bodduv commented on PR #14500: URL: https://github.com/apache/iceberg/pull/14500#issuecomment-3542324121
> I'm digging into the RFCs exact language but I also came across https://stackoverflow.com/questions/79489893/does-the-uuid-comparison-in-java-violate-the-uuid-standard it sounds like this sorting behavior is particularly defined for UUIDV6/7 and sorting is not prescribed in V4? At the same time, the openJDK folks acknowledge that this is a bug, so I'm not sure which yet (like I said, still digging into it) https://bugs.openjdk.org/browse/JDK-7025832 I have addressed how RFC defined ordering among UUIDs in #14216. If you take a closer look at [RFC 4122](https://datatracker.ietf.org/doc/html/rfc4122), page 4, paragraph heading "Rules for Lexical Equivalence". RFC 4122 specifies versions 1 through 4. It would be surprising if UUID v6/7 have different comparison semantics from v1-4? [RFC 9562](https://dl.acm.org/doi/pdf/10.17487/RFC9562#page=4.11) section "6.1.1. Sorting" mentions "UUID formats created by this specification are intended to be lexicographically sortable while in the textual representation." > if there's a good argument to retroactively work from the Java reference implementation behavior Maintaining backward compatibility is an argument in this case. But on the other hand, looking forward into the future, is there a possibility to communicate breaking changes and provide table migration strategy(ies) in release docs? I suppose its a topic for the mailing list. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
