While "max_connections: 0" was correctly allocated as unlimited, the session reservation could still lock that same provider for other sessions. This caused the path to erroneously fall into the generic "exhausted" handling later on. (#677)
* Sorce ordinal by category and channel index
* Parsing UserApiRequest from request body or query
* CSS stylings
* media tools refactoring
* xtream info parsing now number/string fields
* fix: Scheduler preserve forced_targets across config reloads.
hot reload could cause loosing scheduler task targets.
warn logs for:
- process_source: warns when all inputs are disabled for given targets
- process_sources: warns when no sources were processed at all
* Fix: Logger initialization
* Fix: TCP USER TIMEOUT for linux. If a client disrupts the connection, the tcp layer will hold it approx for 5-15 minutes. With this setting we can close in 30 seconds under linux
- Fix: resolve task database problem
- Fix: m3u parsing when staged input delivers virtual id
- provider id extraction from url refactored
- some optimizations for hashutils
* Fix HLS session handling and disable retry for adaptive segment requests
* Resolve / probe vod task uses last modified date
* Resolve expiration date in input form
* fixed m3u parsing
* fixed wasm opt for frontend build
* Fixed issues with probe and resolve tasks
* hls and catchup session fixes
* fixed problem with grace and shared stream
- added fix for missing episode title
* keep user table info in frontend when switched to custom video stream
* Fixed problems caused by yew upgrade because of UseStateHandle and UseReducerHandle PartialEq change.
* reformatted
* - Refactored Shared-Stream hot path from Arc<Bytes> to Bytes (leveraging internal reference counting).
- Switched Metadata Trigger deduplication from Arc<Mutex<HashSet<_>>> to Arc<DashSet<_>> to reduce lock contention.
- Eliminated expensive key cloning in the InputMetadataUpdatesCompleted handler.
- Changed active_target_inputs to HashSet<Arc<str>> to avoid per-event to_string() allocations.
- Implemented bounded Cleanup-Queue (switched from unbounded to mpsc::channel(4096)) with a dedicated overflow path (try_send + async flush) to prevent uncontrolled memory growth.
- Optimized by_provider storage by switching from Vec<(addr, alloc_id)> to HashSet<(addr, alloc_id)>, reducing removal complexity from $O(n)$ to $O(1)$ average.
- Optimized ActiveConnectionInfo data layout by reordering fields (large fields first) to minimize padding and improve cache efficiency.
* Improve shared-stream reliability and align docs
- make shared connection registration fallible with rollback on failure
- guard shared broadcaster startup behind successful subscription
- optimize shared stream/provider key maps to Arc<str>
- add Tokio macros feature for runtime and dev dependencies
- harden build_resources.sh with fail-fast ffmpeg error handling
- clarify StreamError display messages with variant context
- sync README stream/custom-response sections with current behavior
* low_priority_preempted stream
* Keep custom TS timestamps monotonic across loops
- advance timestamp offset by stream duration on buffer wrap
- avoid resetting PTS/DTS to zero each cycle
- prevents players from treating looped custom streams as ended/corrupt
* low_priority_preemption custom video stream fix
* inc_version uses now main toml file
* Fixed some coderabbit issues
* Optimized build
* timeout secs for custom video response
* init custom video timeout on mode switch
* terminate immediately when custom video buffer is missing
* add provisioning missing-custom timeout regression test
* add missing-custom termination tests for all custom modes
* fix ts continuity handling for discontinuity and adaptation-only packets
* fix ts loop duration estimation to use min/max timestamps
* harden pes timestamp patching and normalize pid cc sequence
* Resource ffmpeg creation without B-Frames and short GOPs
* Discontinuity patch for transport stream buffer
* grace hold stream is now default true
* streaming fixes
* user connection priority over all aliases
* active client stream priority preemption
* Fixed paths override in read config
* Added custom video stream timeout settings
* Scheduler fix
* No template file not found error message
* Setup wizard improvmenets
* path improvements
* staged input fix
* path improvements
* Provider dns resolved ips now writing to a file instead of source.yml
* Playlist update hard kill after 1h
* matadata update tmdb fix
* Web UI fixes
* GeoIp update schedule
* batch:// prefix for aliases
* fix: preserve provider dns.resolved order by provider URL sequence
* fix: preserve provider dns data on web ui source saves
* fix: keep provider source edits and dns resolved on ui save
* fix: preserve provider URLs during panel-api provisioning
* fix: prevent lost updates with config revision checks
* fix: honor disabled dns and track config context in input form
New Features
Background probing for Live streams with configurable interval; new input options to trigger probing and toggle background resolution.
Configuration Changes
metadata_update reorganized into grouped sections: log, resolve, probe, ffprobe, tmdb.
FFprobe controls moved under ffprobe.* and now require explicit time/size units; new live ffprobe settings.
Added metadata_update.tmdb.cooldown (default 7d) and grouped TMDB settings.
Bug Fixes
TMDB no-match cooldown to prevent endless requeues.
Documentation
README and UI help updated to reflect grouped metadata_update fields and explanations.
Was allocating the right XC account but sometimes still storing/reusing the old base stream URL (often the first batch account) for VOD/series sessions, especially around seek/reconnect. (#605)
That caused provider/account mismatch and possible single-connection resets.
