service UUIDs have incorrect endianness

[phooky]

128-bit BLE UUIDs are, as I understand it, little-endian. For some reason bluer appears to have this the other way around. For instance, when I examine a device, btmon will report a service UUID as "11111111-1111-1111-1111-111100000001", whereas the same service will be reported by service.uuid() as "01000000-1111-1111-1111-111111111111". (It's of course possible that btmon is the one that has this backwards; I'm not looking at the raw packets at the moment.)

I haven't checked whether the same issue is present for service uuids, but I wouldn't be surprised if it is.

[phooky]

(It looks like the uuid crate interprets uuids as per RFC 9562, which specifies the binary representation as big-endian, so this is probably just changing a "from_u128" to "from_u128_le" somewhere.)