[
https://issues.apache.org/jira/browse/ARROW-4678?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16777573#comment-16777573
]
Neville Dipale edited comment on ARROW-4678 at 2/26/19 5:10 AM:
----------------------------------------------------------------
I'm the one who added box_syntax and box_patterns. If there's a way of matching
datatypes in a ListArray(Box<DataType>) that can be done without the box
pattern, we can remove the 2 unstable features.
EDIT: looks like try_from is getting stabilised in 1.34
{code:java}
error: the feature `try_from` has been stable since 1.34.0 and no longer
requires an attribute to enable
--> arrow/src/lib.rs:27:12
|
27 | #![feature(try_from)]
| ^^^^^^^^
|
= note: `-D stable-features` implied by `-D warnings`
{code}
was (Author: nevi_me):
I'm the one who added box_syntax and box_patterns. If there's a way of matching
datatypes in a ListArray(Box<DataType>) that can be done without the box
pattern, we can remove the 2 unstable features.
> [Rust] Minimize unstable feature usage
> --------------------------------------
>
> Key: ARROW-4678
> URL: https://issues.apache.org/jira/browse/ARROW-4678
> Project: Apache Arrow
> Issue Type: Improvement
> Components: Rust
> Affects Versions: 0.12.0
> Reporter: Steven Fackler
> Priority: Major
>
> The Rust implementation currently uses quite a few nightly features. This is
> unfortunately a hard blocker on using these crates for many users.
> Here's the list of currently use nightly features:
> * type_ascription: Unused, can be trivially removed.
> * rustc_private: Unused, can be trivially removed.
> * box_syntax: Indefinitely far from stabilization, trivially replaceable
> with Box::new.
> * box_patterns: Indefinitely far from stabilization, replaceable with some
> minor restructuring of a couple of matches.
> * serde's alloc feature: Unused, can be trivially removed.
> * try_from: Scheduled for stabilization in Rust 1.35.
> * specialization: Actively being worked on - maybe ~1 year timeframe?
> * packed_simd: Actively being worked on - maybe ~1 year timeframe?
> The first set of features are easy enough to get rid of - I'll make a PR to
> do that (https://github.com/sfackler/arrow/tree/more-stable). I'm a bit less
> sure of what to do with specialization and packed_simd, though.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
