Add test coverage for new BLEInterface cleanup functionality:
- _cleanup_pending_identity_connections: timeout handling for non-Reticulum devices
- _process_pending_detaches: delayed interface detachment with grace period
- _validate_spawned_interfaces: orphaned interface cleanup
- Multi-address disconnect handling (keeping interface alive for MAC rotation)
- Thread-safe _spawn_peer_interface with locking
- Central disconnect callback with address fallback
These tests cover the 195 lines of new code in BLEInterface.py to improve
patch coverage for PR #35.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Fix mock_ble_driver.py import path (src/RNS/Interfaces -> src/ble_reticulum)
- Add address_to_interface, _pending_detach, _pending_detach_grace_period to test fixture
- Update test_identity_cache to expect grace period behavior (deferred cleanup)
- Update test_v2_2_mac_sorting to use renamed _cleanup_stale_address function
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Rewrite _validate_spawned_interfaces() with 3-pass approach:
- Pass 1: Collect orphaned addresses
- Pass 2: Clean up address mappings, track interfaces to detach
- Pass 3: Only detach interfaces with zero connected addresses
- Fragmenters only cleaned up when interface fully detached
- Enhance _spawn_peer_interface() reuse logic:
- Update address_to_identity and identity_to_address when reusing
- Cancel pending detach for the identity
- Mark interface as online
- Fix disconnect callbacks to preserve fragmenters:
- _device_disconnected_callback: defer fragmenter cleanup to grace period
- handle_central_disconnected: same fragmenter preservation
- _process_pending_detaches: clean up fragmenters on actual detach
- Rename _cleanup_stale_interface() to _cleanup_stale_address():
- No longer detaches interface during MAC rotation
- Only cleans up stale address-specific mappings
- Interface preserved for reuse with new address
Fixes orphaned peer interfaces and "No fragmenter for peer" warnings
during BLE MAC address rotation.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Connections from non-Reticulum BLE devices (AirTags, BLE scanners, etc.)
that connect to our GATT server but never complete the identity handshake
are now automatically disconnected after 30 seconds.
Changes:
- Track pending identity connections with timestamps in _pending_identity_connections
- Add _cleanup_pending_identity_connections() to disconnect stale connections
- Remove from pending tracking when identity is provided in callback
- Add debug logging for cleanup timer operations
This prevents non-protocol devices from appearing indefinitely in the
BLE connections list.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
BLE peer interfaces weren't being cleaned up when connections dropped
if the identity-to-address mapping wasn't available at disconnect time.
This caused orphaned interfaces to persist (peer interfaces shown with
zero active connections).
Changes:
- Add address_to_interface mapping for direct address-based cleanup
- Update _device_disconnected_callback with dual-index approach:
try identity lookup first, fall back to address_to_interface
- Update handle_central_disconnected with same dual-index approach
- Add _validate_spawned_interfaces() periodic validation (every 30s)
that cross-checks interfaces against driver.connected_peers
- Update _cleanup_stale_interface and _address_changed_callback to
maintain the new mapping
- Clear address_to_interface on detach()
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The previous patch changed RNS.log in the test's namespace, but
BLEInterface.py imports RNS at module load time. To capture log
calls from _address_changed_callback(), we need to patch RNS.log
where it's used: ble_reticulum.BLEInterface.RNS.log
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The test was using mock_rns.log.assert_called() but the real
BLEInterface code calls the actual RNS.log function, not the mock.
Fixed by patching RNS.log at module level to capture actual log calls,
then asserting the "no identity found" warning was logged.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add 14 tests that exercise the REAL BLEInterface code:
- TestIdentityCacheOnDisconnect: verify caching on disconnect
- TestIdentityCacheOnDataReceive: verify cache restoration on data
- TestAddressChangedCallback: verify address migration
- TestCacheTTL: verify 60-second TTL behavior
- TestReassemblyCodePath: verify cache in reassembly path
- TestEdgeCases: concurrent access, multiple disconnects
Tests skip locally if RNS not installed but run in CI to provide
actual line coverage on the identity cache changes.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add TestIdentityCacheIntegration class that imports and tests actual
BLEInterface methods instead of just mocking the logic. This should
provide codecov coverage on the changed lines.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add CODECOV_TOKEN to codecov-action uploads (required for v4)
- Change unit test coverage from single file to entire package
- Add codecov.yml with coverage thresholds and flags
Note: CODECOV_TOKEN secret must be added to repository settings.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
When Python's disconnect callback fires but the driver layer (Android/Kotlin)
maintains or quickly re-establishes the GATT connection, data was being
dropped because address_to_identity was cleared.
Changes:
- Add _identity_cache with 60-second TTL to preserve identities after disconnect
- Cache identity in _device_disconnected_callback before cleanup
- Check cache in _handle_ble_data and restore identity if found
- Add on_address_changed callback for dual connection deduplication
- Add _address_changed_callback to migrate identity mappings
- Support driver.request_identity_resync() for fallback recovery
This fixes the "no identity for peer X, dropping data" warning that occurred
when the Python layer lost track of a peer that was still connected at the
driver level.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Update install.sh to copy from src/ble_reticulum/
- Update test files with new source paths
- Update GitHub workflows for new package structure
- Remove temporary refactoring helper scripts
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The sed replacement was too aggressive - it replaced the import for
the base Interface class from the Reticulum package itself.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Fixes namespace collision with Reticulum's own RNS.Interfaces package.
When both packages were installed, the collision caused import issues
and prevented BLE discovery between devices.
Changes:
- Rename src/RNS/Interfaces/ to src/ble_reticulum/
- Update pyproject.toml package configuration
- Update all imports in source and test files
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The peer_identity parameter is already the identity hash received from
the BLE handshake. Calling RNS.Identity.full_hash() on it again produced
a completely different value, causing identity mismatch between peers.
This caused "no reassembler for X" errors because the sending peer's
identity didn't match what the receiving peer computed.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
* fix(ble): Increase D-Bus monitoring intervals to prevent HCI errors
The D-Bus monitoring threads were polling too frequently (0.5s and 30s),
causing HCI command collisions on BCM43xx single-radio chips. These chips
cannot handle concurrent BLE operations, and the frequent D-Bus activity
was interfering with scan/advertise cycles.
Changes:
- Increase D-Bus disconnect monitor interval from 0.5s to 5s
- Increase stale connection poll interval from 30s to 120s
This eliminates HCI errors (Opcode 0x2005/0x2006) while preserving
disconnect detection functionality with slightly higher latency.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
* refactor(ble): Convert D-Bus monitoring to event-driven approach
Replace polling-based D-Bus monitoring with true event-driven pattern:
1. D-Bus monitor thread:
- Use asyncio.Event instead of periodic sleep
- Store loop reference for thread-safe shutdown
- Use call_soon_threadsafe to wake loop on stop
2. Stale poll thread:
- Replace busy-wait loop (240 x 0.5s) with single Event.wait()
- Increase interval from 120s to 300s (safety net only)
- Immediate response to stop signal
Benefits:
- Zero CPU usage while waiting (no periodic wakeups)
- Immediate shutdown response (ms instead of 5s)
- Cleaner, simpler code
- Maintains disconnect detection via D-Bus signals
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
* test(ble): Add comprehensive unit tests for HCI error fixes
Add 26 new unit tests for the event-driven D-Bus monitoring fixes
that eliminated HCI errors on BCM43xx single-radio chips.
Test coverage:
- TestEventDrivenDBusMonitor: Tests asyncio.Event usage, immediate
wake response, call_soon_threadsafe cross-thread signaling
- TestStalePollImprovements: Tests threading.Event.wait() usage,
300s interval, immediate stop response
- TestStopShutdownBehavior: Tests stop() async signaling, RuntimeError
handling, shutdown latency improvement
- TestIntegrationScenarios: Tests full lifecycle, multiple stop calls
- TestCodeVerification: Verifies actual source patterns match expected
All 26 tests pass without requiring pytest-asyncio plugin.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
* fix(tests): Use threading.RLock instead of asyncio.Lock in test fixtures
In Python 3.8/3.9, asyncio.Lock() requires a running event loop. When
test_hci_error_fixes.py runs first (alphabetically) and uses asyncio.run(),
it closes the event loop after each test. Subsequent test fixtures that
create asyncio.Lock() then fail with "no current event loop" errors.
Since these are mock fixtures that don't need async semantics, use
threading.RLock() instead.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
* fix(tests): Replace all asyncio.Lock() with threading.RLock() in test mocks
asyncio.Lock() requires a running event loop in Python 3.8/3.9. When
test files using asyncio.run() execute first, the event loop is closed,
causing subsequent test fixtures to fail when creating asyncio.Lock().
Fixed in:
- test_peripheral_disconnect_cleanup.py (mock_gatt_server fixture)
- test_bluez_state_cleanup.py (mock_driver fixture)
- test_ble_peer_interface.py (create_mock_peer_interface helper)
- conftest.py (create_mock_interface helper)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
---------
Co-authored-by: torlando-tech <torlando-tech@users.noreply.github.com>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
The _cleanup_stale_interface() method was cleaning up identity_to_address
but not the reverse mapping address_to_identity. This caused:
- Memory leak: stale entries accumulate over time
- Inconsistent state: bidirectional mappings become out of sync
This fix matches the cleanup pattern in _disconnected_callback() which
properly cleans up both directions of the mapping.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
pip extracts package metadata from the wheel filename, which must follow
PEP 427 format: {package}-{version}-{pythontag}-{abi}-{platform}.whl
The previous filename (dbus_fast-armv6l-$$.whl) caused pip to fail with
"Invalid wheel filename (wrong number of parts)" on 32-bit ARM systems.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Detect and handle RNS installations via `uv tool install rns`:
- Add uv detection when python3 can import RNS (path contains /uv/tools/)
- Add shebang-based detection when system python3 differs from tool's python
- Install dependencies using `uv pip install --python`
- Handle uv Python path for setcap Bluetooth permissions
This fixes "Could not determine installation mode" errors for users
who install Reticulum with uv instead of pip/pipx.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Resolved conflicts:
- BLEInterface.py, test_v2_2_mac_sorting.py, CLAUDE.md: Keep release
(HW_MTU fix, pending_mtu, MAC rotation recovery/bypass, corrected tests)
- install.sh, README.md: Keep main (JustWorksRepairing auto-configuration)
- CHANGELOG.md: Consolidated detailed notes from main into [0.2.2] section
- Added FUNDING.yml from main
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add TestMACRotationBypassesSorting class with 4 tests for the MAC rotation fix
- Fix MockInterface to properly mock RNS.Interfaces.Interface module
- Add get_config_obj() to MockInterface for BLEInterface initialization
- Fix inverted test expectations in test_sequential_mac_addresses and
test_mac_sorting_with_multiple_peers (lower MAC initiates connection)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
After detecting MAC rotation and cleaning up the stale interface,
immediately add the peer to scored_peers and continue, bypassing
the MAC sorting check. This ensures reconnection always happens
after MAC rotation regardless of which device has the higher MAC.
Bug: After MAC rotation, peer interface wasn't recreated because
the code fell through to MAC sorting check which could skip the
peer if local MAC > peer MAC.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Enable JustWorksRepairing = always for automatic BLE pairing
- Disable GATT caching (Cache = no) to prevent Service Changed
subscription pairing prompts on Android 15+ devices
- Track config changes and only restart BlueZ when needed
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Reorder operations in handle_peripheral_data() to create
fragmenter/reassembler BEFORE spawning peer interface. This
prevents data from being dropped during the brief window when
the interface exists but the reassembler doesn't.
Also adds unit tests to verify the fix and prevent regression.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Reorder operations in handle_peripheral_data() to create
fragmenter/reassembler BEFORE spawning peer interface. This
prevents data from being dropped during the brief window when
the interface exists but the reassembler doesn't.
Also adds unit tests to verify the fix and prevent regression.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
When a peer reconnects from a new MAC address (common with Android MAC
rotation, also possible on Linux after adapter reset), the old interface
state may become stale. This change adds:
1. _cleanup_stale_interface() method:
- Detaches old interface
- Cleans up identity mappings
- Removes fragmenter/reassembler for old address
- Clears pending MTU for old address
2. MAC rotation detection in _select_peers_to_connect():
- If interface exists for identity but old address is not in peers
(connection dead), clean up stale state and allow reconnection
- If old connection still active, skip as before (no change)
This is defensive code that handles edge cases like:
- Android MAC rotation (~15 min intervals)
- Linux adapter reset while peer state is tracked
- Connection timeout without proper disconnect callback
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Two critical fixes ported from Android testing:
1. HW_MTU instance attribute fix:
- Base Interface.__init__() sets self.HW_MTU = None
- BLEInterface.HW_MTU = 500 is a CLASS attribute, not instance
- After super().__init__(), self.HW_MTU is None (instance shadows class)
- BLEPeerInterface copies: self.HW_MTU = parent.HW_MTU (gets None)
- When HW_MTU is None, Transport.py truncates packet.data incorrectly
- Result: Link establishment fails with mismatched link_id
- Fix: Explicitly set self.HW_MTU = BLEInterface.HW_MTU after super().__init__()
2. MTU/identity race condition fix:
- MTU negotiation can complete before identity is received
- Previously: Warning logged, fragmenter not created
- Now: Store pending MTU, create fragmenter when identity arrives
- Adds pending_mtu dict to track deferred MTU values
- _device_connected_callback checks for pending MTU after storing identity
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Update CHANGELOG.md with release notes for v0.2.2
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Removed .github/workflows/deploy.yaml which was a byte-for-byte duplicate
of deploy.yml. Both workflows were executing independently, causing:
- Duplicate deployments to Raspberry Pi devices
- Potential race conditions from concurrent SSH sessions
- Unnecessary resource usage
Kept deploy.yml as it uses the more standard extension in the GitHub
Actions ecosystem.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Removed .github/workflows/deploy.yaml which was a byte-for-byte duplicate
of deploy.yml. Both workflows were executing independently, causing:
- Duplicate deployments to Raspberry Pi devices
- Potential race conditions from concurrent SSH sessions
- Unnecessary resource usage
Kept deploy.yml as it uses the more standard extension in the GitHub
Actions ecosystem.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Adds Step 5C to install.sh that automatically configures BlueZ's
JustWorksRepairing setting to "always" in /etc/bluetooth/main.conf.
This enables automatic pairing for peer-initiated BLE connections,
which is required for zero-touch mesh networking.
Changes:
- install.sh: Add Step 5C to configure JustWorksRepairing
- Checks current setting and only modifies if needed
- Handles both root and non-root execution
- Restarts bluetooth service to apply changes
- Container-friendly with non-fatal restart operations
- README.md: Document JustWorksRepairing in manual installation steps
- README.md: Add troubleshooting section for JustWorksRepairing warning
Fixes#13🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
