hugomossberg/Babel v2.5.7-beta

Babel v2.5.7-beta

on GitHub

What's New in v2.5.7-beta

This maintenance release resolves an extraction cache race condition under concurrent requests, improves source preparation visibility across the dashboard and library views, and adds targeted regression suites.

Bug Fixes

• Extraction Cache Handoff: Recheck the persistent subtitle cache after acquiring the extraction semaphore to prevent redundant physical extraction scans when concurrent requests target the same embedded subtitle track.
• Accurate Source Preparation Status: Distinguish media track inspection, subtitle resolving, and SDH cleaning from active translation. The UI now displays informative preparation stages (Preparing source: inspecting media, resolving subtitles, cleaning subtitles) instead of an inaccurate TRANSLATING 0% indicator.
• Active Job Library Transport: Propagate current_batch across library API endpoints and Alpine component helpers, ensuring season and movie badges reflect current source resolution state with null-safe bindings.

Improvements

• Lifecycle and Cache Regression Coverage: Added targeted regression test suites for concurrent extraction admission, cancellation safety, source preparation activity preservation across checkpoints, and Alpine template lifecycle state.

Upgrade Notes

• Upgrading requires no database schema migration. Existing configuration, provider API keys, and translation job histories are fully preserved.
• Standard Docker Compose upgrade:

  docker compose pull
  docker compose up -d

• Versioned image: ghcr.io/hugomossberg/babel:v2.5.7-beta
• Rolling beta channel: ghcr.io/hugomossberg/babel:beta
• To pin this version, reference babel:v2.5.7-beta in your Compose file. This beta release does not update the stable latest channel. Platform: Linux amd64.

Validation and Known Limitations

• 1,917 tests passed in Python test suite execution (including 34 targeted regression tests) with 0 failures, 0 errors, and 0 skips. Static analysis (ruff) and Alpine lifecycle contract checks passed cleanly.
• Verified on a production-like Docker media stack with live Sonarr media automation (Married at First Sight S20E02, 1,453 cues processed automatically with QA PASS 100/100, 0 dropped lines, 0 ms sync drift).
• The underlying hardware I/O variance causing prolonged extraction times on certain large MKV containers remains under investigation. Caching mitigates repeated extraction, but initial scan time depends on underlying storage throughput.
• All-caps dialogue observed in certain media sources (e.g. Bosch) is preserved by design in this release; sentence-case normalization rules remain open pending source file inspection.

Full Changelog: v2.5.6-beta...v2.5.7-beta