mbutrovich commented on PR #11467: URL: https://github.com/apache/arrow-rs/pull/11467#issuecomment-6403853744
> every new variant (`u32` has a faster split, `i128` will maybe have more) touches all of them and breaks semver What do you mean by the faster split for `u32`? > I tried it on top of your types: default, thin LTO and `opt-level=2`, generic and native, it compiles to the same instructions as the match-once pattern Could you share the prototype? I'd like to see the signature of `map` and what it returns. I'd also like to see how it fits callers that own their loop, such as a closure passed to [`try_binary`](https://github.com/apache/arrow-rs/blob/9d7f294ca6bfa3591616059d11a3c15464a21f58/arrow-arith/src/arity.rs#L254) or DataFusion's [`partition_indices`](https://github.com/apache/datafusion/blob/db83fccb9ac10f843be04daa5826501d2338caf4/datafusion/physical-plan/src/repartition/mod.rs#L961-L980), which scatters row indices by remainder. -- 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]
