The original PR added retry logic on top of BEGIN IMMEDIATE, but this caused
multi-minute hangs because:
1. Connection has busy_timeout=30s set via pragma
2. Each BEGIN IMMEDIATE waits up to 30s before returning SQLITE_BUSY
3. With 5 retries, worst case was 5 × 30s = 150+ seconds
The fix removes the retry loop since SQLite's busy_timeout already handles
retries internally. BEGIN IMMEDIATE still acquires the write lock early,
preventing deadlocks - we just let busy_timeout handle contention.
Root cause analysis in bd-9ldm.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Update withTx to use BEGIN IMMEDIATE with exponential backoff retry
on SQLITE_BUSY errors. This prevents "database is locked" failures
during concurrent operations (daemon + CLI, multi-agent workflows).
Changes:
- withTx now uses beginImmediateWithRetry (same pattern as RunInTransaction)
- Add dbExecutor interface for helper functions that work with both
*sql.Tx and *sql.Conn
- Update all withTx callers to use *sql.Conn
- Refactor DeleteIssue to use withTx (fixes the specific error in auto-import)
- Update markIssuesDirtyTx to accept dbExecutor interface
Affected paths:
- MarkIssuesDirty, ClearDirtyIssuesByID (dirty.go)
- AddDependency, RemoveDependency (dependencies.go)
- executeLabelOperation (labels.go)
- AddComment (events.go)
- ApplyCompaction (compact.go)
- DeleteIssue (queries.go)
Note: Some direct BeginTx calls in queries.go (CloseIssue, UpdateIssue,
ReopenIssue, DeleteIssues) still use the old pattern and could be
refactored in a follow-up.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
When creating issues with explicit IDs via PrefixOverride (for cross-rig
creation), the storage layer was still validating against the local
config prefix, causing errors like:
issue ID 'da-DataArchive-witness' does not match configured prefix 'hq'
The queries.go layer already handled PrefixOverride correctly, but
transaction.go did not. This fix adds the same PrefixOverride handling
to transaction.go:
1. Check if PrefixOverride is set
2. If set, use it as the prefix and skip validation (caller knows best)
3. Otherwise, fall back to existing IDPrefix/configPrefix logic
This allows gt doctor --fix to create agent beads for non-local rigs.
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
fix(daemon): prevent zombie state after database file replacement
Adds checkFreshness() to health check paths (GetMetadata, GetConfig, GetAllConfig) and refactors reconnect() to validate new connection before closing old.
PR-URL: https://github.com/steveyegge/beads/pull/1213
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix(ready): exclude molecule steps from bd ready by default (GH#1239)
Add ID prefix constants (IDPrefixMol, IDPrefixWisp) to types.go as single
source of truth. Update pour.go and wisp.go to use these constants.
GetReadyWork now excludes issues with -mol- in their ID when no explicit
type filter is specified. Users can still see mol steps with --type=task.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
* feat(ready): config-driven ID pattern exclusion (GH#1239)
Add ready.exclude_id_patterns config for excluding IDs from bd ready.
Default patterns: -mol-, -wisp- (molecule steps and wisps).
Changes:
- Add IncludeMolSteps to WorkFilter for internal callers
- Update findGateReadyMolecules and getMoleculeCurrentStep to use it
- Make exclusion patterns config-driven via ready.exclude_id_patterns
- Remove hardcoded MolStepIDPattern() in favor of config
- Add test for custom patterns (e.g., gastown's -role-)
Usage: bd config set ready.exclude_id_patterns "-mol-,-wisp-,-role-"
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
* docs: remove -role- example from ready.go comments
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
* docs: remove GH#1239 references from code comments
Issue references belong in commit messages, not code.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
Fixes stack overflow and database initialization failures when running bd on WSL2 with Docker Desktop bind mounts.
## Problem
The bd CLI crashed with stack overflow when running on WSL2 with repositories on Docker Desktop bind mounts (/mnt/wsl/docker-desktop-bind-mounts/...). SQLite WAL mode fails with 'locking protocol' error on these network filesystems.
## Solution
- Expand WAL mode detection to identify Docker bind mounts at /mnt/wsl/* (in addition to Windows paths at /mnt/[a-zA-Z]/)
- Fall back to DELETE journal mode on these problematic paths
- Add comprehensive unit tests for path detection
Fixes GH #1224, relates to GH #920
Co-authored-by: maphew <matt.wilkie@gmail.com>
Remove Gas Town-specific type constants (TypeMolecule, TypeGate, TypeConvoy,
TypeMergeRequest, TypeSlot, TypeAgent, TypeRole, TypeRig, TypeEvent, TypeMessage)
from internal/types/types.go.
Beads now only has core work types built-in:
- bug, feature, task, epic, chore
All Gas Town types are now purely custom types with no special handling in beads.
Use string literals like "gate" or "molecule" when needed, and configure
types.custom in config.yaml for validation.
Changes:
- Remove Gas Town type constants from types.go
- Remove mr/mol aliases from Normalize()
- Update bd types command to only show core types
- Replace all constant usages with string literals throughout codebase
- Update tests to use string literals
This decouples beads from Gas Town, making it a generic issue tracker.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
During bd init, auto-import fails with "invalid issue type" errors even
when types.custom is defined in config.yaml. This happens because custom
types are read from the database, but the database is being created
during init and doesn't have the config set yet.
Changes:
- Add GetCustomTypesFromYAML() to internal/config/config.go to read
types.custom from config.yaml via viper
- Modify GetCustomTypes() in sqlite/config.go to fallback to config.yaml
when the database doesn't have types.custom configured
- Add tests for GetCustomTypesFromYAML()
This allows fresh clones with custom types defined in config.yaml (e.g.,
Gas Town types like molecule, gate, convoy, agent, event) to successfully
auto-import their JSONL during bd init.
Fixes GH#1225
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
The ncruces/go-sqlite3 driver does not always auto-convert TEXT columns
to time.Time. This caused scan errors on updated_at/created_at fields,
blocking witness startup.
Fix: Scan timestamps into sql.NullString and parse with parseTimeString()
helper that handles RFC3339Nano, RFC3339, and SQLite native formats.
Fixes: bd-4dqmy
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The scanIssues and scanIssuesWithDependencyType helper functions were
not checking rows.Err() after iterating through query results. This
could cause errors during iteration (connection drops, context
cancellation, etc.) to be silently ignored.
Per Go database/sql best practices, rows.Err() should always be checked
after a rows.Next() loop completes to catch any errors that occurred
during iteration.
Co-authored-by: Steven Syrek <steven.syrek@deepl.com>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
Core beads built-in types now only include work types:
- bug, feature, task, epic, chore
Gas Town types (molecule, gate, convoy, merge-request, slot, agent,
role, rig, event, message) are now "well-known custom types":
- Constants still exist for code convenience
- Require types.custom configuration for validation
- bd types command shows core types and configured custom types
Changes:
- types.go: Separate core work types from well-known custom types
- IsValid(): Only accepts core work types
- bd types: Updated to show core types and custom types from config
- memory.go: Use ValidateWithCustom for custom type support
- multirepo.go: Only check core types as built-in
- Updated all tests to configure custom types
This allows Gas Town (and other projects) to define their own types
via config while keeping beads core focused on work tracking.
Closes: bd-find4
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Support --type enhancement as an alias for --type feature when creating
issues. The normalization happens before validation to ensure consistency
across all code paths.
Closes gt-hzanoe
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
When an issue is deleted, issues that depend on it were not being marked
dirty. This caused stale dependency references to persist in JSONL after
the target issue was deleted, because the dependent issues were never
re-exported.
This manifests as FK validation failures during multi-repo hydration:
"foreign key violation: issue X depends on non-existent issue Y"
The fix queries for dependent issues before deleting and marks them dirty
so they get re-exported without the stale dependency reference.
Adds test: TestDeleteIssueMarksDependentsDirty
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Role and rig beads are reference metadata that should never be closed
or appear as actionable work. They are similar to agent beads which
are already excluded.
- Add 'role' and 'rig' to the issue type exclusion list in GetReadyWork
- Update comments to document the excluded types
Fixes confusion where role/rig beads appeared in bd ready output,
leading agents to try to close them as regular work items.
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
Centralizes repository context resolution via RepoContext API, fixing bugs where git commands run in the wrong repo when BEADS_DIR points elsewhere or in worktree scenarios.
When using `bd create --id=<id>` where the ID matches an existing
tombstone (from `bd delete --hard --force`), the creation now succeeds
by first deleting the tombstone and all related records.
This enables use cases like polecat respawn where a worker needs to
recreate an issue with the same ID.
Changes:
- queries.go: Check for tombstone before insert, delete it if found
(cleans up events, labels, dependencies, comments, dirty_issues)
- tombstone_test.go: Add regression test
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
Add multiple layers of defense against misclassified wisps:
- Importer auto-detects -wisp- pattern and sets ephemeral flag
- GetReadyWork excludes -wisp- IDs via SQL LIKE clause
- Doctor check 26d detects misclassified wisps in JSONL
This addresses recurring issue where wisps with missing ephemeral
flag would pollute bd ready output after JSONL import.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add insertIssuesStrict function that uses plain INSERT instead of
INSERT OR IGNORE. Update bulkInsertIssues and transactional CreateIssues
to use the strict variant.
This fixes a race condition where INSERT OR IGNORE could silently skip
duplicate insertions, but the code would still attempt to record events
for those "inserted" issues, causing FOREIGN KEY constraint failures.
The strict INSERT will now fail explicitly if a duplicate is encountered,
which should never happen since checkForExistingIDs runs first within
the same transaction.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add missing reconnectMu.RLock() protection to storage methods that were
vulnerable to the same race condition fixed in GH#607. The FreshnessChecker
can trigger reconnect() which closes s.db while queries are in flight,
causing "database is closed" errors during daemon export operations.
Protected methods:
- labels.go: GetLabelsForIssues (GetLabels intentionally unprotected - called from GetIssue which holds lock)
- comments.go: GetIssueComments, GetCommentsForIssues
- dependencies.go: GetDependencyCounts, GetDependencyRecords, GetAllDependencyRecords, GetDependencyTree, loadDependencyGraph
- config.go: SetConfig, GetConfig, GetAllConfig, DeleteConfig, SetMetadata, GetMetadata
- dirty.go: MarkIssueDirty, GetDirtyIssues, GetDirtyIssueHash, GetDirtyIssueCount
- events.go: GetEvents, GetStatistics, GetMoleculeProgress
- hash.go: All hash methods
- hash_ids.go: GetNextChildID, ensureChildCounterUpdated (getNextChildNumber unprotected - called internally)
Internal helpers called from already-locked contexts intentionally omit
RLock to avoid deadlock (Go's RWMutex doesn't support recursive locking).
Fixes: bd-vx7fp
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
* fix(create): Use prefix from routes.jsonl when creating issues with --rig
When using `bd create --rig <name>`, the prefix from routes.jsonl was
being discarded. This caused issues to be created with the target
database's default prefix instead of the route's prefix.
This is particularly problematic when using the redirect mechanism to
share a single database across multiple rigs - the redirect correctly
routes to the shared database, but the prefix was not being applied.
The fix:
1. Capture the prefix from routing.ResolveBeadsDirForRig()
2. Temporarily override the target database's issue_prefix config
3. Restore the original prefix after issue creation
Example scenario that now works:
- routes.jsonl: {"prefix": "aops-", "path": "src/academicOps"}
- src/academicOps/.beads/redirect points to ~/writing/.beads
- `bd create --rig aops "Test"` now creates aops-xxx instead of ns-xxx
Co-Authored-By: Claude <noreply@anthropic.com>
* fix(create): pass prefix via struct field instead of mutating config
The previous approach temporarily mutated the database's issue_prefix
config during cross-rig issue creation, then restored it afterward.
This was fragile in multi-user scenarios where concurrent operations
could see the wrong prefix.
New approach:
- Add PrefixOverride field to types.Issue
- CreateIssue checks PrefixOverride first, uses it if set
- createInRig sets issue.PrefixOverride instead of mutating config
This passes state as a parameter rather than mutating shared state,
making it safe for concurrent multi-user access.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude <noreply@anthropic.com>
RenameDependencyPrefix updates issue IDs in the dependencies table but
was not rebuilding the blocked_issues_cache, leaving stale IDs in the
cache that no longer exist in the issues table.
Add invalidateBlockedCache() call at the end of RenameDependencyPrefix
to rebuild the cache with the new issue IDs.
Fixes: GH#1016
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Remove duplicate crystallizes column from schema.go
- Add crystallizes to SELECT in transaction.go SearchIssues
- Add crystallizes to SELECT in ready.go GetReadyWork and GetNewlyUnblockedByClose
- Add crystallizes to SELECT in labels.go GetIssuesByLabel
- Add missing placeholder in issues.go INSERT VALUES
- Update migrations_test.go schema to include crystallizes column
Fixes test failures caused by schema/query column count mismatches.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add TypeRig IssueType constant for rig identity beads
- Add TypeRig to IsValid() switch statement
- Add IsBuiltIn() method for multi-repo hydration trust checks
- Add Crystallizes field to Issue struct and ComputeContentHash
Fixes validation rejecting documented issue types like 'rig' and 'agent'.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Adds crystallizes column for work economics (compounds vs evaporates)
per Decision 006. Includes migration 036 and updates to all INSERT/SELECT
queries in issues.go, queries.go, dependencies.go, and transaction.go.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Tests were failing because beads.FindDatabasePath() follows the
project's .beads/redirect file, causing tests to find unexpected
databases. Fixed by:
- Setting BEADS_DIR in tests that need isolation from git repo detection
- Clearing BEADS_DIR in TestMain to prevent global contamination
- Updating migration test schema to include owner column
This ensures tests work correctly in crew directories that have
redirect files pointing to shared .beads directories.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
Merges schema additions from crew/fang, crew/giles, crew/grip, and crew/wolf:
- crystallizes: bool field for work economics (compounds vs evaporates)
- work_type: WorkType field for assignment model (mutex vs open_competition)
- source_system: string field for federation adapter tracking
- quality_score: *float32 for aggregate quality (0.0-1.0)
- delegated-from: new dependency type for work delegation chains
Migrations properly sequenced as 037-040 (after existing 036 owner_column).
Also fixes test compilation errors for removed TypeRig and IsBuiltIn references.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
Add checkForExistingIDs check to transaction-based batch creation
(sqliteTxStorage.CreateIssues) before calling insertIssues. This
prevents INSERT OR IGNORE from silently skipping duplicate IDs,
which would cause FK constraint failures when recording events
for issues that weren't actually inserted.
Also fixes unrelated test bug: renamed parseCommaSeparated to
parseCommaSeparatedList in validators_test.go.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add TypeRig constant to IssueType enum
- Add IsBuiltIn() method to IssueType for multi-repo hydration trust logic
- Fix parseCommaSeparated -> parseCommaSeparatedList function name in test
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add 'owner' field to Issue struct for tracking the human responsible
for the issue, distinct from 'created_by' which tracks the executor.
Owner is populated from git author email (GIT_AUTHOR_EMAIL or git
config user.email), per Decision 008 for CV accumulation.
Changes:
- Add Owner field to types.Issue with omitempty JSON tag
- Include Owner in content hash computation
- Add owner column migration (036_owner_column.go)
- Update all SQL queries to include owner field
- Add getOwner() helper using git author email fallback chain
- Populate owner in bd create command
- Add owner to RPC CreateArgs protocol
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
* fix: respect hierarchy.max-depth config setting (GH#995)
The hierarchy.max-depth config setting was being ignored because storage
implementations had the depth limit hardcoded to 3. This fix:
- Registers hierarchy.max-depth default (3) in config initialization
- Adds hierarchy.max-depth to yaml-only keys for config.yaml storage
- Updates SQLite and Memory storage to read max depth from config
- Adds validation to reject hierarchy.max-depth values < 1
- Adds tests for configurable hierarchy depth
Users can now set deeper hierarchies:
bd config set hierarchy.max-depth 10
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
* refactor: extract shared CheckHierarchyDepth function (GH#995)
- Extract duplicated depth-checking logic to types.CheckHierarchyDepth()
- Update sqlite and memory storage backends to use shared function
- Add t.Cleanup() for proper test isolation in sqlite test
- Add equivalent test coverage for memory storage backend
- Add comprehensive unit tests for CheckHierarchyDepth function
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
Auto-detect WSL2 environment with Windows filesystem paths (/mnt/c/, etc.)
and fall back to DELETE journal mode instead of WAL. SQLite WAL mode
does not work reliably across the WSL2/Windows 9P filesystem boundary
due to shared-memory limitations.
Detection checks:
1. Path matches /mnt/[a-z]/ pattern (Windows drive mount)
2. /proc/version contains microsoft or wsl
Fixes#920
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
The sync command was closing the daemon connection without initializing
the direct store, leaving store=nil. This caused errors in post-checkout
hook when running bd sync --import-only.
Fixed by using fallbackToDirectMode() which properly closes daemon and
initializes the store.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Root cause: CreateIssue used INSERT OR IGNORE which could silently skip
the insert (e.g., on duplicate ID from hash collision), then fail with
FOREIGN KEY constraint error when trying to record the creation event.
Fix: Add insertIssueStrict() that uses plain INSERT (fails on duplicates)
and use it for CreateIssue in both queries.go and transaction.go. The
existing insertIssue() with INSERT OR IGNORE is preserved for import
scenarios where duplicates are expected.
Added test TestCreateIssueDuplicateID to verify duplicate IDs are properly
rejected instead of silently ignored.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Added IsBlocked method to Storage interface that checks if an issue is
in the blocked_issues_cache and returns the blocking issue IDs.
The close command now checks for blockers before allowing an issue to
be closed:
- If an issue has open blockers, closing is blocked with an error message
- The --force flag overrides this check
- Works in both daemon mode (RPC) and direct storage mode
- Also handles cross-rig routed IDs
This addresses the bug where agents could close a bead even when it
depends on an open bug/issue.
Closes#962
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Implements federation trust model for multi-repo type validation:
- Built-in types are validated (catch typos)
- Non-built-in types trusted from source repos
Changes:
- Add IssueType.IsBuiltIn() method
- Add Issue.ValidateForImport() for trust-based validation
- Update upsertIssueInTx to use ValidateForImport
Closes: bd-dqwuf, bd-alpw2 | Epic: bd-9ji4z
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Remove Gas Town-specific issue types (agent, role, rig, convoy, slot)
from beads core. These types are now identified by labels instead:
- gt:agent, gt:role, gt:rig, gt:convoy, gt:slot
Changes:
- internal/types/types.go: Remove TypeAgent, TypeRole, TypeRig, TypeConvoy, TypeSlot constants
- cmd/bd/agent.go: Create agents with TypeTask + gt:agent label
- cmd/bd/merge_slot.go: Create slots with TypeTask + gt:slot label
- internal/storage/sqlite/queries.go, transaction.go: Query convoys by gt:convoy label
- internal/rpc/server_issues_epics.go: Check gt:agent label for role_type/rig label auto-add
- cmd/bd/create.go: Check gt:agent label for role_type/rig label auto-add
- internal/ui/styles.go: Remove agent/role/rig type colors
- cmd/bd/export_obsidian.go: Remove agent/role/rig/convoy type tag mappings
- Update all affected tests
This enables beads to be a generic issue tracker while Gas Town
uses labels for its specific type semantics.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
Add types.custom config key mirroring the status.custom pattern:
- Add CustomTypeConfigKey constant and GetCustomTypes() to storage interface
- Add IssueType.IsValidWithCustom() method for validation
- Add ValidateWithCustom() to Issue for combined status/type validation
- Update all validation call sites to use GetCustomTypes()
- Rename parseCustomStatuses to parseCommaSeparated for reuse
This enables Gas Town to register custom types like agent/role/rig/convoy
without hardcoding them in beads core, supporting the type extraction epic.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
- Extract inode function to platform-specific files (inode_unix.go,
inode_windows.go) to fix syscall.Stat_t compile error on Windows
- Add skipOnWindows helper and skip Unix permission/symlink tests
on Windows where chmod semantics differ
- Increase Windows test timeout from 10m to 20m since full test
suite runs slower without race detector
Fixes Windows CI failures introduced when PR #904 expanded Windows
testing from smoke tests to full test suite.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
* feat(daemon): unify auto-sync config for simpler agent workflows
## Problem
Agents running `bd sync` at session end caused delays in the Claude Code
"event loop", slowing development. The daemon was already auto-exporting
DB→JSONL instantly, but auto-commit and auto-push weren't enabled by
default when sync-branch was configured - requiring manual `bd sync`.
Additionally, having three separate config options (auto-commit, auto-push,
auto-pull) was confusing and could get out of sync.
## Solution
Simplify to two intuitive sync modes:
1. **Read/Write Mode** (`daemon.auto-sync: true` or `BEADS_AUTO_SYNC=true`)
- Enables auto-commit + auto-push + auto-pull
- Full bidirectional sync - eliminates need for manual `bd sync`
- Default when sync-branch is configured
2. **Read-Only Mode** (`daemon.auto-pull: true` or `BEADS_AUTO_PULL=true`)
- Only receives updates from team
- Does NOT auto-publish changes
- Useful for experimental work or manual review before sharing
## Benefits
- **Faster agent workflows**: No more `bd sync` delays at session end
- **Simpler config**: Two modes instead of three separate toggles
- **Backward compatible**: Legacy auto_commit/auto_push settings still work
(treated as auto-sync=true)
- **Adaptive `bd prime`**: Session close protocol adapts when daemon is
auto-syncing (shows simplified 4-step git workflow, no `bd sync`)
- **Doctor warnings**: `bd doctor` warns about deprecated legacy config
## Changes
- cmd/bd/daemon.go: Add loadDaemonAutoSettings() with unified config logic
- cmd/bd/doctor.go: Add CheckLegacyDaemonConfig call
- cmd/bd/doctor/daemon.go: Add CheckDaemonAutoSync, CheckLegacyDaemonConfig
- cmd/bd/init_team.go: Use daemon.auto-sync in team wizard
- cmd/bd/prime.go: Detect daemon auto-sync, adapt session close protocol
- cmd/bd/prime_test.go: Add stubIsDaemonAutoSyncing for testing
* docs: add comprehensive daemon technical analysis
Add daemon-summary.md documenting the beads daemon architecture,
memory analysis (explaining the 30-35MB footprint), platform support
comparison, historical problems and fixes, and architectural guidance
for other projects implementing similar daemon patterns.
Key sections:
- Architecture deep dive with component diagrams
- Memory breakdown (SQLite WASM runtime is the main contributor)
- Platform support matrix (macOS/Linux full, Windows partial)
- Historical bugs and their fixes with reusable patterns
- Analysis of daemon usefulness without database (verdict: low value)
- Expert-reviewed improvement proposals (3 recommended, 3 skipped)
- Technical design patterns for other implementations
* feat: add cross-platform CI matrix and dual-mode test framework
Cross-Platform CI:
- Add Windows, macOS, Linux matrix to catch platform-specific bugs
- Linux: full tests with race detector and coverage
- macOS: full tests with race detector
- Windows: full tests without race detector (performance)
- Catches bugs like GH#880 (macOS path casing) and GH#387 (Windows daemon)
Dual-Mode Test Framework (cmd/bd/dual_mode_test.go):
- Runs tests in both direct mode and daemon mode
- Prevents recurring bug pattern (GH#719, GH#751, bd-fu83)
- Provides DualModeTestEnv with helper methods for common operations
- Includes 5 example tests demonstrating the pattern
Documentation:
- Add dual-mode testing section to CONTRIBUTING.md
- Document RunDualModeTest API and available helpers
Test Fixes:
- Fix sync_local_only_test.go gitPull/gitPush calls
- Add gate_no_daemon_test.go for beads-70c4 investigation
* fix(test): isolate TestFindBeadsDir tests with BEADS_DIR env var
The tests were finding the real project's .beads directory instead of
the temp directory because FindBeadsDir() walks up the directory tree.
Using BEADS_DIR env var provides proper test isolation.
* fix(test): stop daemon before running test suite guard
The test suite guard checks that tests don't modify the real repo's .beads
directory. However, a background daemon running auto-sync would touch
issues.jsonl during test execution, causing false positives.
Changes:
- Set BEADS_NO_DAEMON=1 to prevent daemon auto-start from tests
- Stop any running daemon for the repo before taking the "before" snapshot
- Uses exec to call `bd daemon --stop` to avoid import cycle issues
* chore: revert .beads/issues.jsonl to upstream/main
Per CONTRIBUTING.md, .beads/issues.jsonl should not be modified in PRs.
Add BenchmarkRebuildBlockedCache_Large and _XLarge to measure cache
rebuild performance in isolation.
Results show cache rebuild takes ~773ms at 10K issues and ~1.3s at 20K,
significantly slower than the ~50ms originally estimated in the issue.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Optimize CheckExternalDeps to group refs by project and open each
external DB only once, checking all capabilities in a single query.
Before: If 10 issues depend on external:gastown:cap1, we opened
gastown's DB 10 times.
After: We open each external project's DB once, query for all
capabilities needed from that project, then close.
Also added deduplication in filterByExternalDeps to collect all
unique refs before checking, avoiding redundant checks for the
same ref across multiple issues.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Implements automatic discovery of GitHub workflow run IDs for gates
awaiting CI/CD completion. This enables the Refinery patrol to
auto-populate await_id for gh:run gates that were created without one.
Changes:
- Add `bd gate discover` command that:
- Finds open gh:run gates without await_id
- Queries recent GitHub workflow runs via gh CLI
- Matches runs to gates using heuristics (branch, commit, time)
- Updates gates with discovered run IDs
- Add `--await-id` flag to `bd update` for manual setting
- Add AwaitID to UpdateArgs in RPC protocol
- Add await_id to allowedUpdateFields in storage layer
Matching heuristics (scored, highest match wins):
- Commit SHA match: +100 points
- Branch match: +50 points
- Time proximity (<5min: +30, <10min: +20, <30min: +10)
- In-progress/queued status: +5 points
Usage:
bd gate discover # Auto-discover for all matching gates
bd gate discover --dry-run # Preview without updating
bd gate discover --branch main --limit 10
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
* feat(dates): add due date schema and --due flag
- Add due_at and defer_until columns to issues table via migration 035
- Implement --due flag on create command with ISO date parsing
- Extend RPC protocol and daemon to pass DueAt from CLI to storage
- Display DueAt and DeferUntil in show command output
- Update Issue type with new date fields
Users can now set due dates when creating issues, enabling deadline-based
task management.
* feat(dates): add compact duration parser (+6h, +1d, +2w)
- Create internal/timeparsing package with layered parser architecture
- Implement parseCompactDuration with regex pattern [+-]?\d+[hdwmy]
- Add comprehensive test suite (22 cases) for duration parsing
- Integrate into create.go with fallback to ISO format
Supports hours (h), days (d), weeks (w), months (m), and years (y).
Negative values allowed for past dates.
* feat(dates): add NLP parsing for natural language dates
Integrate olebedev/when library for natural language time expressions.
The layered parser now handles: compact duration → absolute formats → NLP.
Changes:
- Add olebedev/when dependency for NLP parsing
- Implement ParseNaturalLanguage and ParseRelativeTime functions
- Reorder layers: absolute formats before NLP to avoid misinterpretation
- Simplify create.go to use unified ParseRelativeTime
- Add comprehensive NLP test coverage (22 test cases)
Supports: tomorrow, next monday, in 3 days, 3 days ago
* feat(dates): add --defer flag to create/update/defer commands
Add time-based deferral support alongside existing status-based defer.
Issues can now be hidden from bd ready until a specific time.
Changes:
- Add --defer flag to bd create (sets defer_until on creation)
- Add --due and --defer flags to bd update (modify existing issues)
- Add --until flag to bd defer (combines status=deferred with defer_until)
- Add DueAt/DeferUntil fields to UpdateArgs in protocol.go
Supports: +1h, tomorrow, next monday, 2025-01-15
* feat(dates): add defer_until filtering to ready command
Add time-based deferral support to bd ready:
- Add --include-deferred flag to show issues with future defer_until
- Filter out issues where defer_until > now by default
- Update undefer to clear defer_until alongside status change
- Add IncludeDeferred to WorkFilter and RPC ReadyArgs
Part of GH#820: Relative Date Parsing (Phase 5)
* feat(dates): add polish and tests for relative date parsing
Add user-facing warnings when defer date is in the past to help catch
common mistakes. Expand help text with format examples and document
the olebedev/when September parsing quirk.
Tests:
- TestCreateSuite/WithDueAt, WithDeferUntil, WithBothDueAndDefer
- TestReadyWorkDeferUntil (ExcludesFutureDeferredByDefault, IncludeDeferredShowsAll)
Docs:
- CLAUDE.md quick reference updated with new flags
- Help text examples for --due, --defer on create/update
Closes: Phase 6 of beads-820-relative-dates spec
* feat(list): add time-based query filters for defer/due dates
Add --deferred, --defer-before, --defer-after, --due-before, --due-after,
and --overdue flags to bd list command. All date filters now support
relative time expressions (+6h, tomorrow, next monday) via the
timeparsing package.
Filters:
- --deferred: issues with defer_until set
- --defer-before/after: filter by defer_until date range
- --due-before/after: filter by due_at date range
- --overdue: due_at in past AND status != closed
Existing date filters (--created-after, etc.) now also support relative
time expressions through updated parseTimeFlag().
* build(nix): update vendorHash for olebedev/when dependency
The olebedev/when library was added for natural language date parsing
(GH#820). This changes go.sum, requiring an updated vendorHash in the
Nix flake configuration.
