kajisho5/ffmpeg-skill v3.1.3

on GitHub

Automated release: version and notes generated from pull requests merged since 3.1.2.

Fixed

• A command-line argument carrying a line break or a control character (a crafted file name) reached
  error.message, error.hint and the diagnostics verbatim, where a newline could open what reads
  as a new line of instructions for the calling agent. Such an argument is now written escaped
  (\n, \x1b, \u202e) wherever a message prints it, as given, resolved or as a file name, and
  any C0/C1 control, line separator or bidi mark left in a message is escaped too. Ordinary paths,
  Windows backslashes included, and the line breaks a tool writes itself (an ffmpeg stderr tail)
  read as before (#400).
• Every writing tool, not only cut.py, kept one audio track of a source with several without
  saying so (fit, color, loudness, audio, overlay, ...: the shape of #397). emit() now compares
  the first input's audio tracks with the written file's and records an audio_tracks_dropped
  warning and a notes line when the output kept sound but fewer tracks; a run told to drop the sound
  (--no-audio, a picture-only output) reports nothing.
• grid.py (labels, the default) and scenes.py --sheet (timecode stamps) failed on a build
  without drawtext, as look.py did in #394. Both draw the picture without the text and record
  the fallback (grid_no_drawtext, scenes_sheet_no_drawtext) and a notes line.
• Escape control characters from arguments in messages; report dropped audio tracks from every tool; grid/scenes without drawtext (#405)

What's Changed

• Escape control characters from arguments in messages; report dropped audio tracks from every tool; grid/scenes without drawtext by @kajisho5 in #405

Full Changelog: v3.1.2...v3.1.3