Rust 1.13 announcement #136
Conversation
Based on a bunch of stuff @brson wrote
|
r? @aturon (rust_highfive has picked a reviewer for you, use r? to override) |
| !4 = !{i32 2, !"Debug Info Version", i32 3} | ||
| ``` | ||
|
|
||
| You're eyes don't deceive you. This crate contains no code at all, and it's not |
|
|
||
| In some cases this results in dramatic improvements. Build times for the ndarray | ||
| crate [improved by 50%][ev2], and in the (unreleased) [winapi 0.3] crate, which | ||
| is nothing but inline functions, rustc now generates the following LLVM IR: |
There was a problem hiding this comment.
To clarify, it has plenty of things other than inline functions, its just that all the functions are inline.
| the crate they are defined _unless_ they are called directly. This saves the | ||
| cost of rustc converting the function to LLVM IR and LLVM optimizing and | ||
| converting the function to machine code (TODO is this right about converting to | ||
| machine code - were these functions public before?). |
| GitHub. 1448 patches were landed in this release. | ||
|
|
||
| [install]: https://www.rust-lang.org/install.html | ||
| [notes]: |
| ### What's in 1.13.0 stable | ||
|
|
||
| The two big things in Rust 1.13 stable are compiler performance and a new | ||
| operator, `?`. As usual, there's also a number of library stabilizations and |
There was a problem hiding this comment.
Wording a little awk. Maybe "Rust 1.13 comes with two big improvements: a new ? operator that makes error handling easier and fixes to compiler performance improvements of XX%"
| diagnostic](https://github.com/rust-lang/rust/issues/35946), we now have | ||
| the ability to do so. | ||
|
|
||
| #### Performance improvements |
There was a problem hiding this comment.
Needs a graph or two in the release announcement, I think.
There was a problem hiding this comment.
There's a graph linked, but yeah it would be better to be inline.
| let mut f = match f { | ||
| Ok(file) => file, | ||
| Err(e) => return Err(e), | ||
| }; |
There was a problem hiding this comment.
People would write this with try! today, right? Feels like we might as well just write the idiomatic Rust and then show how it got better.
There was a problem hiding this comment.
I tried doing so, but this particular example is too simple, and doesn't look like a compelling justification of ? over try!. I have a more complex example here in comparison.
I do like that this current explains what ? does, not just that it's the same as try!.
| unwraps it and gives the inner value. If it was an `Err`, it returns from | ||
| the function you're currently in. | ||
|
|
||
| Seasoned Rustaceans may recongize that this is the same as the `try!` macro |
There was a problem hiding this comment.
Can we do a quick search of crates.io to see how much try! is used? I'd assume people are using it, but if it's not often used, perhaps it needs this introduction.
| poor error messages, and is very ugly when chained together. Consider: | ||
|
|
||
| ```rust | ||
| try!(try!(try!(foo()).bar()).baz()) |
There was a problem hiding this comment.
Do people write patterns like this? I'd assume this would be a bit more of an expanded form with unnecessary steps, like:
let f = try!(foo());
let g = try!(f.bar());
let h = try!(g.baz());There was a problem hiding this comment.
I wouldn't write that because it's so ugly. Now you can though because it's not ugly!
| !1 = !DIFile(filename: ".\5Csrc\5Clib.rs", directory: "C:\5Cmsys64\5Chome\5CPeter\5Cwinapi-rs") | ||
| !2 = !{} | ||
| !3 = !{i32 2, !"CodeView", i32 1} | ||
| !4 = !{i32 2, !"Debug Info Version", i32 3} |
There was a problem hiding this comment.
Is this llvm? I don't think most people will be able to read this (I know it took me a while looking at it to figure out what was possibly going on). Is there another way we can show this, maybe?
There was a problem hiding this comment.
I picked this because it's one of the interesting examples we had on hand, and I think a little technical peek under the hood is not too risky. Maybe it could be explained better, replaced with something else visually interesting, like an inline graph, or just removed.
|
Since @steveklabnik is out today I opened another PR with edits: #137. I have not yet addressed all @jonathandturner's comments. |
|
|
|
Do we want to talk about diagnostics? There are a few PRs with pretty pictures we could steal. |
|
@brson - true, I could try to dig around and see which new one landed in the beta |
|
@brson - here's an example that improved between stable and the beta: Before:
After:
|
|
I think the giant lists of library and cargo features are overwhelming, already produced in this form in the release notes themselves. Maybe we can just have a short section listing other notable stuff. |
| #### Performance improvements | ||
|
|
||
|
|
||
| There has bee a lot of focus on compiler performance lately. There's good news |
|
This has been moved to a fresh PR: #137 |
Based on a bunch of stuff @brson wrote
/cc @rust-lang/core
I have "allow edits from maintainers" on so you can just push changes to this branch if you'd like to make them.