[
https://issues.apache.org/jira/browse/ARROW-2373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16421484#comment-16421484
]
Andy Grove commented on ARROW-2373:
-----------------------------------
I just got this suggestion on the Rust forum:
"Just make a type that is a 64-byte aligned array of 16 i32s (or just 64 u8s,
although if you are using it as i32s it’s more elegant to use i32s), make a Vec
of that, and then a newtype over that Vec that implements Deref to &i32 by
transmuting."
"Add an extra size field to the newtype if you need to keep track of size at
primitive granularity."
This seems almost ideal to me.
> [Rust] Investigate alternate ways to achieve byte-aligned memory
> ----------------------------------------------------------------
>
> Key: ARROW-2373
> URL: https://issues.apache.org/jira/browse/ARROW-2373
> Project: Apache Arrow
> Issue Type: New Feature
> Components: Rust
> Reporter: Andy Grove
> Priority: Minor
>
> The current Rust implementation of Buffer<T> relies on libc and unsafe code
> to allocate byte-aligned buffers.
> It would be nice to find a safer way to do this.
> Rust 1.25 added support for byte-aligned structs, but that doesn't help.
> I have started a discussion in the Rust develop forum about adding support
> for byte-aligned Vec:
> [https://internals.rust-lang.org/t/byte-aligned-vec/7200]
> Feedback so far is "We’re getting close to stabilizing some of the allocator
> API which will let you make a custom type that does this though.".
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
