Read chapters from an input context and add chapters to an output context.
Note: unlike avformat_new_stream, the equivalent function for chapter is private:
avpriv_new_chapter (part of libavformat/internal.h). I couldn't find any other
solution but re-implementing it in format::context::output::add_chapter.
* Update bindings to newest ffmpeg version for new ffmpeg-sys, which is mostly generated by bindgen
* Bring back removed feature flags
* Fix whitespace formating
* Remove prepended enum names to enum variants
* Remove unneeded allows
