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

N
News and Events Feed by Topic
Malwarebytes
Malwarebytes
Threat Intelligence Blog | Flashpoint
Threat Intelligence Blog | Flashpoint
C
Cybersecurity and Infrastructure Security Agency CISA
F
Future of Privacy Forum
C
Cisco Blogs
T
The Exploit Database - CXSecurity.com
A
Arctic Wolf
S
Securelist
K
Kaspersky official blog
S
Schneier on Security
T
ThreatConnect
T
Tenable Blog
Spread Privacy
Spread Privacy
T
True Tiger Recordings
AWS News Blog
AWS News Blog
F
Fox-IT International blog
量子位
T
Threatpost
V
Vulnerabilities – Threatpost
C
CERT Recently Published Vulnerability Notes
Cisco Talos Blog
Cisco Talos Blog
GbyAI
GbyAI
宝玉的分享
宝玉的分享
腾讯CDC
G
Google Developers Blog
aimingoo的专栏
aimingoo的专栏
Cyberwarzone
Cyberwarzone
有赞技术团队
有赞技术团队
S
SegmentFault 最新的问题
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
V
Visual Studio Blog
U
Unit 42
雷峰网
雷峰网
cs.CV updates on arXiv.org
cs.CV updates on arXiv.org
Simon Willison's Weblog
Simon Willison's Weblog
O
OpenAI News
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
The GitHub Blog
The GitHub Blog
The Register - Security
The Register - Security
MyScale Blog
MyScale Blog
小众软件
小众软件
A
About on SuperTechFans
Last Week in AI
Last Week in AI
Y
Y Combinator Blog
博客园 - 三生石上(FineUI控件)
美团技术团队
Google Online Security Blog
Google Online Security Blog
P
Proofpoint News Feed
MongoDB | Blog
MongoDB | Blog

Recent Commits to openclaw:main

test(e2e): harden multi-node update smoke Clean up browser MCP subprocess tree (#85832) fix(agents): log warnings instead of swallowing subagent errors (#82943) · openclaw/openclaw@907bc03 fix(config): do not suppress recovery retry after failed backup resto… · openclaw/openclaw@5d174a5 chore: release 2026.5.25 fix(installer): support alpine cli installs · openclaw/openclaw@f68ed72 test(agents): keep runtime-plan provider mock current fix(scripts): launch env package scripts on Windows · openclaw/openclaw@4d4ce9e fix(agents): cache fallback provider resolution · openclaw/openclaw@3c8d101 fix(test): make import timing scripts Windows-safe · openclaw/openclaw@8ae9977 fix(telegram): transient Telegram pairing prompts (#85555) · openclaw/openclaw@8209426 fix(test): make max Vitest scripts Windows-safe · openclaw/openclaw@b681d5d fix(doctor): migrate Feishu account bot names (#86081) · openclaw/openclaw@9e8cc7e fix(scripts): prefilter conflict marker scans docs: add ClawSweeper review policy to AGENTS (#86197) · openclaw/openclaw@242e876 fix(installer): avoid before with npm release-age configs (#85491) · openclaw/openclaw@4742db6 fix(e2e): retry Windows kitchen sink probes · openclaw/openclaw@3e275a5 fix(installer): install node with apk on alpine fix(installer): detect musl linux shells · openclaw/openclaw@acfed37 perf(plugins,gateway): thread metadata snapshot + discovery through h… · openclaw/openclaw@8ccb11c fix(ui): split control ui runtime chunks · openclaw/openclaw@8bf4f7d refactor(config): extract GoogleChat schema into zod-schema.providers… · openclaw/openclaw@fe34141 fix(update): suppress internal handoff version warnings · openclaw/openclaw@6cc8244 test(e2e): select installable bundled plugins · openclaw/openclaw@0acc3e3 fix(scripts): harden Windows native opus install · openclaw/openclaw@43252c8 fix(agents): match runtime policy entries when session provider is em… fix(scripts): harden Windows generated formatting · openclaw/openclaw@0a98559 fix(mcp): bound tools/list during catalog discovery (#85063) · openclaw/openclaw@07f500a fix(test): focus plugin binding Docker smoke · openclaw/openclaw@dfa1a51 test(e2e): fail release memory indexing errors test(daemon): fail launchd integration bootstrap errors · openclaw/openclaw@af07769 feat(imessage): support thumb approval reactions (#85952) · openclaw/openclaw@5c7980f fix(crabbox): default macos aws runs on demand fix(scripts): preserve test passthrough args · openclaw/openclaw@e4332f7 fix(e2e): harden Windows plugin assertions fix(test): mount upgrade survivor helper · openclaw/openclaw@5f03154 fix(android): prevent stale chat during session switches fix(android): keep permission setup action visible · openclaw/openclaw@94bc18a style(android): sharpen voice mode surfaces · openclaw/openclaw@c452510 fix(android): hide internal chat content blocks · openclaw/openclaw@d86ed21 style(android): refine list surface spacing · openclaw/openclaw@955909c feat(android): add pair new gateway action · openclaw/openclaw@cc5eb97 Advance iMessage catchup cursor after live handling (#85475) · openclaw/openclaw@102555c fix(scripts): ignore forwarded arg separator · openclaw/openclaw@79ee70c fix(test): fail empty gateway startup samples · openclaw/openclaw@5a8ce6a fix(e2e): harden Windows kitchen sink assertions · openclaw/openclaw@87a2eba fix(e2e): harden Telegram credential paths on Windows · openclaw/openclaw@c643370 fix(android): align setup pairing scopes fix(android): complete qr setup operator handoff · openclaw/openclaw@be9bb77 fix(test): copy cleanup smoke prepare hook · openclaw/openclaw@dbc08f6 fix(secrets): allow hash in exec SecretRef ids (#86072) · openclaw/openclaw@675158c fix(media): use static image compression metadata · openclaw/openclaw@694d45e fix(release): verify large plugin npm packs fix(test): require kitchen sink diagnostic canaries · openclaw/openclaw@7e51f83 fix(scripts): harden Windows upgrade survivor recipe · openclaw/openclaw@483d7be fix(installer): count verify progress stage fix: Refine PR template for review state (#86054) fix(test): repair split agent shard runs · openclaw/openclaw@125d82c fix(codex): harden Windows protocol formatting · openclaw/openclaw@ce48e4c fix(openrouter): use endpoint context limits (#86041) · openclaw/openclaw@dd01a2e test(qa): remove brittle capability flip setup turn fix(telegram): migrate legacy cache sidecars · openclaw/openclaw@eb9b882 fix(telegram): migrate account topic cache sidecars · openclaw/openclaw@5cfb12f fix(scripts): harden Windows ZAI fallback repro · openclaw/openclaw@5be62e7 style(android): sharpen v2 screen rhythm · openclaw/openclaw@400d90a test(qa): extend capability flip setup budget · openclaw/openclaw@c91c3c6 fix(android): simplify gateway status copy fix(android): route offline voice to gateway setup fix(scripts): harden Windows control UI i18n commands · openclaw/openclaw@581c8a6 fix(android): stop operator chat subscription · openclaw/openclaw@5c15859 fix(test): fail missing kitchen sink rss samples test(qa): widen capability flip restart budget · openclaw/openclaw@c7d4e9e fix(android): smooth gateway pairing recovery · openclaw/openclaw@60e6ccd fix(test): suppress rolldown timing noise · openclaw/openclaw@6d9b388 style(android): fix talk mode ktlint formatting · openclaw/openclaw@01b284c fix(telegram): store topic cache in plugin state fix(telegram): store bot info cache in plugin state · openclaw/openclaw@2ed5296 fix(test): sync sparse AWS Crabbox runs from full checkout · openclaw/openclaw@0f82c81 fix(release): harden Windows cross-os command shims · openclaw/openclaw@7154767 fix(test): harden Docker resource ceilings test(telegram): keep startup limiter coverage focused test(telegram): isolate startup probe limiter timing · openclaw/openclaw@04d86e0 test(release): harden plugin prerelease checks · openclaw/openclaw@578e73f fix(telegram): serialize topic dispatch replies (#85709) · openclaw/openclaw@62b51a6 test(release): stabilize plugin prerelease checks · openclaw/openclaw@3679151 fix(test): fail live gateway false greens · openclaw/openclaw@295339d fix(test): build startup artifacts for smoke scripts · openclaw/openclaw@3838e45 test(telegram): wait for startup probe slots · openclaw/openclaw@0a8af67 test(codex): match sandbox exec-server yolo policy · openclaw/openclaw@783290f fix(test): fail missing explicit test targets · openclaw/openclaw@9ff4d36 test(codex): avoid full sandbox exec-server turn run · openclaw/openclaw@558c1bc fix(ci): keep Crabbox pnpm hydration shims writable · openclaw/openclaw@bca1ac0 fix(release): harden Windows release-check npm probes · openclaw/openclaw@75ac11a fix(docker): parse peer-suffixed lockfile packages · openclaw/openclaw@cf46f2e fix(docker): seed lockfile packages before prune · openclaw/openclaw@f799da0 fix(docker): seed lockfile snapshot tarballs before prune · openclaw/openclaw@2cd93f1 test(codex): type thread start mock params · openclaw/openclaw@a4ef3a2 test(codex): avoid full sandbox run in thread-start test · openclaw/openclaw@11bf642 fix(plugins): harden Windows npm package staging · openclaw/openclaw@abdd8a4 test(codex): complete sandbox turn inline · openclaw/openclaw@c14a0c6
fix(compaction): preserve partial summary on mid-chain chunk failure … · openclaw/openclaw@f0061dd
SebTardif · 2026-05-25 · via Recent Commits to openclaw:main

@@ -0,0 +1,262 @@

1+

import type { AgentMessage } from "@earendil-works/pi-agent-core";

2+

import type { ExtensionContext } from "@earendil-works/pi-coding-agent";

3+

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

4+5+

const compactionMocks = vi.hoisted(() => {

6+

function readText(value: unknown): string {

7+

if (typeof value === "string") {

8+

return value;

9+

}

10+

if (Array.isArray(value)) {

11+

return value.map(readText).join("");

12+

}

13+

if (value && typeof value === "object") {

14+

const record = value as { text?: unknown; content?: unknown; arguments?: unknown };

15+

return `${readText(record.text)}${readText(record.content)}${readText(record.arguments)}`;

16+

}

17+

return "";

18+

}

19+

return {

20+

estimateTokens: vi.fn((message: unknown) =>

21+

Math.max(1, Math.ceil(readText(message).length / 4)),

22+

),

23+

generateSummary: vi.fn(),

24+

logWarn: vi.fn(),

25+

};

26+

});

27+28+

vi.mock("@earendil-works/pi-coding-agent", async () => {

29+

const actual = await vi.importActual<typeof import("@earendil-works/pi-coding-agent")>(

30+

"@earendil-works/pi-coding-agent",

31+

);

32+

return {

33+

...actual,

34+

estimateTokens: compactionMocks.estimateTokens,

35+

generateSummary: compactionMocks.generateSummary,

36+

};

37+

});

38+39+

vi.mock("../logging/subsystem.js", () => ({

40+

createSubsystemLogger: () => ({

41+

info: vi.fn(),

42+

warn: compactionMocks.logWarn,

43+

error: vi.fn(),

44+

debug: vi.fn(),

45+

trace: vi.fn(),

46+

raw: vi.fn(),

47+

child: vi.fn().mockReturnThis(),

48+

}),

49+

}));

50+51+

// Mock retryAsync to bypass retry delays while preserving the single-call semantic.

52+

// summarizeChunks wraps generateSummary in retryAsync with 500-5000 ms delays;

53+

// eliminating them keeps tests fast without altering the catch-block behavior under test.

54+

vi.mock("../infra/retry.js", async () => {

55+

const actual = await vi.importActual<typeof import("../infra/retry.js")>("../infra/retry.js");

56+

return {

57+

...actual,

58+

retryAsync: async <T>(fn: () => Promise<T>) => fn(),

59+

};

60+

});

61+62+

let summarizeWithFallback: typeof import("./compaction.js").summarizeWithFallback;

63+64+

beforeAll(async () => {

65+

vi.resetModules();

66+

({ summarizeWithFallback } = await import("./compaction.js"));

67+

});

68+69+

describe("summarizeChunks partial summary preservation (#82952)", () => {

70+

const testModel = {

71+

id: "test",

72+

name: "test",

73+

contextWindow: 200_000,

74+

contextTokens: 200_000,

75+

maxTokens: 8192,

76+

} as unknown as NonNullable<ExtensionContext["model"]>;

77+78+

// Two messages sized to split into two chunks with maxChunkTokens=150.

79+

// Each message is ~100 tokens (400 chars / 4), and effectiveMax = floor(150/1.2) = 125.

80+

const twoChunkMessages: AgentMessage[] = [

81+

{ role: "user", content: "x".repeat(400), timestamp: 1 },

82+

{ role: "user", content: "y".repeat(400), timestamp: 2 },

83+

];

84+85+

function callSummarize(messages = twoChunkMessages) {

86+

return summarizeWithFallback({

87+

messages,

88+

model: testModel,

89+

apiKey: "test-key", // pragma: allowlist secret

90+

signal: new AbortController().signal,

91+

reserveTokens: 1000,

92+

maxChunkTokens: 150,

93+

contextWindow: 200_000,

94+

});

95+

}

96+97+

beforeEach(() => {

98+

compactionMocks.generateSummary.mockReset();

99+

compactionMocks.logWarn.mockClear();

100+

});

101+102+

it("returns partial summary when a later chunk fails with a non-abort error", async () => {

103+

compactionMocks.generateSummary

104+

.mockResolvedValueOnce("Summary of chunk 1")

105+

.mockRejectedValue(new Error("API quota exceeded"));

106+107+

const result = await callSummarize();

108+109+

expect(result).toContain("Summary of chunk 1");

110+

expect(result).toContain("[Partial summary:");

111+

expect(result).toMatch(/chunks 1-1 of 2 were summarized/);

112+

expect(compactionMocks.logWarn).toHaveBeenCalledWith(

113+

"chunk summarization failed after retries; partial summary available",

114+

expect.objectContaining({ err: expect.any(Error) }),

115+

);

116+

});

117+118+

it("re-throws abort errors instead of returning partial summary", async () => {

119+

const abortErr = new Error("aborted");

120+

abortErr.name = "AbortError";

121+122+

compactionMocks.generateSummary

123+

.mockResolvedValueOnce("Summary of chunk 1")

124+

.mockRejectedValue(abortErr);

125+126+

const result = await callSummarize();

127+128+

// Abort error propagates from summarizeChunks; summarizeWithFallback catches it

129+

// and falls through to the final fallback (not the partial summary).

130+

expect(result).not.toBe("Summary of chunk 1");

131+

expect(result).toContain("Context contained");

132+

expect(compactionMocks.logWarn).not.toHaveBeenCalledWith(

133+

"chunk summarization failed after retries; partial summary available",

134+

expect.anything(),

135+

);

136+

});

137+138+

it("re-throws timeout errors instead of returning partial summary", async () => {

139+

const timeoutErr = new Error("request timed out");

140+

timeoutErr.name = "TimeoutError";

141+142+

compactionMocks.generateSummary

143+

.mockResolvedValueOnce("Summary of chunk 1")

144+

.mockRejectedValue(timeoutErr);

145+146+

const result = await callSummarize();

147+148+

expect(result).not.toBe("Summary of chunk 1");

149+

expect(result).toContain("Context contained");

150+

expect(compactionMocks.logWarn).not.toHaveBeenCalledWith(

151+

"chunk summarization failed after retries; partial summary available",

152+

expect.anything(),

153+

);

154+

});

155+156+

it("returns the full final summary when all chunks succeed", async () => {

157+

compactionMocks.generateSummary

158+

.mockResolvedValueOnce("Summary of chunk 1")

159+

.mockResolvedValueOnce("Combined summary of chunks 1+2");

160+161+

const result = await callSummarize();

162+163+

expect(result).toBe("Combined summary of chunks 1+2");

164+

expect(compactionMocks.generateSummary).toHaveBeenCalledTimes(2);

165+

});

166+167+

it("falls back to default when the first chunk fails (no partial to recover)", async () => {

168+

compactionMocks.generateSummary.mockRejectedValue(new Error("network error"));

169+170+

const result = await callSummarize();

171+172+

// With no successful chunk, summarizeChunks rethrows into

173+

// summarizeWithFallback's outer catch -> final fallback path.

174+

expect(result).toContain("Context contained");

175+

expect(result).not.toBe("Summary of chunk 1");

176+

});

177+178+

it("tries oversized-message retry before falling back to partial summary", async () => {

179+

// Scenario: chunk 1 (small) succeeds, chunk 2 (has oversized message) fails.

180+

// summarizeWithFallback should try the non-oversized retry, which may

181+

// recover more content than the partial summary alone.

182+

const mixedMessages: AgentMessage[] = [

183+

// Small message (chunk 1)

184+

{ role: "user", content: "Short question about code", timestamp: 1 },

185+

// Oversized message (will be in chunk 2, triggers the oversized retry)

186+

{ role: "assistant", content: "x".repeat(500_000), timestamp: 2 } as unknown as AgentMessage,

187+

// Small message after oversized (should be recovered by oversized retry)

188+

{ role: "user", content: "Follow-up question", timestamp: 3 },

189+

];

190+191+

compactionMocks.generateSummary

192+

// Call 1: chunk 1 of full attempt (success)

193+

.mockResolvedValueOnce("Summary of chunk 1")

194+

// Call 2: chunk 2 of full attempt (fails - oversized message)

195+

.mockRejectedValueOnce(new Error("context too long"))

196+

// Call 3: oversized retry with small messages only (succeeds!)

197+

.mockResolvedValueOnce("Summary of small messages (oversized retry)");

198+199+

const result = await callSummarize(mixedMessages);

200+201+

// The oversized retry should have recovered more content than

202+

// the partial summary from chunk 1 alone.

203+

expect(result).toContain("Summary of small messages (oversized retry)");

204+

// The partial summary should NOT be the final result because the

205+

// oversized retry succeeded.

206+

expect(result).not.toContain("[Partial summary:");

207+

});

208+209+

it("prefers oversized retry partial summary over full attempt partial", async () => {

210+

// Scenario: full attempt's chunk 1 succeeds, chunk 2 (oversized) fails.

211+

// Oversized retry (small messages only) chunk 1 succeeds, chunk 2 fails.

212+

// The oversized retry's partial summary should be preferred because it

213+

// covers the non-oversized transcript.

214+

const mixedMessages: AgentMessage[] = [

215+

{ role: "user", content: "Short question", timestamp: 1 },

216+

// Oversized message that will be filtered in the retry

217+

{ role: "assistant", content: "x".repeat(500_000), timestamp: 2 } as unknown as AgentMessage,

218+

{ role: "user", content: "a".repeat(400), timestamp: 3 },

219+

{ role: "user", content: "b".repeat(400), timestamp: 4 },

220+

];

221+222+

compactionMocks.generateSummary

223+

// Full attempt: chunk 1 succeeds, chunk 2 fails (oversized message)

224+

.mockResolvedValueOnce("Full attempt chunk 1")

225+

.mockRejectedValueOnce(new Error("context too long"))

226+

// Oversized retry: chunk 1 succeeds, chunk 2 also fails

227+

.mockResolvedValueOnce("Oversized retry chunk 1 (better coverage)")

228+

.mockRejectedValue(new Error("rate limited on retry"));

229+230+

const result = await callSummarize(mixedMessages);

231+232+

// The oversized retry's partial summary should win, with oversized notes

233+

expect(result).toContain("Oversized retry chunk 1 (better coverage)");

234+

expect(result).toContain("[Partial summary:");

235+

expect(result).toContain("[Large assistant");

236+

expect(result).toContain("omitted from summary]");

237+

});

238+239+

it("preserves the latest successful summary in a 3+ chunk chain", async () => {

240+

const threeChunkMessages: AgentMessage[] = [

241+

{ role: "user", content: "a".repeat(400), timestamp: 1 },

242+

{ role: "user", content: "b".repeat(400), timestamp: 2 },

243+

{ role: "user", content: "c".repeat(400), timestamp: 3 },

244+

];

245+246+

compactionMocks.generateSummary

247+

.mockResolvedValueOnce("Summary after chunk 1")

248+

.mockResolvedValueOnce("Summary after chunks 1+2")

249+

.mockRejectedValue(new Error("rate limited"));

250+251+

const result = await callSummarize(threeChunkMessages);

252+253+

// Chunk 3 failed -> partial summary from chunk 2 is returned with marker.

254+

expect(result).toContain("Summary after chunks 1+2");

255+

expect(result).toMatch(/\[Partial summary: chunks 1-2 of 3 were summarized/);

256+

expect(compactionMocks.generateSummary).toHaveBeenCalledTimes(3);

257+

expect(compactionMocks.logWarn).toHaveBeenCalledWith(

258+

"chunk summarization failed after retries; partial summary available",

259+

expect.objectContaining({ completedChunks: 2, totalChunks: 3 }),

260+

);

261+

});

262+

});