[juliusmarminge]

Problem

Every completed one-minute pull-request sweep can remain reachable through Effect stack frames retained by a cached lookup fiber. On an isolated headless server with 31 stopped thread projections linked to one open PR, successive completed sweeps retained one then two groups Maps after full GC.

Change

Use untraced function wrappers with explicit named spans for the two helpers defined inside each sweep. This releases sweep inputs while preserving the existing tracing span names and polling behavior. Add a collection regression that keeps a lookup fiber cached after the sweep finishes.

Scope and approval

Fixes #25172

Triaged bug via the automated issue flow: #25172. Focused server retention fix; no client, contract, authentication, or persisted-state changes.

Verification

• Regression on the baseline implementation: expected zero surviving link objects, actual 30. Exact red command and failure.
• Final head: all 49 reactor tests pass, including the collection regression. Command and output.
• vp run check:changed passed format, lint, both knip checks, GraphQL, affected typechecks, and focused tests. Focused tests passed again after merging current main.
• Two independent reviews returned MERGE without blocking findings. Final proof audit returned SUFFICIENT.

Real headless server measurement	Before	After
Retained sweep Maps after completed sweep 1 / full GC	1	0
Retained sweep Maps after completed sweep 2 / full GC	2	0
Helper tracing span names exported	Present	Preserved

Evidence index, runtime transcript, counters, provenance, and fixture script. Server-only change: visual fixture matrix is empty. Baseline runtime was captured at 0ad399c6c5; fixed runtime at 15b7291404, whose changed server files are byte-identical at final head 57a7bd0a8d. Upstream changes merged afterward were client-only.

Real-data seeding hit a VM database migration-history mismatch, so runtime proof uses synthetic stopped threads. The baseline contained 30 synthetic threads plus the bootstrap thread; after used 31 synthetic threads. This verifies retention and release, not the reporter's full 18,736-link workload or exact growth rate. Raw heaps are not published because they may contain credentials. CI and hosted review gates remain pending; no bugfix label will be added before they pass.

Models: Claude Opus 5.5 and GPT-6.1-sol, through T3 Code's Claude and Codex harnesses.

[macroscopeapp]

Approvability

Verdict: Approved at 57a7bd0

Macroscope's review found this PR approvable — This is a narrowly scoped server memory-retention fix that changes only Effect wrapper instrumentation, preserving existing spans and pull-request synchronization behavior. A focused regression test verifies that completed sweep data can be collected when a cache retains the lookup fiber.

^{You can add or adjust custom eligibility rules. Learn more.}

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	5.0 KiB	5.0 KiB	0 B (0.0%)	6.8 KiB	✅
Codex	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Codex	Live turn WebSocket wire	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	20.9 KiB	20.9 KiB	0 B (0.0%)	29.3 KiB	✅
Codex	Live turn messages	2	2	0 (0.0%)	8	✅
Claude	Total thread wire	5.0 KiB	5.0 KiB	0 B (0.0%)	6.8 KiB	✅
Claude	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Claude	Live turn WebSocket wire	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	21.2 KiB	21.2 KiB	0 B (0.0%)	29.3 KiB	✅
Claude	Live turn messages	2	2	0 (0.0%)	8	✅

Baseline: f4e1407 · PR result: 57a7bd0 · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 108.5 KiB
• Claude decoded thread snapshot: 108.8 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[coderabbitai]

No actionable comments were generated in the recent review.

Recent review info

Run configuration

• Source mode: Auto
• Repository: pingdotgg/t3code
• Branch: fix/issue-25172
• File: .coderabbit.config.ts
• Commit: 57a7bd0
• Layers: Repository TS config
• Review profile: CHILL
• Plan: Team
• Run ID: 0cfa97c0-7f88-4a7c-b95a-a13f487c3002

Commits

Reviewing files that changed from the base of the PR and between f4e1407 and 57a7bd0.

Files selected for processing: 2

• apps/server/src/orchestration-v2/PullRequestSyncReactor.test.ts
• apps/server/src/orchestration-v2/PullRequestSyncReactor.ts

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 0 remain after this review.

Walkthrough

The pull-request sync helpers now use untraced Effect functions with explicit spans. A new test runs a cache-backed sweep, clears the snapshot, and checks that no SweptLink objects remain.

Changes: Sweep retention

Layer / File(s)	Summary
Span handling and retention test apps/server/src/orchestration-v2/PullRequestSyncReactor.ts, apps/server/src/orchestration-v2/PullRequestSyncReactor.test.ts	syncEntry and syncGroup use Effect.fnUntraced with explicit spans. The test runs a cache-backed sweep and checks for remaining SweptLink objects after clearing the snapshot.

Priority: Normal

Estimated code review effort: 2 (Simple) | ~10 minutes

Change: Bug fix · Severity of issue fixed: Medium

Suggested labels: size:S

Suggested reviewers: t3dotgg, maria-rcks

Merge Risk: Minimal · up to 57a7b

No actionable merge-blocking risk was identified; the change is ready for normal checks.

Pre-merge checks: 4

Check name	Status	Explanation
Title check		The title is concise, specific, and accurately describes the server fix that releases completed pull-request sweep data.
Description check		The description includes the required Problem, Change, Scope and approval, and Verification sections. It explains the retention issue, links the triaged issue, documents focused tests and measurements…
Linked Issues check		Issue [#25172] requires completed PullRequestSyncReactor.sweep runs to release the groups map and its thread and link data. The change replaces the sweep-local Effect.fn closures for syncEntry…
Out of Scope Changes check		The reviewed changes are limited to PullRequestSyncReactor.ts and its regression test. The implementation changes Effect tracing and closure retention for the server pull-request sweep. The test val…

Full details: Description check

Explanation

The description includes the required Problem, Change, Scope and approval, and Verification sections. It explains the retention issue, links the triaged issue, documents focused tests and measurements, and states verification limits and pending gates.

Full details: Linked Issues check

Explanation

Issue [#25172] requires completed PullRequestSyncReactor.sweep runs to release the groups map and its thread and link data. The change replaces the sweep-local Effect.fn closures for syncEntry and syncGroup with Effect.fnUntraced wrappers and explicit Effect.withSpan calls. The existing span names remain PullRequestSyncReactor.syncEntry and PullRequestSyncReactor.syncGroup. The new regression keeps a cache lookup fiber alive and verifies that 30 swept link objects become collectible. The reported reactor and focused checks also pass.

Full details: Out of Scope Changes check

Explanation

The reviewed changes are limited to PullRequestSyncReactor.ts and its regression test. The implementation changes Effect tracing and closure retention for the server pull-request sweep. The test validates collection behavior under a retained cache fiber. These changes directly support issue [#25172]. No unrelated client, contract, authentication, or persisted-state changes are present in the reviewed diff.

Finishing Touches

📝 Generate docstrings

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

_{Comment @coderabbitai help to get the list of available commands.}