Reply to your post "RFC 4122/9562: UUID version 7 and 8 implementation"

[sergeyprokhorenko]

Hello Bénédikt,

This is the only available way to contact you and respond to your post RFC 4122/9562: UUID version 7 and 8 implementation. I was unable to reply via https://discuss.python.org/ due to their excessive restrictions.

The best choice is to take uuid-rs/uuid (it is perfect) as an example and add to it an optional formal parameter "timestamp offset" in milliseconds (positive or negative).

See also https://kodraus.github.io/rust/2024/06/24/uuid-v7-counters.html and https://www.reddit.com/r/rust/comments/1dn5ge4/uuid_now_properly_sorts_v7_ids/

[picnixz]

Oh! thank you for your reply. I will add a link to the Python issue as well in the discussion but the reason why I did not go for the rust implementation at first is because I wasn't sure how supported it is. But considering the references you linked and the other threads, I think most people prefer a fixed-bit length counter rather than a monotonous counter (it doesn't really matter to me, I just want to introduce something that is compatible in general).