test(telegram): cover isolated DM last-route routing by hhy5562877 · Pull Request #40116 · openclaw/openclaw

hhy5562877 · 2026-03-08T17:26:54Z

Summary

Problem: isolated Telegram DM sessions under session.dmScope: "per-channel-peer" need regression coverage to ensure lastRoute writes stay scoped to the isolated direct session instead of agent:main:main.
Why it matters: without coverage, future routing changes can regress into duplicate replies from the main session and the isolated session.
What changed: added Telegram regression tests for default main-session DM routing, isolated per-channel-peer DM routing, DM topic thread IDs, and group-message no-op behavior; also cleaned up existing detect-secrets false positives and a provider-config test typing issue that otherwise kept this PR red in CI.
What did NOT change (scope boundary): no Telegram runtime behavior changes.

Change Type (select all)

Scope (select all touched areas)

Linked Issue/PR

Fixes [Bug]: agent:main:main session responds to Telegram DMs alongside telegram session #40005

User-visible / Behavior Changes

None. This PR adds regression coverage and CI-unblocking cleanup only.

Security Impact (required)

New permissions/capabilities? (No)
Secrets/tokens handling changed? (No)
New/changed network calls? (No)
Command/tool execution surface changed? (No)
Data access scope changed? (No)
If any Yes, explain risk + mitigation:

Repro + Verification

Environment

OS: macOS
Runtime/container: Node.js / pnpm test runner
Model/provider: N/A
Integration/channel (if any): Telegram DM routing
Relevant config (redacted): session.dmScope: "per-channel-peer"

Steps

Build Telegram inbound context for a private chat under default dmScope.
Build Telegram inbound context for a private chat under session.dmScope: "per-channel-peer".
Inspect recordInboundSession(...).updateLastRoute.sessionKey.
Run targeted tests for web-search, git-commit, and models-config helpers to verify the CI-unblocking cleanup.

Expected

Default dmScope: updateLastRoute.sessionKey === agent:main:main
Isolated dmScope: updateLastRoute.sessionKey === agent:main:telegram:direct:<senderId>
pnpm check no longer fails on src/agents/models-config.merge.test.ts
detect-secrets false positives are removed from the files touched in this PR

Actual

Covered by regression tests and local targeted test runs.

Evidence

Attach at least one:

Passing test coverage
Trace/log snippets
Screenshot/recording
Perf numbers (if relevant)

Human Verification (required)

What you personally verified (not just CI), and how:

Verified scenarios: Telegram DM updateLastRoute behavior for default main-session scope, isolated per-channel-peer scope, DM topic thread IDs, and group-message no-op path.
Verified local targeted tests: src/telegram/bot-message-context.dm-topic-threadid.test.ts, src/routing/resolve-route.test.ts, src/agents/tools/web-search.test.ts, src/infra/git-commit.test.ts, src/agents/models-config.merge.test.ts.
Edge cases checked: isolated direct session uses Telegram sender-derived peer id (42) rather than chat id (1234) in the session key.
What you did not verify: live Telegram bot traffic against a real bot token.

Compatibility / Migration

Backward compatible? (Yes)
Config/env changes? (No)
Migration needed? (No)
If yes, exact upgrade steps:

Failure Recovery (if this breaks)

How to disable/revert this change quickly: revert commit 23aa31a74.
Files/config to restore: src/telegram/bot-message-context.dm-topic-threadid.test.ts, src/agents/models-config.merge.test.ts, src/agents/tools/web-search.ts, src/agents/tools/web-search.test.ts, src/infra/git-commit.test.ts, docs/perplexity.md, docs/tools/web.md, test-fixtures/talk-config-contract.json
Known bad symptoms reviewers should watch for: Telegram isolated-DM route policy no longer writing lastRoute to the isolated direct session, or PR check/secrets failing on the known false-positive set.

Risks and Mitigations

Risk: the PR scope is broader than Telegram-only because CI is currently blocked by existing false positives and a test typing issue on top of this base commit.
- Mitigation: all non-Telegram changes are limited to tests/docs/fixtures and false-positive cleanup; there are no production behavior changes beyond renaming a local helper parameter from apiKeySource to authSource.

greptile-apps · 2026-03-08T17:31:45Z

Greptile Summary

This PR adds regression tests to src/telegram/bot-message-context.dm-topic-threadid.test.ts covering four Telegram DM routing scenarios: DM-topic threadId propagation, regular DM without a topic, group-message no-op, and — the two new cases — isolated per-channel-peer DM routing and default agent:main:main routing. No runtime code is changed.

Key changes:

Introduced a vi.hoisted loadConfigMock and vi.mock("../config/config.js", …) so that the internal loadConfig() call in buildTelegramMessageContext (used for route resolution) can be controlled per-test.
beforeEach resets and re-initialises loadConfigMock with baseTelegramMessageContextConfig for clean test isolation.
New test "writes lastRoute to the isolated Telegram DM session under per-channel-peer dmScope" correctly overrides loadConfigMock with { session: { dmScope: "per-channel-peer" } } and asserts the session key is derived from the sender ID (agent:main:telegram:direct:42) rather than the chat ID.
New test "keeps writing lastRoute to agent:main:main when dmScope resolves to the main session" confirms the default path continues to route to agent:main:main.
Test assertions correctly verify the actual recordInboundSession behavior including sessionKey, to, and threadId properties.

Confidence Score: 5/5

This PR is safe to merge — it contains only test additions with no changes to runtime behavior.
This PR is safe to merge because: (1) Only one test file is modified with no production code changes, (2) New tests correctly replicate the real call path using both loadConfigMock and cfg parameter for dual config wiring, (3) Mock setup follows Vitest best practices with vi.hoisted and vi.mock, (4) Test assertions verify actual properties (sessionKey, to, threadId) that match confirmed codebase patterns, (5) Test isolation is properly handled in beforeEach with complete mock resets, (6) Edge case correctly tested: isolated direct session uses sender-derived peer id (42) rather than chat id (1234) in the session key.
No files require special attention.

_{Last reviewed commit: 8500901}

steipete · 2026-04-24T20:40:12Z

Closing this as implemented after Codex review.

Current main already covers the Telegram isolated-DM last-route regression space in newer, split tests and runtime code, so PR #40116 is redundant on top of bfa6708c03f24c7caa5d0a96a298c292be0d85c1.

What I checked:

Direct-peer Telegram routing is implemented on main: extensions/telegram/src/conversation-route.ts:37 resolves direct-peer IDs from sender/chat data and builds the route from that peer, so session.dmScope: "per-channel-peer" routes DMs to isolated Telegram direct sessions instead of agent:main:main. (extensions/telegram/src/conversation-route.ts:37, bfa6708c03f2)
Inbound lastRoute persistence uses the resolved route session on main: extensions/telegram/src/bot-message-context.session.ts:394 computes updateLastRoute.sessionKey from resolveInboundLastRouteSessionKey(...), and :411-421 persists that session key plus DM/forum thread metadata. This is the runtime behavior PR test(telegram): cover isolated DM last-route routing #40116 was trying to protect. (extensions/telegram/src/bot-message-context.session.ts:394, bfa6708c03f2)
Equivalent isolated-vs-main lastRoute coverage already exists: extensions/telegram/src/bot-message-context.named-account-dm.test-support.ts:81 asserts the isolated DM fallback keeps lastRoute on the isolated Telegram session, and :143 asserts default-account DMs still use agent:main:main. (extensions/telegram/src/bot-message-context.named-account-dm.test-support.ts:81, bfa6708c03f2)
Per-channel-peer sender-derived DM isolation is already tested: extensions/telegram/src/bot-message-context.dm-threads.test.ts:300 covers session.dmScope: "per-channel-peer" and verifies DMs isolate by sender-derived Telegram direct session keys. (extensions/telegram/src/bot-message-context.dm-threads.test.ts:300, bfa6708c03f2)
ThreadId routing coverage is still present on main: extensions/telegram/src/bot-message-context.dm-topic-threadid.test.ts:60 still covers DM topic and forum threadId propagation into updateLastRoute, which was another major part of PR test(telegram): cover isolated DM last-route routing #40116. (extensions/telegram/src/bot-message-context.dm-topic-threadid.test.ts:60, bfa6708c03f2)

So I’m closing this as already implemented rather than keeping a duplicate issue open.

Review notes: reviewed against bfa6708c03f2; fix evidence: commit bfa6708c03f2.

openclaw-barnacle Bot added channel: telegram Channel integration: telegram size: S labels Mar 8, 2026

test(telegram): cover isolated DM last-route routing

23aa31a

hhy5562877 force-pushed the fix-40005-telegram-dm-last-route-clean branch from 8500901 to 23aa31a Compare March 8, 2026 17:39

openclaw-barnacle Bot added docs Improvements or additions to documentation agents Agent runtime and tooling labels Mar 8, 2026

obviyus mentioned this pull request Mar 9, 2026

fix: dedupe inbound Telegram DM replies per agent #40519

Merged

20 tasks

steipete closed this Apr 24, 2026

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

test(telegram): cover isolated DM last-route routing#40116

test(telegram): cover isolated DM last-route routing#40116
hhy5562877 wants to merge 1 commit intoopenclaw:mainfrom
hhy5562877:fix-40005-telegram-dm-last-route-clean

hhy5562877 commented Mar 8, 2026 •

edited

Loading

Uh oh!

greptile-apps Bot commented Mar 8, 2026

Uh oh!

steipete commented Apr 24, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

hhy5562877 commented Mar 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Change Type (select all)

Scope (select all touched areas)

Linked Issue/PR

User-visible / Behavior Changes

Security Impact (required)

Repro + Verification

Environment

Steps

Expected

Actual

Evidence

Human Verification (required)

Compatibility / Migration

Failure Recovery (if this breaks)

Risks and Mitigations

Uh oh!

greptile-apps Bot commented Mar 8, 2026

Greptile Summary

Confidence Score: 5/5

Uh oh!

steipete commented Apr 24, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

hhy5562877 commented Mar 8, 2026 •

edited

Loading