Use `edition = "2024"` in the compiler #129636

compiler-errors · 2024-08-27T00:00:10Z

Self-explanatory.

This involves changing all of the compiler crates' edition to 2024 (except for codegen_gcc and codegen_cranelift and pattern_analysis), and massively simplifying opaques and their lifetimes. I think the overwhelming majority of functions look significantly better now.

There are some other changes sprinkled in too, such as how env var functions are now unsafe.

rustbot · 2024-08-27T00:00:18Z

r? @albertlarsan68

rustbot has assigned @albertlarsan68.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

traviscross

This all looks right to me.

It's gratifying to see how much the Lifetime Capture Rules 2024 really clean things up, in particular by allowing lifetime elision to be used in many more places.

compiler/rustc_hir_analysis/src/collect/resolve_bound_vars.rs

traviscross · 2024-08-27T08:08:46Z

compiler/rustc_hir_analysis/src/collect/resolve_bound_vars.rs

@@ -1294,7 +1294,7 @@ impl<'a, 'tcx> BoundVarContext<'a, 'tcx> {
        // responsibility of lowering. This may create a mismatch between the resolution
        // AST found (`region_def_id`) which points to HRTB, and what HIR allows.
        // ```
-        // fn foo(x: impl for<'a> Trait<'a, Assoc = impl Copy + 'a>) {}
+        // fn foo(x: impl for<'a> Trait<'a, Assoc = impl Copy >) {}


Suggested change

// fn foo(x: impl for<'a> Trait<'a, Assoc = impl Copy >) {}

// fn foo(x: impl for<'a> Trait<'a, Assoc = impl Copy>) {}

compiler/rustc_hir_analysis/src/variance/mod.rs

compiler/rustc_hir_typeck/src/upvar.rs

compiler/rustc_parse/src/parser/ty.rs

compiler/rustc_pattern_analysis/src/lib.rs

compiler/rustc_trait_selection/src/error_reporting/infer/nice_region_error/static_impl_trait.rs

bors · 2024-08-28T00:28:38Z

☔ The latest upstream changes (presumably #129665) made this pull request unmergeable. Please resolve the merge conflicts.

compiler-errors · 2024-09-18T16:30:44Z

I think this is ready.

rustbot · 2024-09-18T16:30:57Z

This PR changes Stable MIR

cc @oli-obk, @celinval, @ouz-a

Some changes occurred in coverage instrumentation.

cc @Zalathar

rust-analyzer is developed in its own repository. If possible, consider making this change to rust-lang/rust-analyzer instead.

cc @rust-lang/rust-analyzer

rustc_error_messages was changed

cc @davidtwco, @compiler-errors, @TaKO8Ki

Some changes occurred in exhaustiveness checking

cc @Nadrieril

Some changes occurred in coverage instrumentation.

cc @Zalathar

Some changes occurred in compiler/rustc_sanitizers

cc @rust-lang/project-exploit-mitigations, @rcvalle

Some changes occurred to the CTFE / Miri interpreter

cc @rust-lang/miri

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

Some changes occurred to the core trait solver

cc @rust-lang/initiative-trait-system-refactor

compiler-errors · 2024-09-18T16:31:03Z

r? compiler

compiler/rustc_pattern_analysis/Cargo.toml

lcnr

can we remove Captures now that we have explicit use syntax?

I thought tidy requires safety comments for unsafe blocks? 🤔 I feel somewhat :/ about wrapping set_var in an unsafe block without explaining why that's correct. I guess it's always SAFETY: rustc is single threaded up until this point? I am not totally confident that this holds for all unsafe blocks, so alternatively, please open an issue tracking the fact that we may unsoundly use [set|remove]_env in the compiler

lcnr · 2024-09-19T00:16:41Z

apart from that r=me, cc @rust-lang/compiler @rust-lang/compiler-contributors

compiler-errors · 2024-09-19T00:20:00Z

can we remove Captures now that we have explicit use syntax?

I'll have to wait until use<...> is stable for that to be doable. I've hestitated to touch anything having to do with r-a for stability reasons. I can open a draft PR on top of this converting the last crate to ed 2024, and deleting Captures from the pattern analysis crate outright.

I am not totally confident that this holds for all unsafe blocks, so alternatively, please open an issue tracking the fact that we may unsoundly use [set|get]_env in the compiler

That's a good point. Yeah, I'll open a tracking issue for that (and enforcing // SAFETY in the compiler in general). and see if anyone would like to chime in on the safety. I'm not totally qualified to make these sort of libs/safety judgements, but also this is no more unsafe than it was on edition 2021, just more explicit.

saethlin · 2024-09-19T00:33:25Z

I thought tidy requires safety comments for unsafe blocks? 🤔

We have a handful of unsafe blocks without comments, such as

rust/compiler/rustc_index/src/slice.rs

Lines 137 to 138 in a5cf8bb

    
           let ptr = self.raw.as_mut_ptr(); 
        
           unsafe { (&mut *ptr.add(ai), &mut *ptr.add(bi), &mut *ptr.add(ci)) }

rust/compiler/rustc_data_structures/src/tagged_ptr/copy.rs

Lines 153 to 155 in a5cf8bb

    
           pub(super) fn pointer_raw(&self) -> NonNull<P::Target> { 
        
               self.packed.map_addr(|addr| unsafe { NonZero::new_unchecked(addr.get() << T::BITS) }) 
        
           }

Considering the strange way we handle formatting, I wouldn't speculate about what tidy is doing here.

RalfJung · 2024-09-19T06:06:53Z

but also this is no more unsafe than it was on edition 2021, just more explicit.

Well the entire point of this is to give everyone a chance to audit their use of set_env. If not even rustc bothers with thta, we can hardly expect any better from our users...

compiler/rustc_interface/src/passes.rs

RalfJung · 2024-09-19T06:10:08Z

compiler/rustc_llvm/build.rs

@@ -51,9 +51,13 @@ fn detect_llvm_link() -> (&'static str, &'static str) {
 fn restore_library_path() {
    let key = tracked_env_var_os("REAL_LIBRARY_PATH_VAR").expect("REAL_LIBRARY_PATH_VAR");
    if let Some(env) = tracked_env_var_os("REAL_LIBRARY_PATH") {
-        env::set_var(&key, env);
+        unsafe {


This one (and the others that are not immediately obvious) should get a FIXME, like the warning suggests.

bjorn3 · 2024-09-19T08:19:29Z

compiler/rustc_feature/src/tests.rs

@@ -3,7 +3,9 @@ use super::UnstableFeatures;
 #[test]
 fn rustc_bootstrap_parsing() {
    let is_bootstrap = |env, krate| {
-        std::env::set_var("RUSTC_BOOTSTRAP", env);
+        unsafe {
+            std::env::set_var("RUSTC_BOOTSTRAP", env);


This would affect other concurrent tests, right?

compiler/rustc_hir_analysis/src/variance/mod.rs

compiler/rustc_interface/src/passes.rs

bors · 2024-09-21T17:02:59Z

☔ The latest upstream changes (presumably #129283) made this pull request unmergeable. Please resolve the merge conflicts.

compiler-errors · 2024-09-21T18:49:01Z

I've opened #130672 to track those couple sketchy usages of set_var. I think this is otherwise good to go.

@bors r=lcnr

bors · 2024-09-21T18:49:04Z

📌 Commit d33d913 has been approved by lcnr

It is now in the queue for this repository.

lqd · 2024-09-21T22:59:44Z

do you know if the rustc-crates-on-stable rmake test failed as we expect with the pattern analysis changes? I wonder if using some of cargo’s unstable features would make it fail.

The answer is no, unfortunately.

It's also unclear to me whether we can ask cargo to only check the unstable features in the crates we're building rather than all the crates in the workspace. I've tried and couldn't do it: the presence of the unstable edition 2024 feature anywhere + cargo's -Zallow-features= failed validation even on the edition 2021 crates.

TL;DR: to my knowledge, we can't easily enforce on CI to not use cargo unstable features (including edition 2024) in the following crates that need to build on stable:

rustc_type_ir
rustc_next_trait_solver
rustc_pattern_analysis
rustc_lexer

compiler-errors · 2024-09-21T23:08:50Z

Hm... I probably should hold back rustc_lexer from edition 2024 as well.

@bors r-

bors · 2024-09-23T05:14:26Z

☔ The latest upstream changes (presumably #130724) made this pull request unmergeable. Please resolve the merge conflicts.

lqd · 2024-10-21T13:39:59Z

I probably should hold back rustc_lexer from edition 2024

as well as these 2 crates I've just learned r-a also uses (via #131997)

rustc_abi
rustc_parse_format

rustbot assigned albertlarsan68 Aug 27, 2024

compiler-errors unassigned albertlarsan68 Aug 27, 2024

This comment has been minimized.

Sign in to view

compiler-errors force-pushed the ed2024 branch 2 times, most recently from 2d3f18c to 47c874d Compare August 27, 2024 02:09

This comment has been minimized.

Sign in to view

traviscross approved these changes Aug 27, 2024

View reviewed changes

traviscross added the A-edition-2024 Area: The 2024 edition label Aug 27, 2024

traviscross mentioned this pull request Sep 3, 2024

Tracking Issue for Lifetime Capture Rules 2024 (RFC 3498) #117587

Closed

10 tasks

compiler-errors force-pushed the ed2024 branch 2 times, most recently from 8d11dbf to 3675f01 Compare September 18, 2024 16:30

compiler-errors marked this pull request as ready for review September 18, 2024 16:30

rustbot assigned lcnr Sep 18, 2024

compiler-errors commented Sep 18, 2024

View reviewed changes

compiler/rustc_pattern_analysis/Cargo.toml Outdated Show resolved Hide resolved

compiler-errors force-pushed the ed2024 branch 2 times, most recently from d4e9137 to b7420ee Compare September 18, 2024 16:37

lcnr reviewed Sep 19, 2024

View reviewed changes

RalfJung reviewed Sep 19, 2024

View reviewed changes

compiler/rustc_interface/src/passes.rs Show resolved Hide resolved

RalfJung reviewed Sep 19, 2024

View reviewed changes

bjorn3 reviewed Sep 19, 2024

View reviewed changes

WaffleLapkin reviewed Sep 19, 2024

View reviewed changes

compiler/rustc_hir_analysis/src/variance/mod.rs Show resolved Hide resolved

compiler/rustc_interface/src/passes.rs Show resolved Hide resolved

compiler-errors added 5 commits September 21, 2024 13:21

Set compiler crates to edition 2024

ac1d067

Remove unncessary outlives and captures from RPITs

f7c859b

Simplify some lifetimes

c673863

Actually use precise capturing in favor of outlives

93ea92c

Fix builtin lints

d22fb9a

compiler-errors force-pushed the ed2024 branch from 00e426a to d22fb9a Compare September 21, 2024 17:22

Add issue number, fix into_iter FIXME

d33d913

bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Sep 21, 2024

bors added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Sep 21, 2024

oli-obk mentioned this pull request Dec 4, 2024

coverage: Use a separate counter type and simplification step during counter creation #133849

Merged

klensy mentioned this pull request Dec 18, 2024

Use field init shorthand where possible #134443

Merged

Enselic mentioned this pull request Jan 8, 2025

rustc_feature: Avoid unsafe std::env::set_var() in UnstableFeatures tests #135245

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Use `edition = "2024"` in the compiler #129636

Use `edition = "2024"` in the compiler #129636

compiler-errors commented Aug 27, 2024 •

edited

Loading

rustbot commented Aug 27, 2024

This comment has been minimized.

This comment has been minimized.

traviscross left a comment

traviscross Aug 27, 2024

bors commented Aug 28, 2024

compiler-errors commented Sep 18, 2024

rustbot commented Sep 18, 2024

compiler-errors commented Sep 18, 2024

lcnr left a comment •

edited

Loading

lcnr commented Sep 19, 2024

compiler-errors commented Sep 19, 2024 •

edited

Loading

saethlin commented Sep 19, 2024

RalfJung commented Sep 19, 2024

RalfJung Sep 19, 2024

bjorn3 Sep 19, 2024

bors commented Sep 21, 2024

compiler-errors commented Sep 21, 2024

bors commented Sep 21, 2024

lqd commented Sep 21, 2024 •

edited

Loading

compiler-errors commented Sep 21, 2024

bors commented Sep 23, 2024

lqd commented Oct 21, 2024

	// fn foo(x: impl for<'a> Trait<'a, Assoc = impl Copy >) {}
	// fn foo(x: impl for<'a> Trait<'a, Assoc = impl Copy>) {}

Use edition = "2024" in the compiler #129636

Are you sure you want to change the base?

Use edition = "2024" in the compiler #129636

Conversation

compiler-errors commented Aug 27, 2024 • edited Loading

rustbot commented Aug 27, 2024

This comment has been minimized.

This comment has been minimized.

traviscross left a comment

Choose a reason for hiding this comment

traviscross Aug 27, 2024

Choose a reason for hiding this comment

bors commented Aug 28, 2024

compiler-errors commented Sep 18, 2024

rustbot commented Sep 18, 2024

compiler-errors commented Sep 18, 2024

lcnr left a comment • edited Loading

Choose a reason for hiding this comment

lcnr commented Sep 19, 2024

compiler-errors commented Sep 19, 2024 • edited Loading

saethlin commented Sep 19, 2024

RalfJung commented Sep 19, 2024

RalfJung Sep 19, 2024

Choose a reason for hiding this comment

bjorn3 Sep 19, 2024

Choose a reason for hiding this comment

bors commented Sep 21, 2024

compiler-errors commented Sep 21, 2024

bors commented Sep 21, 2024

lqd commented Sep 21, 2024 • edited Loading

compiler-errors commented Sep 21, 2024

bors commented Sep 23, 2024

lqd commented Oct 21, 2024

Use `edition = "2024"` in the compiler #129636

Use `edition = "2024"` in the compiler #129636

compiler-errors commented Aug 27, 2024 •

edited

Loading

lcnr left a comment •

edited

Loading

compiler-errors commented Sep 19, 2024 •

edited

Loading

lqd commented Sep 21, 2024 •

edited

Loading