter adjudication, submit transaction, and the default sink live inside. */ - submit(): void + /** + * THE complexity sink: enter adjudication, submit transaction, and the default sink live inside. + * @param mode - delivery intent retained through asynchronous adjudication and serialization. + */ + submit(mode?: InputSubmitMode): void /** * Surface a notice outside the machine's own effect stream: detached * command results and business notifications render through here. @@ -82,8 +86,8 @@ export interface ComposerKeyboard { readonly snapshot: InputState /** Draft write with the DOM-observed edit shape (narrows occurrence math). */ setDraft(text: string, editRange?: EditRange): void - /** Newline at the selection as a machine transaction (Ctrl+Enter path). */ - newline(selection: EditSelection): void + /** Submit with an explicit delivery mode resolved by the keyboard policy. */ + submit(mode: InputSubmitMode): void undo(): void redo(): void /** Paste over the selection (sync components ride the same transaction). */ @@ -191,7 +195,7 @@ export interface InputState { readonly occurrences: readonly Occurrence[] /** Live paste-match attempt (absent when no paste is matchable). */ readonly paste?: PasteAttemptState - /** Read-only queue projection (session/queued frames + connect snapshot). */ + /** Read-only transient inbox projection (`session/queue`, including pending steering). */ readonly queue: readonly QueuedMessage[] } @@ -206,6 +210,8 @@ export interface SubmitAttempt { readonly signal: AbortSignal /** Draft at enter time; rollback restores it only while the live draft still equals it. */ readonly draftSnapshot: string + /** Default-message delivery intent retained while slash adjudication is pending. */ + readonly mode: InputSubmitMode } /** @@ -217,8 +223,6 @@ export interface SubmitAttempt { export type InputEvent = /** Full next draft from the textarea; editRange narrows the occurrence math (absent → diff scan). */ | { readonly type: 'draft-changed'; readonly draft: string; readonly editRange?: EditRange } - /** Insert '\n' replacing the selection (F1: the execCommand newline path moved into the machine). */ - | { readonly type: 'newline'; readonly selection: EditSelection } | { readonly type: 'begin-command'; readonly claim: CommandClaim; readonly span: TokenSpan } /** Place one U+FFFC at the span and mint the occurrence (scoped insert-reference event payload). */ | { readonly type: 'insert-ref'; readonly reference: ReferenceInsert; readonly span: TokenSpan } @@ -239,7 +243,7 @@ export type InputEvent = | { readonly type: 'paste-upgrade'; readonly attemptId: number; readonly span: TokenSpan; readonly reference: ReferenceInsert } /** Shell-observed attempt killers the machine cannot see itself (caret/selection ops, Slash interaction updates). */ | { readonly type: 'invalidate-paste' } - | { readonly type: 'enter' } + | { readonly type: 'enter'; readonly mode: InputSubmitMode } | { readonly type: 'adjudicated'; readonly attempt: SubmitAttempt; readonly outcome: PickOutcome } | { readonly type: 'adjudication-failed'; readonly attempt: SubmitAttempt; readonly message: string } | { readonly type: 'submit-settled'; readonly attempt: SubmitAttempt; readonly ok: boolean; readonly outcome?: SubmitOutcome; readonly message?: string } @@ -258,5 +262,5 @@ export type InputEvent = export type InputEffect = | { readonly type: 'adjudicate'; readonly attempt: SubmitAttempt; readonly draft: string } | { readonly type: 'begin-submit'; readonly attempt: SubmitAttempt; readonly claim: CommandClaim; readonly args: string } - | { readonly type: 'default-sink'; readonly draft: string } + | { readonly type: 'default-sink'; readonly draft: string; readonly mode: InputSubmitMode } | { readonly type: 'notice'; readonly level: 'info' | 'error'; readonly text: string } diff --git a/packages/client/ui-conversation/src/client/input/facade.ts b/packages/client/ui-conversation/src/client/input/facade.ts index 79a223a43e..79516d04fa 100644 --- a/packages/client/ui-conversation/src/client/input/facade.ts +++ b/packages/client/ui-conversation/src/client/input/facade.ts @@ -16,6 +16,7 @@ import type { EditRange, EditSelection, InputActions, InputEffect, InputNotice, InputState, PasteComponent, QueuedMessage, SessionInput, SubmitAttempt, } from './contract.ts' +import type { InputSubmitMode } from '../contract/composer-submission.ts' import { InputMachine } from './machine.ts' /** Popup face the shell needs (dismissal only; typed structurally to avoid a value import). */ @@ -39,7 +40,7 @@ export interface SessionInputDeps { /** Queue read face; overlaid onto InputState.queue (absent = empty). */ queue?: ObservableSnapshot | undefined /** The plain-message sink (send choreography / materialize fork — the hub owns it). */ - defaultSink(text: string): void + defaultSink(text: string, mode: InputSubmitMode): void } /** Guard tier from the machine phase. */ @@ -68,7 +69,7 @@ export class SessionInputShell implements SessionInput { /** The public provide-channel action face (one stable identity per session — decision 20). */ readonly actions: InputActions = { setDraft: (text) => { this.setDraft(text) }, - submit: () => { this.submit() }, + submit: () => { this.submit('queue') }, } // Real wall clock: the typing-run merge window must actually expire in @@ -106,15 +107,6 @@ export class SessionInputShell implements SessionInput { this.run(this.core.dispatch({ type: 'send-committed' })) } - /** - * Insert a newline at the selection as one machine transaction (the - * execCommand path is gone — a second undo history would fork). - * @param selection - current DOM selection in draft coordinates. - */ - newline(selection: EditSelection): void { - this.run(this.core.dispatch({ type: 'newline', selection })) - } - /** Undo the latest transaction (InputBar intercepts the platform chord). */ undo(): void { this.run(this.core.dispatch({ type: 'undo' })) @@ -152,8 +144,8 @@ export class SessionInputShell implements SessionInput { * (adjudicating/submitting) force-closes the transient layers: the popup * dismisses and the menu tracks frozen. */ - submit(): void { - this.run(this.core.dispatch({ type: 'enter' })) + submit(mode: InputSubmitMode = 'queue'): void { + this.run(this.core.dispatch({ type: 'enter', mode })) const phase = this.snapshot.phase if (phase === 'adjudicating' || phase === 'submitting') { this.deps.popup?.()?.dismiss() @@ -338,7 +330,7 @@ export class SessionInputShell implements SessionInput { return } case 'default-sink': { - this.sinkSerialized(fx.draft) + this.sinkSerialized(fx.draft, fx.mode) return } default: @@ -353,10 +345,10 @@ export class SessionInputShell implements SessionInput { * send — notice + draft and chips retained, never a silent downgrade to * the clipboard text. Chip-free drafts skip the async detour. */ - private sinkSerialized(draft: string): void { + private sinkSerialized(draft: string, mode: InputSubmitMode): void { const occurrences = this.core.state.occurrences if (occurrences.length === 0) { - this.deps.defaultSink(draft.trim()) + this.deps.defaultSink(draft.trim(), mode) return } const slash = this.deps.slash?.() @@ -376,7 +368,7 @@ export class SessionInputShell implements SessionInput { cursor = part.offset + 1 } out += draft.slice(cursor) - this.deps.defaultSink(out.trim()) + this.deps.defaultSink(out.trim(), mode) }, (error: unknown) => { controller.abort() diff --git a/packages/client/ui-conversation/src/client/input/hub.ts b/packages/client/ui-conversation/src/client/input/hub.ts index 1fe2f95e7f..560999aa21 100644 --- a/packages/client/ui-conversation/src/client/input/hub.ts +++ b/packages/client/ui-conversation/src/client/input/hub.ts @@ -12,6 +12,7 @@ import type { ClientContext, ISessions, SessionBinding, SessionFace, SessionId } import type { SlashController } from '@deepseek-ai/dsh-client-ui-slash/client' import { queueReadFaceOf } from '../queue/store.ts' import type { ComposerKeyboard, InputService, SessionInput } from './contract.ts' +import type { InputSubmitMode } from '../contract/composer-submission.ts' import type { PopupDismissFace } from './facade.ts' import { SessionInputShell } from './facade.ts' @@ -56,7 +57,7 @@ export class InputHub implements InputService { slash: () => this.controller(actx), popup: () => this.popup(actx), queue: queueReadFaceOf(session), - defaultSink: (text) => { this.sink(session, text) }, + defaultSink: (text, mode) => { this.sink(session, text, mode) }, }) this.shells.set(id, shell) // The one teardown axis: listeners, shell, and map entries all ride the @@ -123,12 +124,12 @@ export class InputHub implements InputService { * exactly one path; a failed first prompt is an ordinary prompt failure * (error strip via promptError, draft restored only while untouched). */ - private sink(session: SessionFace, text: string): void { + private sink(session: SessionFace, text: string, mode: InputSubmitMode): void { if (text === '') return const shell = this.shells.get(session.sessionId) // Commit, not an editable clear: undo must not resurrect sent content. shell?.commitSend() - void session.prompt([{ type: 'text', text }], 'queue').then( + void session.prompt([{ type: 'text', text }], mode).then( (result) => { if (!result.ok && shell?.snapshot.draft === '') shell.setDraft(text) }, diff --git a/packages/client/ui-conversation/src/client/input/machine.ts b/packages/client/ui-conversation/src/client/input/machine.ts index f32d9e11d7..ec62eb63b4 100644 --- a/packages/client/ui-conversation/src/client/input/machine.ts +++ b/packages/client/ui-conversation/src/client/input/machine.ts @@ -14,6 +14,7 @@ * paste-upgrade all answer their bail events this way). */ import type { CommandClaim, ReferenceInsert, TokenSpan } from '@deepseek-ai/dsh-client-ui-slash/client' +import type { InputSubmitMode } from '../contract/composer-submission.ts' import type { ConsumeTokenGuard, EditRange, EditSelection, InputEffect, InputEvent, InputMachineOptions, InputState, Occurrence, PasteAttemptState, PasteComponent, SubmitAttempt, @@ -149,7 +150,6 @@ export class InputMachine { dispatch(ev: InputEvent): readonly InputEffect[] { switch (ev.type) { case 'draft-changed': return this.onDraftChanged(ev.draft, ev.editRange) - case 'newline': return this.onNewline(ev.selection) case 'begin-command': return this.onBeginCommand(ev.claim, ev.span) case 'insert-ref': return this.onInsertRef(ev.reference, ev.span) case 'consume-token': return this.onConsumeToken(ev.guard) @@ -162,7 +162,7 @@ export class InputMachine { this.paste = undefined return [] } - case 'enter': return this.onEnter() + case 'enter': return this.onEnter(ev.mode) case 'adjudicated': return this.onAdjudicated(ev.attempt, ev.outcome) case 'adjudication-failed': return this.onAdjudicationFailed(ev.attempt, ev.message) case 'submit-settled': return this.onSubmitSettled(ev) @@ -254,19 +254,6 @@ export class InputMachine { return [] } - /** F1: caret newline as an ordinary machine transaction (execCommand path removed). */ - private onNewline(selection: EditSelection): InputEffect[] { - const { start, end } = selection - if (start < 0 || start > end || end > this.draft.length) return [] - this.pushTxn(selection) - this.typingRun = undefined - this.reconcile({ start, end, insertedLength: 1 }) - this.adopt(this.draft.slice(0, start) + '\n' + this.draft.slice(end)) - this.watchClaim() - this.paste = undefined - return [] - } - /** Span CAS: revision equality (content identity follows) plus bounds sanity. */ private casOk(span: TokenSpan): boolean { return span.draftRev === this.draftRev @@ -461,18 +448,18 @@ export class InputMachine { // ---- submit plane ---- /** Mint the next SubmitAttempt and take the in-flight slot. */ - private beginAttempt(): SubmitAttempt { + private beginAttempt(mode: InputSubmitMode): SubmitAttempt { const controller = new AbortController() this.seq += 1 - const attempt: SubmitAttempt = { seq: this.seq, signal: controller.signal, draftSnapshot: this.draft } + const attempt: SubmitAttempt = { seq: this.seq, signal: controller.signal, draftSnapshot: this.draft, mode } this.inflight = { attempt, controller } return attempt } - private onEnter(): InputEffect[] { + private onEnter(mode: InputSubmitMode): InputEffect[] { if (this.phase === 'adjudicating' || this.phase === 'submitting') return [] if (this.phase === 'claimed' && this.claim !== undefined) { - const attempt = this.beginAttempt() + const attempt = this.beginAttempt(mode) this.phase = 'submitting' this.paste = undefined return [{ type: 'begin-submit', attempt, claim: this.claim, args: argsAfter(this.draft, this.claim.token) }] @@ -481,11 +468,11 @@ export class InputMachine { if (trimmed === '') return [] this.paste = undefined if (trimmed.startsWith('/')) { - const attempt = this.beginAttempt() + const attempt = this.beginAttempt(mode) this.phase = 'adjudicating' return [{ type: 'adjudicate', attempt, draft: this.draft }] } - return [{ type: 'default-sink', draft: this.draft }] + return [{ type: 'default-sink', draft: this.draft, mode }] } private onAdjudicated(attempt: SubmitAttempt, outcome: Extract['outcome']): InputEffect[] { @@ -506,7 +493,7 @@ export class InputMachine { this.inflight = undefined this.phase = 'plain' return outcome === undefined - ? [{ type: 'default-sink', draft: attempt.draftSnapshot }] + ? [{ type: 'default-sink', draft: attempt.draftSnapshot, mode: attempt.mode }] : [] } diff --git a/packages/client/ui-conversation/src/client/input/submission-policy.ts b/packages/client/ui-conversation/src/client/input/submission-policy.ts new file mode 100644 index 0000000000..99ce299c52 --- /dev/null +++ b/packages/client/ui-conversation/src/client/input/submission-policy.ts @@ -0,0 +1,72 @@ +/** + * Browser-local Composer submission policy. It owns the persisted busy-Enter + * preference and resolves keyboard gestures into queue/steer delivery modes; + * Host and Agent keep the actual delivery-window authority. + */ +import { createSnapshotStore, type SnapshotStore } from '@deepseek-ai/dsh-client-runtime/client' +import type { + BusyEnterBehavior, ComposerSubmitGesture, InputSubmitMode, +} from '../contract/composer-submission.ts' + +/** localStorage key holding the busy-Enter preference. */ +export const BUSY_ENTER_STORAGE_KEY = 'dsh.conversation.busyEnter' + +/** Default preserves Enter-as-Queue for running conversations. */ +export const DEFAULT_BUSY_ENTER_BEHAVIOR: BusyEnterBehavior = 'queue' + +/** + * Persisted policy used by both the composer inject face and its Settings row. + * Direct `steer` is intentionally best-effort: AgentLoop turns a closed-window + * submission into the next waking Queue item. + */ +export class ComposerSubmissionPolicy { + /** Reactive preference source for the Settings row. */ + readonly busyEnter: SnapshotStore = createSnapshotStore(restoreBusyEnter()) + + /** + * Resolve one keyboard gesture without changing state. + * @param running - whether the addressed agent currently reports busy. + * @param gesture - plain Enter or the Cmd/Ctrl-accelerated chord. + * @returns Queue outside busy state; otherwise the preferred mode or its opposite. + */ + resolve(running: boolean, gesture: ComposerSubmitGesture): InputSubmitMode { + if (!running) return 'queue' + const preferred = this.busyEnter.getSnapshot() + if (gesture === 'enter') return preferred + return preferred === 'queue' ? 'steer' : 'queue' + } + + /** + * Change and persist the plain-Enter behavior used during busy state. + * @param behavior - Queue or Steer. + */ + setBusyEnter(behavior: BusyEnterBehavior): void { + if (this.busyEnter.getSnapshot() === behavior) return + this.busyEnter.set(behavior) + persistBusyEnter(behavior) + } +} + +/** Restore a valid preference; unavailable or corrupt storage uses Queue. */ +function restoreBusyEnter(): BusyEnterBehavior { + if (typeof localStorage === 'undefined') return DEFAULT_BUSY_ENTER_BEHAVIOR + let stored: string | null + try { + stored = localStorage.getItem(BUSY_ENTER_STORAGE_KEY) + } catch { + // Storage access can fail in privacy modes; the default remains usable. + return DEFAULT_BUSY_ENTER_BEHAVIOR + } + if (stored === 'queue' || stored === 'steer') return stored + return DEFAULT_BUSY_ENTER_BEHAVIOR +} + +/** Persist a preference when browser storage is available. */ +function persistBusyEnter(behavior: BusyEnterBehavior): void { + if (typeof localStorage === 'undefined') return + try { + localStorage.setItem(BUSY_ENTER_STORAGE_KEY, behavior) + } catch { + // A storage failure makes the preference session-only; input stays usable. + } +} diff --git a/packages/client/ui-conversation/src/client/locales.ts b/packages/client/ui-conversation/src/client/locales.ts index 21897a529b..1ecb33b854 100644 --- a/packages/client/ui-conversation/src/client/locales.ts +++ b/packages/client/ui-conversation/src/client/locales.ts @@ -23,6 +23,10 @@ export const zh = { 'input.stop': '停止生成', 'input.send': '发送消息', 'input.accessMode': '访问模式,当前:{name}', + 'settings.enter.title': '繁忙时 Enter 键行为', + 'settings.enter.description': '仅在智能体运行时生效;Cmd/Ctrl+Enter 使用另一行为', + 'settings.enter.queue': '排队发送', + 'settings.enter.steer': '插话发送', 'access.confirm.title': '确认启用 Full access?', 'access.confirm.description': '启用 Full access 后,agent 将减少确认步骤,并且可以直接执行更多操作,包括敏感操作、文件修改或外部命令。仅建议在你信任当前任务时使用。', 'access.confirm.acknowledge': '我已了解风险,并愿意继续', @@ -126,6 +130,10 @@ export const en = { 'input.stop': 'Stop generating', 'input.send': 'Send message', 'input.accessMode': 'Access mode, current: {name}', + 'settings.enter.title': 'Enter behavior while busy', + 'settings.enter.description': 'Busy only; Cmd/Ctrl+Enter uses the other behavior', + 'settings.enter.queue': 'Queue', + 'settings.enter.steer': 'Steer', 'access.confirm.title': 'Enable Full access?', 'access.confirm.description': 'Full access reduces confirmation steps and lets the agent perform more actions directly, including sensitive operations, file changes, or external commands. Only use it when you trust the current task.', 'access.confirm.acknowledge': 'I understand the risks and want to continue', diff --git a/packages/client/ui-conversation/src/client/queue/QueueDock.tsx b/packages/client/ui-conversation/src/client/queue/QueueDock.tsx index 10c16e6dc5..5085a6e579 100644 --- a/packages/client/ui-conversation/src/client/queue/QueueDock.tsx +++ b/packages/client/ui-conversation/src/client/queue/QueueDock.tsx @@ -4,7 +4,7 @@ // The 'conversation.input.dock' SlotMap declaration lives in // ../contract/slots.ts beside the other input-region slots. import type { Context } from 'cordis' -import { useEffect, useId, useState } from 'react' +import { useEffect, useId, useMemo, useState } from 'react' import type { PropsLocale, PropsRuntime } from '@deepseek-ai/dsh-client-ui-slots' import type { SessionId } from '@deepseek-ai/dsh-client-runtime/client' import { @@ -29,7 +29,8 @@ export type QueueDockProps = PropsRuntime<'conversation.input.dock'> & QueueDock * collapsible count header; an empty queue renders nothing. */ export function QueueDock({ useSession, updateQueue, notify, t }: QueueDockProps) { - const queue = useSession(s => s.queue) + const inbox = useSession(s => s.queue) + const queue = useMemo(() => inbox.filter(row => row.placement === 'queued'), [inbox]) const running = useSession(s => s.running) const [editing, setEditing] = useState<{ id: QueueItemId; text: string } | null>(null) const [busy, setBusy] = useState(null) diff --git a/packages/client/ui-conversation/src/client/queue/store.ts b/packages/client/ui-conversation/src/client/queue/store.ts index 536523465b..35364383c6 100644 --- a/packages/client/ui-conversation/src/client/queue/store.ts +++ b/packages/client/ui-conversation/src/client/queue/store.ts @@ -1,7 +1,7 @@ /** * Queue read face for the InputState.queue projection (frozen contract in * ../input/contract.ts): a uSES-compatible observable over one session's - * queue rows. The Session snapshot already keeps the queue array + * transient inbox rows. The Session snapshot already keeps the queue array * reference-stable across unrelated snapshot swaps, so this is a pure * projection — no second store, no copy. */ @@ -9,7 +9,7 @@ import type { ObservableSnapshot, SessionFace } from '@deepseek-ai/dsh-client-ru import type { QueuedMessage } from '../input/contract.ts' /** - * Project a session's queue rows as a bare observable (subscribe/getSnapshot). + * Project a session's transient inbox rows as a bare observable (subscribe/getSnapshot). * The wiring layer (T5) overlays this onto InputState.queue; the runtime * QueuedMessage and the input-contract QueuedMessage are structurally * identical. diff --git a/packages/client/ui-conversation/src/client/settings/EnterBehaviorRow.module.css b/packages/client/ui-conversation/src/client/settings/EnterBehaviorRow.module.css new file mode 100644 index 0000000000..9d8b3d7ceb --- /dev/null +++ b/packages/client/ui-conversation/src/client/settings/EnterBehaviorRow.module.css @@ -0,0 +1,56 @@ +/* Composer Enter preference row: title/description plus selector pill. */ + +.row { + display: flex; + align-items: center; + gap: 8px; + padding: 16px 0; + border-bottom: 1px solid var(--dsw-alias-border-l2); +} + +.rowText { + flex: 1; + min-width: 0; + display: flex; + flex-direction: column; + gap: 4px; + padding-right: 48px; +} + +.title { + font-size: 14px; + font-weight: 400; + line-height: 22px; + color: var(--dsw-alias-label-primary); +} + +.desc { + font-size: 12px; + font-weight: 400; + line-height: 18px; + color: var(--dsw-alias-label-tertiary); +} + +.selector { + display: inline-flex; + align-items: center; + gap: 12px; + height: 36px; + padding: 0 14px; + border: none; + border-radius: 18px; + background: var(--dsw-alias-bg-module-platform); + font: inherit; + font-size: 14px; + line-height: 22px; + color: var(--dsw-alias-label-primary); + cursor: pointer; +} + +.selector:hover { + background: var(--dsw-alias-interactive-bg-hover); +} + +.chevron { + flex: none; +} diff --git a/packages/client/ui-conversation/src/client/settings/EnterBehaviorRow.tsx b/packages/client/ui-conversation/src/client/settings/EnterBehaviorRow.tsx new file mode 100644 index 0000000000..2b5552c3cc --- /dev/null +++ b/packages/client/ui-conversation/src/client/settings/EnterBehaviorRow.tsx @@ -0,0 +1,76 @@ +/** General Settings row for the Composer's busy-state Enter preference. */ +import { useState } from 'react' +import type { SnapshotStore } from '@deepseek-ai/dsh-client-runtime/client' +import type { InjectFace, PropsLocale, PropsRuntime } from '@deepseek-ai/dsh-client-ui-slots' +import { IconChevronDownOutline14, Menu } from '@deepseek-ai/dsh-client-ui-primitives' +import type { BusyEnterBehavior } from '../contract/composer-submission.ts' +import type { ConversationKey } from '../locales.ts' +import css from './EnterBehaviorRow.module.css' + +/** Registration-side preference face. */ +export interface EnterBehaviorRowInjected { + hooks: { + /** Persisted busy-state preference bound as useBusyEnter. */ + busyEnter: SnapshotStore + } + /** Change the busy-state plain-Enter behavior. */ + setBusyEnter: (behavior: BusyEnterBehavior) => void +} + +/** Full Settings-row props. */ +export type EnterBehaviorRowProps = + PropsRuntime<'settings.general.item'> + & PropsLocale<'conversation'> + & InjectFace + +const OPTIONS: readonly { + id: BusyEnterBehavior + label: ConversationKey +}[] = [ + { id: 'queue', label: 'settings.enter.queue' }, + { id: 'steer', label: 'settings.enter.steer' }, +] + +/** + * Render the busy-state Enter behavior selector. + * @param props - composed Settings slot props. + * @returns the preference row. + */ +export function EnterBehaviorRow({ useBusyEnter, setBusyEnter, t }: EnterBehaviorRowProps) { + const behavior = useBusyEnter(value => value) + const [open, setOpen] = useState(false) + const selectedLabel = behavior === 'queue' ? 'settings.enter.queue' : 'settings.enter.steer' + + return ( +
+
+
{t('settings.enter.title')}
+
{t('settings.enter.description')}
+
+ { setOpen(false) }} + items={OPTIONS.map(option => ({ id: option.id, label: t(option.label) }))} + selectedId={behavior} + onSelect={(id) => { + setOpen(false) + setBusyEnter(id as BusyEnterBehavior) + }} + align="end" + portal + anchor={( + + )} + /> +
+ ) +} diff --git a/packages/client/ui-conversation/src/client/skeleton/InputBar.tsx b/packages/client/ui-conversation/src/client/skeleton/InputBar.tsx index 43a0f3e1ec..7b8aa87414 100644 --- a/packages/client/ui-conversation/src/client/skeleton/InputBar.tsx +++ b/packages/client/ui-conversation/src/client/skeleton/InputBar.tsx @@ -34,7 +34,7 @@ export interface InputBarError { export type InputBarProps = ComposerBarProps export function InputBar({ - useSession, useInput, inputActions, keyboard, toggleCommandMenu, stop, command, t, + useSession, useInput, inputActions, keyboard, resolveSubmitMode, toggleCommandMenu, stop, command, t, renderSlot, useNotices, useLexicon, useMenuLauncher, useProjection, sessionId, variant, disabled: inert = false, placeholder, accessory, overlay, leftItems, rightItems, footer, }: InputBarProps) { @@ -178,23 +178,10 @@ export function InputBar({ e.preventDefault() return } - if (e.ctrlKey || e.metaKey) { - // Newline as a machine transaction (the machine owns undo history; an - // execCommand write would fork a second, browser-owned history). - e.preventDefault() - if (!machineBusy && !locked) { - const el = e.currentTarget - const sel = selectionOf(el) - keyboard.newline(sel) - const caret = sel.start + 1 - requestAnimationFrame(() => { el.setSelectionRange(caret, caret) }) - } - return - } e.preventDefault() if (e.repeat) return // held-down Enter must not machine-gun sends if (locked || machineBusy) return - inputActions.submit() + keyboard.submit(resolveSubmitMode(running, e.ctrlKey || e.metaKey ? 'accelerated' : 'enter')) } const onChange = (e: ChangeEvent): void => { diff --git a/packages/client/ui-conversation/tests/chat-apply.spec.tsx b/packages/client/ui-conversation/tests/chat-apply.spec.tsx index d3fabd9bee..20e7e54f86 100644 --- a/packages/client/ui-conversation/tests/chat-apply.spec.tsx +++ b/packages/client/ui-conversation/tests/chat-apply.spec.tsx @@ -37,6 +37,7 @@ async function bench() { await runtime.root.declare({ 'conversation': { kind: 'single', scope: 'session-maybe' }, 'details': { kind: 'single', scope: 'session' }, + 'settings.general.item': { kind: 'list', scope: 'root' }, }, (_p: { renderSlot?: unknown }) => null) const feature = await runtime.mount({ inject: [...inject], apply }) @@ -85,6 +86,7 @@ describe('apply wiring', () => { // The hero workspace picker hole rides the conversation entry's children // declaration (the empty-state occupant is gone). expect(b.slots.spec('conversation.hero.workspace')).toEqual({ kind: 'single', scope: 'root' }) + expect(b.slots.entries('settings.general.item').map(entry => entry.options.id)).toEqual(['composer-enter']) await b.runtime.dispose() }) @@ -112,6 +114,7 @@ describe('apply wiring', () => { expect(b.slots.entries('conversation.chat.toolview')).toHaveLength(0) expect(b.slots.spec('conversation.chat.toolview')).toBeUndefined() expect(b.slots.entries('details')).toHaveLength(0) + expect(b.slots.entries('settings.general.item')).toHaveLength(0) expect(b.runtime.ctx.get('conversation')).toBeUndefined() await b.runtime.dispose() }) diff --git a/packages/client/ui-conversation/tests/chat-view.spec.tsx b/packages/client/ui-conversation/tests/chat-view.spec.tsx index 04f412fb25..507b301922 100644 --- a/packages/client/ui-conversation/tests/chat-view.spec.tsx +++ b/packages/client/ui-conversation/tests/chat-view.spec.tsx @@ -251,6 +251,51 @@ describe('ChatView', () => { expect(view.getByText('run a')).toBeTruthy() }) + it('renders Host-pending steering at the flow tail and hands off to the durable node', () => { + const pending = { + id: 'steer-occurrence' as never, + messageId: 'steer-message' as never, + placement: 'steering' as const, + content: [{ type: 'text' as const, text: 'interrupt now' }], + preview: 'interrupt now', + text: 'interrupt now', + } + const queued = { + id: 'queued-occurrence' as never, + messageId: 'queued-message' as never, + placement: 'queued' as const, + content: [{ type: 'text' as const, text: 'later' }], + preview: 'later', + text: 'later', + } + const h = makeHarness({ nodes: [assistant(1, 'working')], queue: [queued, pending], running: true }) + const view = render() + + expect(view.getByText('interrupt now').closest('[data-pending-steering]')).not.toBeNull() + expect(view.queryByText('later')).toBeNull() + expect(view.getByRole('status').compareDocumentPosition(view.getByText('interrupt now')) + & Node.DOCUMENT_POSITION_FOLLOWING).not.toBe(0) + + act(() => { + h.set({ + queue: [queued, pending], + nodes: [ + assistant(1, 'working'), + { + kind: 'steering', messageId: pending.messageId, + seq: 2, time: 2_000, turn: 1, + content: [{ type: 'text', text: 'interrupt now' }], source: null, + }, + ], + }) + }) + expect(view.getAllByText('interrupt now')).toHaveLength(1) + expect(view.container.querySelector('[data-pending-steering]')).toBeNull() + + act(() => { h.set({ queue: [queued] }) }) + expect(view.getAllByText('interrupt now')).toHaveLength(1) + }) + it('animates only the latest unresolved model retry', () => { const retryNode = retry(2) const nextRetry = { ...retry(3), turn: 2, retry: 2 } diff --git a/packages/client/ui-conversation/tests/enter-behavior-row.spec.tsx b/packages/client/ui-conversation/tests/enter-behavior-row.spec.tsx new file mode 100644 index 0000000000..e8d44c9bf3 --- /dev/null +++ b/packages/client/ui-conversation/tests/enter-behavior-row.spec.tsx @@ -0,0 +1,67 @@ +// @vitest-environment jsdom +import { afterEach, describe, expect, it, vi } from 'vitest' +import { act, cleanup, fireEvent, render, screen } from '@testing-library/react' +import { bindSnapshotSelector } from '@deepseek-ai/dsh-client-web-react' +import { createSnapshotStore, type SessionListState, type WorkspaceListState } from '@deepseek-ai/dsh-client-runtime/client' +import { makeTranslate } from '@deepseek-ai/dsh-client-test-runtime' +import { EnterBehaviorRow } from '../src/client/settings/EnterBehaviorRow.tsx' +import type { EnterBehaviorRowProps } from '../src/client/settings/EnterBehaviorRow.tsx' +import { ComposerSubmissionPolicy } from '../src/client/input/submission-policy.ts' +import { en } from '../src/client/locales.ts' + +afterEach(() => { + cleanup() + localStorage.clear() +}) + +function emptySessions() { + return bindSnapshotSelector(createSnapshotStore({ + ids: [], byId: {}, current: undefined, phase: 'ready', subagentsByParent: {}, currentAddress: undefined, + })) +} + +function emptyWorkspaces() { + return bindSnapshotSelector(createSnapshotStore({ + items: [], archivedSessionIds: [], state: 'idle', phase: 'ready', error: null, + baselinesReady: true, recentWorkspaceId: undefined, + })) +} + +function mount() { + const policy = new ComposerSubmissionPolicy() + const setBusyEnter = vi.fn((behavior: 'queue' | 'steer') => { policy.setBusyEnter(behavior) }) + const props: EnterBehaviorRowProps = { + useSessions: emptySessions(), + useWorkspaces: emptyWorkspaces(), + useBusyEnter: bindSnapshotSelector(policy.busyEnter), + setBusyEnter, + t: makeTranslate(en), + } + render() + return { policy, setBusyEnter } +} + +describe('EnterBehaviorRow', () => { + it('explains the busy-only scope and shows Queue by default', () => { + mount() + expect(screen.getByText('Enter behavior while busy')).toBeDefined() + expect(screen.getByText('Busy only; Cmd/Ctrl+Enter uses the other behavior')).toBeDefined() + expect(screen.getByRole('button', { name: /Queue/ }).getAttribute('aria-expanded')).toBe('false') + }) + + it('selects Steer, follows later preference changes, and closes outside', () => { + const b = mount() + const trigger = screen.getByRole('button', { name: /Queue/ }) + fireEvent.click(trigger) + fireEvent.click(screen.getByRole('menuitem', { name: 'Steer' })) + expect(b.setBusyEnter).toHaveBeenCalledWith('steer') + expect(screen.getByRole('button', { name: /Steer/ })).toBeDefined() + + act(() => { b.policy.setBusyEnter('queue') }) + const queueTrigger = screen.getByRole('button', { name: /Queue/ }) + fireEvent.click(queueTrigger) + expect(screen.getByRole('menuitem', { name: 'Steer' })).toBeDefined() + fireEvent.pointerDown(document.body) + expect(screen.queryByRole('menuitem', { name: 'Steer' })).toBeNull() + }) +}) diff --git a/packages/client/ui-conversation/tests/input-bar.spec.tsx b/packages/client/ui-conversation/tests/input-bar.spec.tsx index 2072fd2c90..bfe34316fa 100644 --- a/packages/client/ui-conversation/tests/input-bar.spec.tsx +++ b/packages/client/ui-conversation/tests/input-bar.spec.tsx @@ -1,6 +1,6 @@ // @vitest-environment jsdom // InputBar behavior over the machine wiring: Enter-send semantics (IME guard, -// shift newline, ctrl/meta insert, repeat suppression), queue-cut-1 running +// Shift newline, busy Enter policy, Ctrl/Meta steering, repeat suppression), running // semantics (input stays free; primary turns stop), the machine pending lock, // decoration backdrop, error/notice strips, and the focus-keeping mousedown. @@ -53,6 +53,7 @@ interface BenchOptions { leftItems?: React.ReactNode rightItems?: React.ReactNode commandMenuOpen?: boolean + busyEnter?: 'queue' | 'steer' toggleCommandMenu?: (selection: { start: number; end: number }) => void } @@ -107,6 +108,11 @@ function bench(over?: BenchOptions) { useInput: bindSnapshotSelector(shell.state), inputActions: shell.actions, keyboard: shell, + resolveSubmitMode: (running, gesture) => { + if (!running) return 'queue' + const preferred = over?.busyEnter ?? 'queue' + return gesture === 'enter' ? preferred : preferred === 'queue' ? 'steer' : 'queue' + }, toggleCommandMenu: over?.toggleCommandMenu ?? vi.fn(), useNotices: bindSnapshotSelector(shell.notices), useLexicon: bindSnapshotSelector(shell.lexicon), @@ -137,7 +143,7 @@ describe('Enter semantics', () => { it('plain Enter submits queue mode through the machine; repeat and empty are suppressed', () => { const { textarea, sink } = bench({ draft: 'hello' }) fireEvent.keyDown(textarea, { key: 'Enter' }) - expect(sink).toHaveBeenCalledWith('hello') + expect(sink).toHaveBeenCalledWith('hello', 'queue') fireEvent.keyDown(textarea, { key: 'Enter', repeat: true }) expect(sink).toHaveBeenCalledTimes(1) const empty = bench({ draft: ' ' }) @@ -159,12 +165,18 @@ describe('Enter semantics', () => { expect(sink).not.toHaveBeenCalled() // and not preventDefault'd: native newline }) - it('Ctrl/Meta+Enter inserts a newline through the machine (no browser execCommand)', () => { - const { textarea, shell, sink } = bench({ draft: 'hello' }) - textarea.setSelectionRange(5, 5) - fireEvent.keyDown(textarea, { key: 'Enter', ctrlKey: true }) - expect(shell.snapshot.draft).toBe('hello\n') - expect(sink).not.toHaveBeenCalled() + it('Ctrl/Meta+Enter sends normally while idle and steers while running', () => { + const idle = bench({ draft: 'hello' }) + fireEvent.keyDown(idle.textarea, { key: 'Enter', metaKey: true }) + expect(idle.sink).toHaveBeenCalledWith('hello', 'queue') + + const busyCtrl = bench({ running: true, draft: 'steer with ctrl' }) + fireEvent.keyDown(busyCtrl.textarea, { key: 'Enter', ctrlKey: true }) + expect(busyCtrl.sink).toHaveBeenCalledWith('steer with ctrl', 'steer') + + const busyMeta = bench({ running: true, draft: 'steer with cmd' }) + fireEvent.keyDown(busyMeta.textarea, { key: 'Enter', metaKey: true }) + expect(busyMeta.sink).toHaveBeenCalledWith('steer with cmd', 'steer') }) it('platform undo/redo chords route to the machine, never the browser stack', () => { @@ -205,12 +217,28 @@ describe('running and lock semantics (queue cut 1)', () => { expect(textarea.disabled).toBe(false) // running no longer locks fireEvent.change(textarea, { target: { value: '排队消息2' } }) fireEvent.keyDown(textarea, { key: 'Enter' }) - expect(sink).toHaveBeenCalledWith('排队消息2') + expect(sink).toHaveBeenCalledWith('排队消息2', 'queue') expect(button.getAttribute('aria-label')).toBe('停止生成') fireEvent.click(button) expect(stop).toHaveBeenCalledTimes(1) }) + it('running plain Enter follows the busy-state Steer preference', () => { + const { textarea, sink } = bench({ running: true, busyEnter: 'steer', draft: '直接插话' }) + fireEvent.keyDown(textarea, { key: 'Enter' }) + expect(sink).toHaveBeenCalledWith('直接插话', 'steer') + }) + + it('running Cmd/Ctrl+Enter uses the opposite of the busy-state Enter preference', () => { + const meta = bench({ running: true, busyEnter: 'steer', draft: '排到下一轮' }) + fireEvent.keyDown(meta.textarea, { key: 'Enter', metaKey: true }) + expect(meta.sink).toHaveBeenCalledWith('排到下一轮', 'queue') + + const ctrl = bench({ running: true, busyEnter: 'steer', draft: 'also queue' }) + fireEvent.keyDown(ctrl.textarea, { key: 'Enter', ctrlKey: true }) + expect(ctrl.sink).toHaveBeenCalledWith('also queue', 'queue') + }) + it('running subagent primary admits a follow-up instead of exposing Stop', () => { const { button, sink, stop } = bench({ running: true, @@ -226,7 +254,7 @@ describe('running and lock semantics (queue cut 1)', () => { }) expect(button.getAttribute('aria-label')).toBe('发送消息') fireEvent.click(button) - expect(sink).toHaveBeenCalledWith('后续消息') + expect(sink).toHaveBeenCalledWith('后续消息', 'queue') expect(stop).not.toHaveBeenCalled() const empty = bench({ @@ -253,7 +281,7 @@ describe('running and lock semantics (queue cut 1)', () => { it('idle primary sends and disables on empty draft', () => { const { button, sink } = bench({ draft: 'go' }) fireEvent.click(button) - expect(sink).toHaveBeenCalledWith('go') + expect(sink).toHaveBeenCalledWith('go', 'queue') const empty = bench() expect(empty.button.disabled).toBe(true) }) diff --git a/packages/client/ui-conversation/tests/input-machine.spec.ts b/packages/client/ui-conversation/tests/input-machine.spec.ts index 510cb28076..89737d7106 100644 --- a/packages/client/ui-conversation/tests/input-machine.spec.ts +++ b/packages/client/ui-conversation/tests/input-machine.spec.ts @@ -40,9 +40,9 @@ function effectAt( } /** Drive plain → adjudicating and hand back the minted attempt. */ -function enterAdjudicating(m: InputMachine, draft: string): SubmitAttempt { +function enterAdjudicating(m: InputMachine, draft: string, mode: 'queue' | 'steer' = 'queue'): SubmitAttempt { m.dispatch({ type: 'draft-changed', draft }) - const fx = m.dispatch({ type: 'enter' }) + const fx = m.dispatch({ type: 'enter', mode }) return effectAt(fx, 0, 'adjudicate').attempt } @@ -52,35 +52,42 @@ function enterSubmitting(m: InputMachine, name: string, args: string): { attempt m.dispatch({ type: 'draft-changed', draft: `/${name.slice(0, 2)}` }) m.dispatch({ type: 'begin-command', claim, span: spanOf(m, 0, m.state.draft.length) }) m.dispatch({ type: 'draft-changed', draft: claim.token + args }) - const fx = m.dispatch({ type: 'enter' }) + const fx = m.dispatch({ type: 'enter', mode: 'queue' }) return { attempt: effectAt(fx, 0, 'begin-submit').attempt, claim } } function staleAttempt(): SubmitAttempt { - return { seq: 9999, signal: new AbortController().signal, draftSnapshot: '' } + return { seq: 9999, signal: new AbortController().signal, draftSnapshot: '', mode: 'queue' } } describe('input-machine: plain × enter', () => { it('empty and whitespace-only drafts produce nothing', () => { const m = new InputMachine() - expect(m.dispatch({ type: 'enter' })).toEqual([]) + expect(m.dispatch({ type: 'enter', mode: 'queue' })).toEqual([]) m.dispatch({ type: 'draft-changed', draft: ' \n ' }) - expect(m.dispatch({ type: 'enter' })).toEqual([]) + expect(m.dispatch({ type: 'enter', mode: 'queue' })).toEqual([]) expect(m.state.phase).toBe('plain') }) it('non-command text falls to the default sink', () => { const m = new InputMachine() m.dispatch({ type: 'draft-changed', draft: 'hello world' }) - expect(m.dispatch({ type: 'enter' })) - .toEqual([{ type: 'default-sink', draft: 'hello world' }]) + expect(m.dispatch({ type: 'enter', mode: 'queue' })) + .toEqual([{ type: 'default-sink', draft: 'hello world', mode: 'queue' }]) expect(m.state.phase).toBe('plain') }) + it('retains an explicit steer mode on the default sink effect', () => { + const m = new InputMachine() + m.dispatch({ type: 'draft-changed', draft: 'steer now' }) + expect(m.dispatch({ type: 'enter', mode: 'steer' })) + .toEqual([{ type: 'default-sink', draft: 'steer now', mode: 'steer' }]) + }) + it('leading "/" enters adjudicating with a minted attempt carrying the draft snapshot', () => { const m = new InputMachine() m.dispatch({ type: 'draft-changed', draft: '/goal x' }) - const fx = m.dispatch({ type: 'enter' }) + const fx = m.dispatch({ type: 'enter', mode: 'queue' }) const eff = effectAt(fx, 0, 'adjudicate') expect(eff.draft).toBe('/goal x') expect(eff.attempt.draftSnapshot).toBe('/goal x') @@ -91,14 +98,14 @@ describe('input-machine: plain × enter', () => { it('leading is judged after trim including newlines', () => { const m = new InputMachine() m.dispatch({ type: 'draft-changed', draft: '\n\n/goal x' }) - expect(m.dispatch({ type: 'enter' })[0]?.type).toBe('adjudicate') + expect(m.dispatch({ type: 'enter', mode: 'queue' })[0]?.type).toBe('adjudicate') }) it('a non-whitespace prefix before "/" is not leading — default sink', () => { const m = new InputMachine() m.dispatch({ type: 'draft-changed', draft: '第一行\n/goal x' }) - expect(m.dispatch({ type: 'enter' })) - .toEqual([{ type: 'default-sink', draft: '第一行\n/goal x' }]) + expect(m.dispatch({ type: 'enter', mode: 'queue' })) + .toEqual([{ type: 'default-sink', draft: '第一行\n/goal x', mode: 'queue' }]) }) }) @@ -126,9 +133,9 @@ describe('input-machine: adjudication outcomes', () => { it('undefined outcome falls back to the default sink', () => { const m = new InputMachine() - const attempt = enterAdjudicating(m, '/unknown thing') + const attempt = enterAdjudicating(m, '/unknown thing', 'steer') expect(m.dispatch({ type: 'adjudicated', attempt, outcome: undefined })) - .toEqual([{ type: 'default-sink', draft: '/unknown thing' }]) + .toEqual([{ type: 'default-sink', draft: '/unknown thing', mode: 'steer' }]) expect(m.state.phase).toBe('plain') }) @@ -152,7 +159,7 @@ describe('input-machine: adjudication outcomes', () => { it('enter is a no-op while adjudicating (pending lock)', () => { const m = new InputMachine() enterAdjudicating(m, '/goal x') - expect(m.dispatch({ type: 'enter' })).toEqual([]) + expect(m.dispatch({ type: 'enter', mode: 'queue' })).toEqual([]) expect(m.state.phase).toBe('adjudicating') }) @@ -344,31 +351,6 @@ describe('input-machine: occurrence reconciliation on draft edits', () => { }) }) -describe('input-machine: newline transaction (F1)', () => { - it('inserts \\n at the caret and shifts trailing occurrences', () => { - const m = new InputMachine() - m.dispatch({ type: 'draft-changed', draft: 'ab @wor' }) - m.dispatch({ type: 'insert-ref', reference: refOf('w'), span: spanOf(m, 3, 7) }) - m.dispatch({ type: 'newline', selection: { start: 2, end: 2 } }) - expect(m.state.draft).toBe(`ab\n ${P} `) - expect(m.state.occurrences[0]?.offset).toBe(4) - m.dispatch({ type: 'undo' }) - expect(m.state.draft).toBe(`ab ${P} `) - }) - - it('replaces a selection, breaks the claim prefix when leading, and rejects out-of-bounds', () => { - const m = new InputMachine() - m.dispatch({ type: 'draft-changed', draft: '/go' }) - m.dispatch({ type: 'begin-command', claim: claimOf('goal'), span: spanOf(m, 0, 3) }) - expect(m.dispatch({ type: 'newline', selection: { start: 0, end: 99 } })).toEqual([]) - expect(m.state.phase).toBe('claimed') - m.dispatch({ type: 'newline', selection: { start: 0, end: 0 } }) - expect(m.state.draft).toBe('\n/goal ') - expect(m.state.phase).toBe('plain') - expect(m.state.claim).toBeUndefined() - }) -}) - describe('input-machine: consume-token guards', () => { it('span guard: CAS pass deletes the token — success observable as a draftRev advance', () => { const m = new InputMachine() @@ -587,7 +569,7 @@ describe('input-machine: paste plane', () => { const b = new InputMachine() b.dispatch({ type: 'paste-begin', text: 'plain text', selection: { start: 0, end: 0 } }) - b.dispatch({ type: 'enter' }) + b.dispatch({ type: 'enter', mode: 'queue' }) expect(b.state.paste).toBeUndefined() }) @@ -755,7 +737,7 @@ describe('input-machine: submitting transaction', () => { it('enter and begin-command are locked; draft-changed is recorded without leaving submitting', () => { const m = new InputMachine() enterSubmitting(m, 'goal', 'x') - expect(m.dispatch({ type: 'enter' })).toEqual([]) + expect(m.dispatch({ type: 'enter', mode: 'queue' })).toEqual([]) expect(m.dispatch({ type: 'draft-changed', draft: '/goal y' })).toEqual([]) expect(m.state).toMatchObject({ phase: 'submitting', draft: '/goal y' }) }) @@ -768,7 +750,7 @@ describe('input-machine: submitting transaction', () => { m.dispatch({ type: 'draft-changed', draft: '/go', editRange: { start: 0, end: 1, insertedLength: 0 } }) m.dispatch({ type: 'begin-command', claim: claimOf('goal'), span: spanOf(m, 0, 3) }) m.dispatch({ type: 'draft-changed', draft: '/goal go' }) - const attempt = effectAt(m.dispatch({ type: 'enter' }), 0, 'begin-submit').attempt + const attempt = effectAt(m.dispatch({ type: 'enter', mode: 'queue' }), 0, 'begin-submit').attempt const fx = m.dispatch({ type: 'submit-settled', attempt, ok: true, outcome: { kind: 'success', text: 'goal set' } }) expect(fx).toEqual([{ type: 'notice', level: 'info', text: 'goal set' }]) expect(m.state).toMatchObject({ phase: 'plain', draft: '', occurrences: [] }) @@ -809,7 +791,7 @@ describe('input-machine: submitting transaction', () => { const m = new InputMachine() const { attempt: first } = enterSubmitting(m, 'goal', 'x') m.dispatch({ type: 'submit-settled', attempt: first, ok: false, message: 'retry' }) - const second = effectAt(m.dispatch({ type: 'enter' }), 0, 'begin-submit').attempt + const second = effectAt(m.dispatch({ type: 'enter', mode: 'queue' }), 0, 'begin-submit').attempt expect(second.seq).not.toBe(first.seq) expect(m.dispatch({ type: 'submit-settled', attempt: first, ok: true })).toEqual([]) expect(m.state.phase).toBe('submitting') diff --git a/packages/client/ui-conversation/tests/input-matrix.spec.tsx b/packages/client/ui-conversation/tests/input-matrix.spec.tsx index 45ce41c3b8..af8dbb8e66 100644 --- a/packages/client/ui-conversation/tests/input-matrix.spec.tsx +++ b/packages/client/ui-conversation/tests/input-matrix.spec.tsx @@ -47,6 +47,7 @@ function mountBar(shell: SessionInputShell, over?: { running?: boolean; disabled useInput: bindSnapshotSelector(shell.state), inputActions: shell.actions, keyboard: shell, + resolveSubmitMode: () => 'queue', toggleCommandMenu: vi.fn(), useNotices: bindSnapshotSelector(shell.notices), useLexicon: bindSnapshotSelector(shell.lexicon), @@ -88,7 +89,7 @@ describe('matrix row: plain', () => { fireEvent.change(textarea, { target: { value: '普通消息' } }) expect(shell.snapshot.claim).toBeUndefined() fireEvent.keyDown(textarea, { key: 'Enter' }) - expect(sink).toHaveBeenCalledWith('普通消息') + expect(sink).toHaveBeenCalledWith('普通消息', 'queue') expect(shell.snapshot.phase).toBe('plain') }) }) @@ -187,7 +188,7 @@ describe('matrix row: locked (session disabled)', () => { expect((textarea).disabled).toBe(false) fireEvent.change(textarea, { target: { value: '排队' } }) fireEvent.keyDown(textarea, { key: 'Enter' }) - expect(sink).toHaveBeenCalledWith('排队') + expect(sink).toHaveBeenCalledWith('排队', 'queue') }) }) diff --git a/packages/client/ui-conversation/tests/input-scenarios.spec.tsx b/packages/client/ui-conversation/tests/input-scenarios.spec.tsx index a4eb2e668f..f3ef3410ff 100644 --- a/packages/client/ui-conversation/tests/input-scenarios.spec.tsx +++ b/packages/client/ui-conversation/tests/input-scenarios.spec.tsx @@ -133,6 +133,7 @@ async function scopedBench(register?: (slash: SlashService) => void) { useInput: bindSnapshotSelector(shell.state), inputActions: shell.actions, keyboard: shell, + resolveSubmitMode: () => 'queue', toggleCommandMenu: (selection) => { const snapshot = shell.snapshot controller.toggleSource('command', { @@ -235,7 +236,7 @@ describe('scenario D: execute-kind /compact', () => { act(() => { b2.shell.setDraft('/compact 现在') }) fireEvent.keyDown(b2.textarea, { key: 'Enter' }) // execute with trailing → matchEnter answers undefined → default sink. - await vi.waitFor(() => { expect(b2.sink).toHaveBeenCalledWith('/compact 现在') }) + await vi.waitFor(() => { expect(b2.sink).toHaveBeenCalledWith('/compact 现在', 'queue') }) expect(b2.executed).toHaveLength(0) }) }) @@ -289,7 +290,7 @@ describe('scenario I: unknown /xyz + enter', () => { const b = await bench() act(() => { b.shell.setDraft('/xyz 干点啥') }) fireEvent.keyDown(b.textarea, { key: 'Enter' }) - await vi.waitFor(() => { expect(b.sink).toHaveBeenCalledWith('/xyz 干点啥') }) + await vi.waitFor(() => { expect(b.sink).toHaveBeenCalledWith('/xyz 干点啥', 'queue') }) expect(b.shell.snapshot.phase).toBe('plain') expect(b.execute).not.toHaveBeenCalled() }) diff --git a/packages/client/ui-conversation/tests/queue-dock.spec.tsx b/packages/client/ui-conversation/tests/queue-dock.spec.tsx index eb5c293501..4195488a53 100644 --- a/packages/client/ui-conversation/tests/queue-dock.spec.tsx +++ b/packages/client/ui-conversation/tests/queue-dock.spec.tsx @@ -23,7 +23,11 @@ const SID = 's1' as SessionId const iid = (id: string): QueueItemId => id as QueueItemId function row(id: string, text: string | null, preview = text ?? '[image]'): QueuedMessage { - return { id: iid(id), preview, text } + return { + id: iid(id), messageId: `message-${id}` as never, placement: 'queued', + content: text === null ? [{ type: 'image', data: 'x' } as never] : [{ type: 'text', text }], + preview, text, + } } function snapshotWith(queue: QueuedMessage[]): ConversationSnapshot { @@ -85,6 +89,14 @@ describe('QueueDock', () => { expect(container.innerHTML).toBe('') }) + it('leaves pending steering to the conversation flow', () => { + const steering = { ...row('s-1', 'interrupt'), placement: 'steering' as const } + const snap = snapshotWith([steering]) + const source = liveSession(snap) + const { container } = render() + expect(container.innerHTML).toBe('') + }) + it('renders one row directly and defaults multiple rows to a collapsible count header', () => { const single = snapshotWith([row('i-1', 'one')]) const source = liveSession(single) diff --git a/packages/client/ui-conversation/tests/skeleton.spec.tsx b/packages/client/ui-conversation/tests/skeleton.spec.tsx index eff655d229..1642425930 100644 --- a/packages/client/ui-conversation/tests/skeleton.spec.tsx +++ b/packages/client/ui-conversation/tests/skeleton.spec.tsx @@ -164,6 +164,7 @@ function mount( useInput={useInput} inputActions={inputActions} keyboard={wiring} + resolveSubmitMode={() => 'queue'} toggleCommandMenu={vi.fn()} useNotices={bindSnapshotSelector(wiring.notices)} useLexicon={bindSnapshotSelector(wiring.lexicon)} @@ -220,7 +221,7 @@ describe('ConversationRoot resident composer', () => { fireEvent.change(box, { target: { value: 'ordinary revised' } }) expect(b.chat.store.getSnapshot().draft).toBe('ordinary revised') fireEvent.keyDown(box, { key: 'Enter' }) - expect(b.sink).toHaveBeenCalledWith('ordinary revised') + expect(b.sink).toHaveBeenCalledWith('ordinary revised', 'queue') expect((b.view.getByRole('button', { name: 'Child' }) as HTMLButtonElement).disabled).toBe(true) expect(b.view.queryByText('Root')).toBeNull() }) diff --git a/packages/client/ui-conversation/tests/submission-policy.spec.ts b/packages/client/ui-conversation/tests/submission-policy.spec.ts new file mode 100644 index 0000000000..485840fc2e --- /dev/null +++ b/packages/client/ui-conversation/tests/submission-policy.spec.ts @@ -0,0 +1,65 @@ +// @vitest-environment jsdom +import { afterEach, describe, expect, it, vi } from 'vitest' +import { + BUSY_ENTER_STORAGE_KEY, ComposerSubmissionPolicy, DEFAULT_BUSY_ENTER_BEHAVIOR, +} from '../src/client/input/submission-policy.ts' + +afterEach(() => { + vi.unstubAllGlobals() + localStorage.clear() +}) + +describe('ComposerSubmissionPolicy', () => { + it('defaults to Queue and only applies the preference while running', () => { + const policy = new ComposerSubmissionPolicy() + expect(policy.busyEnter.getSnapshot()).toBe(DEFAULT_BUSY_ENTER_BEHAVIOR) + expect(policy.resolve(false, 'enter')).toBe('queue') + expect(policy.resolve(false, 'accelerated')).toBe('queue') + expect(policy.resolve(true, 'enter')).toBe('queue') + expect(policy.resolve(true, 'accelerated')).toBe('steer') + + const changed = vi.fn() + policy.busyEnter.subscribe(changed) + policy.setBusyEnter('steer') + expect(changed).toHaveBeenCalledTimes(1) + expect(policy.resolve(true, 'enter')).toBe('steer') + expect(policy.resolve(true, 'accelerated')).toBe('queue') + expect(policy.resolve(false, 'enter')).toBe('queue') + expect(policy.resolve(false, 'accelerated')).toBe('queue') + expect(localStorage.getItem(BUSY_ENTER_STORAGE_KEY)).toBe('steer') + }) + + it('restores a valid preference and leaves an identical write untouched', () => { + localStorage.setItem(BUSY_ENTER_STORAGE_KEY, 'steer') + const write = vi.spyOn(Storage.prototype, 'setItem') + const policy = new ComposerSubmissionPolicy() + expect(policy.busyEnter.getSnapshot()).toBe('steer') + policy.setBusyEnter('steer') + expect(write).not.toHaveBeenCalled() + write.mockRestore() + }) + + it('uses Queue for invalid, unavailable, or unreadable storage', () => { + localStorage.setItem(BUSY_ENTER_STORAGE_KEY, 'invalid') + expect(new ComposerSubmissionPolicy().busyEnter.getSnapshot()).toBe('queue') + + vi.stubGlobal('localStorage', undefined) + expect(new ComposerSubmissionPolicy().busyEnter.getSnapshot()).toBe('queue') + + vi.stubGlobal('localStorage', { + getItem: () => { throw new Error('blocked') }, + setItem: vi.fn(), + }) + expect(new ComposerSubmissionPolicy().busyEnter.getSnapshot()).toBe('queue') + }) + + it('keeps the in-memory preference when persistence throws', () => { + vi.stubGlobal('localStorage', { + getItem: () => null, + setItem: () => { throw new Error('quota') }, + }) + const policy = new ComposerSubmissionPolicy() + policy.setBusyEnter('steer') + expect(policy.busyEnter.getSnapshot()).toBe('steer') + }) +}) diff --git a/packages/host/apiproxy/src/api-proxy.ts b/packages/host/apiproxy/src/api-proxy.ts index 7d1c033393..e9808518fe 100644 --- a/packages/host/apiproxy/src/api-proxy.ts +++ b/packages/host/apiproxy/src/api-proxy.ts @@ -818,29 +818,28 @@ export function createApiProxy(ctx: Context, defaults: ApiProxyDefaults): ApiPro sessionId, items: items.map(item => ({ id: item.id, + placement: item.placement, message: item.message, })), }) } ctx.effect(() => { - const retire = (agent: Agent, item: InboxItem): boolean => { - const entries = queuedMirror.get(agent.id) - if (entries === undefined) { - rememberUnseen(agent.id, item.id, { kind: 'terminal' }) - return false - } - const index = entries.findIndex(entry => entry.id === item.id) - if (index === -1) { - rememberUnseen(agent.id, item.id, { kind: 'terminal' }) - return false - } + const retireKnown = (sessionId: SessionId, itemId: InboxItemId): boolean => { + const entries = queuedMirror.get(sessionId) + if (entries === undefined) return false + const index = entries.findIndex(entry => entry.id === itemId) + if (index === -1) return false entries.splice(index, 1) - if (entries.length === 0) queuedMirror.delete(agent.id) + if (entries.length === 0) queuedMirror.delete(sessionId) return true } + const retire = (agent: Agent, item: InboxItem): boolean => { + if (retireKnown(agent.id, item.id)) return true + rememberUnseen(agent.id, item.id, { kind: 'terminal' }) + return false + } const disposers = [ ctx.on('agent/inbox/enqueue', (agent: Agent, item: InboxItem) => { - if (item.placement !== 'queued') return const unseen = takeUnseen(agent.id, item.id) if (unseen?.kind === 'terminal') return let entries = queuedMirror.get(agent.id) @@ -866,7 +865,24 @@ export function createApiProxy(ctx: Context, defaults: ApiProxyDefaults): ApiPro publishQueue(agent.id) }), ctx.on('agent/inbox/dequeue', (agent: Agent, item: InboxItem) => { - if (retire(agent, item)) publishQueue(agent.id) + if (item.placement === 'steering') { + // AgentLoop appends the durable steering/message synchronously after + // this claim. Retain and retire the mirror row in the following + // microtask so any re-entrant snapshot and the Host's linear mux + // stream keep it visible until the durable event exists. + const present = queuedMirror.get(agent.id)?.some(entry => entry.id === item.id) === true + if (!present) { + retire(agent, item) + return + } + queueMicrotask(() => { + if (retireKnown(agent.id, item.id)) publishQueue(agent.id) + }) + } else if (retire(agent, item)) { + // Queued claims have no durable same-message handoff to order. + // Publish retirement synchronously as before. + publishQueue(agent.id) + } }), ctx.on('agent/inbox/discard', (agent: Agent, items: InboxItem[]) => { let changed = false @@ -2509,6 +2525,7 @@ export function createApiProxy(ctx: Context, defaults: ApiProxyDefaults): ApiPro sessionId, items: items.map(item => ({ id: item.id, + placement: item.placement, message: item.message, })), })) diff --git a/packages/host/apiproxy/src/api/events.schema.ts b/packages/host/apiproxy/src/api/events.schema.ts index 0fd97b346e..3d5e94c2c1 100644 --- a/packages/host/apiproxy/src/api/events.schema.ts +++ b/packages/host/apiproxy/src/api/events.schema.ts @@ -54,6 +54,7 @@ export const muxFrameSchema = z.discriminatedUnion('type', [ sessionId: sessionIdSchema, items: z.array(z.object({ id: inboxItemIdSchema, + placement: z.union([z.literal('queued'), z.literal('steering')]), message: messageSchema, })), }), diff --git a/packages/host/apiproxy/src/api/events.ts b/packages/host/apiproxy/src/api/events.ts index 43bebaa595..f4460fadc2 100644 --- a/packages/host/apiproxy/src/api/events.ts +++ b/packages/host/apiproxy/src/api/events.ts @@ -32,10 +32,12 @@ export type ToolEventView = | { for: 'call'; view: ToolCallView } | { for: 'result'; view: ToolResultView } -/** One pending queued occurrence in an authoritative queue snapshot. */ +/** One pending inbox occurrence in the authoritative `session/queue` snapshot. */ export interface QueuedInboxItem { - /** Agent-owned occurrence identity used by queue mutations. */ + /** Agent-owned occurrence identity; queue mutations address only `queued` items. */ id: InboxItemId + /** Agent-resolved FIFO placement; clients render queued and steering items on different surfaces. */ + placement: 'queued' | 'steering' /** Complete pending message; it is not durable until the Agent claims it. */ message: Message } @@ -71,11 +73,12 @@ export type MuxFrame = | { type: 'question/requested'; sessionId: SessionId; questions: AskUserQuestionItem[] } | { type: 'question/resolved'; sessionId: SessionId; questionRpcId: RpcId; outcome: 'answered' | 'cancelled' } /** - * Complete transient queue state after every enqueue, mutation, claim, or + * Complete transient inbox state after every enqueue, mutation, claim, or * discard. Pending work is not model-visible and therefore has no durable * session event; the whole snapshot makes edit, deletion, cancel, and - * reconnect converge through one authoritative signal. Pending steering is - * outside this Web queue projection. + * reconnect converge through one authoritative signal. `session/queue` + * covers both resolved placements: queued items render + * in QueueDock, while pending steering renders at the conversation tail. */ | { type: 'session/queue'; sessionId: SessionId; items: QueuedInboxItem[] } /** diff --git a/packages/host/apiproxy/tests/api-proxy-commands.spec.ts b/packages/host/apiproxy/tests/api-proxy-commands.spec.ts index 61e4c6f4d3..473d452327 100644 --- a/packages/host/apiproxy/tests/api-proxy-commands.spec.ts +++ b/packages/host/apiproxy/tests/api-proxy-commands.spec.ts @@ -379,7 +379,7 @@ describe('session/queue frames', () => { const liveFrames = (await collected).filter(frame => frame.type === 'session/queue') expect(liveFrames.map(frame => frame.items)).toEqual([ - [{ id: edited.id, message: edited.message }], + [{ id: edited.id, placement: edited.placement, message: edited.message }], ]) const replay = new AbortController() const replayFrames = await collect( @@ -393,8 +393,8 @@ describe('session/queue frames', () => { const agent = stubAgent(ctx) const live = new AbortController() const liveStream = api.events.mux({ rpcId: RpcId('t-mux-live'), payload: {} }, live.signal) - // subscribed baseline + one queued snapshot; pending steering stays off this wire. - const liveCollected = collect(liveStream, 2, live) + // subscribed baseline + one snapshot per accepted inbox occurrence. + const liveCollected = collect(liveStream, 3, live) const queued = inboxItem('i-1', inboxMessage('m-1', 'queued prompt'), 'queued') const steering = inboxItem('i-2', inboxMessage('m-2', 'steering prompt'), 'steering') @@ -406,7 +406,15 @@ describe('session/queue frames', () => { { type: 'session/queue', sessionId: agent.id, - items: [{ id: queued.id, message: queued.message }], + items: [{ id: queued.id, placement: 'queued', message: queued.message }], + }, + { + type: 'session/queue', + sessionId: agent.id, + items: [ + { id: queued.id, placement: 'queued', message: queued.message }, + { id: steering.id, placement: 'steering', message: steering.message }, + ], }, ]) @@ -414,7 +422,88 @@ describe('session/queue frames', () => { const replay = new AbortController() const replayFrames = await collect( api.events.mux({ rpcId: RpcId('t-mux-replay'), payload: {} }, replay.signal), 2, replay) - expect(replayFrames.filter(f => f.type === 'session/queue')).toEqual([liveFrames[0]]) + expect(replayFrames.filter(f => f.type === 'session/queue')).toEqual([liveFrames[1]]) + }) + + it('publishes the durable steering event before retiring its transient row', async () => { + const ctx = await harness() + const api = createApiProxy(ctx, DEFAULTS) + const agent = stubAgent(ctx) + const abort = new AbortController() + const collected = collect( + api.events.mux({ rpcId: RpcId('t-steering-order'), payload: {} }, abort.signal), 5, abort) + const steering = inboxItem('i-steering', inboxMessage('m-steering', 'interrupt now'), 'steering') + + agent.session.append('turn/start', { + turn: 1, + trigger: { kind: 'message', source: { kind: 'user' } }, + }) + ctx.emit('agent/inbox/enqueue', agent, steering) + ctx.emit('agent/inbox/dequeue', agent, steering) + agent.session.append('steering/message', { + turn: 1, + message: steering.message, + }, { surfaceOp: 'append' }) + + const frames = await collected + expect(frames.map(frame => frame.type)).toEqual([ + 'session/subscribed', + 'session/event', + 'session/queue', + 'session/event', + 'session/queue', + ]) + expect(frames[2]).toMatchObject({ + type: 'session/queue', + items: [{ id: steering.id, placement: 'steering' }], + }) + expect(frames[3]).toMatchObject({ + type: 'session/event', + event: { type: 'steering/message', data: { message: { id: steering.message.id } } }, + }) + expect(frames[4]).toMatchObject({ type: 'session/queue', items: [] }) + }) + + it('retains claimed steering in re-entrant snapshots until its durable event', async () => { + const ctx = await harness() + const api = createApiProxy(ctx, DEFAULTS) + const agent = stubAgent(ctx) + const steering = inboxItem('i-steering', inboxMessage('m-steering', 'interrupt now'), 'steering') + const queued = inboxItem('i-reentrant', inboxMessage('m-reentrant', 'later'), 'queued') + ctx.on('agent/inbox/dequeue', (subject, item) => { + if (subject === agent && item.id === steering.id) ctx.emit('agent/inbox/enqueue', agent, queued) + }) + const abort = new AbortController() + const collected = collect( + api.events.mux({ rpcId: RpcId('t-steering-reentrant-order'), payload: {} }, abort.signal), 6, abort) + + agent.session.append('turn/start', { + turn: 1, + trigger: { kind: 'message', source: { kind: 'user' } }, + }) + ctx.emit('agent/inbox/enqueue', agent, steering) + ctx.emit('agent/inbox/dequeue', agent, steering) + agent.session.append('steering/message', { + turn: 1, + message: steering.message, + }, { surfaceOp: 'append' }) + + const frames = await collected + expect(frames[3]).toMatchObject({ + type: 'session/queue', + items: [ + { id: steering.id, placement: 'steering' }, + { id: queued.id, placement: 'queued' }, + ], + }) + expect(frames[4]).toMatchObject({ + type: 'session/event', + event: { type: 'steering/message', data: { message: { id: steering.message.id } } }, + }) + expect(frames[5]).toMatchObject({ + type: 'session/queue', + items: [{ id: queued.id, placement: 'queued' }], + }) }) it('publishes edits in place in the authoritative order', async () => { @@ -434,10 +523,16 @@ describe('session/queue frames', () => { const frames = (await collected).filter(frame => frame.type === 'session/queue') expect(frames.map(frame => frame.items)).toEqual([ - [{ id: first.id, message: first.message }], - [{ id: first.id, message: first.message }, { id: second.id, message: second.message }], - [{ id: first.id, message: first.message }, { id: edited.id, message: edited.message }], - [{ id: first.id, message: first.message }], + [{ id: first.id, placement: first.placement, message: first.message }], + [ + { id: first.id, placement: first.placement, message: first.message }, + { id: second.id, placement: second.placement, message: second.message }, + ], + [ + { id: first.id, placement: first.placement, message: first.message }, + { id: edited.id, placement: edited.placement, message: edited.message }, + ], + [{ id: first.id, placement: first.placement, message: first.message }], ]) }) diff --git a/packages/host/apiproxy/tests/rpc-schemas.spec.ts b/packages/host/apiproxy/tests/rpc-schemas.spec.ts index 541ddab1c8..5c721b33d1 100644 --- a/packages/host/apiproxy/tests/rpc-schemas.spec.ts +++ b/packages/host/apiproxy/tests/rpc-schemas.spec.ts @@ -481,7 +481,7 @@ describe('events frame schemas', () => { { type: 'question/requested', sessionId: 's', questions: [{ id: 'q', question: 'Q?', options: [{ label: 'L' }], multiSelect: true }] }, { type: 'question/resolved', sessionId: 's', questionRpcId: 'r', outcome: 'answered' }, { type: 'session/queue', sessionId: 's', items: [ - { id: 'i1', message: { id: 'm1', role: 'user', content: [{ type: 'text', text: 'queued prompt' }], source: { kind: 'user', rpcId: 'r9' } } }, + { id: 'i1', placement: 'steering', message: { id: 'm1', role: 'user', content: [{ type: 'text', text: 'queued prompt' }], source: { kind: 'user', rpcId: 'r9' } } }, ] }, { type: 'session/projection', sessionId: 's', key: 'todos', value: [{ content: 'x', status: 'pending' }], seq: 7 }, { type: 'stream/error', error: { code: 'internal', message: 'm', details: {} } }, From 3a9f78c55eba0d27971492c84f085649637d2365 Mon Sep 17 00:00:00 2001 From: imccyu Date: Sun, 2 Aug 2026 15:50:05 +0800 Subject: [PATCH 06/13] feat(web): restore actions after steering consumption --- ...-29-addressable-queue-operations.i18n.yaml | 4 +-- ...2026-07-29-addressable-queue-operations.md | 2 +- ...6-07-29-addressable-queue-operations.zh.md | 2 +- ...026-07-30-web-queue-steer-action.i18n.yaml | 4 +-- .../2026-07-30-web-queue-steer-action.md | 10 +++---- .../2026-07-30-web-queue-steer-action.zh.md | 10 +++---- .../snapshots/steering/mid-steer.expected.md | 2 ++ .../snapshots/steering/settled.expected.md | 6 +++- .../client/ui-conversation/README.i18n.yaml | 4 +-- packages/client/ui-conversation/README.md | 2 +- packages/client/ui-conversation/README.zh.md | 2 +- .../src/client/chat/MessageIconActions.tsx | 26 +++++++++-------- .../src/client/chat/MessageItem.tsx | 28 ++++++++++++++----- .../tests/chat-branch-tails.spec.tsx | 16 +++++++++-- .../ui-conversation/tests/chat-view.spec.tsx | 15 ++++++++++ 15 files changed, 91 insertions(+), 42 deletions(-) diff --git a/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.i18n.yaml b/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.i18n.yaml index 217bcd32db..75c93214c7 100644 --- a/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.i18n.yaml +++ b/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.md -2026-07-29-addressable-queue-operations.md: e5fbacfa7de244daa1c5f8e504ee98a4a7730393 -2026-07-29-addressable-queue-operations.zh.md: a532d18d1b29fcebc7e9d8222da9eb5a162b8ec6 +2026-07-29-addressable-queue-operations.md: 04bcaa26be2f4f70c4003f6cc2fce4798499d205 +2026-07-29-addressable-queue-operations.zh.md: a16bf8667294de8d08da54ddb0d317bf9938e4eb diff --git a/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.md b/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.md index e5fbacfa7d..04bcaa26be 100644 --- a/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.md +++ b/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.md @@ -18,7 +18,7 @@ The Web queue rendered pending messages but could not edit or delete one row. `M **Queue addresses require a live ordinary-session Agent.** `session.updateQueue` queries only the mounted Agent registry and never resumes a cold session: an `InboxItemId` is process-local and cannot name work after restart or disposal. A session-backed subagent returns `agent-busy` before inbox access and retains its continuation owner; for ordinary sessions, a missing Agent and a driver-claimed occurrence both return `queue-item-not-found`. -**Web actions address Queue only.** The Host excludes pending steering from `session/queue`; steering retains its existing durable transcript path after consumption. QueueDock hides while empty, renders one pending occurrence directly, and defaults two or more occurrences to a collapsed `" 条排队消息"` header that expands or collapses the complete list. The header exposes `aria-expanded` and `aria-controls`; the expanded list scrolls within a 180px height bound. An active edit or mutation keeps its rows visible, and emptying the queue restores the collapsed default for the next queue. Visible rows expose edit, delete, and a running-only strict-steer action. The UI derives queue row and mutation types from the runtime `SessionFace` contract rather than importing the connection plugin, so plugin cooperation continues through services and snapshots. Edit is available only when all content blocks are text; the editor cannot silently drop non-text blocks. An editing row exposes only save and cancel, with Enter and Escape as their keyboard equivalents. Delete removes the exact occurrence, while strict steer preserves every content block and retires the row only through the authoritative snapshot. The Web stop action preserves pending Queue work; AgentLoop claims the next waking occurrence only after the interrupted turn reaches quiescence, and its dequeue event retires that row without a browser resend. The [Web Queue steer action](2026-07-30-web-queue-steer-action.md) owns the strict transfer contract. +**Web actions address Queue only.** The placement-aware `session/queue` snapshot carries both queued and pending-steering occurrences; QueueDock selects only queued items, while ChatView projects steering and retains the existing durable transcript path after consumption. QueueDock hides while empty, renders one pending occurrence directly, and defaults two or more occurrences to a collapsed `" 条排队消息"` header that expands or collapses the complete list. The header exposes `aria-expanded` and `aria-controls`; the expanded list scrolls within a 180px height bound. An active edit or mutation keeps its rows visible, and emptying the queue restores the collapsed default for the next queue. Visible rows expose edit, delete, and a running-only strict-steer action. The UI derives queue row and mutation types from the runtime `SessionFace` contract rather than importing the connection plugin, so plugin cooperation continues through services and snapshots. Edit is available only when all content blocks are text; the editor cannot silently drop non-text blocks. An editing row exposes only save and cancel, with Enter and Escape as their keyboard equivalents. Delete removes the exact occurrence, while strict steer preserves every content block and retires the row only through the authoritative snapshot. The Web stop action preserves pending Queue work; AgentLoop claims the next waking occurrence only after the interrupted turn reaches quiescence, and its dequeue event retires that row without a browser resend. The [Web Queue steer action](2026-07-30-web-queue-steer-action.md) owns the strict transfer and pending-projection contract. ## Alternatives considered diff --git a/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.zh.md b/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.zh.md index a532d18d1b..a16bf86672 100644 --- a/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.zh.md +++ b/.agents/notes/implemented/feature/2026-07-29-addressable-queue-operations.zh.md @@ -18,7 +18,7 @@ Web 队列能够渲染待处理消息,但无法编辑或删除其中某一行 **Queue 寻址要求普通会话的 Agent 存活。** `session.updateQueue` 只查询已挂载的 Agent 注册表,绝不恢复冷会话:`InboxItemId` 属于进程本地标识,无法在重启或资源释放后继续指向工作。由会话支撑的 subagent 会在访问 inbox 前返回 `agent-busy`,并保留其继续执行 owner;对于普通会话,Agent 缺失和单次入队项已被驱动器认领这两种情况都返回 `queue-item-not-found`。 -**Web 操作只面向 Queue。** Host 从 `session/queue` 中排除待处理 steering;steering 消费后仍沿用既有的持久 transcript(文本记录)路径。QueueDock 在队列为空时隐藏,只有一个待处理项时直接渲染该行,存在两个或更多待处理项时则默认收起为可展开或收起完整列表的 `" 条排队消息"` 表头。表头暴露 `aria-expanded` 和 `aria-controls`;展开后的列表以 180px 为高度上限,并可滚动。存在进行中的编辑或变更时,列表行会保持可见;队列清空后,下一次出现队列时会恢复默认收起状态。可见行暴露编辑、删除以及仅在运行期间可用的严格 steering 操作。UI 从运行时 `SessionFace` 契约派生队列行与变更类型,而不是导入连接插件,因此插件仍通过服务和快照协作。仅当所有内容块都是文本时才提供编辑功能;编辑器不能静默丢弃非文本块。编辑中的行只展示保存和取消操作,对应的键盘操作分别是 Enter 和 Escape。删除会移除对应的精确入队项;严格 steering 会保留每个内容块,并且只通过权威快照退役该行。Web 停止操作会保留待处理 Queue 工作;只有在被中断轮次达到完全停稳后,AgentLoop 才认领下一个可唤醒入队项,其出队事件会退役该行,无需浏览器重发。[Web Queue steering 操作](2026-07-30-web-queue-steer-action.md)负责定义严格转移契约。 +**Web 操作只面向 Queue。** 带 placement 的 `session/queue` 会同时携带 queued 和待处理 steering;QueueDock 只选择 queued 项,ChatView 则投影 steering,并在消费后沿用既有的持久 transcript(文本记录)路径。QueueDock 在队列为空时隐藏,只有一个待处理项时直接渲染该行,存在两个或更多待处理项时则默认收起为可展开或收起完整列表的 `" 条排队消息"` 表头。表头暴露 `aria-expanded` 和 `aria-controls`;展开后的列表以 180px 为高度上限,并可滚动。存在进行中的编辑或变更时,列表行会保持可见;队列清空后,下一次出现队列时会恢复默认收起状态。可见行暴露编辑、删除以及仅在运行期间可用的严格 steering 操作。UI 从运行时 `SessionFace` 契约派生队列行与变更类型,而不是导入连接插件,因此插件仍通过服务和快照协作。仅当所有内容块都是文本时才提供编辑功能;编辑器不能静默丢弃非文本块。编辑中的行只展示保存和取消操作,对应的键盘操作分别是 Enter 和 Escape。删除会移除对应的精确入队项;严格 steering 会保留每个内容块,并且只通过权威快照退役该行。Web 停止操作会保留待处理 Queue 工作;只有在被中断轮次达到完全停稳后,AgentLoop 才认领下一个可唤醒入队项,其出队事件会退役该行,无需浏览器重发。[Web Queue steering 操作](2026-07-30-web-queue-steer-action.md)负责定义严格转移与待处理投影契约。 ## 考虑过的替代方案 diff --git a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml index 1220cc39fa..555e2fb434 100644 --- a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml +++ b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md -2026-07-30-web-queue-steer-action.md: b84ee5acf472ba8482f6dfb7f466a905ddfd8f4f -2026-07-30-web-queue-steer-action.zh.md: fd4b4f8f8ab18db8a5372ff1a419fa186e29f399 +2026-07-30-web-queue-steer-action.md: 78d8c4d50daa4d162ff17c1efc03c32face8019a +2026-07-30-web-queue-steer-action.zh.md: 6b8af563ce527b8122b9f5e783a6ef700dcdd401 diff --git a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md index b84ee5acf4..78d8c4d50d 100644 --- a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md +++ b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md @@ -6,7 +6,7 @@ English | [中文](2026-07-30-web-queue-steer-action.zh.md) ## Problem -The Web composer originally queued every Enter submission while an agent ran. QueueDock already gives each pending message an addressable row, and the durable transcript already renders consumed `steering/message` events as user-style bubbles without message actions, but Web had neither an action connecting those two surfaces nor a direct composer gesture for choosing current-turn steering. +The Web composer originally queued every Enter submission while an agent ran. QueueDock already gives each pending message an addressable row, and the durable transcript already renders consumed `steering/message` events as user-style bubbles, but Web had neither an action connecting those two surfaces nor a direct composer gesture for choosing current-turn steering. Implementing the row action as a client-side delete followed by `session.prompt(mode: 'steer')` would split one user intent across two RPCs. Driver claim could win between them, the steer could fail after deletion, or the existing best-effort `agent.steer()` fallback could silently append a new Queue item after the original occurrence was removed. A send-now action must therefore distinguish current-turn steering from Queue promotion and preserve the original row when steering is no longer possible. @@ -16,7 +16,7 @@ Implementing the row action as a client-side delete followed by `session.prompt( Each non-editing QueueDock row exposes the upward-arrow action as “插话发送”. The action is enabled only while the session reports a running agent; mixed-content messages remain eligible because steering forwards the complete immutable `UserMessage` rather than the row's text projection. Edit and delete keep their existing behavior. -Activating the action requests strict current-turn steering for that exact `InboxItemId`. Success removes the Queue row through the authoritative Host snapshot. When AgentLoop drains it, the existing durable `steering/message` event renders the same user-style bubble without a separate durable presentation path. +Activating the action requests strict current-turn steering for that exact `InboxItemId`. Success removes the Queue row through the authoritative Host snapshot and immediately projects the same pending steering after the `Deep diving...` running-status row; that bubble offers Copy but no Fork because the message has no durable event sequence yet. Once AgentLoop drains it, the existing durable `steering/message` event takes over the same user-style bubble and restores its clock, Copy, and Fork without a separate durable presentation path. The running bit is only an interaction hint. AgentLoop's `acceptsNextStep` value is authoritative at the synchronous mutation boundary. If that window has closed, the operation leaves the Queue occurrence unchanged and returns a typed `steer-unavailable` error; if the driver already claimed the occurrence, it returns the existing `queue-item-not-found` error. The UI reports either race without optimistically removing the row. @@ -34,9 +34,9 @@ The action does not run `agent/prompt-submit`: choosing steering intentionally c `session.updateQueue` carries the `steer` action and maps the two negative outcomes to typed RPC errors. The conversion is one synchronous Agent operation; the Host never reconstructs it by combining remove and prompt calls. -The Host's existing `queuedMirror` remains the sole transient inbox authority. Its `session/queue` snapshot carries every live occurrence with `placement: 'queued' | 'steering'`: QueueDock renders only queued rows, while ChatView renders pending steering at the conversation tail without edit or delete actions. Reconnect replays the same snapshot, so this visibility does not require client optimism or a second registry. +The Host's existing `queuedMirror` remains the sole transient inbox authority. Its `session/queue` snapshot carries every live occurrence with `placement: 'queued' | 'steering'`: QueueDock renders only queued rows, while ChatView renders pending steering at the conversation tail after the `Deep diving...` running-status row, with Copy but without Fork, edit, or delete actions. Reconnect replays the same snapshot, so this visibility does not require client optimism or a second registry. -When AgentLoop claims pending steering, it emits `agent/inbox/dequeue` immediately before synchronously appending `steering/message`. The Host retires that steering row on the following microtask, allowing the durable session event to enter the linear mux stream first. ChatView matches the shared `MessageId` and suppresses the transient projection as soon as the durable node exists, so one bubble changes authority without a visible gap or duplicate; an append failure still retires the claimed row. +When AgentLoop claims pending steering, it emits `agent/inbox/dequeue` immediately before synchronously appending `steering/message`. The Host retires that steering row on the following microtask, allowing the durable session event to enter the linear mux stream first. ChatView matches the shared `MessageId` and suppresses the transient projection as soon as the durable node exists, so one bubble changes authority without a visible gap or duplicate; the durable projection restores the clock, Copy, and Fork against its logged event time and sequence, while an append failure still retires the claimed row. The existing `session.prompt(mode: 'steer')` contract remains best-effort for new input: outside the next-step window it becomes a waking follow-up. The composer carries an explicit `queue | steer` mode through slash adjudication and reference serialization before calling that contract. A browser-local submission policy owns the persisted busy-Enter preference and resolves plain versus accelerated Enter as complementary gestures; the Settings row and InputBar share that policy without duplicating storage or delivery-window authority. Only the Queue row action is strict, because failure can safely leave its already-pending message untouched. @@ -44,7 +44,7 @@ The existing `session.prompt(mode: 'steer')` contract remains best-effort for ne AgentLoop contract coverage holds prompt admission open, converts one exact queued occurrence, and proves the replacement steering occurrence keeps the message value, drains as `steering/message`, and never starts its former independent turn. It also pins unavailable-window retention, claimed-address rejection, and re-entrant cancellation lifecycle conservation. -Host schema and proxy tests cover the new action, both typed errors, placement-aware snapshots and reconnect replay, plus durable-before-retirement ordering. QueueDock tests cover running-state enablement, complete-content eligibility, failure retention, authoritative success retirement, and filtering of steering occurrences. ChatView tests cover the transient bubble and its single-copy handoff to the durable node. +Host schema and proxy tests cover the new action, both typed errors, placement-aware snapshots and reconnect replay, plus durable-before-retirement ordering. QueueDock tests cover running-state enablement, complete-content eligibility, failure retention, authoritative success retirement, and filtering of steering occurrences. ChatView tests and Web ARIA snapshots cover pending steering after the running-status row with Copy alone, plus its single-copy handoff to a durable node with clock, Copy, and Fork. The keyless Web steering scenario queues a message through the real composer while the first response streams, activates the row arrow, then uses `ask_user_question` as a stable pending-steering barrier. It proves the Host-backed pending bubble appears before admission, hands off to one durable interjection after the answer, and affects the next model request. Assembled composer scenarios prove default-mode Cmd+Enter reaches the same pending and durable path without creating a Queue row, while Steer-mode Cmd+Enter creates a Queue row instead. Settings and submission-policy coverage pin the default, persistence, busy-only scope, and complementary gesture mapping; Queue edit/delete scenarios continue to prove those actions are unchanged. diff --git a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md index fd4b4f8f8a..6b8af563ce 100644 --- a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md +++ b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md @@ -6,7 +6,7 @@ Status: implemented ## 问题 -Web composer 原本会在 agent 运行期间把所有 Enter 提交作为 Queue 入队。QueueDock 已经为每条待处理消息提供可寻址的行,持久 transcript(文本记录)也已能把消费后的 `steering/message` 事件渲染为不带消息操作的用户样式气泡,但 Web 既没有连接这两个界面的操作,也没有让用户从 composer 直接选择当前轮次 steering 的手势。 +Web composer 原本会在 agent 运行期间把所有 Enter 提交作为 Queue 入队。QueueDock 已经为每条待处理消息提供可寻址的行,持久 transcript(文本记录)也已能把消费后的 `steering/message` 事件渲染为用户样式气泡,但 Web 既没有连接这两个界面的操作,也没有让用户从 composer 直接选择当前轮次 steering 的手势。 如果 Web 先在客户端删除该行,再调用 `session.prompt(mode: 'steer')`,就会把用户的一次意图拆分到两个 RPC 中。驱动器可能在两次调用之间先认领该项,steering 投递也可能在删除后失败;现有尽力而为的 `agent.steer()` 回退还可能在原单次入队项被移除后,静默追加一个新的 Queue 项。因此,立即发送操作必须区分当前轮次 steering 与 Queue 前移,并在 steering 已不可用时保留原行。 @@ -16,7 +16,7 @@ Web composer 原本会在 agent 运行期间把所有 Enter 提交作为 Queue 每个非编辑态的 QueueDock 行都会提供名为“插话发送”的向上箭头操作。仅当会话报告 agent 正在运行时,该操作才会启用;包含混合内容的消息仍可使用,因为 steering 会转发完整且不可变的 `UserMessage`,而非该行的文本投影。编辑和删除保持现有行为。 -触发该操作会针对对应的 `InboxItemId` 请求严格的当前轮次 steering。操作成功后,权威 Host 快照会移除 Queue 行。AgentLoop 排空该项时,现有持久 `steering/message` 事件会渲染相同的用户样式气泡,无需另建持久展示路径。 +触发该操作会针对对应的 `InboxItemId` 请求严格的当前轮次 steering。操作成功后,权威 Host 快照会移除 Queue 行,并在 `Deep diving...` 运行状态行之后立即投影同一条待处理 steering;该气泡提供复制,但消息尚无持久事件序号,因此不提供 fork。AgentLoop 排空该项后,现有持久 `steering/message` 事件会接管同一个用户样式气泡,并恢复时钟、复制和 fork,无需另建持久展示路径。 running 标志位只用于提示交互状态。在同步变更边界上,AgentLoop 的 `acceptsNextStep` 值才是权威依据。如果该窗口已经关闭,操作会保持 Queue 单次入队项不变,并返回类型化的 `steer-unavailable` 错误;如果驱动器已经认领该项,则返回现有的 `queue-item-not-found` 错误。UI 会报告任一竞态,不会乐观地移除该行。 @@ -34,9 +34,9 @@ Composer 对新输入采用另一套尽力而为契约。所寻址会话空闲 `session.updateQueue` 会携带 `steer` 操作,并把两种负面结果映射为类型化 RPC 错误。这项转换是一次同步 Agent 操作;Host 绝不会通过组合移除和提示词调用来重建它。 -Host 仍以现有 `queuedMirror` 作为唯一的瞬态 inbox 权威。`session/queue` 快照会携带所有存活单次入队项及其 `placement: 'queued' | 'steering'`:QueueDock 只渲染 queued 行,ChatView 则在会话流末尾渲染待处理 steering,且不提供编辑或删除操作。重连会重放同一份快照,因此这项可见性既不依赖客户端乐观展示,也不需要第二个 registry。 +Host 仍以现有 `queuedMirror` 作为唯一的瞬态 inbox 权威。`session/queue` 快照会携带所有存活单次入队项及其 `placement: 'queued' | 'steering'`:QueueDock 只渲染 queued 行,ChatView 则在会话流末尾、`Deep diving...` 运行状态行之后渲染待处理 steering,提供复制操作,但不提供 fork、编辑或删除操作。重连会重放同一份快照,因此这项可见性既不依赖客户端乐观展示,也不需要第二个 registry。 -AgentLoop 认领待处理 steering 时,会在同步追加 `steering/message` 之前立即发出 `agent/inbox/dequeue`。Host 会等到下一个微任务才退役该 steering 行,让持久 session 事件先进入线性 mux 流。ChatView 会匹配两边共享的 `MessageId`,并在持久节点出现时立即抑制瞬态投影,因此同一个气泡切换权威时不会产生可见空档或重复;如果追加失败,已认领行仍会退役。 +AgentLoop 认领待处理 steering 时,会在同步追加 `steering/message` 之前立即发出 `agent/inbox/dequeue`。Host 会等到下一个微任务才退役该 steering 行,让持久 session 事件先进入线性 mux 流。ChatView 会匹配两边共享的 `MessageId`,并在持久节点出现时立即抑制瞬态投影,因此同一个气泡切换权威时不会产生可见空档或重复;持久投影会根据已记录的事件时间与序号恢复时钟、复制与 fork 操作,而追加失败时已认领行仍会退役。 现有 `session.prompt(mode: 'steer')` 对新输入仍采用尽力而为的契约:在 next-step 窗口之外,它会变为唤醒 agent 的后续轮次。Composer 会让显式 `queue | steer` 模式经过 slash 裁决与引用序列化,再调用该契约。浏览器本地的提交策略拥有持久化的繁忙态 Enter 偏好,并把普通 Enter 与加速 Enter 解析为互补手势;Settings 行和 InputBar 共享该策略,不重复实现存储或投递窗口权威。只有 Queue 行操作采用严格语义,因为失败时可以安全地保留其已经待处理的消息。 @@ -44,7 +44,7 @@ AgentLoop 认领待处理 steering 时,会在同步追加 `steering/message` AgentLoop 契约覆盖保持提示词接纳窗口打开,转换一个精确的 queued 单次入队项,并证明替代它的 steering 单次入队项保留消息值、以 `steering/message` 的形式排空,且绝不启动原本的独立轮次。该覆盖还钉住窗口不可用时保留原项、拒绝已被认领的地址,以及可重入取消下的生命周期守恒。 -Host schema 和代理测试覆盖新操作、两种类型化错误、带 placement 的快照与重连重放,以及先持久化再退役的顺序。QueueDock 测试覆盖按运行状态启用、混合内容消息仍可完整投递、失败时保留原行、成功后由权威快照退役,以及过滤 steering 单次入队项。ChatView 测试覆盖瞬态气泡及其只保留一份的持久节点交接。 +Host schema 和代理测试覆盖新操作、两种类型化错误、带 placement 的快照与重连重放,以及先持久化再退役的顺序。QueueDock 测试覆盖按运行状态启用、混合内容消息仍可完整投递、失败时保留原行、成功后由权威快照退役,以及过滤 steering 单次入队项。ChatView 测试与 Web ARIA 快照覆盖待处理阶段位于运行状态行之后且仅有复制,并覆盖它向带时钟、复制和 fork 的持久节点进行单份交接。 无密钥 Web steering 场景在第一次响应流式输出期间,通过真实 composer 排队一条消息并触发行上的箭头,再用 `ask_user_question` 作为稳定的待处理 steering 屏障。该场景证明 Host 支撑的待处理气泡会在准入前出现,在回答后交接为唯一一条持久插话,并影响下一次模型请求。组装后的 composer 场景证明默认模式下的 Cmd+Enter 无需创建 Queue 行,也会进入同一条待处理与持久路径;Steer 模式下的 Cmd+Enter 则会创建 Queue 行。Settings 与提交策略覆盖会固定默认值、持久化、仅限繁忙态的作用域和互补手势映射;Queue 编辑/删除场景继续证明这些操作没有变化。 diff --git a/apps/web/tests/snapshots/steering/mid-steer.expected.md b/apps/web/tests/snapshots/steering/mid-steer.expected.md index 9f077b21a5..61f48ffbab 100644 --- a/apps/web/tests/snapshots/steering/mid-steer.expected.md +++ b/apps/web/tests/snapshots/steering/mid-steer.expected.md @@ -24,6 +24,8 @@ - text: Ask question waiting - status: Deep diving... - text: "Interjection: include the word BANANA in your final reply." +- button "Copy": + - img - region "Ready to continue?": - text: Checkpoint - heading "Ready to continue?" [level=2] diff --git a/apps/web/tests/snapshots/steering/settled.expected.md b/apps/web/tests/snapshots/steering/settled.expected.md index 72e28598e5..be28e218d3 100644 --- a/apps/web/tests/snapshots/steering/settled.expected.md +++ b/apps/web/tests/snapshots/steering/settled.expected.md @@ -21,7 +21,11 @@ - img - img - text: Ask question 1/1 answered -- text: "Interjection: include the word BANANA in your final reply." +- text: "Interjection: include the word BANANA in your final reply. {{clock}}" +- button "Copy": + - img +- button "Branch into a new conversation": + - img - button "Think The user selected \"Yes\" and wants me to include the word \"BANANA\" in my final reply. Let me acknowledge their answer.": - img - img diff --git a/packages/client/ui-conversation/README.i18n.yaml b/packages/client/ui-conversation/README.i18n.yaml index f3f3c43124..fcf4c0b84e 100644 --- a/packages/client/ui-conversation/README.i18n.yaml +++ b/packages/client/ui-conversation/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/client/ui-conversation/README.md -README.md: c80ca786fb289055fcb3ad2b5a7aab64127ca485 -README.zh.md: 601139a00d2d40fb34e0cdcb8554984f7db9a724 +README.md: 61056fc36e2e7f5f85726e158b9fd27e06c21941 +README.zh.md: 3a1a74a4106321415ee343908c999a2cf41615a0 diff --git a/packages/client/ui-conversation/README.md b/packages/client/ui-conversation/README.md index c80ca786fb..61056fc36e 100644 --- a/packages/client/ui-conversation/README.md +++ b/packages/client/ui-conversation/README.md @@ -36,7 +36,7 @@ The todo surfaces are two registrations over that shape, both plain registrant p `QueueDock` is the terminal input-dock entry at `order: 20`. It hides while empty, renders one pending row directly, and defaults two or more rows to a collapsed `" 条排队消息"` header whose button expands or collapses the complete list. The header exposes `aria-expanded` and `aria-controls`; the expanded list scrolls within a 180px height bound. An active edit or mutation keeps its rows visible, and emptying the queue restores the collapsed default for the next queue. Each visible row remains a single-line preview with its exact-occurrence edit, delete, and strict-steer actions; an unavailable steering window leaves the Queue occurrence in place and reports the failure. -The Host's placement-aware `session/queue` snapshot also carries pending steering. QueueDock filters it out, while ChatView projects it as a user-style bubble at the conversation tail. The Host delays steering retirement until the durable `steering/message` has entered the mux stream, and ChatView deduplicates the two projections by their shared `MessageId`; the bubble therefore hands off without a gap or duplicate, while reconnect restores pending state from the same authority. +The Host's placement-aware `session/queue` snapshot also carries pending steering. QueueDock filters it out, while ChatView projects it as a user-style bubble with Copy at the conversation tail; Fork stays absent because the message has not entered a durable turn. The Host delays steering retirement until the durable `steering/message` has entered the mux stream, and ChatView deduplicates the two projections by their shared `MessageId`; the bubble therefore hands off without a gap or duplicate, restores Copy and Fork from the durable node, and survives reconnect from the same authority. Keyboard message submission resolves delivery from the addressed session's running state. While idle, Enter and Cmd/Ctrl+Enter both perform an ordinary Queue send. While running, the browser-persisted General Settings preference assigns plain Enter to `Queue` (the default) or `Steer`, and Cmd/Ctrl+Enter performs the other behavior; Shift+Enter remains a newline. The preference affects only the busy-state gesture pair, and the send button and non-keyboard submit actions remain Queue. Composer Steer uses the existing best-effort `session.prompt(mode: 'steer')` contract: if the current next-step window closes before acceptance, AgentLoop admits the message as the next waking Queue turn without surfacing a failure or losing the draft transaction. diff --git a/packages/client/ui-conversation/README.zh.md b/packages/client/ui-conversation/README.zh.md index 601139a00d..3a1a74a410 100644 --- a/packages/client/ui-conversation/README.zh.md +++ b/packages/client/ui-conversation/README.zh.md @@ -36,7 +36,7 @@ todo 两个面就是在该形状上的两个注册项,都是普通注册方插 `QueueDock` 是 `order: 20` 的末端 input-dock 条目。队列为空时隐藏;只有一个待处理项时直接渲染该行;存在两个或更多待处理项时,默认收起为 `" 条排队消息"` 表头,其按钮可展开或收起完整列表。表头暴露 `aria-expanded` 和 `aria-controls`;展开后的列表以 180px 为高度上限,并可滚动。存在进行中的编辑或变更时,列表行会保持可见;队列清空后,下一次出现队列时会恢复默认收起状态。每条可见行仍是单行预览,并提供针对精确单次入队项的编辑、删除和严格 steering(中途引导)操作;steering 窗口不可用时,Queue 单次入队项会留在原处并显示失败。 -Host 带 placement 的 `session/queue` 快照也会携带待处理 steering。QueueDock 会将其过滤掉,ChatView 则把它投影为会话流末尾的用户样式气泡。Host 会等持久 `steering/message` 进入 mux 流之后再退役 steering,ChatView 则按两份投影共享的 `MessageId` 去重;气泡交接时因而不会产生空档或重复,重连也能从同一权威恢复待处理状态。 +Host 带 placement 的 `session/queue` 快照也会携带待处理 steering。QueueDock 会将其过滤掉,ChatView 则把它投影为会话流末尾带复制操作的用户样式气泡;消息尚未进入持久轮次,因此不显示 fork。Host 会等持久 `steering/message` 进入 mux 流之后再退役 steering,ChatView 则按两份投影共享的 `MessageId` 去重;气泡交接时因而不会产生空档或重复,会从持久节点恢复复制与 fork 操作,并能在重连后从同一权威恢复。 键盘消息提交会根据所寻址会话的运行状态解析投递方式。空闲时,Enter 和 Cmd/Ctrl+Enter 都执行普通 Queue 发送。运行期间,浏览器持久化的 General Settings 偏好会把普通 Enter 分配为 `Queue`(默认值)或 `Steer`,Cmd/Ctrl+Enter 则执行另一种行为;Shift+Enter 仍然换行。该偏好只影响繁忙态下这对手势,发送按钮与非键盘提交操作仍使用 Queue。Composer Steer 复用现有尽力而为的 `session.prompt(mode: 'steer')` 契约:如果当前 next-step 窗口在接纳前关闭,AgentLoop 会把消息接纳为下一条唤醒 Queue 轮次,不显示失败,也不会丢失草稿事务。 diff --git a/packages/client/ui-conversation/src/client/chat/MessageIconActions.tsx b/packages/client/ui-conversation/src/client/chat/MessageIconActions.tsx index cd6b0a36ce..91ba066544 100644 --- a/packages/client/ui-conversation/src/client/chat/MessageIconActions.tsx +++ b/packages/client/ui-conversation/src/client/chat/MessageIconActions.tsx @@ -1,5 +1,5 @@ -// Shared IconActions chrome for user and assistant messages: copy live, -// branch wired through onBranch, date-aware clock. +// Shared IconActions chrome for user, steering, and assistant messages: copy +// live, optional branch wiring, and an optional date-aware clock. import { useCallback } from 'react' import { @@ -13,12 +13,14 @@ import css from './MessageIconActions.module.css' export interface MessageIconActionsProps { /** Plain text the copy action writes. */ text: string - /** Unix epoch ms for the clock label. */ - time: number + /** Unix epoch ms for the clock label; omitted for transient messages. */ + time?: number | undefined /** Clock before icons (user) or after (assistant). */ clock: 'start' | 'end' /** Fork the session at this message. */ onBranch?: (() => void) | undefined + /** Whether to render the branch action; defaults to true. */ + showBranch?: boolean | undefined /** Parent layout class composed onto the actions row. */ className?: string | undefined /** The owning view's locale seat, passed down as a plain prop. */ @@ -31,13 +33,13 @@ export interface MessageIconActionsProps { * @returns The actions row element. */ export function MessageIconActions({ - text, time, clock, onBranch, className, t, + text, time, clock, onBranch, showBranch = true, className, t, }: MessageIconActionsProps) { const day = useCalendarDay() const onCopy = useCallback(() => { void writeClipboard(text) }, [text]) - const clockEl = ( + const clockEl = time === undefined ? null : ( {formatMessageClock(time, t, day)} @@ -50,11 +52,13 @@ export function MessageIconActions({ - - - + {showBranch && ( + + + + )} {clock === 'end' ? clockEl : null}
) diff --git a/packages/client/ui-conversation/src/client/chat/MessageItem.tsx b/packages/client/ui-conversation/src/client/chat/MessageItem.tsx index 5783d0211d..6e542bf468 100644 --- a/packages/client/ui-conversation/src/client/chat/MessageItem.tsx +++ b/packages/client/ui-conversation/src/client/chat/MessageItem.tsx @@ -1,6 +1,6 @@ -// MessageItem: simple chat nodes — user bubble (right-aligned, with -// clock + copy / branch IconActions), steering (same bubble, no actions), -// context injection, compaction marker, retry disclosure, and +// MessageItem: simple chat nodes — user and consumed-steering bubbles +// (right-aligned, with clock + copy / branch IconActions), pending steering +// (copy only), context injection, compaction marker, retry disclosure, and // unknown-surface JSON rows. import { memo, useEffect, useMemo, useState } from 'react' @@ -167,7 +167,7 @@ function projectUserText(text: string): ReactNode { return <>{parts} } -/** Right-aligned bubble shared by user and steering rows (steering has no actions). */ +/** Right-aligned bubble shared by user and steering rows. */ function UserStyleBubble({ content, actions, pending = false, t, }: { @@ -201,7 +201,22 @@ export function PendingSteeringBubble({ content, t }: { content: readonly unknown[] t: ChatViewSlotProps['t'] }): ReactNode { - return + return ( + ( + + )} + /> + ) } export const MessageItem = memo(function MessageItem({ @@ -210,6 +225,7 @@ export const MessageItem = memo(function MessageItem({ const truncated = (total: number): string => t('json.truncated', { total }) switch (node.kind) { case 'user': + case 'steering': return ( ) - case 'steering': - return case 'context': return ( diff --git a/packages/client/ui-conversation/tests/chat-branch-tails.spec.tsx b/packages/client/ui-conversation/tests/chat-branch-tails.spec.tsx index a829f6f04a..4074996063 100644 --- a/packages/client/ui-conversation/tests/chat-branch-tails.spec.tsx +++ b/packages/client/ui-conversation/tests/chat-branch-tails.spec.tsx @@ -102,18 +102,28 @@ describe('MessageItem arms', () => { fireEvent.click(screen.getByRole('button', { name: '复制' })) }) - it('steering bubbles render text and non-text rest blocks, without user actions or a badge', () => { + it('consumed steering renders copy and branch actions without a badge', () => { + const writeText = vi.fn().mockResolvedValue(undefined) + Object.defineProperty(navigator, 'clipboard', { + configurable: true, + value: { writeText }, + }) + const fork = vi.fn() const view = render( , ) expect(view.queryByText('插话')).toBeNull() expect(view.getByText('steer!')).toBeTruthy() expect(view.getByText(/附加内容块/)).toBeTruthy() - expect(view.queryByRole('button', { name: '复制' })).toBeNull() + fireEvent.click(view.getByRole('button', { name: '复制' })) + expect(writeText).toHaveBeenCalledWith('steer!') + fireEvent.click(view.getByRole('button', { name: '在新对话中分支' })) + expect(fork).toHaveBeenCalledWith(2) }) it('context uses the Tool calls disclosure chrome and keeps its JSON collapsed by default', () => { diff --git a/packages/client/ui-conversation/tests/chat-view.spec.tsx b/packages/client/ui-conversation/tests/chat-view.spec.tsx index 507b301922..b0e55736a4 100644 --- a/packages/client/ui-conversation/tests/chat-view.spec.tsx +++ b/packages/client/ui-conversation/tests/chat-view.spec.tsx @@ -252,6 +252,11 @@ describe('ChatView', () => { }) it('renders Host-pending steering at the flow tail and hands off to the durable node', () => { + const writeText = vi.fn().mockResolvedValue(undefined) + Object.defineProperty(navigator, 'clipboard', { + configurable: true, + value: { writeText }, + }) const pending = { id: 'steer-occurrence' as never, messageId: 'steer-message' as never, @@ -273,6 +278,11 @@ describe('ChatView', () => { expect(view.getByText('interrupt now').closest('[data-pending-steering]')).not.toBeNull() expect(view.queryByText('later')).toBeNull() + const pendingBubble = view.getByText('interrupt now').closest('[data-pending-steering]') + expect(pendingBubble).not.toBeNull() + fireEvent.click(within(pendingBubble as HTMLElement).getByRole('button', { name: '复制' })) + expect(writeText).toHaveBeenCalledWith('interrupt now') + expect(within(pendingBubble as HTMLElement).queryByRole('button', { name: '在新对话中分支' })).toBeNull() expect(view.getByRole('status').compareDocumentPosition(view.getByText('interrupt now')) & Node.DOCUMENT_POSITION_FOLLOWING).not.toBe(0) @@ -291,6 +301,11 @@ describe('ChatView', () => { }) expect(view.getAllByText('interrupt now')).toHaveLength(1) expect(view.container.querySelector('[data-pending-steering]')).toBeNull() + expect(view.getAllByRole('button', { name: '复制' })).toHaveLength(2) + const branchButtons = view.getAllByRole('button', { name: '在新对话中分支' }) + expect(branchButtons).toHaveLength(2) + fireEvent.click(branchButtons[1]!) + expect(h.forkAt).toHaveBeenCalledWith(2) act(() => { h.set({ queue: [queued] }) }) expect(view.getAllByText('interrupt now')).toHaveLength(1) From 53531b9af2a4cb84ab814aa7551fa1539c72f415 Mon Sep 17 00:00:00 2001 From: imccyu Date: Sun, 2 Aug 2026 16:24:23 +0800 Subject: [PATCH 07/13] test(snapshot): refresh Cordis steering API fixture --- .../tests/snapshots/cordis-inspect-jsdoc/session.jsonl | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/acp-agent/tests/snapshots/cordis-inspect-jsdoc/session.jsonl b/examples/acp-agent/tests/snapshots/cordis-inspect-jsdoc/session.jsonl index 65d3fbc960..2bbe6dd86b 100644 --- a/examples/acp-agent/tests/snapshots/cordis-inspect-jsdoc/session.jsonl +++ b/examples/acp-agent/tests/snapshots/cordis-inspect-jsdoc/session.jsonl @@ -13,7 +13,7 @@ {"type":"assistant/chunk","seq":11,"time":1785487622703,"data":{"turn":1,"step":1,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} {"type":"assistant/message","seq":12,"time":1785487622703,"data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"tool-call","id":"inspect-tools-api","name":"cordis_inspect","arguments":"{\"what\":\"api\",\"name\":\"tools\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"7c207b09-7f6e-4e53-a5d2-77e0d2bbb474"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[7,8,9,10,11],"surfaceOp":"append"} {"type":"tool/call","seq":13,"time":1785487622703,"data":{"turn":1,"step":1,"callId":"inspect-tools-api","name":"cordis_inspect","arguments":"{\"what\":\"api\",\"name\":\"tools\"}"}} -{"type":"tool/result","seq":14,"time":1785487622726,"data":{"turn":1,"step":1,"message":{"source":{"kind":"tool","callId":"inspect-tools-api"},"content":[{"type":"tool-result","toolCallId":"inspect-tools-api","content":[{"type":"text","text":"## api\n- tools — Tool registry and execution pipeline.\n /**\n * Register globally or in the calling agent scope. Scoped tools shadow\n * globals; duplicates within one layer and the reserved `run_code` name fail.\n * @param definition - tool schema, execution, and optional finalization/presentation callbacks.\n * @returns the exact disposer that unregisters the tool.\n */\n register(definition: ToolDefinition): () => void\n /**\n * Restrict global tools for the calling agent scope. Empty filters, unknown\n * names, scope-local names, and reserved transport names fail. Restrictions\n * intersect; scoped registrations remain visible.\n * @param filter - global-surface mask: `allow` (keep only) and/or `deny` (remove).\n * @returns the exact disposer that lifts this restriction.\n */\n restrict(filter: ToolRestriction): () => void\n /**\n * Register a monotonic guard after the extensible `tools/pre-execute`\n * waterfall. A plain-context guard applies globally; one registered through\n * `agent.ctx` applies only to that agent. Any matching guard may deny by\n * returning a reason, while no guard can force-allow a call another guard\n * denied. The exact effect disposer is returned for ordered ownership and\n * HMR cleanup.\n * @param guard - synchronous check; a returned string denies the execution.\n * @returns the exact disposer that unregisters the guard.\n */\n guard(guard: ToolGuard): () => void\n /**\n * Look up a tool as one scope sees it (scoped\n * shadows global; a restricted-away global reads as absent). Presenters pass\n * the calling agent so the rendered card matches the definition that\n * actually executed.\n * @param name - the tool name as registered.\n * @param scope - the viewing scope (the agent); omitted = the global view.\n * @returns the definition the scope resolves, or undefined when none is visible.\n */\n get(name: string, scope?: ScopeKey): ToolDefinition | undefined\n /**\n * Project visible definitions onto the allowlisted model-facing schema fields,\n * excluding execution and presentation callbacks.\n * @param scope - the viewing scope (the agent); omitted = the global view.\n * @returns one deep-cloned schema per visible tool.\n */\n schemas(scope?: ScopeKey): ToolSchema[]\n /**\n * Classify a pending call through the caller's visible tool definition. Only\n * an exact `true` is parallel; unknown, hidden, undeclared, invalid, or\n * throwing classifiers are exclusive.\n * @param exec - call name, parsed arguments, and optional agent scope.\n * @returns the fail-closed scheduling mode.\n */\n executionMode(exec: ToolExecutionInput): ToolExecutionMode\n /**\n * Execute through pre-policy, guards, around-dispatch, post-policy,\n * definition-owned content finalization, and final notification. Tool and\n * listener failures resolve as materialized error results; an invisible tool\n * reports `UNKNOWN_TOOL`. The returned outcome is the same lossless, frozen\n * snapshot final observers receive. Cancellation\n * arriving after entry and before final result materialization skips a\n * not-yet-started body with `ABORTED_BEFORE_DISPATCH` or replaces a\n * successful started outcome with `ABORTED`; already-started work is still\n * drained and may retain a tool-owned structured error.\n * @param exec - the typed same-process call input. The registry assigns its\n * correlation token before policy begins.\n * @returns the materialized final result.\n */\n async execute(exec: ToolExecutionInput): Promise\ntype shapes (referenced by the signatures above — read these before assuming a field is a string):\n export interface Agent {\n readonly id: SessionId;\n readonly options: AgentOptions;\n readonly session: Session;\n readonly status: AgentStatus;\n readonly acceptsNextStep: boolean;\n readonly ctx: Context;\n send(message: UserMessage, options: SendOptions): void;\n reserveTurnAdmission(): (() => void) | undefined;\n updateInbox(id: InboxItemId, action: InboxAction): InboxActionResult;\n cancel(cause: AgentCancelCause, options?: CancelOptions): void;\n whenIdle(): Promise;\n followup(message: UserMessage): void;\n steer(message: UserMessage): SteeringReceipt;\n inject(message: UserMessage): void;\n }\n export type AgentCancelCause = {\n readonly kind: 'user';\n } | {\n readonly kind: 'parent';\n };\n export interface AgentOptions {\n provider?: string;\n model?: string;\n maxTokens?: number;\n }\n export type AgentStatus = 'idle' | 'running';\n export interface AssistantMessage extends Message {\n readonly role: 'assistant';\n readonly source: ModelMessageSource;\n }\n export interface AssistantProvenance {\n provider: string;\n model: string;\n replayState?: unknown;\n }\n export type Branded = string & {\n readonly [BRAND]: B;\n };\n export type CallId = Branded<'CallId'>;\n export interface CancelOptions {\n keepInbox?: boolean;\n }\n export interface ContentBlockMap {\n 'text': TextBlock;\n 'reasoning': ReasoningBlock;\n 'tool-call': ToolCallBlock;\n 'tool-result': ToolResultBlock;\n }\n export type ContentBlockType = keyof ContentBlockMap;\n export interface DiffCallView {\n card: 'diff';\n title: string;\n diffs: FileDiff[];\n locations?: FileLocation[];\n }\n export interface DiffResultView {\n card: 'diff';\n title?: string;\n diffs: FileDiff[];\n }\n export interface EpochHeader {\n config: LlmCallConfig;\n adapterDefaults?: LlmCallConfigAdapterDefaults;\n system?: string;\n tools?: ToolSchema[];\n }\n export interface FileDiff {\n path: string;\n oldText: string | null;\n newText: string;\n }\n export interface FileLocation {\n path: string;\n line?: number;\n }\n export type FinishReason = FinishReasonMap[keyof FinishReasonMap];\n export interface FinishReasonMap {\n 'stop': {\n kind: 'stop';\n };\n 'tool-calls': {\n kind: 'tool-calls';\n };\n 'max-tokens': {\n kind: 'max-tokens';\n };\n 'aborted': {\n kind: 'aborted';\n failure: LlmFailure;\n };\n 'error': {\n kind: 'error';\n failure: LlmFailure;\n };\n }\n export interface GenericCallView {\n card: 'generic';\n title: string;\n kind?: ToolCallKind;\n rawInput?: unknown;\n content?: ContentBlock[];\n locations?: FileLocation[];\n }\n export interface GenericResultView {\n card: 'generic';\n title?: string;\n content?: ContentBlock[];\n }\n export type InboxAction = {\n readonly kind: 'edit';\n readonly content: ContentBlock[];\n } | {\n readonly kind: 'remove';\n };\n export type InboxActionResult = 'applied' | 'not-found';\n export type InboxItemId = Branded<'InboxItemId'>;\n export interface JsonSchemaNode {\n type?: JsonSchemaType;\n oneOf?: JsonSchemaNode[];\n properties?: Record;\n required?: string[];\n additionalProperties?: boolean;\n items?: JsonSchemaNode;\n enum?: JsonSchemaScalar[];\n const?: JsonSchemaScalar;\n description?: string;\n title?: string;\n default?: JsonValue;\n examples?: JsonValue;\n }\n export type JsonSchemaScalar = string | number | boolean | null;\n export type JsonSchemaType = 'object' | 'array' | 'string' | 'number' | 'integer' | 'boolean' | 'null';\n export type JsonValue = null | boolean | number | string | JsonValue[] | {\n [key: string]: JsonValue;\n };\n export interface LlmCallConfig {\n provider: string;\n model: string;\n reasoningEffort?: ReasoningEffortId;\n temperature?: number;\n maxTokens?: number;\n stop?: string[];\n }\n export interface LlmCallConfigAdapterDefaults {\n reasoningEffort?: true;\n maxTokens?: true;\n }\n export interface LlmFailure {\n readonly message: string;\n readonly code: string;\n readonly status?: number;\n readonly providerRetryAfterMs?: number;\n readonly requestId?: ProviderRequestId;\n }\n export interface Message {\n readonly id: MessageId;\n readonly role: 'system' | 'user' | 'assistant';\n readonly content: ContentBlock[];\n readonly source: MessageSource;\n }\n export type MessageId = Branded<'MessageId'>;\n export type MessageSource = MessageSourceMap[keyof MessageSourceMap];\n export interface MessageSourceMap {\n user: {\n kind: 'user';\n };\n plugin: {\n kind: 'plugin';\n plugin: string;\n };\n model: ModelMessageSource;\n tool: ToolMessageSource;\n }\n export interface ModelMessageSource extends AssistantProvenance {\n kind: 'model';\n }\n export type ProviderRequestId = Branded<'ProviderRequestId'>;\n export interface ReadFileLine {\n number: number;\n text: string;\n }\n export interface ReadResultView {\n card: 'read';\n title?: string;\n path: string;\n offset: number;\n lines: ReadFileLine[];\n totalLines: number;\n lang?: string;\n content?: ContentBlock[];\n }\n export interface ReasoningBlock {\n type: 'reasoning';\n text: string;\n }\n export type ReasoningEffortId = Branded<'ReasoningEffortId'>;\n export interface RequestContext {\n provider: string;\n model: string;\n contextWindow?: number;\n }\n export type RequestHeaderReason = 'initial' | 'resume' | 'change';\n export type ScopeKey = object;\n export interface SearchFileMatches {\n path: string;\n matches: SearchLineMatch[];\n }\n export interface SearchLineMatch {\n lineNumber: number;\n line: string;\n }\n export interface SearchMatchesResultView {\n card: 'search';\n shape: 'matches';\n title?: string;\n files: SearchFileMatches[];\n truncated: boolean;\n total: number;\n }\n export interface SearchPathsResultView {\n card: 'search';\n shape: 'paths';\n title?: string;\n paths: string[];\n truncated: boolean;\n total: number;\n }\n export type SearchResultView = SearchMatchesResultView | SearchPathsResultView;\n export interface SendOptions {\n target: SendTarget;\n wakeup: boolean;\n }\n export type SendTarget = 'next-turn' | 'next-step';\n export class Session {\n get surface(): SessionSurface;\n readonly header: SessionHeader;\n get id(): SessionId;\n readonly firstLiveSeq: number;\n constructor(id: SessionId, seed?: readonly SessionEvent[], header?: SessionHeader);\n get events(): readonly SessionEvent[];\n get seq(): number;\n append(type: T, data: SessionEventMap[T], ...opts: T extends SurfaceEventType ? [\n opts: SurfaceIntent\n ] : [\n ]): SessionEvent;\n requestHeader(): EpochHeader | undefined;\n requestContext(): RequestContext | undefined;\n deriveMessages(): Message[];\n deriveEventMessage(event: SessionEvent): Message | null;\n }\n export type SessionEvent = {\n [K in SessionEventType]: {\n type: K;\n seq: number;\n time: number;\n data: SessionEventMap[K];\n } & (K extends SurfaceEventType ? {\n sourceEventSeqs?: number[];\n surfaceOp?: SurfaceOp;\n } : object);\n }[T];\n export interface SessionEventMap {\n 'turn/start': {\n turn: number;\n trigger: TurnTrigger;\n };\n 'turn/end': {\n turn: number;\n reason: TurnEndReason;\n };\n 'step/start': {\n turn: number;\n step: number;\n };\n 'step/end': {\n turn: number;\n step: number;\n };\n 'user/message': UserMessage;\n 'assistant/chunk': {\n turn: number;\n step: number;\n chunk: StreamChunk;\n };\n 'assistant/message': {\n turn: number;\n step: number;\n message: AssistantMessage;\n usage?: TokenUsage;\n };\n 'tool/call': {\n turn: number;\n step: number;\n callId: CallId;\n name: string;\n arguments: string;\n };\n 'tool/result': {\n turn: number;\n step: number;\n message: ToolResultMessage;\n error?: {\n name: string;\n code: string;\n };\n meta?: JsonValue;\n };\n 'steering/message': {\n turn: number;\n message: UserMessage;\n };\n 'todo/write': {\n todos: TodoItem[];\n };\n 'request/header': {\n header: EpochHeader;\n reason: RequestHeaderReason;\n };\n 'request/context': RequestContext;\n 'session/end-seed': Record;\n }\n export type SessionEventType = keyof SessionEventMap;\n export interface SessionHeader {\n readonly version: number;\n readonly id: SessionId;\n readonly createdAt: number;\n readonly cwd?: string;\n readonly parentSession?: SessionId;\n readonly seedLength?: number;\n readonly origin?: 'subagent';\n readonly delegationDepth?: number;\n }\n export type SessionId = Branded<'SessionId'>;\n export interface SessionSurface {\n readonly nodes: readonly number[];\n readonly replaceGeneration: number;\n }\n export type SteeringOutcome = {\n readonly status: 'admitted';\n readonly turn: number;\n readonly step: number;\n } | {\n readonly status: 'rejected';\n };\n export interface SteeringReceipt {\n readonly outcome: Promise;\n }\n export type StreamChunk = {\n type: 'block-start';\n index: number;\n blockType: ContentBlockType;\n } | {\n type: 'text-delta';\n index: number;\n text: string;\n } | {\n type: 'reasoning-delta';\n index: number;\n text: string;\n } | {\n type: 'tool-call-delta';\n index: number;\n id: CallId;\n name?: string;\n argumentsDelta: string;\n } | {\n type: 'block-end';\n index: number;\n block: ContentBlock;\n } | {\n type: 'usage';\n usage: TokenUsage;\n } | {\n type: 'finish';\n reason: FinishReason;\n replayState?: unknown;\n };\n export type SurfaceEventType = 'user/message' | 'assistant/message' | 'tool/result' | 'steering/message';\n export interface SurfaceIntent {\n surfaceOp: SurfaceOp;\n sourceEventSeqs?: number[];\n }\n export type SurfaceOp = 'append' | {\n op: 'replace';\n start: number;\n end: number;\n };\n export interface TerminalCallView {\n card: 'terminal';\n title: string;\n description?: string;\n cwd?: string;\n }\n export interface TerminalResultView {\n card: 'terminal';\n title?: string;\n output?: string;\n exitCode?: number;\n signal?: string;\n }\n export interface TodoItem {\n content: string;\n status: 'pending' | 'in_progress' | 'completed';\n }\n export interface TokenUsage {\n inputTokens: number;\n outputTokens: number;\n cacheReadTokens?: number;\n cacheWriteTokens?: number;\n reasoningTokens?: number;\n }\n export interface ToolCallBlock {\n type: 'tool-call';\n id: CallId;\n name: string;\n arguments: string;\n }\n export type ToolCallKind = 'read' | 'edit' | 'delete' | 'move' | 'search' | 'execute' | 'fetch' | 'other';\n export type ToolCallView = GenericCallView | TerminalCallView | DiffCallView;\n export interface ToolDefinition extends ToolSchema {\n readonly output: ToolOutputDefinition;\n execute(args: unknown, exec: ToolRunContext): Promise;\n finalizeContent?(exec: Readonly, result: Readonly): ContentBlock[] | undefined;\n timeoutMs?: number;\n isConcurrencySafe?(args: unknown): boolean;\n presentCall?(args: unknown): ToolCallView | undefined;\n presentResult?(args: unknown, result: ToolResult): ToolResultView | undefined;\n }\n export interface ToolErrorInfo {\n name: string;\n code: string;\n }\n export interface ToolExecution extends ToolExecutionInput {\n readonly token: ToolExecutionToken;\n }\n export interface ToolExecutionFailure {\n readonly isError: true;\n readonly error: ToolFailure;\n readonly value?: never;\n readonly content: ContentBlock[];\n readonly meta?: JsonValue;\n readonly additionalContexts?: UserMessage[];\n readonly concludesTurn?: never;\n }\n export interface ToolExecutionInput {\n readonly callId: CallId;\n readonly name: string;\n readonly arguments: unknown;\n readonly agent?: Agent;\n readonly parent?: ToolExecutionToken;\n readonly signal: AbortSignal;\n }\n export type ToolExecutionMode = {\n kind: 'parallel';\n } | {\n kind: 'exclusive';\n };\n export type ToolExecutionResult = ToolExecutionSuccess | ToolExecutionFailure;\n export interface ToolExecutionSuccess {\n readonly isError: false;\n readonly value: JsonValue;\n readonly content: ContentBlock[];\n readonly error?: never;\n readonly meta?: JsonValue;\n readonly additionalContexts?: UserMessage[];\n readonly concludesTurn?: true;\n }\n export type ToolExecutionToken = symbol & {\n readonly [toolExecutionTokenBrand]: true;\n };\n export interface ToolFailure {\n message: string;\n info?: ToolErrorInfo;\n }\n export type ToolGuard = (execution: Readonly) => string | undefined;\n export interface ToolMessageSource {\n kind: 'tool';\n callId: CallId;\n }\n export interface ToolOutputDefinition {\n readonly schema: JsonSchemaNode;\n render(args: unknown, value: JsonValue): ContentBlock[];\n presentationMeta?(args: unknown, value: JsonValue): JsonValue;\n }\n export interface ToolRestriction {\n readonly allow?: readonly string[];\n readonly deny?: readonly string[];\n }\n export interface ToolResult {\n content: ContentBlock[];\n isError: boolean;\n meta?: JsonValue;\n }\n export interface ToolResultBlock {\n type: 'tool-result';\n toolCallId: CallId;\n content: ContentBlock[];\n isError?: boolean;\n }\n export interface ToolResultMessage extends Message {\n readonly role: 'user';\n readonly content: [\n ToolResultBlock\n ];\n readonly source: ToolMessageSource;\n }\n export type ToolResultView = GenericResultView | TerminalResultView | DiffResultView | SearchResultView | ReadResultView | WebResultView;\n export interface ToolRunContext extends ToolExecution {\n deferContext(context: UserMessage): void;\n concludeTurn(): void;\n }\n export interface ToolSchema {\n name: string;\n description: string;\n parameters: Record;\n }\n export type TurnEndReason = TurnEndReasonMap[keyof TurnEndReasonMap];\n export interface TurnEndReasonMap {\n completed: {\n kind: 'completed';\n };\n aborted: {\n kind: 'aborted';\n };\n error: {\n kind: 'error';\n step: number;\n } & ({\n failure: LlmFailure;\n message?: never;\n code?: never;\n } | {\n message: string;\n code?: string;\n failure?: never;\n });\n disposed: {\n kind: 'disposed';\n };\n 'max-tokens': {\n kind: 'max-tokens';\n };\n interrupted: {\n kind: 'interrupted';\n };\n }\n export type TurnTrigger = TurnTriggerMap[keyof TurnTriggerMap];\n export interface TurnTriggerMap {\n message: {\n kind: 'message';\n source: MessageSource;\n };\n retry: {\n kind: 'retry';\n };\n injection: {\n kind: 'injection';\n source: MessageSource;\n };\n }\n export interface UserMessage extends Message {\n readonly role: 'user';\n }\n export interface WebFetchResultView {\n card: 'web';\n kind: 'fetch';\n title?: string;\n url: string;\n statusCode: number;\n truncated: boolean;\n }\n export type WebResultView = WebSearchResultView | WebFetchResultView;\n export interface WebSearchResultView {\n card: 'web';\n kind: 'search';\n title?: string;\n sources: WebSource[];\n answer?: string;\n truncated: boolean;\n }\n export interface WebSource {\n url: string;\n title?: string;\n snippet?: string;\n publishedAt?: string;\n }"}],"isError":false}],"role":"user","id":"a4ec9786-5e3f-45b2-a6de-efdf953287f9"}},"sourceEventSeqs":[13],"surfaceOp":"append"} +{"type":"tool/result","seq":14,"time":1785487622726,"data":{"turn":1,"step":1,"message":{"source":{"kind":"tool","callId":"inspect-tools-api"},"content":[{"type":"tool-result","toolCallId":"inspect-tools-api","content":[{"type":"text","text":"## api\n- tools — Tool registry and execution pipeline.\n /**\n * Register globally or in the calling agent scope. Scoped tools shadow\n * globals; duplicates within one layer and the reserved `run_code` name fail.\n * @param definition - tool schema, execution, and optional finalization/presentation callbacks.\n * @returns the exact disposer that unregisters the tool.\n */\n register(definition: ToolDefinition): () => void\n /**\n * Restrict global tools for the calling agent scope. Empty filters, unknown\n * names, scope-local names, and reserved transport names fail. Restrictions\n * intersect; scoped registrations remain visible.\n * @param filter - global-surface mask: `allow` (keep only) and/or `deny` (remove).\n * @returns the exact disposer that lifts this restriction.\n */\n restrict(filter: ToolRestriction): () => void\n /**\n * Register a monotonic guard after the extensible `tools/pre-execute`\n * waterfall. A plain-context guard applies globally; one registered through\n * `agent.ctx` applies only to that agent. Any matching guard may deny by\n * returning a reason, while no guard can force-allow a call another guard\n * denied. The exact effect disposer is returned for ordered ownership and\n * HMR cleanup.\n * @param guard - synchronous check; a returned string denies the execution.\n * @returns the exact disposer that unregisters the guard.\n */\n guard(guard: ToolGuard): () => void\n /**\n * Look up a tool as one scope sees it (scoped\n * shadows global; a restricted-away global reads as absent). Presenters pass\n * the calling agent so the rendered card matches the definition that\n * actually executed.\n * @param name - the tool name as registered.\n * @param scope - the viewing scope (the agent); omitted = the global view.\n * @returns the definition the scope resolves, or undefined when none is visible.\n */\n get(name: string, scope?: ScopeKey): ToolDefinition | undefined\n /**\n * Project visible definitions onto the allowlisted model-facing schema fields,\n * excluding execution and presentation callbacks.\n * @param scope - the viewing scope (the agent); omitted = the global view.\n * @returns one deep-cloned schema per visible tool.\n */\n schemas(scope?: ScopeKey): ToolSchema[]\n /**\n * Classify a pending call through the caller's visible tool definition. Only\n * an exact `true` is parallel; unknown, hidden, undeclared, invalid, or\n * throwing classifiers are exclusive.\n * @param exec - call name, parsed arguments, and optional agent scope.\n * @returns the fail-closed scheduling mode.\n */\n executionMode(exec: ToolExecutionInput): ToolExecutionMode\n /**\n * Execute through pre-policy, guards, around-dispatch, post-policy,\n * definition-owned content finalization, and final notification. Tool and\n * listener failures resolve as materialized error results; an invisible tool\n * reports `UNKNOWN_TOOL`. The returned outcome is the same lossless, frozen\n * snapshot final observers receive. Cancellation\n * arriving after entry and before final result materialization skips a\n * not-yet-started body with `ABORTED_BEFORE_DISPATCH` or replaces a\n * successful started outcome with `ABORTED`; already-started work is still\n * drained and may retain a tool-owned structured error.\n * @param exec - the typed same-process call input. The registry assigns its\n * correlation token before policy begins.\n * @returns the materialized final result.\n */\n async execute(exec: ToolExecutionInput): Promise\ntype shapes (referenced by the signatures above — read these before assuming a field is a string):\n export interface Agent {\n readonly id: SessionId;\n readonly options: AgentOptions;\n readonly session: Session;\n readonly status: AgentStatus;\n readonly acceptsNextStep: boolean;\n readonly ctx: Context;\n send(message: UserMessage, options: SendOptions): void;\n reserveTurnAdmission(): (() => void) | undefined;\n updateInbox(id: InboxItemId, action: InboxAction): InboxActionResult;\n cancel(cause: AgentCancelCause, options?: CancelOptions): void;\n whenIdle(): Promise;\n followup(message: UserMessage): void;\n steer(message: UserMessage): SteeringReceipt;\n inject(message: UserMessage): void;\n }\n export type AgentCancelCause = {\n readonly kind: 'user';\n } | {\n readonly kind: 'parent';\n };\n export interface AgentOptions {\n provider?: string;\n model?: string;\n maxTokens?: number;\n }\n export type AgentStatus = 'idle' | 'running';\n export interface AssistantMessage extends Message {\n readonly role: 'assistant';\n readonly source: ModelMessageSource;\n }\n export interface AssistantProvenance {\n provider: string;\n model: string;\n replayState?: unknown;\n }\n export type Branded = string & {\n readonly [BRAND]: B;\n };\n export type CallId = Branded<'CallId'>;\n export interface CancelOptions {\n keepInbox?: boolean;\n }\n export interface ContentBlockMap {\n 'text': TextBlock;\n 'reasoning': ReasoningBlock;\n 'tool-call': ToolCallBlock;\n 'tool-result': ToolResultBlock;\n }\n export type ContentBlockType = keyof ContentBlockMap;\n export interface DiffCallView {\n card: 'diff';\n title: string;\n diffs: FileDiff[];\n locations?: FileLocation[];\n }\n export interface DiffResultView {\n card: 'diff';\n title?: string;\n diffs: FileDiff[];\n }\n export interface EpochHeader {\n config: LlmCallConfig;\n adapterDefaults?: LlmCallConfigAdapterDefaults;\n system?: string;\n tools?: ToolSchema[];\n }\n export interface FileDiff {\n path: string;\n oldText: string | null;\n newText: string;\n }\n export interface FileLocation {\n path: string;\n line?: number;\n }\n export type FinishReason = FinishReasonMap[keyof FinishReasonMap];\n export interface FinishReasonMap {\n 'stop': {\n kind: 'stop';\n };\n 'tool-calls': {\n kind: 'tool-calls';\n };\n 'max-tokens': {\n kind: 'max-tokens';\n };\n 'aborted': {\n kind: 'aborted';\n failure: LlmFailure;\n };\n 'error': {\n kind: 'error';\n failure: LlmFailure;\n };\n }\n export interface GenericCallView {\n card: 'generic';\n title: string;\n kind?: ToolCallKind;\n rawInput?: unknown;\n content?: ContentBlock[];\n locations?: FileLocation[];\n }\n export interface GenericResultView {\n card: 'generic';\n title?: string;\n content?: ContentBlock[];\n }\n export type InboxAction = {\n readonly kind: 'edit';\n readonly content: ContentBlock[];\n } | {\n readonly kind: 'remove';\n } | {\n readonly kind: 'steer';\n };\n export type InboxActionResult = 'applied' | 'not-found' | 'steer-unavailable';\n export type InboxItemId = Branded<'InboxItemId'>;\n export interface JsonSchemaNode {\n type?: JsonSchemaType;\n oneOf?: JsonSchemaNode[];\n properties?: Record;\n required?: string[];\n additionalProperties?: boolean;\n items?: JsonSchemaNode;\n enum?: JsonSchemaScalar[];\n const?: JsonSchemaScalar;\n description?: string;\n title?: string;\n default?: JsonValue;\n examples?: JsonValue;\n }\n export type JsonSchemaScalar = string | number | boolean | null;\n export type JsonSchemaType = 'object' | 'array' | 'string' | 'number' | 'integer' | 'boolean' | 'null';\n export type JsonValue = null | boolean | number | string | JsonValue[] | {\n [key: string]: JsonValue;\n };\n export interface LlmCallConfig {\n provider: string;\n model: string;\n reasoningEffort?: ReasoningEffortId;\n temperature?: number;\n maxTokens?: number;\n stop?: string[];\n }\n export interface LlmCallConfigAdapterDefaults {\n reasoningEffort?: true;\n maxTokens?: true;\n }\n export interface LlmFailure {\n readonly message: string;\n readonly code: string;\n readonly status?: number;\n readonly providerRetryAfterMs?: number;\n readonly requestId?: ProviderRequestId;\n }\n export interface Message {\n readonly id: MessageId;\n readonly role: 'system' | 'user' | 'assistant';\n readonly content: ContentBlock[];\n readonly source: MessageSource;\n }\n export type MessageId = Branded<'MessageId'>;\n export type MessageSource = MessageSourceMap[keyof MessageSourceMap];\n export interface MessageSourceMap {\n user: {\n kind: 'user';\n };\n plugin: {\n kind: 'plugin';\n plugin: string;\n };\n model: ModelMessageSource;\n tool: ToolMessageSource;\n }\n export interface ModelMessageSource extends AssistantProvenance {\n kind: 'model';\n }\n export type ProviderRequestId = Branded<'ProviderRequestId'>;\n export interface ReadFileLine {\n number: number;\n text: string;\n }\n export interface ReadResultView {\n card: 'read';\n title?: string;\n path: string;\n offset: number;\n lines: ReadFileLine[];\n totalLines: number;\n lang?: string;\n content?: ContentBlock[];\n }\n export interface ReasoningBlock {\n type: 'reasoning';\n text: string;\n }\n export type ReasoningEffortId = Branded<'ReasoningEffortId'>;\n export interface RequestContext {\n provider: string;\n model: string;\n contextWindow?: number;\n }\n export type RequestHeaderReason = 'initial' | 'resume' | 'change';\n export type ScopeKey = object;\n export interface SearchFileMatches {\n path: string;\n matches: SearchLineMatch[];\n }\n export interface SearchLineMatch {\n lineNumber: number;\n line: string;\n }\n export interface SearchMatchesResultView {\n card: 'search';\n shape: 'matches';\n title?: string;\n files: SearchFileMatches[];\n truncated: boolean;\n total: number;\n }\n export interface SearchPathsResultView {\n card: 'search';\n shape: 'paths';\n title?: string;\n paths: string[];\n truncated: boolean;\n total: number;\n }\n export type SearchResultView = SearchMatchesResultView | SearchPathsResultView;\n export interface SendOptions {\n target: SendTarget;\n wakeup: boolean;\n }\n export type SendTarget = 'next-turn' | 'next-step';\n export class Session {\n get surface(): SessionSurface;\n readonly header: SessionHeader;\n get id(): SessionId;\n readonly firstLiveSeq: number;\n constructor(id: SessionId, seed?: readonly SessionEvent[], header?: SessionHeader);\n get events(): readonly SessionEvent[];\n get seq(): number;\n append(type: T, data: SessionEventMap[T], ...opts: T extends SurfaceEventType ? [\n opts: SurfaceIntent\n ] : [\n ]): SessionEvent;\n requestHeader(): EpochHeader | undefined;\n requestContext(): RequestContext | undefined;\n deriveMessages(): Message[];\n deriveEventMessage(event: SessionEvent): Message | null;\n }\n export type SessionEvent = {\n [K in SessionEventType]: {\n type: K;\n seq: number;\n time: number;\n data: SessionEventMap[K];\n } & (K extends SurfaceEventType ? {\n sourceEventSeqs?: number[];\n surfaceOp?: SurfaceOp;\n } : object);\n }[T];\n export interface SessionEventMap {\n 'turn/start': {\n turn: number;\n trigger: TurnTrigger;\n };\n 'turn/end': {\n turn: number;\n reason: TurnEndReason;\n };\n 'step/start': {\n turn: number;\n step: number;\n };\n 'step/end': {\n turn: number;\n step: number;\n };\n 'user/message': UserMessage;\n 'assistant/chunk': {\n turn: number;\n step: number;\n chunk: StreamChunk;\n };\n 'assistant/message': {\n turn: number;\n step: number;\n message: AssistantMessage;\n usage?: TokenUsage;\n };\n 'tool/call': {\n turn: number;\n step: number;\n callId: CallId;\n name: string;\n arguments: string;\n };\n 'tool/result': {\n turn: number;\n step: number;\n message: ToolResultMessage;\n error?: {\n name: string;\n code: string;\n };\n meta?: JsonValue;\n };\n 'steering/message': {\n turn: number;\n message: UserMessage;\n };\n 'todo/write': {\n todos: TodoItem[];\n };\n 'request/header': {\n header: EpochHeader;\n reason: RequestHeaderReason;\n };\n 'request/context': RequestContext;\n 'session/end-seed': Record;\n }\n export type SessionEventType = keyof SessionEventMap;\n export interface SessionHeader {\n readonly version: number;\n readonly id: SessionId;\n readonly createdAt: number;\n readonly cwd?: string;\n readonly parentSession?: SessionId;\n readonly seedLength?: number;\n readonly origin?: 'subagent';\n readonly delegationDepth?: number;\n }\n export type SessionId = Branded<'SessionId'>;\n export interface SessionSurface {\n readonly nodes: readonly number[];\n readonly replaceGeneration: number;\n }\n export type SteeringOutcome = {\n readonly status: 'admitted';\n readonly turn: number;\n readonly step: number;\n } | {\n readonly status: 'rejected';\n };\n export interface SteeringReceipt {\n readonly outcome: Promise;\n }\n export type StreamChunk = {\n type: 'block-start';\n index: number;\n blockType: ContentBlockType;\n } | {\n type: 'text-delta';\n index: number;\n text: string;\n } | {\n type: 'reasoning-delta';\n index: number;\n text: string;\n } | {\n type: 'tool-call-delta';\n index: number;\n id: CallId;\n name?: string;\n argumentsDelta: string;\n } | {\n type: 'block-end';\n index: number;\n block: ContentBlock;\n } | {\n type: 'usage';\n usage: TokenUsage;\n } | {\n type: 'finish';\n reason: FinishReason;\n replayState?: unknown;\n };\n export type SurfaceEventType = 'user/message' | 'assistant/message' | 'tool/result' | 'steering/message';\n export interface SurfaceIntent {\n surfaceOp: SurfaceOp;\n sourceEventSeqs?: number[];\n }\n export type SurfaceOp = 'append' | {\n op: 'replace';\n start: number;\n end: number;\n };\n export interface TerminalCallView {\n card: 'terminal';\n title: string;\n description?: string;\n cwd?: string;\n }\n export interface TerminalResultView {\n card: 'terminal';\n title?: string;\n output?: string;\n exitCode?: number;\n signal?: string;\n }\n export interface TodoItem {\n content: string;\n status: 'pending' | 'in_progress' | 'completed';\n }\n export interface TokenUsage {\n inputTokens: number;\n outputTokens: number;\n cacheReadTokens?: number;\n cacheWriteTokens?: number;\n reasoningTokens?: number;\n }\n export interface ToolCallBlock {\n type: 'tool-call';\n id: CallId;\n name: string;\n arguments: string;\n }\n export type ToolCallKind = 'read' | 'edit' | 'delete' | 'move' | 'search' | 'execute' | 'fetch' | 'other';\n export type ToolCallView = GenericCallView | TerminalCallView | DiffCallView;\n export interface ToolDefinition extends ToolSchema {\n readonly output: ToolOutputDefinition;\n execute(args: unknown, exec: ToolRunContext): Promise;\n finalizeContent?(exec: Readonly, result: Readonly): ContentBlock[] | undefined;\n timeoutMs?: number;\n isConcurrencySafe?(args: unknown): boolean;\n presentCall?(args: unknown): ToolCallView | undefined;\n presentResult?(args: unknown, result: ToolResult): ToolResultView | undefined;\n }\n export interface ToolErrorInfo {\n name: string;\n code: string;\n }\n export interface ToolExecution extends ToolExecutionInput {\n readonly token: ToolExecutionToken;\n }\n export interface ToolExecutionFailure {\n readonly isError: true;\n readonly error: ToolFailure;\n readonly value?: never;\n readonly content: ContentBlock[];\n readonly meta?: JsonValue;\n readonly additionalContexts?: UserMessage[];\n readonly concludesTurn?: never;\n }\n export interface ToolExecutionInput {\n readonly callId: CallId;\n readonly name: string;\n readonly arguments: unknown;\n readonly agent?: Agent;\n readonly parent?: ToolExecutionToken;\n readonly signal: AbortSignal;\n }\n export type ToolExecutionMode = {\n kind: 'parallel';\n } | {\n kind: 'exclusive';\n };\n export type ToolExecutionResult = ToolExecutionSuccess | ToolExecutionFailure;\n export interface ToolExecutionSuccess {\n readonly isError: false;\n readonly value: JsonValue;\n readonly content: ContentBlock[];\n readonly error?: never;\n readonly meta?: JsonValue;\n readonly additionalContexts?: UserMessage[];\n readonly concludesTurn?: true;\n }\n export type ToolExecutionToken = symbol & {\n readonly [toolExecutionTokenBrand]: true;\n };\n export interface ToolFailure {\n message: string;\n info?: ToolErrorInfo;\n }\n export type ToolGuard = (execution: Readonly) => string | undefined;\n export interface ToolMessageSource {\n kind: 'tool';\n callId: CallId;\n }\n export interface ToolOutputDefinition {\n readonly schema: JsonSchemaNode;\n render(args: unknown, value: JsonValue): ContentBlock[];\n presentationMeta?(args: unknown, value: JsonValue): JsonValue;\n }\n export interface ToolRestriction {\n readonly allow?: readonly string[];\n readonly deny?: readonly string[];\n }\n export interface ToolResult {\n content: ContentBlock[];\n isError: boolean;\n meta?: JsonValue;\n }\n export interface ToolResultBlock {\n type: 'tool-result';\n toolCallId: CallId;\n content: ContentBlock[];\n isError?: boolean;\n }\n export interface ToolResultMessage extends Message {\n readonly role: 'user';\n readonly content: [\n ToolResultBlock\n ];\n readonly source: ToolMessageSource;\n }\n export type ToolResultView = GenericResultView | TerminalResultView | DiffResultView | SearchResultView | ReadResultView | WebResultView;\n export interface ToolRunContext extends ToolExecution {\n deferContext(context: UserMessage): void;\n concludeTurn(): void;\n }\n export interface ToolSchema {\n name: string;\n description: string;\n parameters: Record;\n }\n export type TurnEndReason = TurnEndReasonMap[keyof TurnEndReasonMap];\n export interface TurnEndReasonMap {\n completed: {\n kind: 'completed';\n };\n aborted: {\n kind: 'aborted';\n };\n error: {\n kind: 'error';\n step: number;\n } & ({\n failure: LlmFailure;\n message?: never;\n code?: never;\n } | {\n message: string;\n code?: string;\n failure?: never;\n });\n disposed: {\n kind: 'disposed';\n };\n 'max-tokens': {\n kind: 'max-tokens';\n };\n interrupted: {\n kind: 'interrupted';\n };\n }\n export type TurnTrigger = TurnTriggerMap[keyof TurnTriggerMap];\n export interface TurnTriggerMap {\n message: {\n kind: 'message';\n source: MessageSource;\n };\n retry: {\n kind: 'retry';\n };\n injection: {\n kind: 'injection';\n source: MessageSource;\n };\n }\n export interface UserMessage extends Message {\n readonly role: 'user';\n }\n export interface WebFetchResultView {\n card: 'web';\n kind: 'fetch';\n title?: string;\n url: string;\n statusCode: number;\n truncated: boolean;\n }\n export type WebResultView = WebSearchResultView | WebFetchResultView;\n export interface WebSearchResultView {\n card: 'web';\n kind: 'search';\n title?: string;\n sources: WebSource[];\n answer?: string;\n truncated: boolean;\n }\n export interface WebSource {\n url: string;\n title?: string;\n snippet?: string;\n publishedAt?: string;\n }"}],"isError":false}],"role":"user","id":"a4ec9786-5e3f-45b2-a6de-efdf953287f9"}},"sourceEventSeqs":[13],"surfaceOp":"append"} {"type":"step/end","seq":15,"time":1785487622726,"data":{"turn":1,"step":1}} {"type":"step/start","seq":16,"time":1785487622735,"data":{"turn":1,"step":2}} {"type":"assistant/chunk","seq":17,"time":1784449176734,"data":{"turn":1,"step":2,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} From a31b3eb7b803fba22f15c4084e48fafdd04dca27 Mon Sep 17 00:00:00 2001 From: imccyu Date: Sun, 2 Aug 2026 16:42:11 +0800 Subject: [PATCH 08/13] test(host): cover queue mirror event expiry --- .../apiproxy/tests/api-proxy-commands.spec.ts | 27 +++++++++++++++++++ 1 file changed, 27 insertions(+) diff --git a/packages/host/apiproxy/tests/api-proxy-commands.spec.ts b/packages/host/apiproxy/tests/api-proxy-commands.spec.ts index 473d452327..4cf975dc99 100644 --- a/packages/host/apiproxy/tests/api-proxy-commands.spec.ts +++ b/packages/host/apiproxy/tests/api-proxy-commands.spec.ts @@ -387,6 +387,33 @@ describe('session/queue frames', () => { expect(replayFrames.filter(frame => frame.type === 'session/queue')).toEqual(liveFrames) }) + it('expires unmatched mutations after the synchronous re-entry window', async () => { + const ctx = await harness() + const agent = stubAgent(ctx) + const api = createApiProxy(ctx, DEFAULTS) + const original = inboxItem('i-stale-edit', inboxMessage('m-stale-edit', 'original'), 'queued') + const staleEdit = inboxItem('i-stale-edit', inboxMessage('m-stale-edit', 'stale edit'), 'queued') + const staleTerminal = inboxItem('i-stale-terminal', inboxMessage('m-stale-terminal', 'keep me'), 'queued') + + ctx.emit('agent/inbox/update', agent, staleEdit) + ctx.emit('agent/inbox/discard', agent, [staleTerminal]) + await Promise.resolve() + ctx.emit('agent/inbox/enqueue', agent, original) + ctx.emit('agent/inbox/enqueue', agent, staleTerminal) + + const replay = new AbortController() + const frames = await collect( + api.events.mux({ rpcId: RpcId('t-mux-expired-unseen'), payload: {} }, replay.signal), 2, replay) + expect(frames.filter(frame => frame.type === 'session/queue')).toEqual([{ + type: 'session/queue', + sessionId: agent.id, + items: [ + { id: original.id, placement: original.placement, message: original.message }, + { id: staleTerminal.id, placement: staleTerminal.placement, message: staleTerminal.message }, + ], + }]) + }) + it('publishes complete live snapshots and replays the latest snapshot on reconnect', async () => { const ctx = await harness() const api = createApiProxy(ctx, DEFAULTS) From d59ab5596eb9278c5ce7f123b0a8f024d89f4757 Mon Sep 17 00:00:00 2001 From: imccyu Date: Sun, 2 Aug 2026 16:42:20 +0800 Subject: [PATCH 09/13] docs(host): align queue mirror contract --- packages/host/apiproxy/README.i18n.yaml | 4 ++-- packages/host/apiproxy/README.md | 2 +- packages/host/apiproxy/README.zh.md | 2 +- packages/host/apiproxy/src/api-proxy.ts | 9 +++++---- 4 files changed, 9 insertions(+), 8 deletions(-) diff --git a/packages/host/apiproxy/README.i18n.yaml b/packages/host/apiproxy/README.i18n.yaml index 5f3bec73b0..4f0e10cb87 100644 --- a/packages/host/apiproxy/README.i18n.yaml +++ b/packages/host/apiproxy/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/host/apiproxy/README.md -README.md: 87c477362dc159621a38a1ff923bebf07cbd3c6f -README.zh.md: 7316f3a6d93d30e101537c40608037fedfa77321 +README.md: d4af267ef46f327bdabf31429b291ffd04308203 +README.zh.md: f148cfa0c023b54d62016c5211205d391b634544 diff --git a/packages/host/apiproxy/README.md b/packages/host/apiproxy/README.md index 87c477362d..d4af267ef4 100644 --- a/packages/host/apiproxy/README.md +++ b/packages/host/apiproxy/README.md @@ -22,7 +22,7 @@ Session model routing is a session-domain contract. `session.models` returns the Generic Agent-bound session, command, and goal operations serve ordinary sessions only. They return `agent-busy` for a session-backed subagent instead of resuming or driving it; explicit-id `session.create` adoption and the attached-only queue controls enforce the same ownership boundary. Subagent conversation reads and continuation use the dedicated `subagent.*` domain, which retains catalog-mode and direct-parent authorization. -Pending queued input is a live control-plane contract, not session history. The gateway mirrors queued `InboxItem` occurrences from `agent/inbox/*` and broadcasts authoritative `session/queue` snapshots on every queued change and reconnect; pending steering stays outside this Web projection. `session.updateQueue` addresses one `InboxItemId`: edit replaces pending content, remove discards it, and strict steer transfers its complete message into the current next-step window. A closed window returns `steer-unavailable` without changing the row. `session.cancel` aborts only the active turn and preserves pending inbox work; after cancellation reaches quiescence and the closing turn flushes, AgentLoop claims the next waking occurrence in FIFO order. The browser never resends or promotes that occurrence. A driver claim wins races by retiring the address before admission; a later operation returns `queue-item-not-found`. Queue operations query only an attached ordinary-session Agent and never resume a cold session because process-local inbox identities do not survive restart or disposal. The client never infers retirement from turn or status events. +Pending inbox input is a live control-plane contract, not session history. The gateway mirrors `InboxItem` occurrences from `agent/inbox/*` with their `queued` or `steering` placement and broadcasts authoritative `session/queue` snapshots on every change and reconnect. A steering occurrence remains in this projection until the corresponding durable `steering/message` has been published, preserving the Host's linear event order during the handoff. `session.updateQueue` addresses one `InboxItemId`: edit replaces pending content, remove discards it, and strict steer transfers its complete message into the current next-step window. A closed window returns `steer-unavailable` without changing the row. `session.cancel` aborts only the active turn and preserves pending inbox work; after cancellation reaches quiescence and the closing turn flushes, AgentLoop claims the next waking occurrence in FIFO order. The browser never resends or promotes that occurrence. A driver claim wins races by retiring the address before admission; a later operation returns `queue-item-not-found`. Queue operations query only an attached ordinary-session Agent and never resume a cold session because process-local inbox identities do not survive restart or disposal. The client never infers retirement from turn or status events. Workspace and Session lists are separate reconnect baselines. `workspace.create({ name })` creates a uniquely titled directory under the configured root, while `workspace.create({ path })` adopts an existing canonical directory and permits basename-derived titles to repeat. `workspace.delete` removes only the Workspace registration, `session.create` accepts an optional preallocated Session id, and `host/workspace-changed`, `host/workspace-removed`, plus `host/session-added` carry committed increments in either arrival order. `workspace.archiveSession` adds one session to the registry-global archive set and answers the full updated set; `workspace.list` carries that set as the reconnect baseline and `host/archived-sessions-changed` pushes the full snapshot after every durable change. Archiving hides the session from grouping surfaces without touching its log or its workspace account; a session neither live nor persisted fails with `session-not-found`. Registration deletion preserves the directory and session logs; its Sessions remain in `session.list` and become Ungrouped. `SessionSummary.blank` and the `host/session-added` frame carry the derived zero-events bit: clients hide blank sessions and reuse them per workspace, flip blank on the first `host/session-status(running:true)`, and treat `session.list` as the reconnect authority; cold summaries are never blank because lazy persistence keeps never-appended sessions out of `list()`. The `session.list` summaries and `host/session-added` frames also carry the optional durable `origin: 'subagent'` classification so navigation can suppress duplicate child rows immediately and after reconnect; that bit is never continuation authority. diff --git a/packages/host/apiproxy/README.zh.md b/packages/host/apiproxy/README.zh.md index 7316f3a6d9..f148cfa0c0 100644 --- a/packages/host/apiproxy/README.zh.md +++ b/packages/host/apiproxy/README.zh.md @@ -22,7 +22,7 @@ 绑定到 Agent 的通用会话、命令与目标操作只服务普通会话。对于由会话支撑的 subagent,它们会返回 `agent-busy`,而不是恢复或驱动它;显式 id 的 `session.create` 接纳与仅针对已附加会话的队列控件也会执行同一所有权边界。subagent 对话读取与继续执行使用专用的 `subagent.*` 领域,该领域保留目录 mode 与直接 parent 授权。 -待处理的 queued 输入属于实时控制平面契约,而非会话历史。网关镜像来自 `agent/inbox/*` 的 queued `InboxItem` 入队项,并在每次 queued 变更和重连时广播权威的 `session/queue` 快照;待处理 steering(中途引导)不进入此 Web 投影。`session.updateQueue` 通过 `InboxItemId` 寻址单个项:编辑会替换待处理内容,移除会将其丢弃,严格 steering 会把其完整消息转移到当前 next-step 窗口。窗口关闭时返回 `steer-unavailable`,且不改变该行。`session.cancel` 仅中止活动轮次,并保留待处理 inbox 工作;取消达到完全停稳且结束中的轮次完成 flush 后,AgentLoop 按 FIFO 顺序认领下一个可唤醒入队项。浏览器绝不重发或提升该入队项。驱动器在接纳前退役寻址标识,因此认领会赢得竞态;之后的操作返回 `queue-item-not-found`。队列操作只查询当前已挂载的普通会话 Agent,绝不恢复冷会话,因为进程本地 inbox 标识无法在重启或资源释放后存活。客户端绝不根据轮次或状态事件推断项已退役。 +待处理的 inbox 输入属于实时控制平面契约,而非会话历史。网关镜像来自 `agent/inbox/*` 的 `InboxItem` 入队项及其 `queued` 或 `steering` placement,并在每次变更和重连时广播权威的 `session/queue` 快照。steering 入队项会一直保留在该投影中,直到对应的持久 `steering/message` 已发布,从而在交接期间保持 Host 的线性事件顺序。`session.updateQueue` 通过 `InboxItemId` 寻址单个项:编辑会替换待处理内容,移除会将其丢弃,严格 steering 会把其完整消息转移到当前 next-step 窗口。窗口关闭时返回 `steer-unavailable`,且不改变该行。`session.cancel` 仅中止活动轮次,并保留待处理 inbox 工作;取消达到完全停稳且结束中的轮次完成 flush 后,AgentLoop 按 FIFO 顺序认领下一个可唤醒入队项。浏览器绝不重发或提升该入队项。驱动器在接纳前退役寻址标识,因此认领会赢得竞态;之后的操作返回 `queue-item-not-found`。队列操作只查询当前已挂载的普通会话 Agent,绝不恢复冷会话,因为进程本地 inbox 标识无法在重启或资源释放后存活。客户端绝不根据轮次或状态事件推断项已退役。 Workspace 列表与 Session 列表是相互独立的重连基线。`workspace.create({ name })` 会在配置根目录下创建显示标题唯一的目录,而 `workspace.create({ path })` 会接纳已有的规范目录,并允许由 basename 派生的标题重复。`workspace.delete` 只移除 Workspace 注册记录,`session.create` 接受可选的预分配 Session id,`host/workspace-changed`、`host/workspace-removed` 与 `host/session-added` 则以任意到达顺序携带已提交的增量。`workspace.archiveSession` 向注册表级全局归档集合添加一个会话,并应答完整的更新后集合;`workspace.list` 携带该集合作为重连基线,`host/archived-sessions-changed` 在每次持久变更后推送完整快照。归档只把会话从各分组视图中隐藏,不触碰其日志和 workspace 记账;既非实时也未持久化的会话以 `session-not-found` 失败。删除注册记录会保留目录和会话日志;相关 Session 仍留在 `session.list` 中,并进入 Ungrouped。`SessionSummary.blank` 与 `host/session-added` 帧携带派生的零事件位:客户端隐藏空白会话并按 workspace 复用它们,在首个 `host/session-status(running:true)` 时翻转 blank,并以 `session.list` 作为重连权威;冷会话摘要永远不是空白:惰性持久化让从未追加过事件的会话根本不出现在 `list()` 中。`session.list` 摘要与 `host/session-added` 帧还会携带可选的持久化分类 `origin: 'subagent'`,使导航在实时创建与重连后都能隐藏重复的 child 行;该标记绝不是继续执行的权威依据。 diff --git a/packages/host/apiproxy/src/api-proxy.ts b/packages/host/apiproxy/src/api-proxy.ts index e9808518fe..124f3a5fd0 100644 --- a/packages/host/apiproxy/src/api-proxy.ts +++ b/packages/host/apiproxy/src/api-proxy.ts @@ -776,10 +776,11 @@ export function createApiProxy(ctx: Context, defaults: ApiProxyDefaults): ApiPro }) /** - * Per-session queued-occurrence mirror serving the mux-open queue snapshot - * (the same refresh-recovery baseline as pending questions). Each terminal - * queue event retires one matching occurrence, so repeated sends of the same - * identified message remain visible until every occurrence is claimed. + * Per-session pending-occurrence mirror serving live and mux-open + * `session/queue` snapshots. It carries both queued and steering placements. + * Each terminal inbox event retires one matching occurrence, so repeated + * sends of the same identified message remain visible until every occurrence + * is claimed. */ const queuedMirror = new Map() type UnseenQueueEvent = From cf918fa05dc755f4166bba03de8a46866a846d4b Mon Sep 17 00:00:00 2001 From: imccyu Date: Sun, 2 Aug 2026 16:43:16 +0800 Subject: [PATCH 10/13] test(web): explain steering replay pace --- apps/web/tests/steering.e2e.ts | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/apps/web/tests/steering.e2e.ts b/apps/web/tests/steering.e2e.ts index 794a92b74b..1f9672d74a 100644 --- a/apps/web/tests/steering.e2e.ts +++ b/apps/web/tests/steering.e2e.ts @@ -48,7 +48,8 @@ describe('web e2e: mid-turn steering lands durably and visibly', () => { const sessionEvents: SessionEvent[] = [] beforeAll(async () => { - // The slower replay keeps the Queue action available until the recorded question barrier arrives. + // The 100 ms replay pace is test-only headroom that keeps the Queue action + // available until the recorded question barrier arrives, not a product timing contract. scaffold = await launchWebScaffold(MODE === 'record' ? {} : { replayFixture: FIXTURE, paceMs: 100 }) scaffold.ctx.on('session/event', (_session, event) => { sessionEvents.push(event) }) browser = await chromium.launch() From 7d08e43720f9a2f0e4123271ecbffe6d72721d96 Mon Sep 17 00:00:00 2001 From: imccyu Date: Sun, 2 Aug 2026 17:50:00 +0800 Subject: [PATCH 11/13] fix(web): address steering review feedback --- ...026-07-30-web-queue-steer-action.i18n.yaml | 4 +- .../2026-07-30-web-queue-steer-action.md | 20 +++++----- .../2026-07-30-web-queue-steer-action.zh.md | 20 +++++----- apps/web/tests/steering.e2e.ts | 17 +++++--- packages/client/runtime/README.i18n.yaml | 4 +- packages/client/runtime/README.md | 2 +- packages/client/runtime/README.zh.md | 2 +- .../runtime/src/client/sessions/session.ts | 11 +++++ .../client/runtime/tests/queue-store.spec.ts | 40 ++++++++++++++++++- .../client/ui-conversation/README.i18n.yaml | 4 +- packages/client/ui-conversation/README.md | 6 +-- packages/client/ui-conversation/README.zh.md | 6 +-- .../ui-conversation/src/client/apply.ts | 6 ++- .../src/client/chat/ChatView.tsx | 8 ++-- .../src/client/contract/slots.ts | 6 ++- .../src/client/input/submission-policy.ts | 11 +++-- .../ui-conversation/src/client/locales.ts | 4 +- .../src/client/queue/QueueDock.tsx | 11 ++--- .../ui-conversation/src/client/service.ts | 6 ++- .../src/client/skeleton/InputBar.tsx | 6 ++- .../ui-conversation/tests/chat-view.spec.tsx | 27 +++++++++++-- .../ui-conversation/tests/input-bar.spec.tsx | 22 +++++++++- .../ui-conversation/tests/queue-dock.spec.tsx | 29 ++++++++++++-- .../tests/service-orchestration.spec.ts | 23 +++++++++++ .../tests/submission-policy.spec.ts | 18 +++++---- packages/core/agent-loop/src/agent.ts | 7 +++- .../tests/contract-regressions.spec.ts | 6 ++- 27 files changed, 248 insertions(+), 78 deletions(-) diff --git a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml index 555e2fb434..70b35c2ff0 100644 --- a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml +++ b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md -2026-07-30-web-queue-steer-action.md: 78d8c4d50daa4d162ff17c1efc03c32face8019a -2026-07-30-web-queue-steer-action.zh.md: 6b8af563ce527b8122b9f5e783a6ef700dcdd401 +2026-07-30-web-queue-steer-action.md: b5bc3547e0f72ba8211f4338fb4c1f5223a3d0cc +2026-07-30-web-queue-steer-action.zh.md: 6edde8e2cc2bee0a52a0b05df8f62f77b3d9ccfa diff --git a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md index 78d8c4d50d..b5bc3547e0 100644 --- a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md +++ b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.md @@ -14,19 +14,19 @@ Implementing the row action as a client-side delete followed by `session.prompt( ### Product contract -Each non-editing QueueDock row exposes the upward-arrow action as “插话发送”. The action is enabled only while the session reports a running agent; mixed-content messages remain eligible because steering forwards the complete immutable `UserMessage` rather than the row's text projection. Edit and delete keep their existing behavior. +Each non-editing ordinary-session QueueDock row exposes the upward-arrow action as “插话发送”. The action is enabled only while the session reports a running agent; mixed-content messages remain eligible because steering forwards the complete immutable `UserMessage` rather than the row's text projection. An addressed subagent keeps its Queue projection read-only because its continuation transport does not expose queue mutation. Activating the action requests strict current-turn steering for that exact `InboxItemId`. Success removes the Queue row through the authoritative Host snapshot and immediately projects the same pending steering after the `Deep diving...` running-status row; that bubble offers Copy but no Fork because the message has no durable event sequence yet. Once AgentLoop drains it, the existing durable `steering/message` event takes over the same user-style bubble and restores its clock, Copy, and Fork without a separate durable presentation path. -The running bit is only an interaction hint. AgentLoop's `acceptsNextStep` value is authoritative at the synchronous mutation boundary. If that window has closed, the operation leaves the Queue occurrence unchanged and returns a typed `steer-unavailable` error; if the driver already claimed the occurrence, it returns the existing `queue-item-not-found` error. The UI reports either race without optimistically removing the row. +The running bit is only an interaction hint. AgentLoop's `acceptsNextStep` value is authoritative at the synchronous mutation boundary. If that window has closed, the operation leaves the Queue occurrence unchanged and returns a typed `steer-unavailable` error, after which the original waking occurrence proceeds through Queue. If the driver already claimed the occurrence, it returns the existing `queue-item-not-found` error and independent-turn delivery is already underway. The UI treats both races as converged Queue delivery without a failure notice; transport and unknown errors still surface. -The composer uses a separate best-effort contract for newly typed input. While the addressed session is idle, Enter and Cmd/Ctrl+Enter both perform an ordinary Queue send. While it is running, a General Settings preference assigns plain Enter to Queue (the default) or Steer, and Cmd/Ctrl+Enter performs the other behavior; Shift+Enter inserts a newline. The browser persists that preference, and it affects only the busy-state gesture pair. If a direct composer Steer misses the current next-step window, AgentLoop automatically admits it as the next waking Queue turn and the Web does not report a failure. +The composer uses a separate best-effort contract for newly typed input. While the addressed session is idle, Enter and Cmd/Ctrl+Enter both perform an ordinary Queue send. While a primary session is running, a General Settings preference assigns plain Enter to Queue (the default) or Steer, and Cmd/Ctrl+Enter performs the other behavior; Shift+Enter inserts a newline. An addressed subagent keeps both gestures on its Queue-only continuation transport. The browser persists the preference, and it affects only the steer-capable busy-state gesture pair. If a direct composer Steer misses the current next-step window, AgentLoop automatically admits it as the next waking Queue turn and the Web does not report a failure. ### Agent and lifecycle boundary `InboxAction` gains a consumer-backed `{ kind: 'steer' }` operation alongside edit and remove. `Agent.updateInbox()` handles it only after locating the queued occurrence and proving `acceptsNextStep`; it never delegates to the best-effort `agent.steer()` alias. -An applied action ends the queued occurrence and accepts the same immutable `UserMessage` as a new steering occurrence. The steering occurrence receives a new `InboxItemId` and truthful `placement: 'steering'`, while the message retains its `MessageId`, content, and source. AgentLoop installs the new outbox entry before publishing lifecycle events, then emits its enqueue before the old occurrence's discard so re-entrant cancellation cannot observe or retire an unannounced item. The existing inbox conservation invariant therefore continues to require one enqueue and one terminal dequeue or discard for each occurrence. +An applied action ends the queued occurrence and accepts the same immutable `UserMessage` as a new steering occurrence. The steering occurrence receives a new `InboxItemId` and truthful `placement: 'steering'`, while the message retains its `MessageId`, content, source, and any pending `SteeringReceipt` delivery controller. AgentLoop installs the new outbox entry before publishing lifecycle events, then emits its enqueue before the old occurrence's discard so re-entrant cancellation cannot observe or retire an unannounced item. The existing inbox conservation invariant therefore continues to require one enqueue and one terminal dequeue or discard for each occurrence. The action does not run `agent/prompt-submit`: choosing steering intentionally changes delivery from an independently admitted turn to current-turn next-step input. It neither cancels current work nor reorders the remaining Queue. @@ -36,15 +36,15 @@ The action does not run `agent/prompt-submit`: choosing steering intentionally c The Host's existing `queuedMirror` remains the sole transient inbox authority. Its `session/queue` snapshot carries every live occurrence with `placement: 'queued' | 'steering'`: QueueDock renders only queued rows, while ChatView renders pending steering at the conversation tail after the `Deep diving...` running-status row, with Copy but without Fork, edit, or delete actions. Reconnect replays the same snapshot, so this visibility does not require client optimism or a second registry. -When AgentLoop claims pending steering, it emits `agent/inbox/dequeue` immediately before synchronously appending `steering/message`. The Host retires that steering row on the following microtask, allowing the durable session event to enter the linear mux stream first. ChatView matches the shared `MessageId` and suppresses the transient projection as soon as the durable node exists, so one bubble changes authority without a visible gap or duplicate; the durable projection restores the clock, Copy, and Fork against its logged event time and sequence, while an append failure still retires the claimed row. +When AgentLoop claims pending steering, it emits `agent/inbox/dequeue` immediately before synchronously appending `steering/message`. The Host retires that steering row on the following microtask, allowing the durable session event to enter the linear mux stream first. On the accepted live event, the client Session retires the first matching current steering occurrence before publishing its snapshot; history replay does not consume a later occurrence that reused the same `MessageId`. ChatView therefore renders one authority at a time without scanning durable history, and the durable projection restores the clock, Copy, and Fork against its logged event time and sequence. An append failure still retires the claimed row. -The existing `session.prompt(mode: 'steer')` contract remains best-effort for new input: outside the next-step window it becomes a waking follow-up. The composer carries an explicit `queue | steer` mode through slash adjudication and reference serialization before calling that contract. A browser-local submission policy owns the persisted busy-Enter preference and resolves plain versus accelerated Enter as complementary gestures; the Settings row and InputBar share that policy without duplicating storage or delivery-window authority. Only the Queue row action is strict, because failure can safely leave its already-pending message untouched. +The existing `session.prompt(mode: 'steer')` contract remains best-effort for new primary-session input: outside the next-step window it becomes a waking follow-up. The composer carries an explicit `queue | steer` mode through slash adjudication and reference serialization before calling that contract. A browser-local submission policy owns the persisted busy-Enter preference and resolves plain versus accelerated Enter as complementary gestures only for steer-capable sessions; the Settings row and InputBar share that policy without duplicating storage or delivery-window authority. Only the Queue row action is strict, because either negative result converges through the original Queue occurrence. ### Verification -AgentLoop contract coverage holds prompt admission open, converts one exact queued occurrence, and proves the replacement steering occurrence keeps the message value, drains as `steering/message`, and never starts its former independent turn. It also pins unavailable-window retention, claimed-address rejection, and re-entrant cancellation lifecycle conservation. +AgentLoop contract coverage holds prompt admission open, converts one exact queued occurrence, and proves the replacement steering occurrence keeps the message value and delivery receipt, drains as `steering/message`, and never starts its former independent turn. It also pins unavailable-window retention, claimed-address rejection, and re-entrant cancellation lifecycle conservation. -Host schema and proxy tests cover the new action, both typed errors, placement-aware snapshots and reconnect replay, plus durable-before-retirement ordering. QueueDock tests cover running-state enablement, complete-content eligibility, failure retention, authoritative success retirement, and filtering of steering occurrences. ChatView tests and Web ARIA snapshots cover pending steering after the running-status row with Copy alone, plus its single-copy handoff to a durable node with clock, Copy, and Fork. +Host schema and proxy tests cover the new action, both typed errors, placement-aware snapshots and reconnect replay, plus durable-before-retirement ordering. Client tests cover silent convergence of both semantic races, genuine error reporting, read-only subagent rows and Queue-only subagent gestures. Runtime and ChatView tests cover occurrence-aware pending-to-durable handoff, including repeated `MessageId` values, while Web ARIA snapshots cover pending steering after the running-status row with Copy alone and the durable node with clock, Copy, and Fork. The keyless Web steering scenario queues a message through the real composer while the first response streams, activates the row arrow, then uses `ask_user_question` as a stable pending-steering barrier. It proves the Host-backed pending bubble appears before admission, hands off to one durable interjection after the answer, and affects the next model request. Assembled composer scenarios prove default-mode Cmd+Enter reaches the same pending and durable path without creating a Queue row, while Steer-mode Cmd+Enter creates a Queue row instead. Settings and submission-policy coverage pin the default, persistence, busy-only scope, and complementary gesture mapping; Queue edit/delete scenarios continue to prove those actions are unchanged. @@ -54,7 +54,7 @@ The keyless Web steering scenario queues a message through the real composer whi **Restore Queue promotion under the upward arrow.** Rejected because moving an item to the front still creates an independent admitted turn. The control promises current-turn steering, not priority within Queue. -**Use the existing best-effort `agent.steer()` behavior for the Queue row.** Rejected for that action because a closed next-step window would silently turn the selected row back into queued work, possibly at a different position and identity. Strict failure preserves the original occurrence and makes the semantic race visible. Newly typed composer input has no existing Queue occurrence to preserve, so it intentionally uses the best-effort behavior. +**Use the existing best-effort `agent.steer()` behavior for the Queue row.** Rejected for that action because a closed next-step window would create a new queued occurrence, possibly at a different position and identity. Strict refusal preserves the original occurrence so the UI can treat it as the same accepted Queue delivery. Newly typed composer input has no existing Queue occurrence to preserve, so it intentionally uses the best-effort behavior. **Change `agent.steer()` to be strict for every caller.** Rejected because TUI and plugin callers use its safe follow-up fallback for newly submitted input. A queued row has recoverable state that those callers do not. @@ -66,6 +66,6 @@ The keyless Web steering scenario queues a message through the real composer whi ## Consequences -`session/queue` describes a placement-aware transient inbox snapshot rather than a Queue-only list, so every consumer must filter by placement. Pending steering survives reconnect and appears immediately, but remains non-durable until `steering/message` commits. The running bit can also remain true briefly after the strict next-step window closes, so the button may be enabled for an operation that correctly returns `steer-unavailable`. +`session/queue` describes a placement-aware transient inbox snapshot rather than a Queue-only list, so every consumer must filter by placement. Pending steering survives reconnect and appears immediately, but remains non-durable until `steering/message` commits. The running bit can also remain true briefly after the strict next-step window closes, so an enabled action may internally return `steer-unavailable` while the product continues through Queue without reporting failure. The explicit action changes delivery from an independently admitted turn to current-turn steering, so prompt-admission plugins do not process the converted message. Enqueue-before-discard lifecycle publication remains required for re-entrant cancellation safety; focused regression coverage protects that ordering. diff --git a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md index 6b8af563ce..6edde8e2cc 100644 --- a/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md +++ b/.agents/notes/implemented/feature/2026-07-30-web-queue-steer-action.zh.md @@ -14,19 +14,19 @@ Web composer 原本会在 agent 运行期间把所有 Enter 提交作为 Queue ### 产品契约 -每个非编辑态的 QueueDock 行都会提供名为“插话发送”的向上箭头操作。仅当会话报告 agent 正在运行时,该操作才会启用;包含混合内容的消息仍可使用,因为 steering 会转发完整且不可变的 `UserMessage`,而非该行的文本投影。编辑和删除保持现有行为。 +普通会话中每个非编辑态的 QueueDock 行都会提供名为“插话发送”的向上箭头操作。仅当会话报告 agent 正在运行时,该操作才会启用;包含混合内容的消息仍可使用,因为 steering 会转发完整且不可变的 `UserMessage`,而非该行的文本投影。已寻址 subagent 的 Queue 投影保持只读,因为其继续执行传输不提供 Queue 变更。 触发该操作会针对对应的 `InboxItemId` 请求严格的当前轮次 steering。操作成功后,权威 Host 快照会移除 Queue 行,并在 `Deep diving...` 运行状态行之后立即投影同一条待处理 steering;该气泡提供复制,但消息尚无持久事件序号,因此不提供 fork。AgentLoop 排空该项后,现有持久 `steering/message` 事件会接管同一个用户样式气泡,并恢复时钟、复制和 fork,无需另建持久展示路径。 -running 标志位只用于提示交互状态。在同步变更边界上,AgentLoop 的 `acceptsNextStep` 值才是权威依据。如果该窗口已经关闭,操作会保持 Queue 单次入队项不变,并返回类型化的 `steer-unavailable` 错误;如果驱动器已经认领该项,则返回现有的 `queue-item-not-found` 错误。UI 会报告任一竞态,不会乐观地移除该行。 +running 标志位只用于提示交互状态。在同步变更边界上,AgentLoop 的 `acceptsNextStep` 值才是权威依据。如果该窗口已经关闭,操作会保持 Queue 单次入队项不变并返回类型化的 `steer-unavailable` 错误,随后原唤醒单次入队项会经 Queue 继续执行。如果驱动器已经认领该项,则返回现有的 `queue-item-not-found` 错误,且独立轮次投递已经开始。UI 会把两种竞态都视为已收敛的 Queue 投递,不显示失败通知;传输和未知错误仍会显示。 -Composer 对新输入采用另一套尽力而为契约。所寻址会话空闲时,Enter 和 Cmd/Ctrl+Enter 都执行普通 Queue 发送。会话运行期间,General Settings 偏好会把普通 Enter 分配为 Queue(默认值)或 Steer,Cmd/Ctrl+Enter 则执行另一种行为;Shift+Enter 用于换行。浏览器会持久化该偏好,并且它只影响繁忙态下这对手势。如果 composer 直接发出的 Steer 错过当前 next-step 窗口,AgentLoop 会自动将其接纳为下一条唤醒 Queue 轮次,Web 不显示失败。 +Composer 对新输入采用另一套尽力而为契约。所寻址会话空闲时,Enter 和 Cmd/Ctrl+Enter 都执行普通 Queue 发送。主会话运行期间,General Settings 偏好会把普通 Enter 分配为 Queue(默认值)或 Steer,Cmd/Ctrl+Enter 则执行另一种行为;Shift+Enter 用于换行。已寻址 subagent 会让这两个手势都使用其仅支持 Queue 的继续执行传输。浏览器会持久化该偏好,并且它只影响支持 steering 的繁忙态手势对。如果 composer 直接发出的 Steer 错过当前 next-step 窗口,AgentLoop 会自动将其接纳为下一条唤醒 Queue 轮次,Web 不显示失败。 ### Agent 与生命周期边界 `InboxAction` 会在编辑和移除之外,新增由实际消费方支撑的 `{ kind: 'steer' }` 操作。`Agent.updateInbox()` 只有在找到 queued 单次入队项并确认 `acceptsNextStep` 后才会处理该操作,绝不会委托给尽力而为的 `agent.steer()` 别名。 -操作成功应用后,系统会结束 queued 单次入队项,并把同一个不可变 `UserMessage` 接受为新的 steering 单次入队项。steering 单次入队项会获得新的 `InboxItemId` 和如实反映投递方式的 `placement: 'steering'`,消息则保留其 `MessageId`、内容和来源。AgentLoop 会先安装新的 outbox 项,再发布生命周期事件;随后先发出新单次入队项的 enqueue,再发出旧单次入队项的 discard,确保可重入取消无法观察或退役一个尚未公布的项。因此,现有 inbox 守恒不变量仍然要求每个单次入队项恰好对应一个 enqueue,以及一个终态 dequeue 或 discard。 +操作成功应用后,系统会结束 queued 单次入队项,并把同一个不可变 `UserMessage` 接受为新的 steering 单次入队项。steering 单次入队项会获得新的 `InboxItemId` 和如实反映投递方式的 `placement: 'steering'`,消息则保留其 `MessageId`、内容、来源和任何待处理 `SteeringReceipt` 投递控制器。AgentLoop 会先安装新的 outbox 项,再发布生命周期事件;随后先发出新单次入队项的 enqueue,再发出旧单次入队项的 discard,确保可重入取消无法观察或退役一个尚未公布的项。因此,现有 inbox 守恒不变量仍然要求每个单次入队项恰好对应一个 enqueue,以及一个终态 dequeue 或 discard。 该操作不会运行 `agent/prompt-submit`:选择 steering 会有意把投递方式从经独立接纳的轮次改为当前轮次的 next-step 输入。它既不会取消当前工作,也不会重新排序 Queue 中的剩余项。 @@ -36,15 +36,15 @@ Composer 对新输入采用另一套尽力而为契约。所寻址会话空闲 Host 仍以现有 `queuedMirror` 作为唯一的瞬态 inbox 权威。`session/queue` 快照会携带所有存活单次入队项及其 `placement: 'queued' | 'steering'`:QueueDock 只渲染 queued 行,ChatView 则在会话流末尾、`Deep diving...` 运行状态行之后渲染待处理 steering,提供复制操作,但不提供 fork、编辑或删除操作。重连会重放同一份快照,因此这项可见性既不依赖客户端乐观展示,也不需要第二个 registry。 -AgentLoop 认领待处理 steering 时,会在同步追加 `steering/message` 之前立即发出 `agent/inbox/dequeue`。Host 会等到下一个微任务才退役该 steering 行,让持久 session 事件先进入线性 mux 流。ChatView 会匹配两边共享的 `MessageId`,并在持久节点出现时立即抑制瞬态投影,因此同一个气泡切换权威时不会产生可见空档或重复;持久投影会根据已记录的事件时间与序号恢复时钟、复制与 fork 操作,而追加失败时已认领行仍会退役。 +AgentLoop 认领待处理 steering 时,会在同步追加 `steering/message` 之前立即发出 `agent/inbox/dequeue`。Host 会等到下一个微任务才退役该 steering 行,让持久 session 事件先进入线性 mux 流。客户端 Session 接纳该实时事件时,会在发布快照前退役第一个匹配的当前 steering 单次入队项;历史回放不会消费后来复用同一 `MessageId` 的单次入队项。因此,ChatView 无需扫描持久历史就能每次只渲染一份权威,持久投影则会根据已记录的事件时间与序号恢复时钟、复制与 fork 操作。追加失败时,已认领行仍会退役。 -现有 `session.prompt(mode: 'steer')` 对新输入仍采用尽力而为的契约:在 next-step 窗口之外,它会变为唤醒 agent 的后续轮次。Composer 会让显式 `queue | steer` 模式经过 slash 裁决与引用序列化,再调用该契约。浏览器本地的提交策略拥有持久化的繁忙态 Enter 偏好,并把普通 Enter 与加速 Enter 解析为互补手势;Settings 行和 InputBar 共享该策略,不重复实现存储或投递窗口权威。只有 Queue 行操作采用严格语义,因为失败时可以安全地保留其已经待处理的消息。 +现有 `session.prompt(mode: 'steer')` 对主会话新输入仍采用尽力而为的契约:在 next-step 窗口之外,它会变为唤醒 agent 的后续轮次。Composer 会让显式 `queue | steer` 模式经过 slash 裁决与引用序列化,再调用该契约。浏览器本地的提交策略拥有持久化的繁忙态 Enter 偏好,并且只为支持 steering 的会话把普通 Enter 与加速 Enter 解析为互补手势;Settings 行和 InputBar 共享该策略,不重复实现存储或投递窗口权威。只有 Queue 行操作采用严格语义,因为任一种负面结果都会经原 Queue 单次入队项收敛。 ### 验证 -AgentLoop 契约覆盖保持提示词接纳窗口打开,转换一个精确的 queued 单次入队项,并证明替代它的 steering 单次入队项保留消息值、以 `steering/message` 的形式排空,且绝不启动原本的独立轮次。该覆盖还钉住窗口不可用时保留原项、拒绝已被认领的地址,以及可重入取消下的生命周期守恒。 +AgentLoop 契约覆盖保持提示词接纳窗口打开,转换一个精确的 queued 单次入队项,并证明替代它的 steering 单次入队项保留消息值和投递回执、以 `steering/message` 的形式排空,且绝不启动原本的独立轮次。该覆盖还钉住窗口不可用时保留原项、拒绝已被认领的地址,以及可重入取消下的生命周期守恒。 -Host schema 和代理测试覆盖新操作、两种类型化错误、带 placement 的快照与重连重放,以及先持久化再退役的顺序。QueueDock 测试覆盖按运行状态启用、混合内容消息仍可完整投递、失败时保留原行、成功后由权威快照退役,以及过滤 steering 单次入队项。ChatView 测试与 Web ARIA 快照覆盖待处理阶段位于运行状态行之后且仅有复制,并覆盖它向带时钟、复制和 fork 的持久节点进行单份交接。 +Host schema 和代理测试覆盖新操作、两种类型化错误、带 placement 的快照与重连重放,以及先持久化再退役的顺序。客户端测试覆盖两种语义竞态的静默收敛、真实错误报告、只读 subagent 行和仅支持 Queue 的 subagent 手势。运行时与 ChatView 测试覆盖按单次入队项完成的待处理到持久交接,包括重复的 `MessageId` 值;Web ARIA 快照则覆盖位于运行状态行之后且仅有复制的待处理 steering,以及带时钟、复制和 fork 的持久节点。 无密钥 Web steering 场景在第一次响应流式输出期间,通过真实 composer 排队一条消息并触发行上的箭头,再用 `ask_user_question` 作为稳定的待处理 steering 屏障。该场景证明 Host 支撑的待处理气泡会在准入前出现,在回答后交接为唯一一条持久插话,并影响下一次模型请求。组装后的 composer 场景证明默认模式下的 Cmd+Enter 无需创建 Queue 行,也会进入同一条待处理与持久路径;Steer 模式下的 Cmd+Enter 则会创建 Queue 行。Settings 与提交策略覆盖会固定默认值、持久化、仅限繁忙态的作用域和互补手势映射;Queue 编辑/删除场景继续证明这些操作没有变化。 @@ -54,7 +54,7 @@ Host schema 和代理测试覆盖新操作、两种类型化错误、带 placeme **恢复向上箭头对应的 Queue 前移操作。** 不予采纳,因为把某个项移到队首仍然会创建一个独立接纳的轮次。该控件承诺的是当前轮次 steering,而不是 Queue 内的优先级。 -**为 Queue 行使用现有尽力而为的 `agent.steer()` 行为。** 不予采纳,因为关闭的 next-step 窗口会静默地把选中行重新变成 queued 工作,而且位置和标识可能不同。严格失败会保留原单次入队项,并让这项语义竞态明确可见。新输入的 composer 消息没有需要保留的现有 Queue 单次入队项,因此有意采用尽力而为行为。 +**为 Queue 行使用现有尽力而为的 `agent.steer()` 行为。** 不予采纳,因为关闭的 next-step 窗口会创建新的 queued 单次入队项,而且位置和标识可能不同。严格拒绝会保留原单次入队项,让 UI 能将其视为同一次已接纳的 Queue 投递。新输入的 composer 消息没有需要保留的现有 Queue 单次入队项,因此有意采用尽力而为行为。 **让每个调用方使用的 `agent.steer()` 都采用严格语义。** 不予采纳,因为 TUI 和插件调用方会针对新提交的输入使用其安全的后续轮次回退。queued 行具有这些调用方不具备的可恢复状态。 @@ -66,6 +66,6 @@ Host schema 和代理测试覆盖新操作、两种类型化错误、带 placeme ## 后果 -`session/queue` 表示带 placement 的瞬态 inbox 快照,而不只是 Queue 列表,因此每个消费方都必须按 placement 过滤。待处理 steering 会在界面中立即出现并能在重连后恢复,但在 `steering/message` 提交前仍不持久。严格 next-step 窗口关闭后,running 标志位仍可能短暂保持为 true,因此按钮可能会为一个最终正确返回 `steer-unavailable` 的操作保持启用。 +`session/queue` 表示带 placement 的瞬态 inbox 快照,而不只是 Queue 列表,因此每个消费方都必须按 placement 过滤。待处理 steering 会在界面中立即出现并能在重连后恢复,但在 `steering/message` 提交前仍不持久。严格 next-step 窗口关闭后,running 标志位仍可能短暂保持为 true,因此已启用的操作可能会在内部返回 `steer-unavailable`,而产品仍经 Queue 继续执行且不显示失败。 这项显式操作会把投递方式从经独立接纳的轮次改为当前轮次 steering,因此提示词接纳插件不会处理转换后的消息。为保证可重入取消安全,生命周期事件仍必须先发布 enqueue 再发布 discard;有针对性的回归覆盖会保护这一顺序。 diff --git a/apps/web/tests/steering.e2e.ts b/apps/web/tests/steering.e2e.ts index 1f9672d74a..1ea43f99db 100644 --- a/apps/web/tests/steering.e2e.ts +++ b/apps/web/tests/steering.e2e.ts @@ -25,6 +25,11 @@ const FIXTURE = join(SNAPSHOT_DIR, 'session.jsonl') const MID_EXPECTED = join(SNAPSHOT_DIR, 'mid-steer.expected.md') const SETTLED_EXPECTED = join(SNAPSHOT_DIR, 'settled.expected.md') const MODE = webSnapshotMode() +// The question composer replaces the textarea, so fill → Queue row → Steer +// must finish inside the first replay chunk window. At 15 ms that window is +// shorter than Playwright's round trips; 100 ms supplies test-only headroom, +// while larger values lengthen all three replay scenarios linearly. +const REPLAY_PACE_MS = 100 const PROMPT = 'Use the ask_user_question tool to ask me exactly one question with id "checkpoint", question "Ready to continue?", header "Checkpoint", and options labeled "Yes" and "No". After I answer, reply with one short sentence acknowledging my answer and stop.' const STEER = 'Interjection: include the word BANANA in your final reply.' @@ -48,9 +53,9 @@ describe('web e2e: mid-turn steering lands durably and visibly', () => { const sessionEvents: SessionEvent[] = [] beforeAll(async () => { - // The 100 ms replay pace is test-only headroom that keeps the Queue action - // available until the recorded question barrier arrives, not a product timing contract. - scaffold = await launchWebScaffold(MODE === 'record' ? {} : { replayFixture: FIXTURE, paceMs: 100 }) + scaffold = await launchWebScaffold(MODE === 'record' + ? {} + : { replayFixture: FIXTURE, paceMs: REPLAY_PACE_MS }) scaffold.ctx.on('session/event', (_session, event) => { sessionEvents.push(event) }) browser = await chromium.launch() page = await newEnglishPage(browser) @@ -89,6 +94,8 @@ describe('web e2e: mid-turn steering lands durably and visibly', () => { await expect.poll(() => steerButton.isEnabled(), { timeout: 10_000 }).toBe(true) await steerButton.click({ timeout: 10_000 }) const pendingSteering = page.locator('[data-pending-steering]').filter({ hasText: STEER }) + // A timeout while the Queue row remains means strict steer lost to a + // closing window (`steer-unavailable`); inspect replay pacing first. await pendingSteering.waitFor({ timeout: 10_000 }) // The blocked composer keeps steering pending long enough to observe the @@ -158,7 +165,7 @@ describe('web e2e: composer shortcut steers directly', () => { const sessionEvents: SessionEvent[] = [] beforeAll(async () => { - scaffold = await launchWebScaffold({ replayFixture: FIXTURE, paceMs: 100 }) + scaffold = await launchWebScaffold({ replayFixture: FIXTURE, paceMs: REPLAY_PACE_MS }) scaffold.ctx.on('session/event', (_session, event) => { sessionEvents.push(event) }) browser = await chromium.launch() page = await newEnglishPage(browser) @@ -216,7 +223,7 @@ describe('web e2e: composer shortcut follows the swapped busy behavior', () => { const sessionEvents: SessionEvent[] = [] beforeAll(async () => { - scaffold = await launchWebScaffold({ replayFixture: FIXTURE, paceMs: 100 }) + scaffold = await launchWebScaffold({ replayFixture: FIXTURE, paceMs: REPLAY_PACE_MS }) scaffold.ctx.on('session/event', (_session, event) => { sessionEvents.push(event) }) browser = await chromium.launch() page = await newEnglishPage(browser) diff --git a/packages/client/runtime/README.i18n.yaml b/packages/client/runtime/README.i18n.yaml index 0bbb962352..6f31ca3beb 100644 --- a/packages/client/runtime/README.i18n.yaml +++ b/packages/client/runtime/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/client/runtime/README.md -README.md: a98e97d796ca4e4be07d8d8d25ebc0a24d066a8b -README.zh.md: a973c4fbbf16633fed11d137ae548604516d8a51 +README.md: eca7db1f9b2d5c7e28fa86a363ca4408703b99df +README.zh.md: 6a2e8c6085d06a9f04c1270e5976452b995a7e77 diff --git a/packages/client/runtime/README.md b/packages/client/runtime/README.md index a98e97d796..eca7db1f9b 100644 --- a/packages/client/runtime/README.md +++ b/packages/client/runtime/README.md @@ -22,7 +22,7 @@ SlotsService gives the renderer separate bare observables for `useSessions` and ## Pending queue projection -`ConversationSnapshot.queue` is the Host's authoritative transient Queue snapshot; pending steering stays outside this projection. Each row carries its `InboxItemId`, complete editable text when every content block is text, and a flattened preview. `session/queue` replaces the whole projection; reconnect buffering retains only the latest snapshot, and neither durable turn events nor running-status changes guess that an item was claimed. `Session.updateQueue()` sends edit, remove, and strict-steer operations without optimistic mutation, so the next Host snapshot is the sole visible commit; claim and closed-window races surface `queue-item-not-found` and `steer-unavailable`. +`ConversationSnapshot.queue` is the Host's authoritative transient inbox snapshot and carries both queued and pending-steering occurrences with their resolved placement. Each row carries its `InboxItemId`, stable `MessageId`, complete editable text when every content block is text, and a flattened preview. `session/queue` replaces the whole projection, while an accepted live `steering/message` event retires only the first matching current steering occurrence so the durable node can take over before the following Host snapshot; history replay never consumes a later occurrence that reused the same `MessageId`. Reconnect buffering retains only the latest snapshot, and neither ordinary durable turn events nor running-status changes guess that an item was claimed. `Session.updateQueue()` sends edit, remove, and strict-steer operations without optimistic mutation; claim and closed-window races surface `queue-item-not-found` and `steer-unavailable`. ## The human transcript diff --git a/packages/client/runtime/README.zh.md b/packages/client/runtime/README.zh.md index a973c4fbbf..6a2e8c6085 100644 --- a/packages/client/runtime/README.zh.md +++ b/packages/client/runtime/README.zh.md @@ -22,7 +22,7 @@ SlotsService 分别为 renderer 提供 `useSessions` 与 `useWorkspaces` 的裸 ## 待处理队列投影 -`ConversationSnapshot.queue` 是 Host 提供的权威瞬态 Queue 快照;待处理 steering(中途引导)不进入此投影。每行都携带其 `InboxItemId`、所有内容块均为文本时的完整可编辑文本,以及扁平化预览。`session/queue` 会整体替换该投影;重连缓冲只保留最新快照,持久轮次事件和 running 状态变化都不会猜测某个项已被认领。`Session.updateQueue()` 发送编辑、移除和严格 steering 操作,不进行乐观更新,因此下一份 Host 快照是唯一可见的提交结果;认领与窗口关闭竞态分别会返回 `queue-item-not-found` 和 `steer-unavailable`。 +`ConversationSnapshot.queue` 是 Host 提供的权威瞬态 inbox 快照,携带 queued 与待处理 steering(中途引导)单次入队项及其已解析 placement。每行都携带其 `InboxItemId`、稳定的 `MessageId`、所有内容块均为文本时的完整可编辑文本,以及扁平化预览。`session/queue` 会整体替换该投影;已接纳的实时 `steering/message` 事件则只退役第一个匹配的当前 steering 单次入队项,让持久节点能在下一份 Host 快照之前接管,而历史回放绝不会消费后来复用同一 `MessageId` 的单次入队项。重连缓冲只保留最新快照,普通持久轮次事件和 running 状态变化都不会猜测某个项已被认领。`Session.updateQueue()` 发送编辑、移除和严格 steering 操作,不进行乐观更新;认领与窗口关闭竞态分别会返回 `queue-item-not-found` 和 `steer-unavailable`。 ## 面向人的 transcript(文本记录) diff --git a/packages/client/runtime/src/client/sessions/session.ts b/packages/client/runtime/src/client/sessions/session.ts index 490693a6a4..58d3028fba 100644 --- a/packages/client/runtime/src/client/sessions/session.ts +++ b/packages/client/runtime/src/client/sessions/session.ts @@ -650,9 +650,20 @@ export class Session implements SessionFace { this.events.push(event) this.views.push(view) this.transcript.append(event, view) + this.handoffPendingSteering(event) this.applyEventSideEffects(event, view) } + /** Retire the first matching live steering occurrence when its durable event takes over. */ + private handoffPendingSteering(event: SessionEvent): void { + if (event.type !== 'steering/message') return + const index = this.queued.findIndex(item => + item.placement === 'steering' && item.messageId === event.data.message.id) + if (index === -1) return + this.queued = this.queued.filter((_item, candidate) => candidate !== index) + this.queueRev++ + } + /** Land a live session/event (open/repair in flight -> buffer; overlapping seq -> drop; * a seq gap -> buffer + tail-page repull instead of appending a hole (audit S3: a gap is an * expected reconnect-window artifact, repaired by refetch). The window stays one contiguous diff --git a/packages/client/runtime/tests/queue-store.spec.ts b/packages/client/runtime/tests/queue-store.spec.ts index cc2f3d0e53..82d422aadc 100644 --- a/packages/client/runtime/tests/queue-store.spec.ts +++ b/packages/client/runtime/tests/queue-store.spec.ts @@ -5,7 +5,8 @@ */ import { describe, expect, it } from 'vitest' import { createUserMessage } from '@deepseek-ai/dsh-llm' -import type { ContentBlock } from '@deepseek-ai/dsh-llm/types' +import type { ContentBlock, UserMessage } from '@deepseek-ai/dsh-llm/types' +import type { SessionEvent } from '@deepseek-ai/dsh-session/types' import type { InboxItemId, MuxFrame, RpcId, SessionId, } from '@deepseek-ai/dsh-client-connection/client' @@ -23,6 +24,7 @@ interface QueueFixture { body: string content?: ContentBlock[] placement?: 'queued' | 'steering' + message?: UserMessage } /** Build one authoritative queue snapshot. */ @@ -33,7 +35,7 @@ function queueFrame(items: QueueFixture[]): MuxFrame { items: items.map(item => ({ id: iid(item.id), placement: item.placement ?? 'queued', - message: createUserMessage({ + message: item.message ?? createUserMessage({ content: item.content ?? text(item.body), source: { kind: 'user', rpcId: rid(`rpc-${item.id}`) } as never, }), @@ -134,6 +136,40 @@ describe('queue snapshot intake', () => { { id: 's-now', placement: 'steering', content: text('interrupt now') }, ]) }) + + it('hands off exactly one current occurrence when live steering becomes durable', async () => { + const session = makeSession() + await session.open() + const message = createUserMessage({ + content: text('same message'), + source: { kind: 'user' }, + }) + session.handleMuxEnvelope(rid('env-same-id'), queueFrame([ + { id: 's-first', body: '', placement: 'steering', message }, + { id: 's-second', body: '', placement: 'steering', message }, + ])) + const durable = { + seq: 0, + time: 1_700_000_000_000, + type: 'steering/message', + surfaceOp: 'append', + data: { turn: 1, message }, + } as SessionEvent + + session.handleMuxEnvelope(rid('env-durable'), { + type: 'session/event', sessionId: SID, event: durable, + }) + expect(session.getSnapshot().queue.map(item => item.id)).toEqual(['s-second']) + expect(session.getSnapshot().nodes.filter(node => node.kind === 'steering')).toHaveLength(1) + + session.handleMuxEnvelope(rid('env-reused-id'), queueFrame([ + { id: 's-later', body: '', placement: 'steering', message }, + ])) + session.handleMuxEnvelope(rid('env-replayed-durable'), { + type: 'session/event', sessionId: SID, event: durable, + }) + expect(session.getSnapshot().queue.map(item => item.id)).toEqual(['s-later']) + }) }) describe('queue operation transport', () => { diff --git a/packages/client/ui-conversation/README.i18n.yaml b/packages/client/ui-conversation/README.i18n.yaml index fcf4c0b84e..a78b80035b 100644 --- a/packages/client/ui-conversation/README.i18n.yaml +++ b/packages/client/ui-conversation/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/client/ui-conversation/README.md -README.md: 61056fc36e2e7f5f85726e158b9fd27e06c21941 -README.zh.md: 3a1a74a4106321415ee343908c999a2cf41615a0 +README.md: ea48725b02ad0c440984af4aadfec17d0e63791d +README.zh.md: 654901caca68762e313dda456c1e0df23faf734a diff --git a/packages/client/ui-conversation/README.md b/packages/client/ui-conversation/README.md index 61056fc36e..ea48725b02 100644 --- a/packages/client/ui-conversation/README.md +++ b/packages/client/ui-conversation/README.md @@ -34,11 +34,11 @@ Tool rows are slots too — the standalone tool ring (`ToolViewRegistry`/`ctx.to The todo surfaces are two registrations over that shape, both plain registrant plugins with `inject: ['slots', 'conversation']`. `TodoRow` takes the `'conversation.chat.toolview'` key `todo_write` and summarizes what the call attempted (`/ 已完成 · ` parsed from its args, falling back to the generic summary on malformed or wrongly-shaped model JSON, and keeping the generic dot for non-ok execution states so a cancelled call never reads as a completed update). `TodoDock` takes the `'conversation.input.dock'` list slot at `order: 0` — before Goal and Queue — and is the plan strip: it reads the host-computed `todos` projection via `useProjection` (standing plan: latest `todo/write` with no later `turn/start`) and renders `TodoPanel`, which takes the plain list, hides itself while the list is empty, and starts collapsed as a header of title plus `"/ tasks · in progress"` (status glyphs are the figma check / progress / dashed-pending set). The dock adapter owns the selection so the panel stays a pure function of its props; the standing list lives here rather than in the row so the row stays one line. Anything the input-zone composer chain hides (a `conversation.composer` takeover such as ui-question's) hides the whole dock, this strip included. -`QueueDock` is the terminal input-dock entry at `order: 20`. It hides while empty, renders one pending row directly, and defaults two or more rows to a collapsed `" 条排队消息"` header whose button expands or collapses the complete list. The header exposes `aria-expanded` and `aria-controls`; the expanded list scrolls within a 180px height bound. An active edit or mutation keeps its rows visible, and emptying the queue restores the collapsed default for the next queue. Each visible row remains a single-line preview with its exact-occurrence edit, delete, and strict-steer actions; an unavailable steering window leaves the Queue occurrence in place and reports the failure. +`QueueDock` is the terminal input-dock entry at `order: 20`. It hides while empty, renders one pending row directly, and defaults two or more rows to a collapsed `" 条排队消息"` header whose button expands or collapses the complete list. The header exposes `aria-expanded` and `aria-controls`; the expanded list scrolls within a 180px height bound. An active edit or mutation keeps its rows visible, and emptying the queue restores the collapsed default for the next queue. Each visible ordinary-session row remains a single-line preview with its exact-occurrence edit, delete, and strict-steer actions; addressed subagents retain the rows as a read-only projection because their continuation transport does not expose queue mutation. If strict steer loses to a closed window, the original occurrence remains queued for normal delivery; if the driver already claimed it, normal delivery is already underway. Neither converged race displays a failure, while transport and unknown failures do. -The Host's placement-aware `session/queue` snapshot also carries pending steering. QueueDock filters it out, while ChatView projects it as a user-style bubble with Copy at the conversation tail; Fork stays absent because the message has not entered a durable turn. The Host delays steering retirement until the durable `steering/message` has entered the mux stream, and ChatView deduplicates the two projections by their shared `MessageId`; the bubble therefore hands off without a gap or duplicate, restores Copy and Fork from the durable node, and survives reconnect from the same authority. +The Host's placement-aware `session/queue` snapshot also carries pending steering. QueueDock filters it out, while ChatView projects it as a user-style bubble with Copy at the conversation tail; Fork stays absent because the message has not entered a durable turn. The Host delays steering retirement until the durable `steering/message` has entered the mux stream. On that accepted live event, the client runtime retires the first matching current steering occurrence before publishing the snapshot; historical events cannot hide later occurrences that reuse the same `MessageId`. The bubble therefore hands off without a gap or duplicate, restores Copy and Fork from the durable node, and survives reconnect from the same authority. -Keyboard message submission resolves delivery from the addressed session's running state. While idle, Enter and Cmd/Ctrl+Enter both perform an ordinary Queue send. While running, the browser-persisted General Settings preference assigns plain Enter to `Queue` (the default) or `Steer`, and Cmd/Ctrl+Enter performs the other behavior; Shift+Enter remains a newline. The preference affects only the busy-state gesture pair, and the send button and non-keyboard submit actions remain Queue. Composer Steer uses the existing best-effort `session.prompt(mode: 'steer')` contract: if the current next-step window closes before acceptance, AgentLoop admits the message as the next waking Queue turn without surfacing a failure or losing the draft transaction. +Keyboard message submission resolves delivery from the addressed session's running state and steering capability. While idle, Enter and Cmd/Ctrl+Enter both perform an ordinary Queue send. While a primary session is running, the browser-persisted General Settings preference assigns plain Enter to `Queue` (the default) or `Steer`, and Cmd/Ctrl+Enter performs the other behavior; Shift+Enter remains a newline. Addressed subagents keep both gestures on their Queue-only continuation transport even while running. The preference affects only the steer-capable busy-state gesture pair, and the send button and non-keyboard submit actions remain Queue. Composer Steer uses the existing best-effort `session.prompt(mode: 'steer')` contract: if the current next-step window closes before acceptance, AgentLoop admits the message as the next waking Queue turn without surfacing a failure or losing the draft transaction. Per-session UI state for selection and the active view lives in the declared chat store (`stores.ts` `createChatStore`); the InputHub owns the composer state machine and mirrors its draft into that store for persistence. Apply passes one store handle to the strict session subtree, chat view, and details registrations, so each session shares one instance and the framework owns its lifecycle. Components are pure: the framework standard kit supplies `useSession`/`sessionId`, global `useSessions`/`useWorkspaces`, and the input machine's `useInput`/`inputActions`; store faces and inject factories supply the remaining state and callbacks. diff --git a/packages/client/ui-conversation/README.zh.md b/packages/client/ui-conversation/README.zh.md index 3a1a74a410..654901caca 100644 --- a/packages/client/ui-conversation/README.zh.md +++ b/packages/client/ui-conversation/README.zh.md @@ -34,11 +34,11 @@ todo 两个面就是在该形状上的两个注册项,都是普通注册方插件,`inject: ['slots', 'conversation']`。`TodoRow` 占用 `'conversation.chat.toolview'` 的 `todo_write` key,摘要该次调用「试图写入」的内容(从其 args 解析出 `<已完成>/<总数> 已完成 · <进行中条目>`;模型 JSON 残缺或形状不对时回落到通用摘要;非 ok 执行状态保留通用状态点,使被取消的调用绝不读成一次已完成的更新)。`TodoDock` 以 `order: 0` 占用 `'conversation.input.dock'` 列表 slot(位于 Goal 与 Queue 之前),是计划条:它经 `useProjection` 读取 host 计算的 `todos` 投影(站立计划:其后没有更晚 `turn/start` 的最近一次 `todo/write`)并渲染 `TodoPanel`,后者接收纯列表,在列表为空时自我隐藏;列表非空时面板初始折叠,表头显示标题加 `"<已完成>/<总数> tasks · in progress"`(状态图标为 figma 的勾选/进行中/虚线未开始一组)。选取由 dock 适配器负责,因此面板保持为其 props 的纯函数;站立列表放在此处而非行内,行才能保持单行。输入区 composer 链隐藏的一切(例如 ui-question 对 `conversation.composer` 的接管)也会隐藏整个 dock,包括这条计划条。 -`QueueDock` 是 `order: 20` 的末端 input-dock 条目。队列为空时隐藏;只有一个待处理项时直接渲染该行;存在两个或更多待处理项时,默认收起为 `" 条排队消息"` 表头,其按钮可展开或收起完整列表。表头暴露 `aria-expanded` 和 `aria-controls`;展开后的列表以 180px 为高度上限,并可滚动。存在进行中的编辑或变更时,列表行会保持可见;队列清空后,下一次出现队列时会恢复默认收起状态。每条可见行仍是单行预览,并提供针对精确单次入队项的编辑、删除和严格 steering(中途引导)操作;steering 窗口不可用时,Queue 单次入队项会留在原处并显示失败。 +`QueueDock` 是 `order: 20` 的末端 input-dock 条目。队列为空时隐藏;只有一个待处理项时直接渲染该行;存在两个或更多待处理项时,默认收起为 `" 条排队消息"` 表头,其按钮可展开或收起完整列表。表头暴露 `aria-expanded` 和 `aria-controls`;展开后的列表以 180px 为高度上限,并可滚动。存在进行中的编辑或变更时,列表行会保持可见;队列清空后,下一次出现队列时会恢复默认收起状态。普通会话中的每条可见行仍是单行预览,并提供针对精确单次入队项的编辑、删除和严格 steering(中途引导)操作;已寻址 subagent 则保留只读行,因为其继续执行传输不提供 Queue 变更。如果严格 steering 输给已关闭的窗口,原单次入队项会留在 Queue 中正常投递;如果驱动器已经认领该项,正常投递就已开始。这两种已收敛的竞态都不显示失败,传输和未知错误仍会显示。 -Host 带 placement 的 `session/queue` 快照也会携带待处理 steering。QueueDock 会将其过滤掉,ChatView 则把它投影为会话流末尾带复制操作的用户样式气泡;消息尚未进入持久轮次,因此不显示 fork。Host 会等持久 `steering/message` 进入 mux 流之后再退役 steering,ChatView 则按两份投影共享的 `MessageId` 去重;气泡交接时因而不会产生空档或重复,会从持久节点恢复复制与 fork 操作,并能在重连后从同一权威恢复。 +Host 带 placement 的 `session/queue` 快照也会携带待处理 steering。QueueDock 会将其过滤掉,ChatView 则把它投影为会话流末尾带复制操作的用户样式气泡;消息尚未进入持久轮次,因此不显示 fork。Host 会等持久 `steering/message` 进入 mux 流之后再退役 steering。客户端运行时接纳该实时事件时,会在发布快照前退役第一个匹配的当前 steering 单次入队项;历史事件无法隐藏后来复用同一 `MessageId` 的单次入队项。气泡交接时因而不会产生空档或重复,会从持久节点恢复复制与 fork 操作,并能在重连后从同一权威恢复。 -键盘消息提交会根据所寻址会话的运行状态解析投递方式。空闲时,Enter 和 Cmd/Ctrl+Enter 都执行普通 Queue 发送。运行期间,浏览器持久化的 General Settings 偏好会把普通 Enter 分配为 `Queue`(默认值)或 `Steer`,Cmd/Ctrl+Enter 则执行另一种行为;Shift+Enter 仍然换行。该偏好只影响繁忙态下这对手势,发送按钮与非键盘提交操作仍使用 Queue。Composer Steer 复用现有尽力而为的 `session.prompt(mode: 'steer')` 契约:如果当前 next-step 窗口在接纳前关闭,AgentLoop 会把消息接纳为下一条唤醒 Queue 轮次,不显示失败,也不会丢失草稿事务。 +键盘消息提交会根据所寻址会话的运行状态和 steering 能力解析投递方式。空闲时,Enter 和 Cmd/Ctrl+Enter 都执行普通 Queue 发送。主会话运行期间,浏览器持久化的 General Settings 偏好会把普通 Enter 分配为 `Queue`(默认值)或 `Steer`,Cmd/Ctrl+Enter 则执行另一种行为;Shift+Enter 仍然换行。已寻址 subagent 即使正在运行,也会让这两个手势都使用其仅支持 Queue 的继续执行传输。该偏好只影响支持 steering 的繁忙态手势对,发送按钮与非键盘提交操作仍使用 Queue。Composer Steer 复用现有尽力而为的 `session.prompt(mode: 'steer')` 契约:如果当前 next-step 窗口在接纳前关闭,AgentLoop 会把消息接纳为下一条唤醒 Queue 轮次,不显示失败,也不会丢失草稿事务。 逐 Session UI 状态中的选择与活跃视图位于已声明的聊天 store(`stores.ts` `createChatStore`)中;InputHub 拥有输入区状态机,并将草稿镜像到该 store 以便持久化。apply 将同一个 store handle 传给严格限定于会话的子树、聊天视图和详情注册,因此每个会话内共享一个实例,框架拥有其生命周期。组件保持纯粹:框架标准工具包提供 `useSession`/`sessionId`、全局 `useSessions`/`useWorkspaces`,以及输入状态机的 `useInput`/`inputActions`;store 表层与 inject factory 提供其余状态和回调。 diff --git a/packages/client/ui-conversation/src/client/apply.ts b/packages/client/ui-conversation/src/client/apply.ts index bb34164a3f..ec1f9f7ab9 100644 --- a/packages/client/ui-conversation/src/client/apply.ts +++ b/packages/client/ui-conversation/src/client/apply.ts @@ -222,7 +222,8 @@ export function apply(ctx: Context): void { if (sessionId === undefined) { return { keyboard: undefined, - resolveSubmitMode: (running, gesture) => submissionPolicy.resolve(running, gesture), + resolveSubmitMode: (running, gesture, steeringAvailable) => + submissionPolicy.resolve(running, gesture, steeringAvailable), toggleCommandMenu: undefined, stop: undefined, command: undefined, @@ -233,7 +234,8 @@ export function apply(ctx: Context): void { const slash = inputHub.slash(sessionId) return { keyboard: shell, - resolveSubmitMode: (running, gesture) => submissionPolicy.resolve(running, gesture), + resolveSubmitMode: (running, gesture, steeringAvailable) => + submissionPolicy.resolve(running, gesture, steeringAvailable), toggleCommandMenu: slash === undefined ? undefined : (selection) => { diff --git a/packages/client/ui-conversation/src/client/chat/ChatView.tsx b/packages/client/ui-conversation/src/client/chat/ChatView.tsx index 4f237c2d0c..989f56bba2 100644 --- a/packages/client/ui-conversation/src/client/chat/ChatView.tsx +++ b/packages/client/ui-conversation/src/client/chat/ChatView.tsx @@ -249,10 +249,10 @@ export function ChatView({ const selectedCallId = useStore(s => s.selection?.callId) const items = useMemo(() => deriveChatFlow(nodes), [nodes]) - const pendingSteering = useMemo(() => { - const durable = new Set(nodes.flatMap(node => node.kind === 'steering' ? [node.messageId] : [])) - return inbox.filter(item => item.placement === 'steering' && !durable.has(item.messageId)) - }, [inbox, nodes]) + const pendingSteering = useMemo( + () => inbox.filter(item => item.placement === 'steering'), + [inbox], + ) const activeRetry = useMemo(() => activeRetrySeq(nodes, running), [nodes, running]) // Only the last content assistant of each turn owns IconActions; mid-turn // text (before tools) omits `time` so AssistantMarkdown stays chrome-free. diff --git a/packages/client/ui-conversation/src/client/contract/slots.ts b/packages/client/ui-conversation/src/client/contract/slots.ts index 16f789f70c..52dc8dd2d9 100644 --- a/packages/client/ui-conversation/src/client/contract/slots.ts +++ b/packages/client/ui-conversation/src/client/contract/slots.ts @@ -287,7 +287,11 @@ export interface ComposerBarInjected { /** The InputBar-exclusive keyboard/DOM command face (decision 20 private plane); absent with the session. */ keyboard: ComposerKeyboard | undefined /** Resolve one keyboard submission gesture against the current running state and persisted preference. */ - resolveSubmitMode: (running: boolean, gesture: ComposerSubmitGesture) => InputSubmitMode + resolveSubmitMode: ( + running: boolean, + gesture: ComposerSubmitGesture, + steeringAvailable: boolean, + ) => InputSubmitMode /** Toggle the shared slash menu with only its command source; absent without ui-slash or a session. */ toggleCommandMenu: ((selection: EditSelection) => void) | undefined /** Cancel the in-flight turn; absent with the session. */ diff --git a/packages/client/ui-conversation/src/client/input/submission-policy.ts b/packages/client/ui-conversation/src/client/input/submission-policy.ts index 99ce299c52..6ef87e42c8 100644 --- a/packages/client/ui-conversation/src/client/input/submission-policy.ts +++ b/packages/client/ui-conversation/src/client/input/submission-policy.ts @@ -27,10 +27,15 @@ export class ComposerSubmissionPolicy { * Resolve one keyboard gesture without changing state. * @param running - whether the addressed agent currently reports busy. * @param gesture - plain Enter or the Cmd/Ctrl-accelerated chord. - * @returns Queue outside busy state; otherwise the preferred mode or its opposite. + * @param steeringAvailable - whether this session transport supports steering. + * @returns Queue outside steer-capable busy state; otherwise the preferred mode or its opposite. */ - resolve(running: boolean, gesture: ComposerSubmitGesture): InputSubmitMode { - if (!running) return 'queue' + resolve( + running: boolean, + gesture: ComposerSubmitGesture, + steeringAvailable: boolean, + ): InputSubmitMode { + if (!running || !steeringAvailable) return 'queue' const preferred = this.busyEnter.getSnapshot() if (gesture === 'enter') return preferred return preferred === 'queue' ? 'steer' : 'queue' diff --git a/packages/client/ui-conversation/src/client/locales.ts b/packages/client/ui-conversation/src/client/locales.ts index 1ecb33b854..87252fbc77 100644 --- a/packages/client/ui-conversation/src/client/locales.ts +++ b/packages/client/ui-conversation/src/client/locales.ts @@ -97,7 +97,7 @@ export const zh = { 'queue.steer.unavailable': '仅运行中可插话发送', 'queue.editFailed': '编辑失败:这条消息可能已经开始发送。', 'queue.removeFailed': '删除失败:这条消息可能已经开始发送。', - 'queue.steerFailed': '插话失败:当前回复已结束,或这条消息已经开始发送。', + 'queue.steerFailed': '插话发送失败,请重试。', 'terminal.signal': '信号 {signal}', 'terminal.exitCode': '退出码 {code}', 'terminal.running': '运行中', @@ -204,7 +204,7 @@ export const en = { 'queue.steer.unavailable': 'Steering is available only while the agent is running', 'queue.editFailed': 'Edit failed: this message may have already started sending.', 'queue.removeFailed': 'Removal failed: this message may have already started sending.', - 'queue.steerFailed': 'Steering failed: the current response ended or this message already started sending.', + 'queue.steerFailed': 'Steering failed. Try again.', 'terminal.signal': 'signal {signal}', 'terminal.exitCode': 'exit code {code}', 'terminal.running': 'Running', diff --git a/packages/client/ui-conversation/src/client/queue/QueueDock.tsx b/packages/client/ui-conversation/src/client/queue/QueueDock.tsx index 5085a6e579..67d6153519 100644 --- a/packages/client/ui-conversation/src/client/queue/QueueDock.tsx +++ b/packages/client/ui-conversation/src/client/queue/QueueDock.tsx @@ -32,6 +32,7 @@ export function QueueDock({ useSession, updateQueue, notify, t }: QueueDockProps const inbox = useSession(s => s.queue) const queue = useMemo(() => inbox.filter(row => row.placement === 'queued'), [inbox]) const running = useSession(s => s.running) + const queueMutable = useSession(s => s.subagent === null) const [editing, setEditing] = useState<{ id: QueueItemId; text: string } | null>(null) const [busy, setBusy] = useState(null) const [collapsed, setCollapsed] = useState(true) @@ -39,12 +40,12 @@ export function QueueDock({ useSession, updateQueue, notify, t }: QueueDockProps useEffect(() => { if (queue.length === 0 && !collapsed) setCollapsed(true) - if (editing !== null && !queue.some(row => row.id === editing.id)) setEditing(null) - }, [collapsed, editing, queue]) + if (editing !== null && (!queueMutable || !queue.some(row => row.id === editing.id))) setEditing(null) + }, [collapsed, editing, queue, queueMutable]) if (queue.length === 0) return null - const interactionActive = editing !== null || busy !== null + const interactionActive = queueMutable && (editing !== null || busy !== null) const expanded = !collapsed || interactionActive const listVisible = queue.length === 1 || expanded @@ -116,7 +117,7 @@ export function QueueDock({ useSession, updateQueue, notify, t }: QueueDockProps /> ) : {row.preview}} -
+ {queueMutable &&
{editing?.id === row.id ? ( <> @@ -190,7 +191,7 @@ export function QueueDock({ useSession, updateQueue, notify, t }: QueueDockProps )} -
+
} ))} diff --git a/packages/client/ui-conversation/src/client/service.ts b/packages/client/ui-conversation/src/client/service.ts index 4ecf41fd10..decca00e01 100644 --- a/packages/client/ui-conversation/src/client/service.ts +++ b/packages/client/ui-conversation/src/client/service.ts @@ -34,7 +34,7 @@ export interface IConversation { * Apply one edit, remove, or strict steer operation to a pending queue occurrence. * @param itemId - agent-owned inbox occurrence identity. * @param action - requested queue operation. - * @returns completion; business failures reject. + * @returns completion; converged strict-steer races resolve, while other failures reject. */ updateQueue(itemId: QueueItemId, action: QueueAction): Promise /** @@ -82,6 +82,10 @@ export class ConversationService extends Service implements IConversation { const session = this.scopedSession('updateQueue') const result = await session.updateQueue(itemId, action) if (!result.ok) { + if ( + action.kind === 'steer' + && (result.error.code === 'steer-unavailable' || result.error.code === 'queue-item-not-found') + ) return throw new Error(`conversation.updateQueue failed: ${result.error.code}: ${result.error.message}`) } } diff --git a/packages/client/ui-conversation/src/client/skeleton/InputBar.tsx b/packages/client/ui-conversation/src/client/skeleton/InputBar.tsx index 7b8aa87414..c2a718b84a 100644 --- a/packages/client/ui-conversation/src/client/skeleton/InputBar.tsx +++ b/packages/client/ui-conversation/src/client/skeleton/InputBar.tsx @@ -181,7 +181,11 @@ export function InputBar({ e.preventDefault() if (e.repeat) return // held-down Enter must not machine-gun sends if (locked || machineBusy) return - keyboard.submit(resolveSubmitMode(running, e.ctrlKey || e.metaKey ? 'accelerated' : 'enter')) + keyboard.submit(resolveSubmitMode( + running, + e.ctrlKey || e.metaKey ? 'accelerated' : 'enter', + subagent === null, + )) } const onChange = (e: ChangeEvent): void => { diff --git a/packages/client/ui-conversation/tests/chat-view.spec.tsx b/packages/client/ui-conversation/tests/chat-view.spec.tsx index b0e55736a4..04630cf1f6 100644 --- a/packages/client/ui-conversation/tests/chat-view.spec.tsx +++ b/packages/client/ui-conversation/tests/chat-view.spec.tsx @@ -288,7 +288,7 @@ describe('ChatView', () => { act(() => { h.set({ - queue: [queued, pending], + queue: [queued], nodes: [ assistant(1, 'working'), { @@ -306,9 +306,30 @@ describe('ChatView', () => { expect(branchButtons).toHaveLength(2) fireEvent.click(branchButtons[1]!) expect(h.forkAt).toHaveBeenCalledWith(2) + }) - act(() => { h.set({ queue: [queued] }) }) - expect(view.getAllByText('interrupt now')).toHaveLength(1) + it('keeps a later pending occurrence visible when it reuses a durable MessageId', () => { + const pending = { + id: 'steer-occurrence-later' as never, + messageId: 'shared-steer-message' as never, + placement: 'steering' as const, + content: [{ type: 'text' as const, text: 'same steering' }], + preview: 'same steering', + text: 'same steering', + } + const h = makeHarness({ + queue: [pending], + nodes: [{ + kind: 'steering', messageId: pending.messageId, + seq: 2, time: 2_000, turn: 1, + content: pending.content, source: null, + }], + running: true, + }) + const view = render() + + expect(view.getAllByText('same steering')).toHaveLength(2) + expect(view.container.querySelectorAll('[data-pending-steering]')).toHaveLength(1) }) it('animates only the latest unresolved model retry', () => { diff --git a/packages/client/ui-conversation/tests/input-bar.spec.tsx b/packages/client/ui-conversation/tests/input-bar.spec.tsx index bfe34316fa..180b41b73b 100644 --- a/packages/client/ui-conversation/tests/input-bar.spec.tsx +++ b/packages/client/ui-conversation/tests/input-bar.spec.tsx @@ -108,8 +108,8 @@ function bench(over?: BenchOptions) { useInput: bindSnapshotSelector(shell.state), inputActions: shell.actions, keyboard: shell, - resolveSubmitMode: (running, gesture) => { - if (!running) return 'queue' + resolveSubmitMode: (running, gesture, steeringAvailable) => { + if (!running || !steeringAvailable) return 'queue' const preferred = over?.busyEnter ?? 'queue' return gesture === 'enter' ? preferred : preferred === 'queue' ? 'steer' : 'queue' }, @@ -271,6 +271,24 @@ describe('running and lock semantics (queue cut 1)', () => { expect(empty.button.disabled).toBe(true) }) + it('keeps both running subagent Enter gestures on Queue transport', () => { + const subagent = { + address: { + parentSessionId: 'parent' as SessionId, + childSessionId: SID, + mode: 'continuable' as const, + }, + parentAvailable: true, + } + const plain = bench({ running: true, busyEnter: 'steer', draft: 'plain', subagent }) + fireEvent.keyDown(plain.textarea, { key: 'Enter' }) + expect(plain.sink).toHaveBeenCalledWith('plain', 'queue') + + const accelerated = bench({ running: true, draft: 'accelerated', subagent }) + fireEvent.keyDown(accelerated.textarea, { key: 'Enter', metaKey: true }) + expect(accelerated.sink).toHaveBeenCalledWith('accelerated', 'queue') + }) + it('disabled (session removed) locks the textarea and chrome', () => { const { textarea, view } = bench({ disabled: true }) expect(textarea.disabled).toBe(true) diff --git a/packages/client/ui-conversation/tests/queue-dock.spec.tsx b/packages/client/ui-conversation/tests/queue-dock.spec.tsx index 4195488a53..371766db5c 100644 --- a/packages/client/ui-conversation/tests/queue-dock.spec.tsx +++ b/packages/client/ui-conversation/tests/queue-dock.spec.tsx @@ -305,11 +305,34 @@ describe('QueueDock', () => { expect(rendered.getByLabelText('插话发送').getAttribute('title')).toBe('仅运行中可插话发送') }) - it('keeps the row and reports a strict steer race', async () => { + it('renders a session-backed subagent Queue without unsupported actions', () => { + const snap = { + ...snapshotWith([row('i-subagent', 'pending child follow-up')]), + subagent: { + address: { + parentSessionId: 'parent' as SessionId, + childSessionId: SID, + mode: 'continuable' as const, + }, + parentAvailable: true, + }, + } + const source = liveSession(snap) + const view = render( + , + ) + + expect(view.getByText('pending child follow-up')).toBeTruthy() + expect(view.queryByLabelText('编辑排队消息')).toBeNull() + expect(view.queryByLabelText('删除排队消息')).toBeNull() + expect(view.queryByLabelText('插话发送')).toBeNull() + }) + + it('keeps the row and reports a genuine steer failure', async () => { const snap = snapshotWith([row('i-steer-race', 'pending steer')]) const source = liveSession(snap) const notify = vi.fn() - const updateQueue = vi.fn(() => Promise.reject(new Error('steer unavailable'))) + const updateQueue = vi.fn(() => Promise.reject(new Error('transport failed'))) const { getByLabelText, getByText } = render( , ) @@ -318,7 +341,7 @@ describe('QueueDock', () => { await waitFor(() => { expect(notify).toHaveBeenCalledWith( 'error', - '插话失败:当前回复已结束,或这条消息已经开始发送。', + '插话发送失败,请重试。', ) }) expect(getByText('pending steer')).toBeTruthy() diff --git a/packages/client/ui-conversation/tests/service-orchestration.spec.ts b/packages/client/ui-conversation/tests/service-orchestration.spec.ts index 41a21b86e3..ebd51f8408 100644 --- a/packages/client/ui-conversation/tests/service-orchestration.spec.ts +++ b/packages/client/ui-conversation/tests/service-orchestration.spec.ts @@ -50,6 +50,29 @@ describe('ConversationService', () => { await expect(b.scoped.send('x')).rejects.toThrow('conversation.send failed: agent-busy: busy') b.cancel.mockResolvedValueOnce({ ok: false, error: { code: 'internal', message: 'nope', details: {} } } as never) await expect(b.scoped.cancel()).rejects.toThrow('conversation.cancel failed: internal: nope') + b.updateQueue.mockResolvedValueOnce({ + ok: false, error: { code: 'internal', message: 'broken', details: {} }, + } as never) + await expect(b.scoped.updateQueue('item-1' as never, { kind: 'steer' })) + .rejects.toThrow('conversation.updateQueue failed: internal: broken') + await b.runtime.dispose() + }) + + it('treats strict-steer races as converged Queue delivery', async () => { + const b = await bench() + b.updateQueue.mockResolvedValueOnce({ + ok: false, error: { code: 'steer-unavailable', message: 'closed', details: {} }, + } as never) + await expect(b.scoped.updateQueue('item-1' as never, { kind: 'steer' })).resolves.toBeUndefined() + b.updateQueue.mockResolvedValueOnce({ + ok: false, error: { code: 'queue-item-not-found', message: 'claimed', details: {} }, + } as never) + await expect(b.scoped.updateQueue('item-2' as never, { kind: 'steer' })).resolves.toBeUndefined() + b.updateQueue.mockResolvedValueOnce({ + ok: false, error: { code: 'queue-item-not-found', message: 'claimed', details: {} }, + } as never) + await expect(b.scoped.updateQueue('item-3' as never, { kind: 'remove' })) + .rejects.toThrow('conversation.updateQueue failed: queue-item-not-found: claimed') await b.runtime.dispose() }) diff --git a/packages/client/ui-conversation/tests/submission-policy.spec.ts b/packages/client/ui-conversation/tests/submission-policy.spec.ts index 485840fc2e..5b892982ab 100644 --- a/packages/client/ui-conversation/tests/submission-policy.spec.ts +++ b/packages/client/ui-conversation/tests/submission-policy.spec.ts @@ -13,19 +13,21 @@ describe('ComposerSubmissionPolicy', () => { it('defaults to Queue and only applies the preference while running', () => { const policy = new ComposerSubmissionPolicy() expect(policy.busyEnter.getSnapshot()).toBe(DEFAULT_BUSY_ENTER_BEHAVIOR) - expect(policy.resolve(false, 'enter')).toBe('queue') - expect(policy.resolve(false, 'accelerated')).toBe('queue') - expect(policy.resolve(true, 'enter')).toBe('queue') - expect(policy.resolve(true, 'accelerated')).toBe('steer') + expect(policy.resolve(false, 'enter', true)).toBe('queue') + expect(policy.resolve(false, 'accelerated', true)).toBe('queue') + expect(policy.resolve(true, 'enter', true)).toBe('queue') + expect(policy.resolve(true, 'accelerated', true)).toBe('steer') + expect(policy.resolve(true, 'enter', false)).toBe('queue') + expect(policy.resolve(true, 'accelerated', false)).toBe('queue') const changed = vi.fn() policy.busyEnter.subscribe(changed) policy.setBusyEnter('steer') expect(changed).toHaveBeenCalledTimes(1) - expect(policy.resolve(true, 'enter')).toBe('steer') - expect(policy.resolve(true, 'accelerated')).toBe('queue') - expect(policy.resolve(false, 'enter')).toBe('queue') - expect(policy.resolve(false, 'accelerated')).toBe('queue') + expect(policy.resolve(true, 'enter', true)).toBe('steer') + expect(policy.resolve(true, 'accelerated', true)).toBe('queue') + expect(policy.resolve(false, 'enter', true)).toBe('queue') + expect(policy.resolve(false, 'accelerated', true)).toBe('queue') expect(localStorage.getItem(BUSY_ENTER_STORAGE_KEY)).toBe('steer') }) diff --git a/packages/core/agent-loop/src/agent.ts b/packages/core/agent-loop/src/agent.ts index 88fcd64ead..7954d31cf3 100644 --- a/packages/core/agent-loop/src/agent.ts +++ b/packages/core/agent-loop/src/agent.ts @@ -264,7 +264,12 @@ export class ReactLoopAgent implements Agent { message: pending.item.message, placement: 'steering', }) - this.outbox.push({ message: item.message, steering: true, item }) + this.outbox.push({ + message: item.message, + steering: true, + item, + ...pending.delivery === undefined ? {} : { delivery: pending.delivery }, + }) // Publish the replacement only after it is owned by the outbox. Its // enqueue precedes the old occurrence's discard so reentrant // cancellation can terminally account for both occurrences. diff --git a/packages/core/agent-loop/tests/contract-regressions.spec.ts b/packages/core/agent-loop/tests/contract-regressions.spec.ts index f61af0833c..672b1aa597 100644 --- a/packages/core/agent-loop/tests/contract-regressions.spec.ts +++ b/packages/core/agent-loop/tests/contract-regressions.spec.ts @@ -179,8 +179,11 @@ describe('addressable inbox operations', () => { const idle = waitForIdle(ctx, agent) send(agent, 'open the turn') + const receipt = agent.steer(createUserMessage({ + content: [{ type: 'text', text: 'steer this message' }], + source: { kind: 'user' }, + })) await entered.promise - send(agent, 'steer this message') const queued = enqueued.find(item => inboxText(item) === 'steer this message')! expect(agent.updateInbox(queued.id, { kind: 'steer' })).toBe('applied') @@ -194,6 +197,7 @@ describe('addressable inbox operations', () => { expect(agent.session.events.flatMap(event => event.type === 'steering/message' ? [event.data.message] : [], )).toEqual([queued.message]) + expect(await receipt.outcome).toEqual({ status: 'admitted', turn: 1, step: 1 }) expect(agent.updateInbox(queued.id, { kind: 'steer' })).toBe('not-found') }) From 975b57b54aec648baf77219fea8d03c1b231478e Mon Sep 17 00:00:00 2001 From: ZiyaZhang <199893125+ZiyaZhang@users.noreply.github.com> Date: Sun, 2 Aug 2026 05:29:25 -0700 Subject: [PATCH 12/13] feat(web): follow live reasoning tail while collapsed --- ...6-08-02-web-thinking-tail-scroll.i18n.yaml | 6 ++++ .../2026-08-02-web-thinking-tail-scroll.md | 31 ++++++++++++++++ .../2026-08-02-web-thinking-tail-scroll.zh.md | 31 ++++++++++++++++ apps/web/tests/lifecycle-chrome.e2e.ts | 19 ++++++++-- .../client/ui-conversation/README.i18n.yaml | 4 +-- packages/client/ui-conversation/README.md | 2 ++ packages/client/ui-conversation/README.zh.md | 2 ++ .../src/client/chat/AssistantMarkdown.tsx | 9 ++++- .../src/client/chat/ToolRow.module.css | 5 +++ .../src/client/chat/ToolRow.tsx | 23 +++++++++--- .../tests/chat-tool-row.spec.tsx | 36 +++++++++++++++++++ 11 files changed, 158 insertions(+), 10 deletions(-) create mode 100644 .agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.i18n.yaml create mode 100644 .agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.md create mode 100644 .agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.zh.md diff --git a/.agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.i18n.yaml b/.agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.i18n.yaml new file mode 100644 index 0000000000..f4f0ef3891 --- /dev/null +++ b/.agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.i18n.yaml @@ -0,0 +1,6 @@ +# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each +# side as of the last confirmed-consistent state. Both languages carry equal authority; +# after editing either side, bring the other along and re-record with: +# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.md +2026-08-02-web-thinking-tail-scroll.md: c45840731153627b4ce460ee140257ba33d2c007 +2026-08-02-web-thinking-tail-scroll.zh.md: b8d0444d62294e123bec1d26cb4c07538bbf966f diff --git a/.agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.md b/.agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.md new file mode 100644 index 0000000000..c458407311 --- /dev/null +++ b/.agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.md @@ -0,0 +1,31 @@ +# Agent Note: Web thinking tail scroll — collapsed reasoning follows live output + +Status: implemented + +English | [中文](2026-08-02-web-thinking-tail-scroll.zh.md) + +## Problem + +The Web Think row rendered the first reasoning line as its collapsed summary for both settled and streaming blocks. Once that first line existed, every later reasoning delta changed hidden body text only. A fast model therefore looked stationary while it was thinking, and the user had to expand the full chain of thought to verify that output was still moving. The product backlog already called for “thinking: scrolling chain-of-thought updates, expandable”; the current row satisfied only the second half. + +## Decision + +Only a collapsed Think row whose reasoning block is the active streaming tail follows live output. Its summary is the latest non-blank line instead of the settled first line, and the existing single-line summary element becomes a programmatic horizontal scrollport pinned to `scrollWidth - clientWidth` after each text update. Direct `scrollLeft` assignment deliberately follows real deltas without inventing an independent marquee speed: fast tokens move fast, a paused model stops, and short text stays still because the scroll range is zero. + +The behavior is owned by the existing presentation components. `AssistantMarkdown` chooses the latest line only while the Think row is running; `ToolRow` already owns collapsed/open state and therefore owns whether its summary should follow the inline end. No session, wire, durable event, or model-visible contract changes. Expanding removes the collapsed summary and renders the complete reasoning body in ordinary page flow. When the row settles, it restores the stable first line and resets the summary to the left edge. Other tool summaries and settled Think rows retain their existing ellipsis behavior. + +## Alternatives considered + +**Animate a CSS marquee independent of streaming.** Rejected: it would keep moving through provider stalls and make a slow model look fast, which breaks the throughput signal the interaction exists to expose. + +**Always show a fixed suffix of the complete reasoning string.** Rejected: character slicing can cut a word or grapheme, discards the current line’s beginning before overflow actually requires it, and jumps rather than moving with each delta. + +**Auto-scroll the expanded reasoning body or the conversation page.** Rejected: expanded content is a reading surface. Forcing it to follow would fight a user who scrolls back; the follower belongs only to the collapsed one-line summary. + +## Consequences + +The collapsed row now communicates provider cadence through content motion as well as the existing sweep, while the settled transcript remains byte-for-byte stable. The scroll update runs only on React renders the streaming accumulator already causes; it adds no timer, animation loop, subscription, durable state, or transport traffic. A long current reasoning line retains its full DOM text and programmatically clips the already-overflowing prefix, so expansion still reveals the complete block and assistive technology reads the same current summary text. + +## Testing + +`packages/client/ui-conversation/tests/chat-tool-row.spec.tsx` pins the latest-line selection, the calculated right-edge scroll position, and the settlement reset to the first line and `scrollLeft = 0`. The keyless assembled Chromium scenario in `apps/web/tests/lifecycle-chrome.e2e.ts` replays real recorded reasoning chunks at observable pacing, narrows the viewport until the summary overflows, and asserts that the live collapsed Think row reaches its actual browser scroll extent. Its settled replay golden remains unchanged, proving the historical summary contract stays stable. diff --git a/.agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.zh.md b/.agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.zh.md new file mode 100644 index 0000000000..b8d0444d62 --- /dev/null +++ b/.agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.zh.md @@ -0,0 +1,31 @@ +# Agent Note:Web 思考尾部滚动 —— 折叠态 reasoning 跟随实时输出 + +Status: implemented + +[English](2026-08-02-web-thinking-tail-scroll.md) | 中文 + +## 问题 + +Web Think 行在结算与流式 block 中都把 reasoning 首行渲染成折叠摘要。首行一旦出现,之后每个 reasoning delta 只会改变隐藏的正文。于是快速模型在思考时看起来静止,用户必须展开完整思维链才能确认输出仍在推进。产品事项表已经要求“thinking:滚动展示思维链更新、可展开”;当前行只满足了后半项。 + +## 决策 + +只有 reasoning block 是当前流式尾部、且仍处于折叠态的 Think 行会跟随实时输出。其摘要使用最新的非空行,而不是结算后的首行;已有单行摘要元素成为程序化横向滚动区,每次文本更新后钉到 `scrollWidth - clientWidth`。这里刻意直接赋值 `scrollLeft`,通过真实 delta 推进而不虚构独立的跑马灯速度:token 快则移动快,模型停顿则停止,短文本因滚动范围为零而保持静止。 + +该行为由已有呈现组件拥有。`AssistantMarkdown` 只在 Think 行运行时选择最新行;`ToolRow` 已经拥有折叠/展开状态,因此由它决定摘要是否追随行内末端。不改变 session、wire、持久事件或模型可见契约。展开会移除折叠摘要,并让完整 reasoning 正文进入普通页面流。该行结算后恢复稳定首行,同时把摘要重置到左端。其他工具摘要与已结算 Think 行保留已有省略号行为。 + +## 曾考虑的替代方案 + +**播放与流式输出无关的 CSS 跑马灯。** 否决:它会在 provider 停顿时继续移动,让慢模型显得很快,破坏该交互本应暴露的吞吐信号。 + +**始终显示完整 reasoning 字符串的固定后缀。** 否决:按字符切片可能截断单词或字素,在内容真正溢出前就丢掉当前行的开头,而且只会跳变,无法随每个 delta 移动。 + +**自动滚动展开的 reasoning 正文或会话页面。** 否决:展开内容是阅读界面,强制跟随会与向上回看的用户争夺滚动;跟随器只属于折叠的单行摘要。 + +## 后果 + +折叠行现在会同时通过内容移动和已有扫光传达 provider 节奏,而结算后的 transcript 保持逐字节稳定。滚动更新只发生在流式累加器本就会触发的 React 渲染中;不会增加计时器、动画循环、订阅、持久状态或传输流量。较长的当前 reasoning 行仍会把完整文本留在 DOM 中,只以编程方式裁掉已经溢出的前缀,因此展开仍能显示完整 block,辅助技术读到的也仍是同一份当前摘要文本。 + +## 测试 + +`packages/client/ui-conversation/tests/chat-tool-row.spec.tsx` 固定最新行选择、算出的右端滚动位置,以及结算后恢复首行和 `scrollLeft = 0`。`apps/web/tests/lifecycle-chrome.e2e.ts` 中的 keyless 完整 Chromium 场景以可观察节奏回放真实录制的 reasoning chunks,把视口收窄到摘要溢出,并断言实时折叠 Think 行到达真实浏览器的滚动边界。其结算态 replay golden 保持不变,证明历史摘要契约仍然稳定。 diff --git a/apps/web/tests/lifecycle-chrome.e2e.ts b/apps/web/tests/lifecycle-chrome.e2e.ts index e285b0f0ac..65098063ad 100644 --- a/apps/web/tests/lifecycle-chrome.e2e.ts +++ b/apps/web/tests/lifecycle-chrome.e2e.ts @@ -158,8 +158,23 @@ describe('web e2e: lifecycle & chrome (workspace flow / reload / dark mode)', () } const settled = scaffold.whenTurnSettled() await input.fill(PROMPT) - await input.press('Enter') - const sessionId = await settled + const observeTurn = async () => { + if (MODE !== 'record') await page.setViewportSize({ width: 640, height: 1000 }) + try { + await input.press('Enter') + if (MODE !== 'record') { + const liveTail = page.locator('[data-variant="think"][data-state="running"] [data-follow-end]') + await expect.poll(async () => await liveTail.evaluate(element => ( + element.scrollWidth > element.clientWidth + && element.scrollLeft >= element.scrollWidth - element.clientWidth - 1 + )), { timeout: 10_000 }).toBe(true) + } + return await settled + } finally { + if (MODE !== 'record') await page.setViewportSize({ width: 1680, height: 1000 }) + } + } + const sessionId = await observeTurn() if (MODE === 'record') { await recordFixture(scaffold, sessionId, FIXTURE) } diff --git a/packages/client/ui-conversation/README.i18n.yaml b/packages/client/ui-conversation/README.i18n.yaml index a78b80035b..d131e31202 100644 --- a/packages/client/ui-conversation/README.i18n.yaml +++ b/packages/client/ui-conversation/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/client/ui-conversation/README.md -README.md: ea48725b02ad0c440984af4aadfec17d0e63791d -README.zh.md: 654901caca68762e313dda456c1e0df23faf734a +README.md: e610b990dd89204fd7e22e8b86f807d10b8ba439 +README.zh.md: 268e05a806db1468ba689608c178646af132fe25 diff --git a/packages/client/ui-conversation/README.md b/packages/client/ui-conversation/README.md index ea48725b02..e610b990dd 100644 --- a/packages/client/ui-conversation/README.md +++ b/packages/client/ui-conversation/README.md @@ -16,6 +16,8 @@ The session header declares and renders the session-scoped `'conversation.sessio Logged non-user messages render as a default-collapsed `上下文注入` disclosure. It shares the Tool calls header geometry and interaction with `ToolRow` through the package-internal `DisclosureRow`, while retaining context semantics: the expanded body follows its content height up to a 141px scrolling cap, shows inline JSON for both `content` and `source`, and synthesizes no tool state, summary, or keyed toolview dispatch ([decision](../../../.agents/notes/implemented/feature/2026-07-30-web-context-injection-disclosure.md)). +A Think row stays collapsed by default and exposes live reasoning throughput without expanding the chain of thought: while its reasoning block is the streaming tail, the summary switches from the settled first line to the latest non-blank line and its one-line scrollport follows each delta to the inline end. Expanding the row removes the moving summary and leaves the full reasoning in ordinary page flow, so page reading never fights an internal follower; settlement restores the stable first-line summary at the left edge ([decision](../../../.agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.md)). + Generic tool rows classify the built-in bash, read, search, write, edit, and run_code names into dedicated visual variants. The filesystem variants render the edit icon and a path summary; that path is a hover-underline link that opens the file with the host OS default application (`host.openPath`, relative paths resolve against the session cwd). Tool rows are not whole-row click targets and do not open the details panel. The code variant summarizes with the model-authored `description` and expands to the program itself; its logged sub-dispatches render as always-visible nested rows through the SAME keyed toolview hole (custom registrations and the GenericToolCard fallback apply to sub-rows unchanged). Cordis lifecycle tools reuse those generic variants while presenting `Inspect`, `Mount temporary Plugin`, and `Unmount temporary Plugin` with a shared Cordis accent; mount keeps the code variant's expandable source rendering. A tool call declaring the `terminal` render intent renders its command output inline, at both conversation render sites, through ui-primitives' `TerminalBlock`. `contract/terminal-card-model.ts` is the single derivation from the snapshot's `callView`/`resultView` pair, so the sites cannot disagree about a command, its cwd, or its exit status; it yields null — the generic path — for any other card tag, including one this client version does not know. Both sites therefore also show the card's run-state dot, which is the same `StateDot` semantic a tool row's leading icon carries, so a row and its own card always agree about one command's state. A multi-line command gets one prompt row per line, with the dot marking the call once on the first row — the exit status is the whole call's, so a dot per line would claim a per-line outcome bash does not report. The keyed `BashRow` carries the card resident below its summary row; since tool rows are no longer details-panel click targets, the card's copy and expand controls are the row's only interactions. The render-site fallback row keeps the card behind its existing expand control. Rows cap at `CHAT_TERMINAL_MAX_LINES` (8) against the panel's 16, which is what keeps a summary surface bounded — the panel stays the single-call reading surface. Inline output is licensed per render intent — the terminal and web cards, each with its own bound; a generic tool's content remains panel-only ([decision](../../../.agents/notes/implemented/feature/2026-07-28-web-terminal-card.md)). diff --git a/packages/client/ui-conversation/README.zh.md b/packages/client/ui-conversation/README.zh.md index 654901caca..268e05a806 100644 --- a/packages/client/ui-conversation/README.zh.md +++ b/packages/client/ui-conversation/README.zh.md @@ -14,6 +14,8 @@ 已记录的非用户消息渲染为默认折叠的 `上下文注入` 展开项。它通过包内部的 `DisclosureRow` 与 `ToolRow` 共享 Tool calls 标题栏的几何与交互,同时保留上下文语义:展开内容区的高度会随内容自适应,最大为 141px,超出后滚动,并以内联 JSON 展示 `content` 和 `source`,且不会合成工具状态、摘要或键控 toolview 分发([决策](../../../.agents/notes/implemented/feature/2026-07-30-web-context-injection-disclosure.md))。 +Think 行默认保持折叠,并在不展开思维链的情况下暴露实时推理吞吐:当 reasoning block 是流式尾部时,摘要从结算后的首行切换到最新的非空行,其单行滚动区会随每个 delta 追到行内末端。展开该行会移除移动摘要,让完整 reasoning 进入普通页面流,因此页面阅读不会与内部跟随器争夺滚动;结算后恢复左对齐的稳定首行摘要([决策](../../../.agents/notes/implemented/feature/2026-08-02-web-thinking-tail-scroll.md))。 + 通用工具行把内置的 bash、read、search、write、edit 和 run_code 名称归入专用视觉变体。文件系统变体会渲染 edit 图标和路径摘要;该路径是悬停下划线链接,点击后通过宿主操作系统的默认应用打开文件(`host.openPath`,相对路径相对会话 cwd 解析)。工具行不再是整行点击目标,也不会打开 details 面板。code 变体以模型撰写的 `description` 作摘要,展开后显示程序本身;其已记录的子调用经由同一个键控 toolview 空位渲染为始终可见的嵌套行(自定义注册和 GenericToolCard fallback 原样适用于子行)。Cordis 生命周期工具复用这些通用变体,同时以统一的 Cordis 强调色呈现 `Inspect`、`Mount temporary Plugin` 和 `Unmount temporary Plugin`;mount 行保留 code 变体的可展开源码渲染。 声明 `terminal` 渲染意图的工具调用,会在两个对话渲染点上都通过 ui-primitives 的 `TerminalBlock` 内联渲染其命令输出。`contract/terminal-card-model.ts` 是从快照的 `callView`/`resultView` 对推导的唯一位置,因此两个渲染点不可能在命令、cwd 或退出状态上产生分歧;对任何其他 card 标签——包括当前客户端版本不认识的标签——它返回 null,落回通用路径。因此两个渲染点也都显示卡片的运行状态点,它与工具行行首图标承载同一套 `StateDot` 语义,所以一行与其自身的卡片对同一条命令的状态总是一致。多行命令的每一行各占一个提示行,状态点只在第一行为整次调用标记一次——退出状态属于整次调用,因此每行一枚就会声称一个 bash 并不报告的逐行结果。键控的 `BashRow` 把卡片常驻在摘要行下方;由于工具行已不再是详情面板的点击目标,卡片的复制与展开控件就是该行唯一的交互。渲染点兜底行则保持其既有的展开控件。行的上限是 `CHAT_TERMINAL_MAX_LINES`(8),面板为 16,正是这一点让摘要面保持有界——面板仍是单次调用的阅读面。内联输出按渲染意图开放——终端卡片与 web 卡片,各有自己的上限;通用工具的内容仍然只在面板中呈现([决策](../../../.agents/notes/implemented/feature/2026-07-28-web-terminal-card.md))。 diff --git a/packages/client/ui-conversation/src/client/chat/AssistantMarkdown.tsx b/packages/client/ui-conversation/src/client/chat/AssistantMarkdown.tsx index 387a7fd82a..f5cbadef8a 100644 --- a/packages/client/ui-conversation/src/client/chat/AssistantMarkdown.tsx +++ b/packages/client/ui-conversation/src/client/chat/AssistantMarkdown.tsx @@ -39,6 +39,13 @@ function firstLine(text: string): string { return nl === -1 ? text : text.slice(0, nl) } +/** Latest non-blank reasoning line while the block is still streaming. */ +function latestLine(text: string): string { + const visible = text.trimEnd() + const nl = visible.lastIndexOf('\n') + return nl === -1 ? visible : visible.slice(nl + 1) +} + /** Joined text blocks for the copy action (reasoning / tool heads stay out). */ function copyText(blocks: readonly AssistantBlock[]): string { const parts: string[] = [] @@ -61,7 +68,7 @@ function ThinkRow({ text, running, t }: { text: string; running: boolean; t: Ass variant="think" icon={} title="Think" - summary={firstLine(text)} + summary={running ? latestLine(text) : firstLine(text)} body={text} state={running ? 'running' : 'ok'} /> diff --git a/packages/client/ui-conversation/src/client/chat/ToolRow.module.css b/packages/client/ui-conversation/src/client/chat/ToolRow.module.css index 81e41b066f..14bfe0e76c 100644 --- a/packages/client/ui-conversation/src/client/chat/ToolRow.module.css +++ b/packages/client/ui-conversation/src/client/chat/ToolRow.module.css @@ -84,6 +84,11 @@ color: var(--dsw-alias-label-tertiary); } +/* Live reasoning follows its one-line summary to the inline end. */ +.summary[data-follow-end] { + text-overflow: clip; +} + /* File-tool path: same geometry as .summary; hover underline + pointer. */ .fileLink { flex: 1 1 auto; diff --git a/packages/client/ui-conversation/src/client/chat/ToolRow.tsx b/packages/client/ui-conversation/src/client/chat/ToolRow.tsx index 46cdbda191..4ec07aa6fc 100644 --- a/packages/client/ui-conversation/src/client/chat/ToolRow.tsx +++ b/packages/client/ui-conversation/src/client/chat/ToolRow.tsx @@ -19,7 +19,7 @@ // independent); an error row's collapsed summary is the failure's first line in // the error color. -import { useState, type KeyboardEvent, type MouseEvent, type ReactNode } from 'react' +import { useLayoutEffect, useRef, useState, type KeyboardEvent, type MouseEvent, type ReactNode } from 'react' import clsx from 'clsx' import { CodeBlock, DiffBlock, ReadBlock, SearchBlock, StateDot, TerminalBlock, WebBlock, @@ -152,6 +152,7 @@ export function ToolRow({ inspect, }: ToolRowProps) { const [expanded, setExpanded] = useState(false) + const summaryRef = useRef(null) const terminalBody = terminal ?? null const diffBody = diff ?? null const readBody = read ?? null @@ -173,6 +174,15 @@ export function ToolRow({ const summaryText = failureLine ?? summary // The failure line is error prose, not the path: no open-file affordance. const fileLink = filePath !== undefined && onOpenFile !== undefined && failureLine === null + const isThink = variant === 'think' + const followSummaryEnd = isThink && state === 'running' && !open + useLayoutEffect(() => { + const summaryElement = summaryRef.current + if (summaryElement === null) return + summaryElement.scrollLeft = followSummaryEnd + ? summaryElement.scrollWidth - summaryElement.clientWidth + : 0 + }, [followSummaryEnd, summaryText]) const toggleExpand = () => { setExpanded(v => !v) } @@ -188,9 +198,8 @@ export function ToolRow({ if (event.key === 'Enter' || event.key === ' ') event.stopPropagation() } // Think reasoning is prose, not an input payload: expanded, it renders as - // plain indented text (no IN/OUT card) and the inline summary — the body's - // own first line — yields to avoid repeating itself. - const isThink = variant === 'think' + // plain indented text (no IN/OUT card) and the inline summary yields to avoid + // repeating the body. // The code variant's program renders through CodeBlock (shiki), so only its // output joins the IN/OUT card; every other variant's input does too. const cardBody = variant === 'code' ? null : body @@ -227,7 +236,11 @@ export function ToolRow({ {summaryText} ) : ( - + {summaryText} )} diff --git a/packages/client/ui-conversation/tests/chat-tool-row.spec.tsx b/packages/client/ui-conversation/tests/chat-tool-row.spec.tsx index e0d01656ba..ff9a2a3f07 100644 --- a/packages/client/ui-conversation/tests/chat-tool-row.spec.tsx +++ b/packages/client/ui-conversation/tests/chat-tool-row.spec.tsx @@ -320,6 +320,42 @@ describe('ToolRow', () => { }) describe('ThinkRow', () => { + it('follows the latest streaming line, scrolls to its end, then restores the settled first line', () => { + const view = render( + , + ) + const summary = view.getByText('Newest reasoning tokens') + Object.defineProperties(summary, { + scrollWidth: { configurable: true, value: 300 }, + clientWidth: { configurable: true, value: 100 }, + }) + + view.rerender( + , + ) + expect(summary.scrollLeft).toBe(200) + expect(summary.getAttribute('data-follow-end')).toBe('true') + + view.rerender( + , + ) + expect(view.getByText('Inspect the session')).toBeTruthy() + expect(summary.scrollLeft).toBe(0) + expect(summary.hasAttribute('data-follow-end')).toBe(false) + }) + it('expands from either Think or the reasoning summary', () => { const view = render( Date: Sun, 2 Aug 2026 06:12:13 -0700 Subject: [PATCH 13/13] test(web): harden live thinking tail observation --- apps/web/tests/lifecycle-chrome.e2e.ts | 7 ++++--- .../client/ui-conversation/src/client/chat/ToolRow.tsx | 3 ++- 2 files changed, 6 insertions(+), 4 deletions(-) diff --git a/apps/web/tests/lifecycle-chrome.e2e.ts b/apps/web/tests/lifecycle-chrome.e2e.ts index 65098063ad..8378bf6f5b 100644 --- a/apps/web/tests/lifecycle-chrome.e2e.ts +++ b/apps/web/tests/lifecycle-chrome.e2e.ts @@ -159,7 +159,8 @@ describe('web e2e: lifecycle & chrome (workspace flow / reload / dark mode)', () const settled = scaffold.whenTurnSettled() await input.fill(PROMPT) const observeTurn = async () => { - if (MODE !== 'record') await page.setViewportSize({ width: 640, height: 1000 }) + const originalViewport = page.viewportSize() ?? { width: 1680, height: 1000 } + if (MODE !== 'record') await page.setViewportSize({ width: 480, height: 1000 }) try { await input.press('Enter') if (MODE !== 'record') { @@ -167,11 +168,11 @@ describe('web e2e: lifecycle & chrome (workspace flow / reload / dark mode)', () await expect.poll(async () => await liveTail.evaluate(element => ( element.scrollWidth > element.clientWidth && element.scrollLeft >= element.scrollWidth - element.clientWidth - 1 - )), { timeout: 10_000 }).toBe(true) + )), { timeout: 10_000, interval: 10 }).toBe(true) } return await settled } finally { - if (MODE !== 'record') await page.setViewportSize({ width: 1680, height: 1000 }) + if (MODE !== 'record') await page.setViewportSize(originalViewport) } } const sessionId = await observeTurn() diff --git a/packages/client/ui-conversation/src/client/chat/ToolRow.tsx b/packages/client/ui-conversation/src/client/chat/ToolRow.tsx index 4ec07aa6fc..fa4be6179a 100644 --- a/packages/client/ui-conversation/src/client/chat/ToolRow.tsx +++ b/packages/client/ui-conversation/src/client/chat/ToolRow.tsx @@ -5,7 +5,8 @@ // Enter / Space, icon→chevron hover preview). The collapsed row is always // one line; every row with body, output, or a card material (terminal, diff, // read, search, web) is expandable; the summary stays inline while open, -// except Think, whose body opens with the same first line and would repeat it. +// except Think, where the running collapsed row follows the latest line at its +// scroll end and the summary yields while open to avoid repeating the body. // The expanded body — an IN/OUT gutter-labeled card (figma 1249:35657) for // text input/output, the run_code program through CodeBlock, or a card // primitive (TerminalBlock, DiffBlock, ReadBlock, SearchBlock, WebBlock) for a