guitar_score/.agent/services/claude-mem/docs/PR-SHIPPING-REPORT.md

Claude-Mem PR Shipping Report

Generated: 2026-02-04

Executive Summary

6 PRs analyzed for shipping readiness. 1 is ready to merge, 4 have conflicts, 1 is too large for easy review.

PR	Title	Status	Recommendation
#856	Idle timeout for zombie processes	✅ MERGEABLE	Ship it
#700	Windows Terminal popup fix	⚠️ Conflicts	Rebase, then ship
#722	In-process worker architecture	⚠️ Conflicts	Rebase, high impact
#657	generate/clean CLI commands	⚠️ Conflicts	Rebase, then ship
#863	Ragtime email investigation	🔍 Needs review	Research pending
#464	Sleep Agent Pipeline (contributor)	🔴 Too large	Request split or dedicated review

---

Ready to Ship

PR #856: Idle Timeout for Zombie Observer Processes

Status: ✅ MERGEABLE (no conflicts)

Metric	Value
Additions	928
Deletions	171
Files	8
Risk	Low-Medium

What it does:

• Adds 3-minute idle timeout to SessionQueueProcessor
• Prevents zombie observer processes that were causing 13.4GB swap usage
• Processes exit gracefully after inactivity instead of waiting forever

Why ship it:

• Fixes real user-reported issue (79 zombie processes)
• Well-tested (11 new tests, 440 lines of test coverage)
• Clean implementation, preventive approach
• Supersedes PR #848's reactive cleanup
• No conflicts, ready to merge

Review notes:

• 1 Greptile bot comment (addressed)
• Race condition fix included
• Enhanced logging added

---

Needs Rebase (Have Conflicts)

PR #700: Windows Terminal Popup Fix

Status: ⚠️ CONFLICTING

Metric	Value
Additions	187
Deletions	399
Files	8
Risk	Medium

What it does:

• Eliminates Windows Terminal popup by removing spawn-based daemon
• Worker start command becomes daemon directly (no child spawn)
• Removes restart command (users do stop then start)
• Net simplification: -212 lines

Breaking changes:

• restart command removed

Review status:

• ✅ 1 APPROVAL from @volkanfirat (Jan 15, 2026)

Action needed: Resolve conflicts, then ready to ship.

---

PR #722: In-Process Worker Architecture

Status: ⚠️ CONFLICTING

Metric	Value
Additions	869
Deletions	4,658
Files	112
Risk	High

What it does:

• Hook processes become the worker (no separate daemon spawning)
• First hook that needs worker becomes the worker
• Eliminates Windows spawn issues ("NO SPAWN" rule)
• 761 tests pass

Architectural impact: HIGH

• Fundamentally changes worker lifecycle
• Hook processes stay alive (they ARE the worker)
• First hook wins port 37777, others use HTTP

Action needed: Resolve conflicts. Consider relationship with PR #700 (both touch worker architecture).

---

PR #657: Generate/Clean CLI Commands

Status: ⚠️ CONFLICTING

Metric	Value
Additions	1,184
Deletions	5,057
Files	104
Risk	Medium

What it does:

• Adds claude-mem generate and claude-mem clean CLI commands
• Fixes validation bugs (deleted folders recreated from stale DB)
• Fixes Windows path handling
• Adds automatic shell alias installation
• Disables subdirectory CLAUDE.md files by default

Breaking changes:

• Default behavior change: folder CLAUDE.md now disabled by default

Action needed: Resolve conflicts, complete Windows testing.

---

Needs Attention

PR #863: Ragtime Email Investigation

Status: 🔍 Research pending

Research agent did not return results. Manual review needed.

---

PR #464: Sleep Agent Pipeline (Contributor: @laihenyi)

Status: 🔴 Too large for effective review

Metric	Value
Additions	15,430
Deletions	469
Files	73
Wait time	37+ days
Risk	High

What it does:

• Sleep Agent Pipeline with memory tiering
• Supersession detection
• Session Statistics API (/api/session/:id/stats)
• StatusLine + PreCompact hooks
• Context Generator improvements
• Self-healing CI workflow

Concerns:

Issue	Details
🔴 Size	15K+ lines is too large for effective review
🔴 SupersessionDetector	Single file with 1,282 additions
🟡 No tests visible	Test plan checkboxes unchecked
🟡 Self-healing CI	Auto-fix workflow could cause infinite commit loops
🟡 Serena config	Adds .serena/ tooling

Recommendation:

1. Option A: Request contributor split into 4-5 smaller PRs
2. Option B: Allocate dedicated review time (several hours)
3. Option C: Cherry-pick specific features (hooks, stats API)

Note: Contributor has been waiting 37+ days. Deserves response either way.

---

Shipping Strategy

Phase 1: Quick Wins (This Week)

1. Merge #856 — Ready now, fixes real user issue
2. Rebase #700 — Has approval, Windows fix needed
3. Rebase #657 — Useful CLI commands

Phase 2: Architecture (Careful Review)

4. Review #722 — High impact, conflicts with #700 approach?
   • Both PRs eliminate spawning but in different ways
   • May need to pick one approach

Phase 3: Contributor PR

5. Respond to #464 — Options:
   • Ask for split
   • Schedule dedicated review
   • Cherry-pick subset

Phase 4: Investigation

6. Manual review #863 — Ragtime email feature

---

Conflict Resolution Order

Since multiple PRs have conflicts, suggested rebase order:

1. #856 (merge first — no conflicts)
2. #700 (rebase onto main after #856)
3. #657 (rebase onto main after #700)
4. #722 (rebase last — may conflict with #700 architecturally)

---

Summary

Ready	Conflicts	Needs Work
1 PR (#856)	3 PRs (#700, #722, #657)	2 PRs (#464, #863)

Immediate action: Merge #856, then rebase the conflict PRs in order.