惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

J
Java Code Geeks
腾讯CDC
博客园 - 聂微东
爱范儿
爱范儿
罗磊的独立博客
P
Proofpoint News Feed
博客园 - Franky
博客园 - 三生石上(FineUI控件)
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
酷 壳 – CoolShell
酷 壳 – CoolShell
Jina AI
Jina AI
Blog — PlanetScale
Blog — PlanetScale
让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
博客园 - 司徒正美
美团技术团队
MongoDB | Blog
MongoDB | Blog
WordPress大学
WordPress大学
A
About on SuperTechFans
I
InfoQ
博客园_首页
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
H
Help Net Security
Microsoft Azure Blog
Microsoft Azure Blog
G
Google Developers Blog

Recent Commits to openclaw:main

test: merge chat side-result checks · openclaw/openclaw@ddd2c2a test: merge cron history checks · openclaw/openclaw@f7eb746 test: merge responsive navigation shell checks · openclaw/openclaw@c2e4b47 docs(changelog): add codex oauth fixes · openclaw/openclaw@628e6cd test: merge navigation routing cases · openclaw/openclaw@5d8cecb Tests: mock channel registry bundled fallback · openclaw/openclaw@2b08233 Secrets: avoid broad web search discovery for single plugin config · openclaw/openclaw@a464f59 test: merge config view browser checks · openclaw/openclaw@20cf511 fix(status): align oauth health with runtime · openclaw/openclaw@eed7116 feat: add macOS screen snapshots for monitor preview (#67954) thanks … · openclaw/openclaw@f377db1 fix: report shared auth scopes in hello-ok (#67810) thanks @BunsDev · openclaw/openclaw@0b6c39b Auto-reply: avoid eager bundled route fallback · openclaw/openclaw@3ea1bf4 Tests: narrow session binding contract setup · openclaw/openclaw@54e4e16 fix(macOS): enable undo/redo in webchat composer text input (#34962) · openclaw/openclaw@00951dc Tests: speed up channel setup promotion · openclaw/openclaw@82b529a Docs: refresh agent instructions · openclaw/openclaw@5775fe2 fix(auth): serialize OAuth refresh across agents to fix #26322 (#67876) · openclaw/openclaw@8e79080 test: allow ollama public surface boundary test · openclaw/openclaw@7d4f1a6 Docs: add test performance guardrails · openclaw/openclaw@89706d3 Tests: restore context-engine usage proof · openclaw/openclaw@e4c4f95 Tests: slim context engine runtime coverage · openclaw/openclaw@74c198f ci: retry failed custom checkouts · openclaw/openclaw@0ee5baf test: trim duplicate provider auth onboarding cases · openclaw/openclaw@1ffc02e matrix: fix sessions_spawn --thread subagent session spawning (#67643) · openclaw/openclaw@1ce2596 test: reduce auth choice fixture churn · openclaw/openclaw@857b9cd test: mock health status config boundaries · openclaw/openclaw@9d5ab4a test: mock onboard config io boundary · openclaw/openclaw@299694d test: mock legacy state plugin boundaries · openclaw/openclaw@2713089 test: mock channel install boundaries · openclaw/openclaw@b945248 test: mock doctor preview channel boundaries · openclaw/openclaw@b1a3ad4
fix(sessions): classify spawn-child sessions correctly · ...
efpiva · 2026-05-14 · via Recent Commits to openclaw:main

@@ -0,0 +1,214 @@

1+

import { afterEach, beforeEach, describe, expect, it, vi } from "vitest";

2+

import type { SessionEntry } from "../config/sessions/types.js";

3+

import {

4+

mockSessionsConfig,

5+

resetMockSessionsConfig,

6+

runSessionsJson,

7+

writeStore,

8+

} from "./sessions.test-helpers.js";

9+10+

/**

11+

* Catalog #19 — `kind` misclassified as `"direct"` for ACP spawn-child sessions.

12+

*

13+

* Bug summary: `classifySessionKey` (defined twice — `src/commands/sessions.ts:136-152`

14+

* and `src/commands/status.summary.runtime.ts:129-145`) classifies a session

15+

* based ONLY on the key shape (`:group:` / `:channel:` substrings) plus

16+

* `entry.chatType`. It ignores `entry.spawnedBy` and `entry.deliveryContext`,

17+

* so ACP spawn-child sessions (e.g., `agent:copilot:acp:<uuid>` with

18+

* `spawnedBy: "agent:main:telegram:group:..."` and

19+

* `deliveryContext: { channel: "telegram", to: <groupId>, threadId: <topic> }`)

20+

* are misclassified as `kind: "direct"` even though they were spawned from a

21+

* group/topic-bound parent.

22+

*

23+

* Available kinds today:

24+

* "global" | "unknown" | "cron" | "group" | "direct"

25+

*

26+

* The fix shape proposed in the catalog is to add a new `"spawn-child"` kind

27+

* (or, alternatively, fall through to the parent's classification — but the

28+

* catalog calls out `"spawn-child"` as the cleanest minimal fix).

29+

*

30+

* NOTE ON DUPLICATION: the same logic lives in two places —

31+

* - `src/commands/sessions.ts:136-152` (called by `sessionsCommand`,

32+

* the path under test here)

33+

* - `src/commands/status.summary.runtime.ts:129-145`

34+

* The eventual fix MUST update both, or extract a shared helper.

35+

*

36+

* NOTE ON SURFACE: `classifySessionKey` is private to each file (not exported),

37+

* so this test drives the classification through the exposed seam:

38+

* `sessionsCommand --json` and inspects the `kind` field of each session row

39+

* (mirroring `src/commands/sessions.test.ts` and `sessions.acp-runtime-metadata.test.ts`).

40+

*/

41+42+

mockSessionsConfig();

43+44+

const { sessionsCommand } = await import("./sessions.js");

45+46+

type SessionRowKind = "global" | "unknown" | "cron" | "group" | "direct" | "spawn-child";

47+48+

type SessionsJsonPayload = {

49+

sessions?: Array<{

50+

key: string;

51+

kind: SessionRowKind;

52+

}>;

53+

};

54+55+

const ACP_SPAWN_CHILD_KEY = "agent:copilot:acp:7de23a0a-799d-4d63-b1b1-a7de9d4cd840";

56+

const ACP_DM_KEY = "agent:copilot:acp:86b7b5af-3773-4a56-b244-069d6c5d3db9";

57+

const TELEGRAM_GROUP_KEY = "agent:main:telegram:group:-1003967207344:topic:1";

58+59+

/**

60+

* SessionEntry shape mirroring the deployed-container record described in

61+

* the catalog (a copilot ACP session spawned by a telegram supergroup parent).

62+

* Only the fields the classifier and the JSON emit path care about are set;

63+

* everything else stays unset / default.

64+

*/

65+

function buildAcpSpawnChildEntry(): SessionEntry {

66+

return {

67+

sessionId: "spawn-child-session-id",

68+

updatedAt: Date.now() - 2 * 60_000,

69+

spawnedBy: TELEGRAM_GROUP_KEY,

70+

deliveryContext: {

71+

channel: "telegram",

72+

to: "-1003967207344",

73+

threadId: 323,

74+

},

75+

// No chatType — ACP spawn-child entries don't carry one. The classifier

76+

// must infer "this came from a group" from spawnedBy / deliveryContext.

77+

};

78+

}

79+80+

/**

81+

* Plain DM-driven ACP session: same key shape (`agent:copilot:acp:<uuid>`)

82+

* but no `spawnedBy` and a direct delivery context. Today's classifier

83+

* correctly reports `"direct"` for this; that behavior should be preserved

84+

* after the fix.

85+

*/

86+

function buildAcpDirectEntry(): SessionEntry {

87+

return {

88+

sessionId: "dm-session-id",

89+

updatedAt: Date.now() - 5 * 60_000,

90+

deliveryContext: {

91+

channel: "telegram",

92+

to: "+15555550123",

93+

},

94+

};

95+

}

96+97+

/**

98+

* Group session with a key that explicitly embeds `:group:` — the

99+

* classifier's existing key-shape branch picks this up correctly today

100+

* and reports `"group"`.

101+

*/

102+

function buildTelegramGroupEntry(): SessionEntry {

103+

return {

104+

sessionId: "group-session-id",

105+

updatedAt: Date.now() - 10 * 60_000,

106+

chatType: "group",

107+

};

108+

}

109+110+

describe("sessionsCommand kind classification (catalog #19)", () => {

111+

beforeEach(() => {

112+

vi.useFakeTimers();

113+

vi.setSystemTime(new Date("2025-12-06T00:00:00Z"));

114+

});

115+116+

afterEach(() => {

117+

resetMockSessionsConfig();

118+

vi.useRealTimers();

119+

});

120+121+

it("RED: ACP spawn-child session must NOT be classified as 'direct'", async () => {

122+

// RED today. The classifier ignores `spawnedBy` and `deliveryContext`,

123+

// so an ACP key with no `:group:` substring and no `chatType` falls

124+

// through to `"direct"`. Operators see this session in

125+

// `openclaw sessions --json` as `kind: "direct"` even though it was

126+

// plainly spawned from a group/topic. See `src/commands/sessions.ts:136-152`.

127+

const store = writeStore(

128+

{ [ACP_SPAWN_CHILD_KEY]: buildAcpSpawnChildEntry() },

129+

"sessions-kind-spawn-child-red",

130+

);

131+132+

const payload = await runSessionsJson<SessionsJsonPayload>(sessionsCommand, store);

133+

const row = payload.sessions?.find((entry) => entry.key === ACP_SPAWN_CHILD_KEY);

134+135+

expect(

136+

row,

137+

`Expected sessionsCommand --json to include a row for ${ACP_SPAWN_CHILD_KEY}; got none.`,

138+

).toBeDefined();

139+

expect(

140+

row?.kind,

141+

`ACP spawn-child session ${ACP_SPAWN_CHILD_KEY} is misclassified: kind="${row?.kind}". ` +

142+

`It carries spawnedBy="${TELEGRAM_GROUP_KEY}" and deliveryContext.channel="telegram", ` +

143+

`which clearly mark it as a non-direct origin. The classifier at ` +

144+

`src/commands/sessions.ts:136-152 ignores these fields and returns "direct".`,

145+

).not.toBe("direct");

146+

});

147+148+

it("RED (fix-shape): ACP spawn-child session should resolve to 'spawn-child'", async () => {

149+

// RED today; flips GREEN once the proposed fix lands.

150+

//

151+

// The catalog's recommended fix introduces a new `"spawn-child"` kind

152+

// checked BEFORE the key-shape branch so spawn-child ACP sessions take

153+

// precedence over the fallback `"direct"` classification.

154+

//

155+

// If the fix author chooses a different label (e.g., `"acp-child"`) or

156+

// a different shape (e.g., fall through to the parent's classification

157+

// and report `"group"`), update this assertion to match. The structural

158+

// point is that `entry.spawnedBy` / `entry.deliveryContext` MUST drive

159+

// the classification for ACP children.

160+

const store = writeStore(

161+

{ [ACP_SPAWN_CHILD_KEY]: buildAcpSpawnChildEntry() },

162+

"sessions-kind-spawn-child-fix-shape",

163+

);

164+165+

const payload = await runSessionsJson<SessionsJsonPayload>(sessionsCommand, store);

166+

const row = payload.sessions?.find((entry) => entry.key === ACP_SPAWN_CHILD_KEY);

167+168+

expect(row).toBeDefined();

169+

expect(

170+

row?.kind,

171+

`ACP spawn-child session ${ACP_SPAWN_CHILD_KEY} should classify as "spawn-child" ` +

172+

`(or whichever non-direct label the fix author chooses). Got "${row?.kind}". ` +

173+

`Fix locations: src/commands/sessions.ts:136-152 AND ` +

174+

`src/commands/status.summary.runtime.ts:129-145 (the same logic is duplicated; ` +

175+

`extract to a shared helper or update both).`,

176+

).toBe("spawn-child");

177+

});

178+179+

it("GREEN control: non-spawn-child ACP DM session resolves to 'direct'", async () => {

180+

// GREEN today. An ACP-keyed session WITHOUT `spawnedBy` and with a

181+

// direct delivery context (or none) correctly resolves to `"direct"`.

182+

// This control proves the test infrastructure exercises the real

183+

// classification path; if it accidentally regressed to a different

184+

// value, that would indicate the test harness was broken.

185+

const store = writeStore(

186+

{ [ACP_DM_KEY]: buildAcpDirectEntry() },

187+

"sessions-kind-acp-direct-control",

188+

);

189+190+

const payload = await runSessionsJson<SessionsJsonPayload>(sessionsCommand, store);

191+

const row = payload.sessions?.find((entry) => entry.key === ACP_DM_KEY);

192+193+

expect(row).toBeDefined();

194+

expect(row?.kind).toBe("direct");

195+

});

196+197+

it("GREEN control: telegram group key with chatType='group' resolves to 'group'", async () => {

198+

// GREEN today. The classifier's key-shape branch (`:group:` substring)

199+

// and the `chatType === "group"` branch both fire for this entry,

200+

// yielding `"group"`. This control proves the existing happy-path

201+

// classification still works and is not silently broken by the test

202+

// harness.

203+

const store = writeStore(

204+

{ [TELEGRAM_GROUP_KEY]: buildTelegramGroupEntry() },

205+

"sessions-kind-group-control",

206+

);

207+208+

const payload = await runSessionsJson<SessionsJsonPayload>(sessionsCommand, store);

209+

const row = payload.sessions?.find((entry) => entry.key === TELEGRAM_GROUP_KEY);

210+211+

expect(row).toBeDefined();

212+

expect(row?.kind).toBe("group");

213+

});

214+

});