fix: share Git objects and prune orphan snapshots - #300
Conversation
Adding .gitkeep for PR creation (default mode). This file will be removed when the task is complete. Issue: #298
Working session summaryI reproduced the defect: tracking an 8 MB repository created another ~8 MB snapshot object store with no alternates. Using Git alternates reduced the unchanged snapshot store to 16 KB while preserving object access. This is a bounded fix. Proposed design:
PR #300 currently has no reviews or comments. Approve this design so I can begin the failing tests and implementation? This summary was automatically extracted from the AI working session output. |
🤖 Solution Draft LogThis log file contains the complete execution trace of the AI solution draft process. 💰 Cost estimation:
📊 Context and tokens usage:
Total: (79.3K + 725.8K cached) input tokens, 8.9K output tokens, $1.482124 cost 🤖 Models used:
📎 Log file uploaded as Gist (833KB)Now working session is ended, feel free to review and add any feedback on the solution draft. |
🔄 Auto-restart 1/5Reason: Merge conflicts detected Starting new session to address the issues. Auto-restart-until-mergeable mode is active. This run will stop after 5 restart iterations in total. |
# Conflicts: # .gitkeep
Configure Git alternates before indexing so unchanged blobs remain in the source repository instead of being duplicated.\n\nRefs #298
Keep the newest, recent, and live snapshot stores while removing old stores whose recorded worktrees no longer exist. Cascade cleanup through project session storage at startup and disposal.\n\nRefs #298
Complete the per-session storage cascade and bound the isolated regression test process. Refs #298.
Let bun add create the clean consumer manifest so changes in the bun init template cannot inject incompatible peer versions.
Working session summaryResolved and finalized PR #300.
This summary was automatically extracted from the AI working session output. |
🔄 Auto-restart-until-mergeable Log 1/5This log file contains the complete execution trace of the AI solution draft process. 💰 Cost estimation:
📊 Context and tokens usage:
Total: (283.5K + 8.3M cached) input tokens, 44.1K output tokens, $10.270603 cost 🤖 Models used:
📎 Log file uploaded as Gist (4897KB)Now working session is ended, feel free to review and add any feedback on the solution draft. |
🎉 Auto-mergedThis pull request has been automatically merged by hive-mind.
Auto-merged by hive-mind with --auto-merge flag |
Summary
Reproduction
Before this change, creating snapshots for multiple project worktrees populated each snapshot repository with its own full object database, and deleting a worktree left its snapshot and associated storage records indefinitely.
Verification
Fixes #298