fix(gateway): reuse subagent registry snapshot in session listing by vincentkoc · Pull Request #75019 · openclaw/openclaw

vincentkoc · 2026-04-30T10:10:33Z

Summary

Problem: sessions listing rebuilt subagent registry reads repeatedly while rendering rows, which can produce inconsistent child-link state during active subagent work.
Why it matters: the session list is used by subagent orchestration and QA visibility; inconsistent snapshots make active and recently completed children brittle.
What changed: preserve the original fix(gateway): reuse subagent registry snapshot in session listing #75013 patch from @anyech by building a request-scoped subagent registry read context and reusing it across session row construction.
What did NOT change (scope boundary): no session storage schema, agent runtime, or subagent spawn policy changes.

Change Type (select all)

Scope (select all touched areas)

Linked Issue/PR

Related fix(gateway): reuse subagent registry snapshot in session listing #75013
This PR fixes a bug or regression

Root Cause (if applicable)

Root cause: session rows queried mutable subagent registry state independently instead of sharing one read snapshot for the list request.
Missing detection / guardrail: row-level tests did not assert snapshot reuse across child session listing.
Contributing context (if known): replacement branch for fix(gateway): reuse subagent registry snapshot in session listing #75013 after its fork branch was accidentally emptied during CI repair.

Regression Test Plan (if applicable)

Coverage level that should have caught this:
- Unit test
- Seam / integration test
- End-to-end test
- Existing coverage already sufficient
Target test or file: src/agents/subagent-registry-read-context.test.ts, src/gateway/session-utils.subagent.test.ts
Scenario the test should lock in: one sessions list request reuses a stable subagent registry snapshot while retaining child link liveness behavior.
Why this is the smallest reliable guardrail: it tests the snapshot seam and gateway row composition without running the full QA lab.
Existing test that already covers this (if any): N/A
If no new test is added, why not: N/A

User-visible / Behavior Changes

None.

Diagram (if applicable)

Before:
sessions.list -> row A reads registry -> row B reads registry -> possible drift

After:
sessions.list -> one registry snapshot -> all rows share the same read context

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

Repro + Verification

Environment

OS: macOS local, GitHub/Blacksmith CI for parity gate
Runtime/container: Node 22 local; CI runner for parity gate
Model/provider: mock OpenAI QA parity lane
Integration/channel (if any): qa-lab / qa-channel
Relevant config (redacted): N/A

Steps

Run the focused registry and gateway session tests.
Run changed checks in CI.
Rerun the parity gate for the replacement branch.

Expected

Session list child links are based on one request-scoped registry snapshot.

Actual

Prior branch hit a QA parity timeout in source-docs-discovery-report; this replacement branch is draft until the new exact-head CI result is known.

Evidence

Failing test/log before + passing after
Trace/log snippets
Screenshot/recording
Perf numbers (if relevant)

Human Verification (required)

Verified scenarios: pnpm test:serial src/agents/subagent-registry-read-context.test.ts src/gateway/session-utils.subagent.test.ts; git diff --check origin/main...HEAD; git diff --check
Edge cases checked: active, ended, stale, and nested child session link selection in the restored patch tests.
What you did not verify: full parity gate has not completed on this replacement PR yet.

Review Conversations

I replied to or resolved every bot review conversation I addressed in this PR.
I left unresolved only the conversations that still need reviewer or maintainer judgment.

Compatibility / Migration

Backward compatible? Yes
Config/env changes? No
Migration needed? No

Risks and Mitigations

Risk: request-scoped snapshot could hide registry mutations within one long list render.
- Mitigation: context is only request-scoped and rebuilt for each sessions list call.

clawsweeper · 2026-04-30T10:17:44Z

Codex review: passed.

Summary
The branch reuses a request-scoped subagent registry read index across Gateway sessions.list spawnedBy filtering and row enrichment, with focused regression tests and a changelog entry.

Reproducibility: yes. On current main, spawnedBy filtering still calls registry read helpers independently from the row-enrichment context, so source inspection gives a high-confidence reproduction path for snapshot drift during active registry mutation.

Next step before merge
No repair job is needed; the branch already contains the focused fix and should proceed through automerge once exact-head checks and merge gates pass.

Security
Cleared: The diff only changes TypeScript Gateway/subagent registry read logic, focused tests, and the changelog; it does not touch workflows, dependencies, secrets, network calls, command execution, or package publishing surfaces.

Review details

Best possible solution:

Land the current PR head after required exact-head checks finish, keeping the registry snapshot request-scoped and avoiding global cache or session schema changes.

Do we have a high-confidence way to reproduce the issue?

Yes. On current main, spawnedBy filtering still calls registry read helpers independently from the row-enrichment context, so source inspection gives a high-confidence reproduction path for snapshot drift during active registry mutation.

Is this the best way to solve the issue?

Yes. The PR uses the existing buildSubagentRunReadIndex/SessionListRowContext path and keeps the snapshot bounded to one list request, which is the narrowest maintainable fix for this behavior.

What I checked:

Current-main reproduction path: filterAndSortSessionEntries reads subagent display and descendant state directly during spawnedBy filtering, while listSessionsFromStore separately builds a row context for enrichment, so one list request can observe different registry snapshots. (src/gateway/session-utils.ts:1857, 08ce17c33ddb)
Existing request-scoped index seam: Current main already has buildSubagentRunReadIndex and SessionListRowContext; the PR adapts the fix to this existing architecture instead of adding persistent/global cache state. (src/gateway/session-utils.ts:483, 08ce17c33ddb)
PR implementation: The PR diff threads rowContext into filterAndSortSessionEntries, lazy-builds one row context per list request, and avoids falling back to a second controller-run registry read when an index is present. (src/gateway/session-utils.ts:381, 23ae6243747e)
Regression coverage: The PR adds pure read-index tests plus a Gateway session-list test that asserts one disk snapshot is reused across spawnedBy filtering and row enrichment, and that raw filters can avoid registry reads when no rows survive. (src/gateway/session-utils.subagent.test.ts:787, 23ae6243747e)
Protected automerge state: The pull API reports authorAssociation MEMBER, labels maintainer and clawsweeper:automerge, mergeable: true, and head 23ae6243747e5e0f55cad8691e78ac6265fb4e00, so this should remain open for automerge/maintainer handling rather than cleanup closure. (23ae6243747e)
Exact-head checks: The checks API returned 89 check runs for the repaired head, with broad CI/security checks mostly successful and a few still in progress at review time; merge should stay gated on those exact-head checks. (23ae6243747e)

Likely related people:

steipete: Recent main history shows repeated work on the session-list hot path, row-context/read-index shape, and subagent session-list tests. (role: recent maintainer and feature-history owner; confidence: high; commits: 3aaf30ffa600, 2b37b383ed9b, 9402bca614d0; files: src/gateway/session-utils.ts, src/gateway/session-utils.subagent.test.ts, src/agents/subagent-registry-read.ts)
vincentkoc: Current-main history includes recent child-link indexing and bounded session-list/transcript work in the affected Gateway session utility file, separate from this PR author role. (role: recent gateway session-list maintainer; confidence: high; commits: 37f8c3806ac9, 694598822f19, ecf6cbf75d3d; files: src/gateway/session-utils.ts)
Takhoffman: Recent subagent query history includes moved-child and restarted-descendant counting fixes that overlap the liveness and descendant-count semantics preserved by this PR. (role: adjacent subagent registry behavior owner; confidence: medium; commits: e48a0b80a81b, c541cde0f66e, e24704d5eb8a; files: src/agents/subagent-registry-queries.ts)

Remaining risk / open question:

Exact-head GitHub checks for 23ae6243747e5e0f55cad8691e78ac6265fb4e00 still had a few in-progress runs at review time, so merge should remain gated on those checks finishing cleanly.

Codex review notes: model gpt-5.5, reasoning high; reviewed against 08ce17c33ddb.

vincentkoc · 2026-04-30T10:31:21Z

CI update: exact head 5da0e2b490baed01cf323c20689efc6cbb08fb7e is green: 76 successful checks, 0 failed, 0 pending. The replacement branch passed the parity gate, including the OpenAI candidate lane that failed on #75013 at source-docs-discovery-report.

vincentkoc · 2026-05-01T11:57:51Z

/clawsweeper automerge

clawsweeper · 2026-05-01T11:58:41Z

🦞🦞
ClawSweeper automerge is enabled.

Head: 23ae6243747e
Label: clawsweeper:automerge
Action: exact-head review queued.
Flow: review this head, repair/rebase only if needed, then re-review the exact repaired head before merge.

Draft PRs stay fix-only until GitHub marks them ready for review. Pause with /clawsweeper stop.

Automerge progress:

2026-05-02 19:37:08 UTC review queued [`5da0e2b490ba`](https://github.com/openclaw/openclaw/commit/5da0e2b490baed01cf323c20689efc6cbb08fb7e) (queued)
2026-05-02 20:01:00 UTC review passed [`5da0e2b490ba`](https://github.com/openclaw/openclaw/commit/5da0e2b490baed01cf323c20689efc6cbb08fb7e) (structured ClawSweeper verdict: pass (sha=5da0e2b490baed01cf323c20689efc6cbb08f...)
2026-05-01 12:04:19 UTC repair queued `5da0e2b490ba` (autonomous) Run: https://github.com/openclaw/clawsweeper/actions/runs/25213633147
2026-05-01 12:38:02 UTC repair queued `5da0e2b490ba` (autonomous) Run: https://github.com/openclaw/clawsweeper/actions/runs/25214528251
2026-05-01 12:41:33 UTC repair queued `5da0e2b490ba` (autonomous) Run: https://github.com/openclaw/clawsweeper/actions/runs/25214629884
2026-05-02 19:57:14 UTC review queued [`5da0e2b490ba`](https://github.com/openclaw/openclaw/commit/5da0e2b490baed01cf323c20689efc6cbb08fb7e) (queued)
2026-05-02 19:41:26 UTC repair queued [`5da0e2b490ba`](https://github.com/openclaw/openclaw/commit/5da0e2b490baed01cf323c20689efc6cbb08fb7e) (autonomous) Run: https://github.com/openclaw/clawsweeper/actions/runs/25260217169
2026-05-02 20:03:31 UTC repair completed [`23ae6243747e`](https://github.com/openclaw/openclaw/commit/23ae6243747e5e0f55cad8691e78ac6265fb4e00) (branch updated) in 19m 23s Run: https://github.com/openclaw/clawsweeper/actions/runs/25260217169 initial automerge rebase is delegated to Codex repair
2026-05-02 20:03:31 UTC review queued [`23ae6243747e`](https://github.com/openclaw/openclaw/commit/23ae6243747e5e0f55cad8691e78ac6265fb4e00) (after repair)
2026-05-02 20:07:00 UTC review passed [`23ae6243747e`](https://github.com/openclaw/openclaw/commit/23ae6243747e5e0f55cad8691e78ac6265fb4e00) (structured ClawSweeper verdict: pass (sha=23ae6243747e5e0f55cad8691e78ac6265fb4...)
2026-05-02 20:07:12 UTC merged [`23ae6243747e`](https://github.com/openclaw/openclaw/commit/23ae6243747e5e0f55cad8691e78ac6265fb4e00) (merged by ClawSweeper automerge)
2026-05-02 20:07:16 UTC review queued [`23ae6243747e`](https://github.com/openclaw/openclaw/commit/23ae6243747e5e0f55cad8691e78ac6265fb4e00) (queued)

Add a request-scoped subagent registry read context and pass it through sessions.list row construction. This avoids rebuilding subagent registry snapshots and descendant indexes for each row while preserving existing subagent display, spawnedBy, and child-session semantics. Cover parity, moved-child, snapshot-boundary, and registry read-count behavior with focused tests.

…enclaw#75019) Summary: - The branch reuses a request-scoped subagent registry read index across Gateway `sessions.list` `spawnedBy` filtering and row enrichment, with focused regression tests and a changelog entry. - Reproducibility: yes. On current main, `spawnedBy` filtering still calls registry read helpers independently ... ce inspection gives a high-confidence reproduction path for snapshot drift during active registry mutation. ClawSweeper fixups: - Included follow-up commit: fix(gateway): reuse subagent registry snapshot in session listing Validation: - ClawSweeper review passed for head 23ae624. - Required merge gates passed before the squash merge. Prepared head SHA: 23ae624 Review: openclaw#75019 (comment) Co-authored-by: anyech <anyech@gmail.com> Co-authored-by: clawsweeper <274271284+clawsweeper[bot]@users.noreply.github.com>

vincentkoc mentioned this pull request Apr 30, 2026

fix(gateway): reuse subagent registry snapshot in session listing #75013

Closed

vincentkoc self-assigned this Apr 30, 2026

openclaw-barnacle Bot added gateway Gateway runtime agents Agent runtime and tooling size: L maintainer Maintainer-authored PR labels Apr 30, 2026

vincentkoc marked this pull request as ready for review April 30, 2026 10:12

clawsweeper Bot added the clawsweeper:automerge Maintainer opted this PR into bounded ClawSweeper-reviewed automerge label May 1, 2026

anyech and others added 2 commits May 2, 2026 20:03

fix(gateway): reuse subagent registry snapshot in session listing

23ae624

clawsweeper Bot force-pushed the ci-heal/75013-restore branch from 5da0e2b to 23ae624 Compare May 2, 2026 20:03

openclaw-barnacle Bot added size: M and removed size: L labels May 2, 2026

clawsweeper Bot merged commit a1dc8c0 into main May 2, 2026
89 of 90 checks passed

clawsweeper Bot deleted the ci-heal/75013-restore branch May 2, 2026 20:07

github-actions Bot mentioned this pull request May 2, 2026

📡 Upstream Digest — 2026-05-02 20:33 UTC curtismercier/openclaw-mods#749

Open

clawsweeper Bot mentioned this pull request May 3, 2026

fix(gateway): cache sessions.list responses with event-driven invalidation #76404

Closed

4 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fix(gateway): reuse subagent registry snapshot in session listing#75019

fix(gateway): reuse subagent registry snapshot in session listing#75019
clawsweeper[bot] merged 2 commits intomainfrom
ci-heal/75013-restore

vincentkoc commented Apr 30, 2026

Uh oh!

clawsweeper Bot commented Apr 30, 2026 •

edited

Loading

Uh oh!

vincentkoc commented Apr 30, 2026

Uh oh!

vincentkoc commented May 1, 2026

Uh oh!

clawsweeper Bot commented May 1, 2026 •

edited

Loading

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

vincentkoc commented Apr 30, 2026

Summary

Change Type (select all)

Scope (select all touched areas)

Linked Issue/PR

Root Cause (if applicable)

Regression Test Plan (if applicable)

User-visible / Behavior Changes

Diagram (if applicable)

Security Impact (required)

Repro + Verification

Environment

Steps

Expected

Actual

Evidence

Human Verification (required)

Review Conversations

Compatibility / Migration

Risks and Mitigations

Uh oh!

clawsweeper Bot commented Apr 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

vincentkoc commented Apr 30, 2026

Uh oh!

vincentkoc commented May 1, 2026

Uh oh!

clawsweeper Bot commented May 1, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

clawsweeper Bot commented Apr 30, 2026 •

edited

Loading

clawsweeper Bot commented May 1, 2026 •

edited

Loading