fix(gateway): guard maxLines compact against active session runs by chinadbo · Pull Request #72765 · openclaw/openclaw

chinadbo · 2026-04-27T10:54:37Z

Summary

sessions.compact with maxLines skips the interruptSessionRunIfActive guard that the AI compaction path uses
When an agent run is active, archiveFileOnDisk renames the live transcript to .bak; the active runner continues writing to the renamed file while writeFileSync creates a new truncated file at the original path — all post-compact agent output is silently lost
The fix inserts the same interruptSessionRunIfActive guard used by the AI compaction path, mirroring the pattern also used by sessions.compaction.restore

Fix

Add interruptSessionRunIfActive call to the maxLines branch of sessions.compact before the transcript file is read and overwritten.

Test plan

sessions.compact with maxLines during an active run now aborts the run first (same behavior as AI compaction)
sessions.compact with maxLines when no run is active works unchanged
pnpm check:changed passes

Thanks @ioodu

greptile-apps · 2026-04-27T10:55:28Z

Greptile Summary

This PR adds a missing interruptSessionRunIfActive guard to the maxLines branch of sessions.compact, preventing data loss when compaction runs against an active session. The fix mirrors the identical guard already present in the AI compaction path and the restore path.

Confidence Score: 5/5

Safe to merge — a minimal, targeted fix with no observed issues.

Single-file change that adds one guard matching an established pattern already used twice in the same handler. Parameters and error handling are identical to the AI compaction path. No new logic, no side-effect changes.

No files require special attention.

_{Reviews (1): Last reviewed commit: "fix(gateway): guard maxLines compact aga..." | Re-trigger Greptile}

chinadbo · 2026-04-27T11:13:38Z

The failing CI checks (checks-fast-contracts-plugins, checks-node-agentic-agents, checks-node-core) are pre-existing failures on main unrelated to this PR:

checks-fast-contracts-plugins: extensions/qwen/vllm missing @mariozechner/pi-ai runtime dep declaration
checks-node-agentic-agents: model-auth.test.ts assertion failure (custom-local vs ollama-local) — pre-existing test flake
checks-node-core: pre-existing

steipete · 2026-04-27T20:04:53Z

Re-reviewed current head.

The earlier chatAbortControllers.values() TypeError looks fixed by the null-safe fallback, and the production code path is the right one: current main still interrupts active runs on the AI compaction path, but not before the maxLines archive/rewrite path.

Before merge, please add regression coverage. This PR currently touches only src/gateway/server-methods/sessions.ts; the risky behavior is transcript mutation during an active run. Suggested test: call sessions.compact with maxLines while the target session has an active run, assert the interrupt path is used before archive/write, and assert an interrupt error / still-active run prevents archiveFileOnDisk() and the transcript rewrite. Existing maxLines tests cover truncation, not this active-run guard.

Live check note: current PR checks are green except the parity gate.

chinadbo · 2026-04-28T02:18:40Z

Added the regression test as requested. The test (sessions.compact with maxLines rejects when an embedded run is active and does not stop) uses embeddedRunMock to simulate an active run that refuses to stop, then asserts:

sessions.compact with maxLines returns ok: false with UNAVAILABLE error code
The transcript file is unchanged (still 10 lines)
No .bak. archive file was created (i.e. archiveFileOnDisk was never called)

All 53 tests in the file pass.

…kedActiveSessionRun

…mbedded run

clawsweeper · 2026-04-30T15:36:54Z

Thanks for the context here. I swept through the related work, and this is now duplicate or superseded.

Close as superseded: the active-run data-loss bug is real, but this branch is now conflicting, lacks real behavior proof, and the clean mergeable replacement PR already carries the same guard in the current sessions code with stronger tests and proof.

Canonical path: Use #91378 as the canonical current-main fix path and close this conflicting branch as superseded.

So I’m closing this here and keeping the remaining discussion on #91378.

Review details

Best possible solution:

Use #91378 as the canonical current-main fix path and close this conflicting branch as superseded.

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

Yes from source inspection. Current main’s maxLines branch reaches transcript tail read, archive, and rewrite without the active-run interrupt used by the non-maxLines compaction branch.

Is this the best way to solve the issue?

No for this branch. The guard shape is right, but the clean replacement PR is the better solution because it applies to current main, preserves agent scoping, adds current-suite coverage, and supplies real behavior proof.

Security review:

Security review cleared: Security review cleared: this stale branch changes gateway session runtime logic and tests only, with no dependency, workflow, package, credential, or supply-chain surface changes.

AGENTS.md: found and applied where relevant.

What I checked:

Current main still has the maxLines gap: Current main calls interruptSessionRunIfActive in the non-maxLines compaction branch, but the maxLines branch proceeds to readRecentSessionTranscriptLines, archiveFileOnDisk, and writeFileSync without that guard. (src/gateway/server-methods/sessions.ts:2517, 66880a5d736c)
Current active-run helper is agent-scoped: The current helper accepts agentId and defaultAgentId, so a rebased fix needs to preserve current agent scoping rather than replay the old branch shape. (src/gateway/server-methods/session-active-runs.ts:63, 66880a5d736c)
This PR is stale relative to current main: Live PR metadata reports this branch as CONFLICTING/DIRTY, with triage: needs-real-behavior-proof and status: needs proof labels still present. (9b95bc649ab0)
Canonical replacement is open and viable: feat(cli): add openclaw sessions compact and fail loudly on CLI /compact (fixes #90640) #91378 is open, clean, mergeable, maintainer-editable, and labeled proof: sufficient / status: ready for maintainer look. (5951a4562037)
Replacement includes the same guard in the current code path: The replacement PR adds interruptSessionRunIfActive to the current maxLines branch with agentId: requestedAgentId and places regression coverage in the current server.sessions.compaction.test.ts suite. (src/gateway/server-methods/sessions.ts:2643, 99878847dfbc)
Replacement proof covers the destructive-write ordering: The replacement PR body and ClawSweeper review include real in-process Gateway proof showing maxLines truncation aborts before archive/truncate when an active run cannot stop. (5951a4562037)

Likely related people:

scoootscooob: Introduced the gateway compaction checkpoint/session compaction surface that this guard modifies. (role: feature introducer; confidence: high; commits: f4fcaa09a358; files: src/gateway/server-methods/sessions.ts, src/gateway/server.sessions.gateway-server-sessions-a.test.ts, src/gateway/session-compaction-checkpoints.ts)
Alix-007: Authored the open clean replacement PR that ports this guard to current main with current-suite tests and real behavior proof. (role: canonical replacement contributor; confidence: high; commits: 99878847dfbc, 5951a4562037; files: src/gateway/server-methods/sessions.ts, src/gateway/server.sessions.compaction.test.ts)
Peter Steinberger: Reviewed this PR’s active-run compaction path and requested regression coverage; also recently touched adjacent gateway sessions tests. (role: recent area reviewer; confidence: medium; commits: f6193687691d; files: src/gateway/server.sessions.gateway-server-sessions-a.test.ts)
Vincent Koc: Recently touched the same gateway sessions test area used by the stale branch. (role: recent adjacent contributor; confidence: medium; commits: 56625a189bf3; files: src/gateway/server.sessions.gateway-server-sessions-a.test.ts)

Codex review notes: model internal, reasoning high; reviewed against 66880a5d736c.

clawsweeper · 2026-05-20T22:55:16Z

ClawSweeper PR egg

🎁 Pass real behavior proof to wake the egg and unlock a hatchable treat.

Where did the egg go?

The egg game starts only after the PR passes the real-behavior proof check.
Before that, no creature or rarity is rolled. The treat waits for real proof.
This is still just collectible flavor: proof affects review readiness, not creature quality.

openclaw-barnacle · 2026-06-04T05:06:51Z

This pull request has been automatically marked as stale due to inactivity.
Please add updates or it will be closed.

…ve runs The non-maxLines (LLM) compact branch interrupts an active session run before compacting, but the maxLines truncate branch read the tail, archived, and overwrote the transcript in place without that guard. Exposing `--max-lines` as a documented CLI command (this PR) would make the active-run data-loss mode tracked by openclaw#72765 easy to trigger from ordinary CLI usage. Run the same interruptSessionRunIfActive guard in the maxLines branch before reading the tail and truncating, matching the LLM compact path. Add gateway regression coverage over a real in-process Gateway: with no active run, the maxLines branch truncates the on-disk transcript 500 -> 50 and preserves the original 500 lines in the .bak archive; with an active embedded run, the maxLines branch fires the same interrupt (abort + wait-for-end) before archiving and truncating.

clawsweeper · 2026-06-14T05:08:44Z

ClawSweeper applied the proposed close for this PR.

Action: closed this PR.
Close reason: duplicate or superseded.
Evidence: durable ClawSweeper review.
Coverage proof: PR B materially carries PR A's core bug fix and review concerns on the current code path with better tests and proof, so PR A has no remaining independent material work. Covering PR: feat(cli): add openclaw sessions compact and fail loudly on CLI /compact (fixes #90640) #91378.

…ve runs The non-maxLines (LLM) compact branch interrupts an active session run before compacting, but the maxLines truncate branch read the tail, archived, and overwrote the transcript in place without that guard. Exposing `--max-lines` as a documented CLI command (this PR) would make the active-run data-loss mode tracked by openclaw#72765 easy to trigger from ordinary CLI usage. Run the same interruptSessionRunIfActive guard in the maxLines branch before reading the tail and truncating, matching the LLM compact path. Add gateway regression coverage over a real in-process Gateway: with no active run, the maxLines branch truncates the on-disk transcript 500 -> 50 and preserves the original 500 lines in the .bak archive; with an active embedded run, the maxLines branch fires the same interrupt (abort + wait-for-end) before archiving and truncating.

…ompact` (fixes #90640) (#91378) * feat(cli): add `sessions compact` command and fail loudly on CLI `/compact` `sessions.compact` was reachable only as an internal Gateway RPC — no CLI command, no docs — and `openclaw agent --message '/compact'` silently no-opped with exit 0 because the slash-command handler rejects CLI-originated senders, so the message fell through to an ordinary agent turn that compacted nothing. - Add `openclaw sessions compact <key>` wrapping the existing `sessions.compact` RPC; exit non-zero on a transport error or an `ok:false` payload so automation never mistakes a silent no-op for success. - Reject `openclaw agent --message '/compact'` with a redirect to the new command and exit 1 instead of a silent exit 0. The shared chat-side `/compact` handler is left untouched (no compatibility / message-delivery blast radius). - Strictly validate `--max-lines` and `--timeout` (positive integers only). - Document the command and the `sessions.compact` RPC in docs/cli/sessions.md. Fixes #90640. * fix(cli): inherit parent `sessions` options for `compact` `openclaw sessions compact <key>` did not merge the parent `sessions` command options the way its sibling subcommands (list/cleanup/info/…) do, so a parent-level `--agent`/`--json` was silently dropped. In particular `openclaw sessions --agent work compact <key>` compacted the default agent's session instead of the work agent's — a wrong-target session-state mutation. Merge the parent options in the compact action (parent `--agent`/`--json`, with the compact-level option taking precedence) and add regression coverage for parent `--agent`, parent `--json`, and the compact-level override. Refs #90640. * fix(cli): report pending Codex compaction and reject unsupported parent options Address two ClawSweeper review findings on the `sessions compact` command: - `sessions-compact.ts`: the Codex app-server `thread/compact/start` path returns `ok:true / compacted:false` with a pending marker, meaning the compaction was *started* asynchronously. The formatter collapsed every non-compacted success into "No compaction needed", so Codex users were told nothing happened. Report it as a started/pending compaction instead. - `register.status-health-sessions.ts`: the parent `sessions` command defines list-only options (`--store`/`--all-agents`/`--active`/`--limit`) that the compact action previously ignored. Silently dropping a parent `--store` is dangerous — the gateway resolves the target store itself, so a user could believe they targeted one store while another is mutated. Reject any unsupported inherited parent option with a clear error and a non-zero exit. Add regression tests for the pending-compaction message and the rejected parent options. Refs #90640. * fix(gateway): guard sessions.compact maxLines truncation against active runs The non-maxLines (LLM) compact branch interrupts an active session run before compacting, but the maxLines truncate branch read the tail, archived, and overwrote the transcript in place without that guard. Exposing `--max-lines` as a documented CLI command (this PR) would make the active-run data-loss mode tracked by #72765 easy to trigger from ordinary CLI usage. Run the same interruptSessionRunIfActive guard in the maxLines branch before reading the tail and truncating, matching the LLM compact path. Add gateway regression coverage over a real in-process Gateway: with no active run, the maxLines branch truncates the on-disk transcript 500 -> 50 and preserves the original 500 lines in the .bak archive; with an active embedded run, the maxLines branch fires the same interrupt (abort + wait-for-end) before archiving and truncating. * docs(cli): move sessions compact section above related links The new "Compact a session" section was inserted between the cleanup section's inline "Related:" list and the page's final "## Related" block, splitting related-link content around the command docs. Move the compact section above the related-links area and merge the orphaned "Session config" link into the single final "## Related" block. * fix(gateway): avoid no-op compact aborts Signed-off-by: sallyom <[email protected]> * fix(gateway): satisfy compact preflight lint Signed-off-by: sallyom <[email protected]> * fix(sessions): preserve compacted transcript structure --------- Signed-off-by: sallyom <[email protected]> Co-authored-by: sallyom <[email protected]> Co-authored-by: Peter Steinberger <[email protected]>

…ompact` (fixes openclaw#90640) (openclaw#91378) * feat(cli): add `sessions compact` command and fail loudly on CLI `/compact` `sessions.compact` was reachable only as an internal Gateway RPC — no CLI command, no docs — and `openclaw agent --message '/compact'` silently no-opped with exit 0 because the slash-command handler rejects CLI-originated senders, so the message fell through to an ordinary agent turn that compacted nothing. - Add `openclaw sessions compact <key>` wrapping the existing `sessions.compact` RPC; exit non-zero on a transport error or an `ok:false` payload so automation never mistakes a silent no-op for success. - Reject `openclaw agent --message '/compact'` with a redirect to the new command and exit 1 instead of a silent exit 0. The shared chat-side `/compact` handler is left untouched (no compatibility / message-delivery blast radius). - Strictly validate `--max-lines` and `--timeout` (positive integers only). - Document the command and the `sessions.compact` RPC in docs/cli/sessions.md. Fixes openclaw#90640. * fix(cli): inherit parent `sessions` options for `compact` `openclaw sessions compact <key>` did not merge the parent `sessions` command options the way its sibling subcommands (list/cleanup/info/…) do, so a parent-level `--agent`/`--json` was silently dropped. In particular `openclaw sessions --agent work compact <key>` compacted the default agent's session instead of the work agent's — a wrong-target session-state mutation. Merge the parent options in the compact action (parent `--agent`/`--json`, with the compact-level option taking precedence) and add regression coverage for parent `--agent`, parent `--json`, and the compact-level override. Refs openclaw#90640. * fix(cli): report pending Codex compaction and reject unsupported parent options Address two ClawSweeper review findings on the `sessions compact` command: - `sessions-compact.ts`: the Codex app-server `thread/compact/start` path returns `ok:true / compacted:false` with a pending marker, meaning the compaction was *started* asynchronously. The formatter collapsed every non-compacted success into "No compaction needed", so Codex users were told nothing happened. Report it as a started/pending compaction instead. - `register.status-health-sessions.ts`: the parent `sessions` command defines list-only options (`--store`/`--all-agents`/`--active`/`--limit`) that the compact action previously ignored. Silently dropping a parent `--store` is dangerous — the gateway resolves the target store itself, so a user could believe they targeted one store while another is mutated. Reject any unsupported inherited parent option with a clear error and a non-zero exit. Add regression tests for the pending-compaction message and the rejected parent options. Refs openclaw#90640. * fix(gateway): guard sessions.compact maxLines truncation against active runs The non-maxLines (LLM) compact branch interrupts an active session run before compacting, but the maxLines truncate branch read the tail, archived, and overwrote the transcript in place without that guard. Exposing `--max-lines` as a documented CLI command (this PR) would make the active-run data-loss mode tracked by openclaw#72765 easy to trigger from ordinary CLI usage. Run the same interruptSessionRunIfActive guard in the maxLines branch before reading the tail and truncating, matching the LLM compact path. Add gateway regression coverage over a real in-process Gateway: with no active run, the maxLines branch truncates the on-disk transcript 500 -> 50 and preserves the original 500 lines in the .bak archive; with an active embedded run, the maxLines branch fires the same interrupt (abort + wait-for-end) before archiving and truncating. * docs(cli): move sessions compact section above related links The new "Compact a session" section was inserted between the cleanup section's inline "Related:" list and the page's final "## Related" block, splitting related-link content around the command docs. Move the compact section above the related-links area and merge the orphaned "Session config" link into the single final "## Related" block. * fix(gateway): avoid no-op compact aborts Signed-off-by: sallyom <[email protected]> * fix(gateway): satisfy compact preflight lint Signed-off-by: sallyom <[email protected]> * fix(sessions): preserve compacted transcript structure --------- Signed-off-by: sallyom <[email protected]> Co-authored-by: sallyom <[email protected]> Co-authored-by: Peter Steinberger <[email protected]>

openclaw-barnacle Bot added gateway Gateway runtime size: XS labels Apr 27, 2026

This was referenced Apr 27, 2026

Discord threads should inherit parent /model override without parent transcript fork #72755

Closed

Thread sessions should inherit parent channel's modelByChannel override #53447

Closed

chinadbo force-pushed the fix/sessions-compact-maxlines-active-run branch from e63091e to 4561004 Compare April 27, 2026 12:47

openclaw-barnacle Bot added size: S channel: discord Channel integration: discord extensions: memory-core Extension: memory-core commands Command implementations size: L and removed size: XS size: S labels Apr 27, 2026

chinadbo force-pushed the fix/sessions-compact-maxlines-active-run branch from c41d0a0 to 0607295 Compare April 27, 2026 13:34

openclaw-barnacle Bot added size: XS and removed channel: discord Channel integration: discord extensions: memory-core Extension: memory-core commands Command implementations size: L labels Apr 27, 2026

chinadbo force-pushed the fix/sessions-compact-maxlines-active-run branch 2 times, most recently from 5cec420 to 237adb0 Compare April 27, 2026 14:19

openclaw-barnacle Bot added size: S and removed size: XS labels Apr 28, 2026

邓波 and others added 3 commits April 28, 2026 11:11

fix(gateway): guard maxLines compact against active session runs

c27898f

fix(gateway): guard chatAbortControllers against undefined in hasTrac…

282b159

…kedActiveSessionRun

test(gateway): add regression test for maxLines compact with active e…

2cf2a8d

…mbedded run

chinadbo force-pushed the fix/sessions-compact-maxlines-active-run branch from 9ede439 to 2cf2a8d Compare April 28, 2026 03:11

ci: retrigger CI

9b95bc6

clawsweeper Bot added rating: 🧂 unranked krab Not merge-ready due to missing proof or serious correctness/safety concerns. status: 📣 needs proof The PR needs real behavior proof before ClawSweeper can clear the contributor ask. labels May 20, 2026

openclaw-barnacle Bot added the triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. label May 20, 2026

openclaw-barnacle Bot added the stale Marked as stale due to inactivity label Jun 4, 2026

clawsweeper Bot added the P1 High-priority user-facing bug, regression, or broken workflow. label Jun 4, 2026

clawsweeper Bot mentioned this pull request Jun 8, 2026

feat(cli): add openclaw sessions compact and fail loudly on CLI /compact (fixes #90640) #91378

Merged

openclaw-barnacle Bot removed the stale Marked as stale due to inactivity label Jun 14, 2026

clawsweeper Bot added the merge-risk: 🚨 session-state 🚨 May lose, corrupt, stale, or mis-associate session, agent, or context state. label Jun 14, 2026

clawsweeper Bot closed this Jun 14, 2026

yetval mentioned this pull request Jun 24, 2026

[BUG]sessions: daily-reset boundary crossed mid-run renames a live transcript, splitting the running turn's output #96546

Closed

This was referenced Jun 24, 2026

fix(session): defer transcript archive when active run crosses daily reset boundary (fixes #96546) (AI-assisted) #96552

Closed

fix(sessions): defer daily rollover during active reply runs #96715

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

fix(gateway): guard maxLines compact against active session runs#72765

fix(gateway): guard maxLines compact against active session runs#72765
chinadbo wants to merge 4 commits into
openclaw:mainfrom
chinadbo:fix/sessions-compact-maxlines-active-run

chinadbo commented Apr 27, 2026

Uh oh!

greptile-apps Bot commented Apr 27, 2026

Uh oh!

chinadbo commented Apr 27, 2026

Uh oh!

steipete commented Apr 27, 2026

Uh oh!

chinadbo commented Apr 28, 2026

Uh oh!

clawsweeper Bot commented Apr 30, 2026 •

edited

Loading

Uh oh!

clawsweeper Bot commented May 20, 2026

Uh oh!

openclaw-barnacle Bot commented Jun 4, 2026

Uh oh!

clawsweeper Bot commented Jun 14, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Uh oh!

Conversation

chinadbo commented Apr 27, 2026

Summary

Fix

Test plan

Uh oh!

greptile-apps Bot commented Apr 27, 2026

Greptile Summary

Confidence Score: 5/5

Uh oh!

chinadbo commented Apr 27, 2026

Uh oh!

steipete commented Apr 27, 2026

Uh oh!

chinadbo commented Apr 28, 2026

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!

clawsweeper Bot commented May 20, 2026

Uh oh!

openclaw-barnacle Bot commented Jun 4, 2026

Uh oh!

clawsweeper Bot commented Jun 14, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

clawsweeper Bot commented Apr 30, 2026 •

edited

Loading