In general, such requests to technical support fall upon deaf ears. I’ve never seen any example of any of the major camera manufacturers directly engaging with any FOSS project, with the slight example of the Sony wireless remote SDK back around when the QX family of cameras launched.
A lot of progress has been made on reverse engineering much of the format (fortunately, Canon used an MPEG-4 container which made things much simpler), but Canon is using a proprietary undocumented compression algorithm for the actual data itself, and reverse engineering this is EXTREMELY difficult.
There are rumors the libraw team may have an internal implementation that may get published as open source sometime in the coming months. I don’t think the Magic Lantern team has made any significant progress here either. (See ML for an example of a team that Canon could cooperate with in order to make their cameras more desirable - but instead Canon often does all they can to keep ML from tinkering with their cameras instead.)