v0.3.55

Removed

• Disabled the subtitle-prestrip pre-pass (the I/O-bound -c copy remux that ran before encoding files with 6+ subs to drop). Users were noticing ~30 s–1 min of dead air before the encode actually started on multi-sub WEB-DLs; question was whether it still earned its keep. Re-tracing the history: the prestrip was added in v0.3.43–v0.3.44 because files with many unmapped sub streams looked like they encoded at ~1× instead of ~5×, with the progress bar pinned. The actual fix landed in those same versions on the parser side — parse_ffmpeg_progress now falls back to frame=N / total_frames when ffmpeg's time= field stalls or reads N/A. Since ffmpeg's speed= is computed as time / wall_clock, the "1×" measurement that justified the prestrip was itself reading the stale time= value the parser fix was about to address — i.e. measurement artefact, not a real encoder slowdown. With the parser bug gone, the prestrip is paying ~30 s–1 min of I/O + a full input-size temp write to fix a problem that was never on ffmpeg's side. Threshold raised from 6 to 9999 (effectively off); _prestrip_subtitles and the call block stay so a single-line revert can re-enable it if a real encoder slowdown does materialise.

Added

• Completed-job report now populates for skipped_larger outcomes (encode finished but was larger than source, original was kept). Previously the early-return at space_saved < 0 && !had_track_removal returned only {success, output_path, space_saved, error, skipped_larger} — no encoding_stats, no ffmpeg_command, no ffmpeg_log — so when the user expanded one of those rows the comparison table was empty. Now mirrors the success-path payload so the same Original-vs-Encoded grid renders, with the encoded-size cell showing (N% larger — discarded) in amber instead of (N% saved) in green when the ratio is negative. Lets the user see which CQ/preset/bitrate produced an unsaving result and tune their threshold rules accordingly, which is the exact use case they asked for.

Docker images

Pinned to this release:

docker pull ghcr.io/i-ial9000/shrinkerr:0.3.55         # portable (amd64 + arm64, CPU)
docker pull ghcr.io/i-ial9000/shrinkerr:0.3.55-nvenc   # x86_64 + NVIDIA GPU (NVENC)
docker pull ghcr.io/i-ial9000/shrinkerr:0.3.55-edge    # as above but ffmpeg master
docker pull ghcr.io/i-ial9000/shrinkerr:0.3.55-edge-nvenc

Floating tags (:latest, :edge, :nvenc, :edge-nvenc) track the most recent release of each lineage.