- Set CrontabSchedule timezone to system timezone for accurate scheduling
- Replace time TextInput with hour/minute Select dropdowns for cleaner UX
- Remove UTC/local time conversion logic (handled by Celery)
- Add tests for timezone functionality in simple and advanced modes
Updated create_backup to use the system's configured timezone for backup
filenames instead of always using UTC. This makes filenames more intuitive
and matches users' local time expectations.
Changes:
- Import pytz and CoreSettings
- Get system timezone from CoreSettings.get_system_time_zone()
- Convert current UTC time to system timezone for filename timestamp
- Fallback to UTC if timezone conversion fails
- Internal metadata timestamps remain UTC for consistency
Example:
- System timezone: America/New_York (EST)
- Created at 3:00 PM EST
- Old filename: dispatcharr-backup-2025.12.09.20.00.00.zip (UTC time)
- New filename: dispatcharr-backup-2025.12.09.15.00.00.zip (local time)
This aligns with the timezone-aware scheduling already implemented.
The list_backups function was creating timezone-naive datetime objects,
which caused the frontend to incorrectly interpret timestamps.
Now uses datetime.UTC when creating timestamps from file modification time
(consistent with other usage in this file on lines 186, 216), so the ISO
string includes timezone info (+00:00). This allows the browser to properly
convert UTC timestamps to the user's local timezone for display.
Before: Backend sends "2025-12-09T14:12:44" (ambiguous timezone)
After: Backend sends "2025-12-09T14:12:44+00:00" (explicit UTC)
The frontend's toLocaleString() will now correctly convert to local time.
Frontend changes:
- Add advanced mode toggle switch for cron expressions
- Show cron expression input with helpful examples when enabled
- Display format hints: "minute hour day month weekday"
- Provide common examples (daily, weekly, every 6 hours, etc.)
- Conditionally render simple or advanced scheduling UI
- Support switching between simple and advanced modes
Backend changes:
- Add cron_expression to schedule settings (SETTING_KEYS, DEFAULTS)
- Update get_schedule_settings to include cron_expression
- Update update_schedule_settings to handle cron_expression
- Extend _sync_periodic_task to parse and use cron expressions
- Parse 5-part cron format: minute hour day_of_month month_of_year day_of_week
- Create CrontabSchedule from cron expression or simple frequency
- Add validation and error handling for invalid cron expressions
This addresses maintainer feedback for "custom scheduler (cron style) for more control".
Users can now schedule backups with full cron flexibility beyond daily/weekly.
Dramatically improve EPG refresh performance by parsing the XML file once
per source instead of once per channel. The new implementation:
- Pre-filters to only process EPG entries mapped to actual channels
- Parses the entire XML file in a single pass
- Uses O(1) set lookups to skip unmapped channel programmes
- Skips non-mapped channels entirely with minimal overhead
For EPG sources with many channels but few mapped (e.g., 10,000 channels
with 100 mapped to channels), this provides approximately:
- 99% reduction in file open operations
- 99% reduction in XML file scans
- Proportional reduction in CPU and I/O overhead
The parse_programs_for_tvg_id() function is retained for single-channel
use cases (e.g., when a new channel is mapped via signals).
Fixes inefficient repeated file parsing that was occurring with large
EPG sources.
- Simplify to database-only backups (remove data directory backup)
- Update UI to match app styling patterns:
- Use ActionIcon with transparent variant for table actions
- Match icon/color conventions (SquareMinus/red.9, RotateCcw/yellow.5, Download/blue.5)
- Use standard button bar layout with Paper/Box/Flex
- Green "Create Backup" button matching "Add" pattern
- Remove Card wrapper, Alert, and Divider for cleaner layout
- Update to Mantine v8 Table syntax
- Use standard ConfirmationDialog (remove unused color prop)
- Update tests to remove get_data_dirs references
- Added m3u_account__name to backend ordering_fields in StreamViewSet
- Implemented field mapping in frontend to convert column IDs to backend field names
- Added sort buttons to both Group and M3U columns with proper icons
- Sort buttons show current sort state (ascending/descending/none)
- Maintains consistent UX with existing Name column sorting
XC (Xtream Codes) clients require integer channel numbers and fail to parse
float values (e.g., 100.5). This change implements collision-free mapping
that converts floats to integers while preserving existing integer channel
numbers where possible.
Changes:
- Implemented two-pass collision detection algorithm that assigns integers
to float channel numbers by incrementing until an unused number is found
- Applied mapping to all XC client interfaces: live streams API, EPG API,
and XMLTV endpoint
- Detection: XC clients identified by authenticated user (user is not None)
- Regular M3U/EPG clients (user is None) continue to receive float channel
numbers without modification
Example: Channels 100, 100.5, 100.9, 101 become 100, 101, 102, 103 for XC
clients, ensuring no duplicate channel numbers and full compatibility.
System Event Logging:
- Add SystemEvent model with 15 event types tracking channel operations, client connections, M3U/EPG activities, and buffering events
- Log detailed metrics for M3U/EPG refresh operations (streams/programs created/updated/deleted)
- Track M3U/EPG downloads with client information (IP address, user agent, profile, channel count)
- Record channel lifecycle events (start, stop, reconnect) with stream and client details
- Monitor client connections/disconnections and buffering events with stream metadata
Event Viewer UI:
- Add SystemEvents component with real-time updates via WebSocket
- Implement pagination, filtering by event type, and configurable auto-refresh
- Display events with color-coded badges and type-specific icons
- Integrate event viewer into Stats page with modal display
- Add event management settings (retention period, refresh rate)
M3U/EPG Endpoint Optimizations:
- Implement content caching with 5-minute TTL to reduce duplicate processing
- Add client-based event deduplication (2-second window) using IP and user agent hashing
- Support HEAD requests for efficient preflight checks
- Cache streamed EPG responses while maintaining streaming behavior for first request
The ChannelSerializer.to_representation() method was not respecting the
ChannelStream.order field when serializing PATCH/PUT responses. This
caused streams to be returned in an arbitrary order rather than the
order specified in the request.
The update() method correctly saves the stream order to the database
using the ChannelStream.order field, and GET requests (with
include_streams=True) correctly return ordered streams via
get_streams(). However, standard PATCH/PUT responses were using
PrimaryKeyRelatedField which doesn't respect the ordering.
This fix ensures that all representations (GET, PATCH, PUT) return
streams ordered by the channelstream__order field.
Impact:
- PATCH/PUT responses now correctly reflect the stream order saved
- Clients can trust the response data without needing a follow-up GET
- No breaking changes - only fixes inconsistent behavior
Tested with:
- PATCH request with ordered stream IDs
- Verified response matches request order
- Verified GET request confirms order persisted to database
