Make sure FFI attrs aren't used on foreign statics
Previously, we verified that FFI attrs were used on foreign items,
but this allowed them on both foreign functions and foreign statics.
This change only allows them on foreign functions.
bors [Tue, 17 Jan 2023 17:39:48 +0000 (17:39 +0000)]
Auto merge of #106294 - Nilstrieb:noundef-everything, r=nikic
Put `noundef` on all scalars that don't allow uninit
Previously, it was only put on scalars with range validity invariants like bool, was uninit was obviously invalid for those.
Since then, we have normatively declared all uninit primitives to be undefined behavior and can therefore put `noundef` on them.
The remaining concern was the `mem::uninitialized` function, which cause quite a lot of UB in the older parts of the ecosystem. After #99182, this function now doesn't return uninit values anymore, making users of it safe from this change.
The only real sources of UB where people could encounter uninit primitives are `MaybeUninit::uninit().assume_init()`, which has always be clear in the docs about being UB and from heap allocations (like reading from the spare capacity of a vec). This is hopefully rare enough to not break anything.
Dylan DPC [Tue, 17 Jan 2023 15:03:03 +0000 (20:33 +0530)]
Rollup merge of #101698 - raldone01:feat/const_cmp_typeid, r=scottmcm
Constify `TypeId` ordering impls
Tracking issue: #101871
Adding const ordering to `TypeId` allows rtti crates to optimize some casting scenarios (without transmuting to `u64`). This would also prevent these crates from breaking if the underlying type is changed from `u64` to something different.
bors [Tue, 17 Jan 2023 11:34:35 +0000 (11:34 +0000)]
Auto merge of #106612 - JakobDegen:cleanup-wf, r=tmiasko
Document wf constraints on control flow in cleanup blocks
Was recently made aware of [this code](https://github.com/rust-lang/rust/blob/a377893da2cd7124e5a18c7116cbb70e16dd5541/compiler/rustc_codegen_ssa/src/mir/analyze.rs#L247-L368), which has this potential ICE: https://github.com/rust-lang/rust/blob/a377893da2cd7124e5a18c7116cbb70e16dd5541/compiler/rustc_codegen_ssa/src/mir/analyze.rs#L308-L314
Roughly speaking, the code there is attempting to partition the cleanup blocks into funclets that satisfy a "unique successor" property, and the ICE is set off if that's not possible. This PR documents the well-formedness constraints that MIR must satisfy to avoid setting off that ICE.
The constraints documented are slightly stronger than the cases in which the ICE would have been set off in that code. This is necessary though, since whether or not that ICE gets set off can depend on iteration order in some graphs.
This sort of constraint is kind of ugly, but I don't know a better alternative at the moment. It's worth knowing that two important optimizations are still correct:
- Removing edges in the cfg: Fewer edges => fewer paths => stronger dominance relations => more contractions, and more contractions can't turn a forest into not-a-forest.
- Contracting an edge u -> v when u only has one successor and v only has one predecessor: u already dominated v, so this contraction was going to happen anyway.
There is definitely a MIR opt somewhere that can run afoul of this, but I don't know where it is. `@saethlin` was able to set it off though, so maybe he'll be able to shed some light on it.
r? `@RalfJung` I suppose, and cc `@tmiasko` who might have insight/opinions on this
bors [Tue, 17 Jan 2023 07:20:32 +0000 (07:20 +0000)]
Auto merge of #106627 - Ezrashaw:no-e0711-without-staged-api, r=Mark-Simulacrum
fix: don't emit `E0711` if `staged_api` not enabled
Fixes #106589
Simple fix, added UI test.
As an aside, it seems a lot of features are susceptible to this, `E0711` stands out to me because it's perma-unstable and we are effectively exposing an implementation detail.
Nilstrieb [Fri, 30 Dec 2022 14:55:05 +0000 (15:55 +0100)]
Put `noundef` on all scalars that don't allow uninit
Previously, it was only put on scalars with range validity invariants
like bool, was uninit was obviously invalid for those.
Since then, we have normatively declared all uninit primitives to be
undefined behavior and can therefore put `noundef` on them.
The remaining concern was the `mem::uninitialized` function, which cause
quite a lot of UB in the older parts of the ecosystem. This function now
doesn't return uninit values anymore, making users of it safe from this
change.
The only real sources of UB where people could encounter uninit
primitives are `MaybeUninit::uninit().assume_init()`, which has always
be clear in the docs about being UB and from heap allocations (like
reading from the spare capacity of a vec. This is hopefully rare enough
to not break anything.
bors [Tue, 17 Jan 2023 04:25:53 +0000 (04:25 +0000)]
Auto merge of #106966 - matthiaskrgr:rollup-e34pevi, r=matthiaskrgr
Rollup of 7 pull requests
Successful merges:
- #106591 (suggestion for attempted integer identifier in patterns)
- #106712 (make error emitted on `impl &Trait` nicer)
- #106829 (Unify `Opaque`/`Projection` handling in region outlives code)
- #106869 (rustdoc: remove redundant item kind class from `.item-decl > pre`)
- #106949 (ConstBlocks are poly if their substs are poly)
- #106953 (Document `EarlyBinder::subst_identity` and `skip_binder`)
- #106958 (Don't add A-bootstrap to PRs modifying Cargo.lock)
Matthias Krüger [Tue, 17 Jan 2023 04:25:23 +0000 (05:25 +0100)]
Rollup merge of #106953 - kylematsuda:early-binder-docs, r=jackh726
Document `EarlyBinder::subst_identity` and `skip_binder`
Finishing implementing #105779 will change several commonly used queries to return `EarlyBinder` by default. This PR adds documentation for two of the methods used to access data inside the `EarlyBinder`. I tried to summarize some of the [discussion from the issue](https://github.com/rust-lang/rust/issues/105779#issuecomment-1375512647) in writing this.
The item decl no longer has a border, and there aren't any kind-specific styles in modern rustdoc's rendering of this UI item.
Most of this PR is updating test cases so that they use `item-decl` to find the `<pre>` tag instead of relying on the fact that the class name had `rust {kind}` in it while other `<pre>` tags only had class `rust`.
Matthias Krüger [Tue, 17 Jan 2023 04:25:21 +0000 (05:25 +0100)]
Rollup merge of #106712 - Ezrashaw:impl-ref-trait, r=estebank
make error emitted on `impl &Trait` nicer
Fixes #106694
Turned out to be simpler than I thought, also added UI test.
Before: ([playground](https://play.rust-lang.org/?version=stable&mode=debug&edition=2021&gist=9bda53271ef3a8886793cf427b8cea91))
```text
error: expected one of `:`, ``@`,` or `|`, found `)`
--> src/main.rs:2:22
|
2 | fn foo(_: impl &Trait) {}
| ^ expected one of `:`, ``@`,` or `|`
|
= note: anonymous parameters are removed in the 2018 edition (see RFC 1685)
help: if this is a parameter name, give it a type
|
2 | fn foo(_: impl Trait: &TypeName) {}
| ~~~~~~~~~~~~~~~~
help: if this is a type, explicitly ignore the parameter name
|
2 | fn foo(_: impl _: &Trait) {}
| ++
error: expected one of `!`, `(`, `)`, `,`, `?`, `for`, `~`, lifetime, or path, found `&`
--> src/main.rs:2:16
|
2 | fn foo(_: impl &Trait) {}
| -^ expected one of 9 possible tokens
| |
| help: missing `,`
error: expected one of `!`, `(`, `,`, `=`, `>`, `?`, `for`, `~`, lifetime, or path, found `&`
--> src/main.rs:3:11
|
3 | fn bar<T: &Trait>(_: T) {}
| ^ expected one of 10 possible tokens
```
After:
```text
error: expected a trait, found type
--> <anon>:2:16
|
2 | fn foo(_: impl &Trait) {}
| -^^^^^
| |
| help: consider removing the indirection
error: expected a trait, found type
--> <anon>:3:11
|
3 | fn bar<T: &Trait>(_: T) {}
| -^^^^^
| |
| help: consider removing the indirection
```
bors [Mon, 16 Jan 2023 19:30:08 +0000 (19:30 +0000)]
Auto merge of #106945 - matthiaskrgr:rollup-c5or8z3, r=matthiaskrgr
Rollup of 5 pull requests
Successful merges:
- #105954 (Update instrument-coverage.md)
- #106835 (new trait solver: rebase impl substs for gats correctly)
- #106912 (check -Z query-dep-graph is enabled if -Z dump-dep-graph (#106736))
- #106940 (Improve a TAIT error and add an error code plus documentation)
- #106942 (Update books)
- fix examples for rustc 1.68.0-nightly (935dc0721 2022-12-19) (#1556) (rust-lang/rustc-dev-guide#1557)
- Update incremental-compilation-in-detail.md (rust-lang/rustc-dev-guide#1553)
- Link to the youtube recording of my talk, not the summary (rust-lang/rustc-dev-guide#1554)
- Change `src/test` to `tests` (rust-lang/rustc-dev-guide#1547)
- add full name for ICE (rust-lang/rustc-dev-guide#1552)
- Fix incorrect links (rust-lang/rustc-dev-guide#1549)
- fix rebase link (rust-lang/rustc-dev-guide#1546)
- Add a section for how to review code more easily (rust-lang/rustc-dev-guide#1538)
Matthias Krüger [Mon, 16 Jan 2023 19:29:37 +0000 (20:29 +0100)]
Rollup merge of #105954 - gftea:patch-1, r=ehuss
Update instrument-coverage.md
explicitly set environment variable LLVM_PROFILE_FILE="default_%m.profraw" when starting cargo test, otherwise, only one `default.profraw ` is generated and will be overwritten if run unnittests and integration tests at the same time.
bors [Mon, 16 Jan 2023 15:11:28 +0000 (15:11 +0000)]
Auto merge of #106853 - TimNN:undo-remap, r=oli-obk
Heuristically undo path prefix mappings.
Because the compiler produces better diagnostics if it can find the source of (potentially remapped) dependencies.
The new test fails without the other changes in this PR. Let me know if you have better suggestions for the test directory. I moved the existing remapping test to be in the same location as the new one.
Some more context: I'm exploring running UI tests with remapped paths by default in https://github.com/rust-lang/rust/pull/105924 and this was one of the issues discovered.
This may also be useful in the context of https://github.com/rust-lang/rfcs/pull/3127 ("New rustc and Cargo options to allow path sanitisation by default").
bors [Mon, 16 Jan 2023 09:37:08 +0000 (09:37 +0000)]
Auto merge of #106872 - dtolnay:nbsp, r=fee1-dead
Emit only one nbsp error per file
Fixes #106101.
See https://github.com/rust-lang/rust/issues/106098 for an explanation of how someone would end up with a large number of these nbsp characters in their source code, which is why I think rustc needs to handle this specific case in a friendlier way.
bors [Mon, 16 Jan 2023 05:55:59 +0000 (05:55 +0000)]
Auto merge of #106395 - compiler-errors:rework-predicates, r=eholk
Rework some `predicates_of`/`{Generic,Instantiated}Predicates` code
1. Make `instantiate_own` return an iterator, since it's a bit more efficient and easier to work with
2. Remove `bound_{explicit,}_predicates_of` -- these `bound_` methods in particular were a bit awkward to work with since `ty::GenericPredicates` *already* acts kinda like an `EarlyBinder` with its own `instantiate_*` methods, and had only a few call sites anyways.
3. Implement `IntoIterator` for `InstantiatedPredicates`, since it's *very* commonly being `zip`'d together.
bors [Sun, 15 Jan 2023 21:35:50 +0000 (21:35 +0000)]
Auto merge of #106914 - matthiaskrgr:rollup-yh0x4gq, r=matthiaskrgr
Rollup of 5 pull requests
Successful merges:
- #106888 (Add tidy check to ensure that rustdoc GUI tests start with a small description)
- #106896 (suggest `is_empty` for collections when casting to `bool`)
- #106900 (Fix regression in `unused_braces` with macros)
- #106906 (remove redundant clones)
- #106909 (Only suggest adding type param if path being resolved was a type)
Matthias Krüger [Sun, 15 Jan 2023 20:17:34 +0000 (21:17 +0100)]
Rollup merge of #106896 - Ezrashaw:str-cast-bool-emptyness, r=compiler-errors
suggest `is_empty` for collections when casting to `bool`
Fixes #106883
Matches on slices, `String` and `str`. It would be nice to do this with something like `Deref<Target=str>` as well, but AFAIK it's not possible in this part of the compiler.
Matthias Krüger [Sun, 15 Jan 2023 20:17:34 +0000 (21:17 +0100)]
Rollup merge of #106888 - GuillaumeGomez:tidy-gui-test, r=notriddle
Add tidy check to ensure that rustdoc GUI tests start with a small description
The first commit comes from https://github.com/rust-lang/rust/pull/106865 to prevent CI to fail.
This PR adds a tidy check to enforce having a small description at the top of the GUI test. Although the format is made to be as easy as possible to read, it's not always obvious what a test is actually checking. I think enforcing this will make it easier for us to come back on these tests if needed.
bors [Sun, 15 Jan 2023 12:10:36 +0000 (12:10 +0000)]
Auto merge of #106171 - compiler-errors:consolidate-extract_callable_info, r=estebank,lcnr
Consolidate two almost duplicated fn info extraction routines
Moves `extract_callable_info` up to trait selection, because it was being (almost) duplicated fully there for similar diagnostic purposes. This also generalizes the diagnostics we can give slightly (see UI test).
bors [Sun, 15 Jan 2023 08:59:55 +0000 (08:59 +0000)]
Auto merge of #105851 - dtolnay:peekmutleak, r=Mark-Simulacrum
Leak amplification for peek_mut() to ensure BinaryHeap's invariant is always met
In the libs-api team's discussion around #104210, some of the team had hesitations around exposing malformed BinaryHeaps of an element type whose Ord and Drop impls are trusted, and which does not contain interior mutability.
For example in the context of this kind of code:
```rust
use std::collections::BinaryHeap;
use std::ops::Range;
use std::slice;
// This is a souped up slice::split_at_mut to split in arbitrary many places.
//
// usize's Ord impl is trusted, so 1 single bounds check guarantees all those
// output slices are non-overlapping and in-bounds
fn chop<T>(slice: &mut [T], mut cut_points: BinaryHeap<usize>) -> Vec<&mut [T]> {
let mut vec = Vec::with_capacity(cut_points.len() + 1);
let max = match cut_points.pop() {
Some(max) => max,
None => {
vec.push(slice);
return vec;
}
};
assert!(max <= slice.len());
let len = slice.len();
let ptr: *mut T = slice.as_mut_ptr();
let get_unchecked_mut = unsafe {
|range: Range<usize>| &mut *slice::from_raw_parts_mut(ptr.add(range.start), range.len())
};
vec.push(get_unchecked_mut(max..len));
let mut end = max;
while let Some(start) = cut_points.pop() {
vec.push(get_unchecked_mut(start..end));
end = start;
}
vec.push(get_unchecked_mut(0..end));
vec
}
```
```rust
let slice = &mut ['0', '1', '2', '3', '4', '5', '6', '7', '8', '9'];
let mut cut_points = BinaryHeap::from(vec![100, 100]);
{
let mut max = cut_points.peek_mut().unwrap();
*max = 0;
std::mem::forget(max);
}
println!("{:?}", chop(slice, cut_points));
```
```console
[['0', '1', '2', '3', '4', '5', '6', '7', '8', '9'], [], ['0', '1', '2', '3', '4', '5', '6', '7', '8', '9', '\u{1}', '\0', '?', '翾', '?', '翾', '\0', '\0', '?', '翾', '?', '翾', '?', '啿', '?', '啿', '?', '啿', '?', '啿', '?', '啿', '?', '翾', '\0', '\0', '', '啿', '\u{5}', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0', '\u{8}', '\0', '`@',` '\0', '\u{1}', '\0', '?', '翾', '?', '翾', '?', '翾', '
thread 'main' panicked at 'index out of bounds: the len is 33 but the index is 33', library/core/src/unicode/unicode_data.rs:319:9
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
```
---
This PR makes `peek_mut()` use leak amplification (https://doc.rust-lang.org/1.66.0/nomicon/leaking.html#drain) to preserve the heap's invariant even in the situation that `PeekMut` gets leaked.
I'll also follow up in the tracking issue of unstable `drain_sorted()` (#59278) and `retain()` (#71503).
bors [Sun, 15 Jan 2023 00:16:09 +0000 (00:16 +0000)]
Auto merge of #106892 - matthiaskrgr:rollup-ohneu8o, r=matthiaskrgr
Rollup of 8 pull requests
Successful merges:
- #106072 (fix: misleading "add dyn keyword before derive macro" suggestion)
- #106859 (Suggestion for type mismatch when we need a u8 but the programmer wrote a char literal)
- #106863 (Remove various double spaces in compiler source comments.)
- #106865 (Add explanation comment for GUI test)
- #106867 (Fix the stability attributes for `std::os::fd`.)
- #106878 (Add regression test for #92157)
- #106879 (Add regression test for #42114)
- #106880 (doc: fix typo)
Matthias Krüger [Sun, 15 Jan 2023 00:01:38 +0000 (01:01 +0100)]
Rollup merge of #106867 - sunfishcode:sunfishcode/std-os-fd-stable-version, r=m-ou-se
Fix the stability attributes for `std::os::fd`.
As `@bjorn3` pointed out [here], I used the wrong stability attribute in #98368 when making `std::os::fd` public. I set it to Rust 1.63, which was when io-safety was stabilized, but it should be Rust 1.66, which was when `std::os::fd` was stabilized.
Matthias Krüger [Sun, 15 Jan 2023 00:01:36 +0000 (01:01 +0100)]
Rollup merge of #106859 - tialaramex:master, r=Nilstrieb
Suggestion for type mismatch when we need a u8 but the programmer wrote a char literal
Today Rust just points out that we have a char and we need a u8, but if I wrote 'A' then I could fix this by just writing b'A' instead. This code should detect the case where we're about to report a type mismatch of this kind, and the programmer wrote a char literal, and the char they wrote is ASCII, so therefore just prefixing b to make a byte literal will do what they meant.
I have definitely written this mistake more than once, it's not difficult to figure out what to do, but the compiler might as well tell us anyway.
I provided a test with two simple examples where the suggestion is appropriate, and one where it is not because the char literal is not ASCII, showing that the suggestion is only triggered in the former cases.
I have contributed only a small typo doc fix before, so this is my first substantive rustc change.
bors [Sat, 14 Jan 2023 20:53:37 +0000 (20:53 +0000)]
Auto merge of #106866 - matthiaskrgr:rollup-r063s44, r=matthiaskrgr
Rollup of 8 pull requests
Successful merges:
- #105526 (libcore: make result of iter::from_generator Clone)
- #106563 (Fix `unused_braces` on generic const expr macro call)
- #106661 (Stop probing for statx unless necessary)
- #106820 (Deprioritize fulfillment errors that come from expansions.)
- #106828 (rustdoc: remove `docblock` class from notable trait popover)
- #106849 (Allocate one less vec while parsing arrays)
- #106855 (rustdoc: few small cleanups)
- #106860 (Remove various double spaces in the libraries.)
The item decl no longer has a border, and there aren't any
kind-specific styles in modern rustdoc's rendering of this UI item.
Most of this commit is updating test cases so that they use `item-decl` to
find the `<pre>` tag instead of relying on the fact that the class name
had `rust {kind}` in it while other `<pre>` tags only had class `rust`.
Dan Gohman [Sat, 14 Jan 2023 17:35:42 +0000 (09:35 -0800)]
Fix the stability attributes for `std::os::fd`.
As @bjorn3 pointed out [here], I used the wrong stability attribute in #98368
when making `std::os::fd` public. I set it to Rust 1.63, which was when
io-safety was stabilized, but it should be Rust 1.66, which was when
`std::os::fd` was stabilized.
projects / rust.git / log