While it's unlikely that more than 2 billion segments or byte ranges spanning 2GB (this one is more likely though...) are used, it would make sense to switch all these two u64 as it costs basically nothing, reduces the need to worry about negative numbers and would allow handling e.g. a playlist that works via byte ranges into a multi-GB file.
If you agree, I can prepare a PR for this. As it would be an API change, it would also make sense to decide about #38 before as that also involves in API change in either case.
While it's unlikely that more than 2 billion segments or byte ranges spanning 2GB (this one is more likely though...) are used, it would make sense to switch all these two
u64
as it costs basically nothing, reduces the need to worry about negative numbers and would allow handling e.g. a playlist that works via byte ranges into a multi-GB file.