v0.9.5

• Fix track language and name misreads on MKV files carrying LANGUAGE or TITLE tags. ffprobe lets those tags hide the real track headers, so affected tracks showed up as Undetermined and language fixes fell into an endless mkvpropedit-to-remux loop. Track metadata is now read from the mkv headers via mkvmerge, with the tag as fallback when the header has no language set (#55)
• Fix "Standardize track names" with an empty template not clearing track names; the clear never reached the conversion plan (#56)
• Show a warning in profile settings when a name template will clear all track names.
• Metadata tag keys in non-MKV containers are now matched case-insensitively.

After updating, run a forced rescan: files stuck on Undetermined only pick up their real language after a fresh scan.

Full Changelog: v0.9.4...v0.9.5

v0.9.4

• Add batch track editing: select files in the library, or everything a filter matches, and change track language, name and flags by position across all of them, with a per-file preview before anything is queued (#39, #50). On Matroska this stays on the in-place metadata path, so fixing the language on a whole show does not remux every file.
• Add a Tracks filter to the library: find files with no, one or multiple audio, subtitle or video tracks (#53)
• Add a {code} track name placeholder for short uppercase language codes: {code} {codec} {channels} names a track "FR AC-3 5.1" (#35)
• "Trim all tracks to video length" now also works on MP4 files, no longer Matroska only.
• Fix MP4 files never taking the "already optimized" skip, doing a pointless full remux when there was nothing to do.
• Fix conversions refused as uncheckable for truncation on files without duration tags whose runaway track ends minutes past everything else. The probe now walks back through the file until it finds the other tracks (#18)
• Fix lists showing an empty page when a filter shrinks the results, and land on the last page instead when rows are deleted. Thanks @stewieoO (#32)
• Show the size saved next to the reduction percentage, so sorting that column makes sense. Thanks @stewieoO (#28)
• Selected files now survive paging and the trip to the batch editor.
• Other small fixes and cleanup.

Full Changelog: v0.9.3...v0.9.4

v0.9.3

Run a force rescan after upgrading. The new options read data older scans never stored, so they do nothing until the files are re-probed.

• Add "Clear file title" profile option, removes the release group name shown as the file's title (#47)
• Add "Remove chapters" profile option (#31)
• Add "Trim all tracks to video length" profile option for files whose audio or subtitles run past the video. Matroska only, and it rides along on a conversion instead of starting one. It cuts real content, so leave it off unless you need it.
• Fix files keeping the profile they were first scanned with. This is why the library listed files as needing track removal when there was nothing to remove.
• Fix conversions rolled back as truncated when the dropped track was the longest one, like a subtitle running past the video (#18, #37)
• Recover the page when the connection to the server is gone, instead of leaving it dead until you close the tab (#48). It's usually a browser extension rewriting the page, which I can't do much about, but at least you're not stuck anymore.
• Skip login on the local network (#43)
• Forced subtitles no longer count as an alternative to SDH (#51)
• "Remove SDH / Accessibility" now also removes audio descriptions
• Show chapters, the file title and the duration next to the tracks, in the preview and the before/after
• Fix crash when scanning a file that is no longer on disk
• Other small fixes and cleanup.

Full Changelog: v0.9.2...v0.9.3

v0.9.2

• Added path mapping support (which I think should fix #40
• Added more preferences when purging multiple subs/audio's of the same language.
  Full Changelog: v0.9.1...v0.9.2

v0.9.1

• Improve migration speed of v0.9.0

Full Changelog: v0.9.0...v0.9.1

v0.9.0

⚠️ This release ships a database schema rewrite. I have done my absolute best to prevent dataloss and hopefully it will succeed in all edge cases!

• Refactor the conversion engine and normalize the database schema (MediaSnapshot / TrackSnapshot tables, ConversionPlan replaces the old JSON columns). Existing queued conversions are migrated automatically. This saves a lot of data and hassle in the future and makes the model more flexible for additions.
• Ignore trailers and sample files during scans and webhook intake (#33)
• Expose the application version on the stats API (#34)
• Track hardlinked files separately, with a library indicator and proper handling when manually queued (#41)
• Improve Sonarr/Radarr file matching so episodes with duplicate filenames map to the right show (#44)
• Add dub detection for audio tracks, plus a "Cancelled" conversion state and the ability to cancel an active conversion.
• Other small fixes and cleanup.

This took a while to release and was planned over 2 months ago, unfortunately I've had a few things going on in my personal life which made it hard to program.

v0.8.1

• Fix some notification providers, add some more.

Full Changelog: v0.8.0...v0.8.1

v0.8.0

• Add notifications (#8)
• Add support for multiple sonarr/radarr instances and change DB structure. (#15)
• Configure automatic scan interval (#16)
• Fix wrong impaired labeling (#17)
• Make conversion timeout configurable
• Other UX fixes / improvements.

• Change links to actual links by @stewieoO in #20

Full Changelog: v0.7.5...v0.8.0

v0.7.5

• Improved the conversions overview by a mile, this could het pretty chaotic with a lot of conversions at the same time.

Full Changelog: v0.7.4...v0.7.5

v0.7.4

• Add profile filter to media view when you have multiple profiles set up
• Ignore OS files that may clutter the media overview
• Add arm64 test runner
• Fix native symlink check on arm64 linux
• Small fixes & improvements to layout

Full Changelog: v0.7.3...v0.7.4