chore(release): version packages

[github-actions]

This PR was opened by the Changesets release GitHub action. When you're ready to do a release, you can merge this and the packages will be published to npm automatically. If you're not ready to do a release yet, that's fine, whenever you add more changesets to main, this PR will be updated.

Releases

ffmpeg_audio@0.4.0

Minor Changes

• 6d7e7dc Thanks @apoint123! - Removed the public TimeBase type and the core::time module. Timestamp handling is now internal to the crate.

• 6d7e7dc Thanks @apoint123! - Changed scan_exact_duration to return the exact duration measured from zero: the time right after the last sample, rather than the span between the earliest and the latest timestamp found while scanning.

• #13 0dff143 Thanks @apoint123! - Updated bundled FFmpeg from 8.1.2 to 9.0.2 (libavcodec 63, libavformat 63, libavutil 61, libswresample 7). APIs removed upstream in this major release (e.g. AVCodecContext.properties, the private fields of AVCodecParser, AVTimebaseSource) are no longer available in the generated bindings.

Patch Changes

• 6d7e7dc Thanks @apoint123! - Documented that SeekMode::Coarse and SeekMode::Accurate start at the nearest reachable position when the container cannot reach positions before the target, and that AudioReader::duration returns the duration declared by the container, which is an estimate.

• 6d7e7dc Thanks @apoint123! - Changed timestamps and durations (AudioFrame::pts, AudioFrame::duration, stream_position, scan_exact_duration) to be computed exactly and rounded down to whole nanoseconds, instead of being rounded to microseconds at every step. Seeking to a reported timestamp now lands on the same sample again.

• 6d7e7dc Thanks @apoint123! - Fixed accurate seeking discarding the preroll trim of the frame it lands on, which could deliver preroll samples when seeking close to the start of a stream with negative timestamps.

• 6d7e7dc Thanks @apoint123! - Fixed accurate seeking trimming one sample too many when a frame timestamp is not a whole number of microseconds (e.g. seeking to 100 ms in a 48 kHz AAC stream now starts exactly at 100 ms instead of 100.02 ms).

• 6676910 Thanks @apoint123! - Fixed scan_exact_duration disturbing reading: after a coarse seek it moved reading back to the start, and mid-stream it could drop samples (e.g. 16 samples in Matroska files, whose timestamps are rounded to milliseconds) or shift frame boundaries. Reading now continues with exactly the frames it would have delivered, and stream_position keeps its value until the next frame.

• 6d7e7dc Thanks @apoint123! - Fixed scan_exact_duration under-reporting the duration of streams whose first samples cannot be reached by seeking (e.g. 892.666 ms instead of 899.667 ms for a Matroska file with negative timestamps), and reporting a 2 s WAV file as 1.999999 s.

• Updated dependencies [2876aba, 2876aba]:

  • ffmpeg_audio_sys@0.2.0

ffmpeg_audio_sys@0.2.0

Minor Changes

• #13 0dff143 Thanks @apoint123! - Updated bundled FFmpeg from 8.1.2 to 9.0.2 (libavcodec 63, libavformat 63, libavutil 61, libswresample 7). APIs removed upstream in this major release (e.g. AVCodecContext.properties, the private fields of AVCodecParser, AVTimebaseSource) are no longer available in the generated bindings.

Patch Changes

• f31bdec Thanks @apoint123! - Fixed the build script reusing FFmpeg sources extracted from a previous version of the bundled archives, so incremental builds could silently keep compiling the old FFmpeg after the archives were updated.