record type inference failing #56
Labels
A-type-system
Area: Type system
Comments
|
Dupe of #50, closing. |
oli-obk
pushed a commit
to oli-obk/rust
that referenced
this issue
Jul 19, 2017
various things that allow miri to crash less and execute more code
kazcw
pushed a commit
to kazcw/rust
that referenced
this issue
Oct 23, 2018
Add assert_instr to the rest of avx.rs
dlrobertson
pushed a commit
to dlrobertson/rust
that referenced
this issue
Nov 29, 2018
I went with "associated break" here to match the terminology given by PR rust-lang#56.
dlrobertson
pushed a commit
to dlrobertson/rust
that referenced
this issue
Nov 29, 2018
Feature/loop break value
rchaser53
pushed a commit
to rchaser53/rust
that referenced
this issue
Jan 19, 2019
Add some codegen related terminology to glossary
Aaron1011
pushed a commit
to Aaron1011/rust
that referenced
this issue
Oct 26, 2020
jacob-hughes
referenced
this issue
in jacob-hughes/alloy
Apr 29, 2021
56: Don't disable incremental GC. r=jacob-hughes a=ltratt If the user wishes to do this they can set the GC_DISABLE_INCREMENTAL=1 environment variable. Co-authored-by: Laurence Tratt <[email protected]>
djtech-dev
pushed a commit
to djtech-dev/rust
that referenced
this issue
Dec 9, 2021
This tries to address concerns raised in rust-lang#56 in different ways with different tradeoffs: 1. Implicit `Deref` to the underlying function pointer is replace with an explicit unsafe method `into_inner`, but in most cases it shouldn't be necessary to use. 2. Provides function `Symbol::call` which guarantees that (a) backing storage will be alive as long as it's required for calling and (b) preserves `unsafe`ty for the public interface. This is done at the cost of less convenient invocation syntax (`symbol.call(a,b,c,...)` instead of `symbol(a,b,c,...)`). 3. Provides `Symbol::into_closure` which allows to cast a `Symbol` into an opaque Rust closure. This also guarantees that the backing storage will be kept alive as long as it's required for calling, and provides a more convenient syntax (you can use normal function calls like `closure(a,b,c,...)`) at the cost of losing `unsafe` marker in the interface of the resulting function (although conversion itself still has `unsafe`). You might argue that this is unacceptable tradeoff for a public API, but, unfortunately, we don't have `UnsafeFn` family of traits yet, and also there is lots of existing precedents in stdlib that do similar "pretend that it's safe" conversions like `Box::from_raw`, `Rc::from_raw`, `slice::from_raw_parts`, `str::from_utf8_unchecked`, etc.
This issue was closed.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
See test rec-auto.rs.
The text was updated successfully, but these errors were encountered: