This is the interface description. It's not the .c file that implements the abstraction. There is lots of code behind this, redundantly implementing the calls using either Win32 or POSIX.
That's not a knock on sqlite, I consider it well implemented and one of my favorite libraries I've ever seen, used and studied.
sed -e 's-/-\\-g' -e 's-^-C:-'This is in part because Windows and POSIX are relatives. Windows is (partly) descended from DOS, and DOS 2 deliberately borrowed a lot of ideas from Xenix (Microsoft’s Unix port). And Windows has continued to borrow ideas from POSIX since, e.g. Winsock is heavily based on Berkeley Sockets.
Compare DOS 2+ handle-based IO to CP/M: CP/M treats files as fixed length 128 byte records instead of bytes, you can only read or write a whole record at a time, and a file’s size on disk must be a whole multiple of 128 [0] - that has more in common with mainframe/minicomputer record-oriented IO than Unix-style byte-oriented filesystems
[0] later on, CP/M added an attribute in the directory entry to store how many bytes were valid in the last record of a file; but it wasn’t actually enforced by the filesystem, apps had to set/get the attribute using a separate API, a lot of apps just ignored it, and even for those who did try to use it, two completely opposite conventions of how to use it coexisted
Yeah, but that is the abstraction. The associated .c files are not the abstraction, because they could be implemented any number of different ways, and it would still be the same abstraction.
It's incredible that even the header file is around 200 LOC. For state-of-the-art, very performant database, I would have thought that all manner of Os-specific IO api's would have to be used.