johno/beads
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

bd sync: 2025-11-07 21:29:59

Browse Source
This commit is contained in:
Steve Yegge
2025-11-07 21:29:59 -08:00
parent 3431cd455a
commit 6f22c9641a
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
.beads/beads.jsonl
View File

@@ -294,7 +294,7 @@
{"id":"bd-mlcz","content_hash":"35434fed7e20e8b67f7b598c46ce5dc65c111de237ea38b36a88a671b023b58b","title":"Implement bd migrate command","description":"Add bd migrate command to move issues between repos with filtering. Should support: filtering by status/priority/labels, dry-run mode, preserving dependencies, handling source_repo field updates.","status":"closed","priority":1,"issue_type":"task","created_at":"2025-11-05T18:04:29.902151-08:00","updated_at":"2025-11-05T18:42:52.536951-08:00","closed_at":"2025-11-05T18:42:52.536951-08:00","source_repo":".","dependencies":[{"issue_id":"bd-mlcz","depends_on_id":"bd-8rd","type":"parent-child","created_at":"2025-11-05T18:04:39.072312-08:00","created_by":"daemon"}]}
{"id":"bd-mn9p","content_hash":"89752e08d5a278ed301c655753e07bc2b564e710ce6d9e7e535bbd9ec48af182","title":"bd-hv01: Brittle string comparison breaks with JSON field reordering","description":"## Problem\ndeletion_tracking.go:125 uses string comparison to detect unchanged issues:\n\n```go\nif leftLine, existsInLeft := leftIndex[id]; existsInLeft \u0026\u0026 leftLine == baseLine {\n deletions = append(deletions, id)\n}\n```\n\nThis breaks if:\n- JSON field order changes (legal in JSON)\n- Timestamps updated by import/export\n- Whitespace/formatting changes\n- Floating point precision varies\n\n## Example Failure\n```json\n// baseLine\n{\"id\":\"bd-1\",\"priority\":1,\"status\":\"open\"}\n// leftLine (same data, different order)\n{\"id\":\"bd-1\",\"status\":\"open\",\"priority\":1}\n```\nThese are semantically identical but string comparison fails.\n\n## Fix\nParse and compare JSON semantically:\n```go\nfunc jsonEquals(a, b string) bool {\n var objA, objB map[string]interface{}\n json.Unmarshal([]byte(a), \u0026objA)\n json.Unmarshal([]byte(b), \u0026objB)\n return reflect.DeepEqual(objA, objB)\n}\n```\n\n## Files Affected\n- cmd/bd/deletion_tracking.go:125\n- cmd/bd/deletion_tracking.go:134-170 (buildIDToLineMap)","status":"closed","priority":1,"issue_type":"bug","created_at":"2025-11-06T18:15:35.090716-08:00","updated_at":"2025-11-06T18:46:55.889888-08:00","closed_at":"2025-11-06T18:46:55.889888-08:00","source_repo":".","dependencies":[{"issue_id":"bd-mn9p","depends_on_id":"bd-rbxi","type":"parent-child","created_at":"2025-11-06T18:19:14.790898-08:00","created_by":"daemon"}]}
{"id":"bd-my64","content_hash":"8f4eb8056f81096e7090813f319b3aa996ada6dc5809d81305271d0584c2f364","title":"Pre-push hook and daemon export produce different JSONL","description":"After committing and pushing, git status shows .beads/beads.jsonl as dirty. Investigation shows:\n\n1. Pre-push hook ran successfully and exported DB → JSONL\n2. Push completed\n3. Shortly after, daemon exported DB → JSONL again with different content\n4. Diff shows comments added to old issues (bd-23a8, bd-6049, bd-87a0)\n\nTimeline:\n- Commit c731c45 \"Update beads JSONL\"\n- Pre-push hook exported JSONL\n- Push succeeded\n- Daemon PID 33314 exported again with different content\n\nQuestions:\n1. Did someone run a command between commit and daemon export?\n2. Is there a timing issue where pre-push hook doesn't capture all DB changes?\n3. Should pre-commit hook flush daemon changes before committing?\n\nThe comments appear to be from Nov 5 (created_at: 2025-11-05T08:38:46Z) but are only appearing in JSONL now. This suggests the DB had these comments but they weren't exported during pre-push.\n\nPossible causes:\n- Pre-push hook uses BEADS_NO_DAEMON=1 which might skip pending writes\n- Daemon has unflushed changes in memory\n- Race condition between pre-push export and daemon's periodic export","notes":"Improved fix based on oracle code review:\n1. Pre-push now flushes pending changes first (prevents debounce race)\n2. Uses git status --porcelain to catch all change types\n3. Handles both beads.jsonl and issues.jsonl\n4. Works even if bd not installed (git-only check)","status":"closed","priority":1,"issue_type":"bug","created_at":"2025-11-06T18:49:54.570993-08:00","updated_at":"2025-11-06T19:01:14.549032-08:00","closed_at":"2025-11-06T18:57:42.710282-08:00","source_repo":"."}
{"id":"bd-ndyz","content_hash":"f3bee1df504d7d508b424b9a01cde89ad8aeedcd036ed1b66d71c4962c4a80cf","title":"GH#243: Recurring stale daemon.lock causes 5s delays","description":"User reports daemon.lock keeps becoming stale after running Claude with beads.\n\nSymptom:\n- bd ready takes 5 seconds (exact)\n- daemon.lock exists but socket is missing\n- bd daemons killall temporarily fixes it\n- Problem recurs after using beads with AI agents\n\nUser on v0.22.0, Macbook M2, 132 issues (89 closed)\n\nHypothesis: Daemon is crashing or exiting uncleanly during agent sessions, leaving stale lock file.\n\nNeed to:\n1. Add crash logging to daemon to understand why it's exiting\n2. Improve cleanup on daemon exit (ensure lock is always removed)\n3. Add automatic stale lock detection/cleanup\n4. Consider making daemon more resilient to crashes","design":"Root cause: 5s delay from slow RPC connect attempts when socket missing but clients retry with long timeouts. Lock file mechanism is fine (OS releases on crash), but missing socket + stale pid cause unnecessary connection attempts.\n\nKey insight: The lock itself isn't stale (OS-managed), but socket cleanup on crash is incomplete, leading clients to wait through full dial timeout.","notes":"Oracle analysis complete. Converting to epic with 5 focused sub-issues:\n1. RPC fast-fail with socket stat + short timeouts (P0)\n2. Standardize daemon detection with lock probe (P1) \n3. Crash recovery improvements (P2)\n4. Self-heal stale artifacts (P2)\n5. Diagnostics and debugging (P3)","status":"in_progress","priority":0,"issue_type":"bug","created_at":"2025-11-07T16:32:23.576171-08:00","updated_at":"2025-11-07T16:41:51.955399-08:00","source_repo":"."}
{"id":"bd-ndyz","content_hash":"f3bee1df504d7d508b424b9a01cde89ad8aeedcd036ed1b66d71c4962c4a80cf","title":"GH#243: Recurring stale daemon.lock causes 5s delays","description":"User reports daemon.lock keeps becoming stale after running Claude with beads.\n\nSymptom:\n- bd ready takes 5 seconds (exact)\n- daemon.lock exists but socket is missing\n- bd daemons killall temporarily fixes it\n- Problem recurs after using beads with AI agents\n\nUser on v0.22.0, Macbook M2, 132 issues (89 closed)\n\nHypothesis: Daemon is crashing or exiting uncleanly during agent sessions, leaving stale lock file.\n\nNeed to:\n1. Add crash logging to daemon to understand why it's exiting\n2. Improve cleanup on daemon exit (ensure lock is always removed)\n3. Add automatic stale lock detection/cleanup\n4. Consider making daemon more resilient to crashes","design":"Root cause: 5s delay from slow RPC connect attempts when socket missing but clients retry with long timeouts. Lock file mechanism is fine (OS releases on crash), but missing socket + stale pid cause unnecessary connection attempts.\n\nKey insight: The lock itself isn't stale (OS-managed), but socket cleanup on crash is incomplete, leading clients to wait through full dial timeout.","notes":"Oracle analysis complete. Converting to epic with 5 focused sub-issues:\n1. RPC fast-fail with socket stat + short timeouts (P0)\n2. Standardize daemon detection with lock probe (P1) \n3. Crash recovery improvements (P2)\n4. Self-heal stale artifacts (P2)\n5. Diagnostics and debugging (P3)","status":"closed","priority":0,"issue_type":"bug","created_at":"2025-11-07T16:32:23.576171-08:00","updated_at":"2025-11-07T21:29:56.009737-08:00","closed_at":"2025-11-07T21:29:56.009737-08:00","source_repo":"."}
{"id":"bd-ng56","content_hash":"adb5cea12f59ff7a9934ddd6f4e1cf09632e162bd4baad9f83ffea96f43d3e88","title":"bd-hv01: Three full JSONL reads on every sync (performance)","description":"Problem: computeAcceptedDeletions reads three JSONL files completely into memory (base, left, merged). For 1000 issues at 1KB each, this is 3MB read and 3000 JSON parse operations.\n\nImpact: Acceptable now (~20-35ms overhead) but will be slow for large repos (10k+ issues).\n\nPossible optimizations: single-pass streaming, memory-mapped files, binary format, incremental snapshots.\n\nFiles: cmd/bd/deletion_tracking.go:101-208","status":"closed","priority":3,"issue_type":"task","created_at":"2025-11-06T18:16:25.653076-08:00","updated_at":"2025-11-06T19:41:04.67733-08:00","closed_at":"2025-11-06T19:41:04.67733-08:00","source_repo":".","dependencies":[{"issue_id":"bd-ng56","depends_on_id":"bd-rbxi","type":"parent-child","created_at":"2025-11-06T18:19:15.148149-08:00","created_by":"daemon"}]}
{"id":"bd-nqes","content_hash":"d7ca10b54b92c464a0c4c1c932f2918a6e0e952a5fd7337acb9160cd524dc59a","title":"bd-hv01: Non-atomic snapshot operations can cause data loss","description":"## Problem\nIn sync.go:146-155 and daemon_sync.go:502-505, snapshot capture failures are logged as warnings but sync continues:\n\n```go\nif err := exportToJSONL(ctx, jsonlPath); err != nil { ... }\nif err := captureLeftSnapshot(jsonlPath); err != nil {\n fmt.Fprintf(os.Stderr, \"Warning: failed to capture snapshot...\")\n}\n```\n\nIf export succeeds but snapshot capture fails, the merge uses stale snapshot data, potentially deleting wrong issues.\n\n## Impact\n- Critical data integrity issue\n- Could delete issues incorrectly during multi-workspace sync\n\n## Fix\nMake snapshot capture mandatory:\n```go\nif err := captureLeftSnapshot(jsonlPath); err != nil {\n return fmt.Errorf(\"failed to capture snapshot (required for deletion tracking): %w\", err)\n}\n```\n\n## Files Affected\n- cmd/bd/sync.go:146-155\n- cmd/bd/daemon_sync.go:502-505","status":"closed","priority":1,"issue_type":"bug","created_at":"2025-11-06T18:15:33.574158-08:00","updated_at":"2025-11-06T18:46:55.874814-08:00","closed_at":"2025-11-06T18:46:55.874814-08:00","source_repo":".","dependencies":[{"issue_id":"bd-nqes","depends_on_id":"bd-rbxi","type":"parent-child","created_at":"2025-11-06T18:19:14.749153-08:00","created_by":"daemon"}]}
{"id":"bd-o1s5","content_hash":"a9d63140e09a3aba769436c54c204e9369912a7d8d51d2aae9e68c8148357d3b","title":"Ready issue","description":"","status":"open","priority":1,"issue_type":"task","created_at":"2025-11-07T19:07:01.984841-08:00","updated_at":"2025-11-07T19:07:01.984841-08:00","source_repo":"."}