RC RANDOM CHAOS

Jellyfin 12.0 Drops the '10', Rebuilds the Database, and Finally Fixes Books

· via Hacker News

Original source

Jellyfin 12.0

Hacker News →

Jellyfin’s new stable release abandons the confusing ‘10.x’ version scheme—what would have been 10.12 is simply 12.0—because the leading ‘10’ never changed and made genuinely major releases (like the database-rewriting 10.11) look like minor updates. Anyone parsing Jellyfin version strings in clients, monitoring, or deployment scripts should account for the new format before upgrading. The release builds on 10.11’s database overhaul with real performance gains: playlists and collections now store each item as its own row instead of one opaque blob, so counting, paging, and single-item edits no longer require loading the entire list—eliminating the freezes that plagued large libraries.

The upgrade is invasive and demands preparation. The schema changes and data is rewritten on first boot, making a full manual backup the only rollback path. Users must be on 10.10.7 or any 10.11.x first, usernames are now case-insensitive (a migration will fail on collisions), and a mandatory full library scan follows the upgrade—it will run long and may surface movies as ‘newly added’ as Jellyfin reconciles every item against files on disk. Third-party plugins built for 10.11 won’t load and should be removed beforehand, and legacy /emby/ and /mediabrowser/ endpoints plus the old sign-in method are gone, breaking very old clients. The release also ships unspecified security fixes, so upgrading is recommended.

On features, alternate versions—previously movies-only—now work for episodes, letting broadcast and extended cuts or 1080p/4K copies group together with version-aware resume data; this is the change driving the required rescan. Book and comic support has been pulled from the Bookshelf plugin into the server itself, with native metadata reading from OPF, ComicInfo, and ComicBookInfo, poster generation for EPUBs and comic archives, filename-based parsing, page-count extraction, and audiobook chapters. Bookshelf’s metadata sources have been split into separate GoogleBooks, ComicVine, and new OpenLibrary providers.

Read the full article

Continue reading at Hacker News →

This is an AI-generated summary. Read the original for the full story.