Re:Non-Lexical Lifetimes in Rust
by steveklabnik on Monday May 02, 2016 @11:39AM
(#11612322)
In an extremely strict sense, no, but in reality yes. MIR makes it so much easier to implement a feature like this that we will wait for it before adding this.
[ Parent
| on HN ]
So is it likely that this will be a Rust 2.x kind of thing?
Re:Non-Lexical Lifetimes in Rust
by GolDDranks on Monday May 02, 2016 @01:31PM
(#11613534)
Not at all – 2.0 means a breaking change, but currently the borrow checker is just too restrictive. Loosening it a bit won't be a breaking change, as it allows strictly more code, not less. The landing of MIR is an implementation detail (a quite large at that, but nevertheless), and it shouldn't have any observable effects. (Except for maybe some bugs getting fixed.)
Re:Non-Lexical Lifetimes in Rust
by steveklabnik on Monday May 02, 2016 @01:28PM
(#11613502)
Not at all. MIR is not user-facing, it's a refactoring of compiler internals, and it's getting close to being finished. The reason Niko wrote this blog post was to start thinking about the eventual RFC to specify exactly how they should work.
There are no current plans for a Rust 2.0.
Nope, nowhere near that far off. MIR will be landing this year, and while we don't have a timeframe for NLL yet it's definitely a priority and it's clear that the devs are giving thought to it, as per this post.