Skip to content

QUEUE.md

Latest commit

 

History

History
58 lines (44 loc) · 2.49 KB

QUEUE.md

File metadata and controls

58 lines (44 loc) · 2.49 KB
Raw

Work Queue

P1 - Refresh history when the mobile app returns to the foreground

Problem

Mobile browsers can suspend the dashboard WebSocket while the share flow or a download runs in the background. The server saves the completed history item and its download timestamp, but the visible dashboard can remain stale until a manual refresh or a later WebSocket reconnect.

Scope

  • Refresh /youtube-dl/history when the page becomes visible again through visibilitychange or pageshow.
  • Refresh history after a WebSocket reconnect without resetting the selected view, filters, sort order, page, or detail item.
  • Debounce overlapping foreground and reconnect refreshes.
  • Keep the existing WebSocket completion update for active foreground sessions.

Acceptance Criteria

  • A download completed while the mobile browser is in the background appears with its download date when the user returns, without a manual refresh.
  • Foreground refreshes do not duplicate history rows or move the user away from the current filtered view unnecessarily.
  • Network failures leave the existing history visible and allow the next foreground or reconnect refresh to retry.
  • The behavior is covered by focused JavaScript or browser regression tests.

P1 - Prevent duplicate history for the same downloaded media

Problem

Submitting the same media repeatedly can create multiple completed history rows that point to the same existing file. Each row receives a new completion time even when yt-dlp reused the file instead of downloading it again.

Scope

  • Identify media by extractor plus media_id, with normalized source URL and final filepath as fallbacks when extractor metadata is unavailable.
  • When a completed record already points to the same existing file, update or reuse that record instead of appending another completed row.
  • Preserve the timestamp of the actual file download when yt-dlp skips an already-downloaded file.
  • Allow a genuinely new download after the previous physical file was deleted.
  • Return a clear Already downloaded result for repeated submissions where practical.

Acceptance Criteria

  • Repeating the same Instagram Reel, YouTube video, or other stable media ID does not create duplicate completed rows for one physical file.
  • The displayed download date represents the actual stored file, not the time of a skipped duplicate request.
  • Different media with the same title remain separate records.
  • Duplicate detection is covered by server regression tests.