ScreenHarbour 0.2.51 — faster return to Home

ScreenHarbour 0.2.51 — prerelease

Returning from movie details now reuses Home’s existing cards, seasonal scenery and channel artwork instead of rebuilding them. The selected item and scroll position are restored to the retained page.

Native rows that already contain items can refresh in the background without blanking Home. Custom collection and streaming-service rows are rebuilt only when their content or configuration changes. Initial loading still waits for the first row batch.

Seasonal timers, animation frames, lazy-background observers and resize work pause while Home is inactive. Returning resumes them around the restored selection, retaining loaded artwork and the row’s expanded state. Channel-logo lookup and image completion also wait to update the page until Home resumes.

Retained Home belongs to the same account and server; profile changes and sign-out do not reuse another user’s page. Saved settings, seasonal dates, row positions, shuffle behavior, expansion, themed headings, frames, doors, parallax and lazy-loaded TV images are preserved. No configuration changes are needed.

Update

Update ScreenHarbour, restart Jellyfin, then fully reload browsers or close and reopen web-based TV clients. Keep File Transformation enabled.

Jellyfin server	Plugin version
10.10.7	0.2.51.1
10.11.x	0.2.51.2
12.x	0.2.51.3

The existing catalogue remains:

https://raw.githubusercontent.com/jampez77/ScreenHarbour/main/manifest.json

Plugin identity, assembly name, existing API routes, storage keys and settings remain unchanged.

Validation

• TypeScript check, 254 unit tests and 801 server checks passed.
• 275 browser checks passed; 14 optional Featured/ElegantFin fixture checks were skipped because their external fixtures are unavailable. New regressions cover retained artwork, remote input during background reads, successive metadata refreshes, account cleanup and deferred collection-tab completion.
• All three target packages passed embedded client/artwork, assembly identity/version and checksum verification. The catalogue retains all 153 historical entries unchanged and selects the correct new build for each server version.
• A Chromium benchmark at 6× CPU slowdown, with 240 collection cards and two fullscreen Nightmare scenes, measured about 45% less main-thread work on warm return and no repeated seasonal-artwork requests. This is a browser measurement, not an LG timing guarantee.
• Physical LG webOS responsiveness still needs confirmation after installation.