Skip to content

Releases: opensensor/lightNVR

Release 0.36.7

Choose a tag to compare

@github-actions github-actions released this 03 Aug 18:08

🤖 AI-Generated Changelog

Fixed

  • Resolved a feedback loop on multi-view dashboards where multiple live view cells would repeatedly trigger shared go2rtc source restarts, causing all viewers to be disrupted simultaneously; retry now only reconnects the local browser player without restarting the shared stream source
  • Added a 30-second cooldown guard on stream refresh requests, coalescing duplicate refresh calls from multiple browser cells to prevent continuous stream disruption
  • Corrected recording trigger type assignment so that always-on (non-scheduled) recordings are now properly labeled as continuous instead of scheduled, fixing incorrect filtering of recordings in the recordings list
  • Fixed the manual recording button showing an incorrect tooltip when manual recording is unavailable due to continuous recording being configured; the tooltip now accurately explains the reason
  • Prevented manual recording from being started when a stream already has continuous recording configured, returning a clear conflict error instead of silently failing
  • Fixed the fullscreen sub-stream fallback logic to also respond to degraded WebRTC connection quality, not only to absence of video data

🔗 Full Changelog: 0.36.6...0.36.7

Release 0.36.6

Choose a tag to compare

@github-actions github-actions released this 30 Jul 21:53

🤖 AI-Generated Changelog

Fixed

  • Resolved recording backlog issue introduced in 0.36.5 that could cause recordings to fall behind or fail to process
  • Made recording metadata migration reversible to allow safe rollback if needed
  • Added section markers to database migration for improved reliability and traceability

🔗 Full Changelog: 0.36.5...0.36.6

Release 0.36.5

Choose a tag to compare

@github-actions github-actions released this 28 Jul 16:48

🤖 AI-Generated Changelog

Fixed

  • Resolved multiple backlog issues affecting live view, timeline, and API functionality (#463, #465, #466, #468)
  • Patched 3 high-severity security vulnerabilities in web dependencies

Changed

  • Updated German (de.json) translation
  • Merged upstream go2rtc improvements: RTSP IPv6 host bracketing support and api/reload endpoint
  • Improved build-from-source documentation with corrected instructions and documented prerequisites
  • Separated E2E tests from unit tests in the web test suite for more reliable npm test runs
  • Corrected stale paths and structure in web test documentation

Security

  • Upgraded sharp and other npm dependencies to address 3 high-severity advisories

🔗 Full Changelog: 0.36.4...0.36.5

Release 0.36.4

Choose a tag to compare

@github-actions github-actions released this 24 Jul 21:41

🤖 AI-Generated Changelog

Added

  • ONVIF smart-detection object classes (e.g., person, vehicle) are now surfaced in detection events
  • Live-view WebRTC connection timeouts are now runtime-configurable without restarting the service
  • Timeline detection legend and tooltips for clearer event identification
  • Link to the Home Assistant add-on repository in documentation

Changed

  • Detection segments on the timeline are now visually distinct for easier differentiation

Fixed

  • RTMP restream now correctly forces H264+AAC encoding so YouTube ingest accepts the stream
  • RTMP restream via go2rtc API now actually starts as expected
  • Recording stops correctly when the recording mode is switched to "No recording" at runtime
  • Mobile UX improvements for the timeline view

Security

  • Forced adm-zip to version ≥0.6.0 to resolve vulnerability GHSA-xcpc-8h2w-3j85
  • Removed unused symbols flagged by CodeQL to reduce attack surface

🔗 Full Changelog: 0.36.3...0.36.4

Release 0.36.3

Choose a tag to compare

@github-actions github-actions released this 18 Jul 11:32

🤖 AI-Generated Changelog

Added

  • RTMP restream support to YouTube and other RTMP endpoints directly from streams
  • Mobile timeline pinch-to-zoom gesture support for easier navigation on touch devices
  • Capacity-based retention policy to automatically manage storage based on disk usage limits
  • Full-disk self-healing to recover from storage-full conditions without manual intervention

Fixed

  • MQTT cleanup lifetime race condition that could cause crashes or undefined behavior on shutdown
  • ONVIF concurrency and parsing safety issues affecting reliability of camera discovery and communication
  • MQTT documentation payload mismatch where documented payloads did not match actual behavior

🔗 Full Changelog: 0.36.2...0.36.3

Release 0.36.2

Choose a tag to compare

@github-actions github-actions released this 06 Jul 05:41

🤖 AI-Generated Changelog

Added

  • Expose ONVIF day/night mode control in the live view interface
  • Include event snapshot images in MQTT detection payloads

Changed

  • Improved space utilization in live view with vertical scrolling support

Fixed

  • Stop ONVIF detection thread from busy-looping on cameras that do not support long polling
  • Align built-in motion detection confidence scoring with the threshold scale shown in the UI

🔗 Full Changelog: 0.36.1...0.36.2

Release 0.36.1

Choose a tag to compare

@github-actions github-actions released this 04 Jul 10:37

🤖 AI-Generated Changelog

Added

  • ONVIF PTZ discovery support, enabling automatic detection and configuration of pan/tilt/zoom capabilities for compatible cameras
  • ONVIF imaging day/night mode control, allowing management of camera day/night switching settings through the ONVIF interface

🔗 Full Changelog: 0.36.0...0.36.1

Release 0.36.0

Choose a tag to compare

@github-actions github-actions released this 03 Jul 19:41

🤖 AI-Generated Changelog

Fixed

  • Resolved a race condition that could occur during login, improving authentication reliability
  • Bounded stream connection attempts with a shared gate to prevent runaway reconnection loops
  • Improved mbedTLS build compatibility for environments where MBEDTLS_VERSION_C is not available

Changed

  • Updated all frontend dependencies to their latest versions
  • Improved Docker Compose configuration for better deployment experience
  • Hardened ffmpeg source validation in go2rtc (merged upstream dev improvements)

🔗 Full Changelog: 0.35.4...0.36.0

Release 0.35.4

Choose a tag to compare

@github-actions github-actions released this 24 Jun 06:46

🤖 AI-Generated Changelog

Added

  • In-process LiteRT/TFLite detection engine for running object detection without an external process
  • Detection memory usage now reported in system statistics
  • Detection grace period is now configurable and exposed in the settings UI
  • Detection URL field available in stream configuration UI
  • Detections are now recorded at frame-arrival time for improved accuracy
  • Optional per-stream detection URL storage and API support

Changed

  • Detection thread reworked with a unified architecture and a secondary detection stream for improved reliability

Fixed

  • Resolved use-after-free bug caused by config reload on settings save
  • Detection stream name no longer printed twice in logs
  • ONVIF SOAP responses with env:Body namespace prefix are now recognized correctly
  • CodeQL quality warnings in frontend JavaScript resolved
  • Build compatibility with GCC 15 (attribute-after-declarator error in XNNPACK)
  • Build compatibility with CMake 4.x via CMAKE_POLICY_VERSION_MINIMUM=3.5

🔗 Full Changelog: 0.35.3...0.35.4

Release 0.35.3

Choose a tag to compare

@github-actions github-actions released this 14 Jun 17:58

🤖 AI-Generated Changelog

Fixed

  • Fixed ONVIF camera discovery to improve detection reliability
  • Fixed MQTT publishing for built-in motion detection events

Changed

  • Improved Korean (한국어) translations and corrected Korean flag image

🔗 Full Changelog: 0.35.2...0.35.3