Regardless of how the discussion in https://github.com/rust-lang/rust/pull/102737 turns out, pinning closure captures is super subtle business and probably worth discussing separately.
bors [Mon, 21 Nov 2022 15:22:54 +0000 (15:22 +0000)]
Auto merge of #104673 - matthiaskrgr:rollup-85f65ov, r=matthiaskrgr
Rollup of 9 pull requests
Successful merges:
- #104420 (Fix doc example for `wrapping_abs`)
- #104499 (rustdoc JSON: Use `Function` everywhere and remove `Method`)
- #104500 (`rustc_ast`: remove `ref` patterns)
- #104511 (Mark functions created for `raw-dylib` on x86 with DllImport storage class)
- #104595 (Add `PolyExistentialPredicate` type alias)
- #104605 (deduplicate constant evaluation in cranelift backend)
- #104628 (Revert "Update CI to use Android NDK r25b")
- #104662 (Streamline deriving on packed structs.)
- #104667 (Revert formatting changes of a test)
Matthias Krüger [Mon, 21 Nov 2022 13:11:13 +0000 (14:11 +0100)]
Rollup merge of #104662 - nnethercote:tweak-deriving-for-packed-non-copy, r=jackh726
Streamline deriving on packed structs.
The current approach to field accesses in derived code:
- Normal case: `&self.0`
- In a packed struct that derives `Copy`: `&{self.0}`
- In a packed struct that doesn't derive `Copy`: `let Self(ref x) = *self`
The `let` pattern used in the third case is equivalent to the simpler field access in the first case. This commit changes the third case to use a field access.
The commit also combines two boolean arguments (`is_packed` and `always_copy`) into a single field (`copy_fields`) earlier, to save passing both around.
The relevant discussion can be found in #103673, where it was agreed that more time is needed to warn the community of the upcoming breakage.
This PR is for the `master` branch, where a conflict was recently introduced due to 6d8160261ff3aee3b6eaacc37ac96cafff530980. The conflict is in `cc_detect.rs`, where the code that corrects the target triple was moved to a new function called `ndk_compiler()`. This puts the old logic in the `ndk_compiler` function, and assumes that it works properly in the other location where that code is being called. I would appreciate review from ``@pietroalbini`` to understand how we can test that the reverted logic is also suitable for the additional use case (seems to be related to setting `cc` and `cxx`). I've confirmed already that with these changes I can compile for `armv7-linux-androideabi`, `aarch64-linux-android`, `i686-linux-android`, and `x86_64-linux-android` using `x.py`.
A separate revert for the `beta` branch will be required, since the original change has already made it to beta. The beta revert is available at https://github.com/alex-pinkus/rust/commit/3fa0d94674fbe37090ebe44ac1f06e2233f3121e, but I'm not sure of the process for staging that PR.
Matthias Krüger [Mon, 21 Nov 2022 13:11:12 +0000 (14:11 +0100)]
Rollup merge of #104605 - RalfJung:clf_consts, r=bjorn3
deduplicate constant evaluation in cranelift backend
The cranelift backend had two matches on `ConstantKind`, which can be avoided, and used this `eval_for_mir` that nothing else uses... this makes things more consistent with the (better-tested) LLVM backend.
I noticed this because cranelift was the only user of `eval_for_mir`. However `try_eval_for_mir` still has one other user in `eval`... the odd thing is that the interpreter has its own `eval_mir_constant` which seems to duplicate the same functionality and does not use `try_eval_for_mir`. No idea what is happening here.
Matthias Krüger [Mon, 21 Nov 2022 13:11:10 +0000 (14:11 +0100)]
Rollup merge of #104511 - dpaoliello:privateglobalworkaround, r=michaelwoerister
Mark functions created for `raw-dylib` on x86 with DllImport storage class
Fix for #104453
## Issue Details
On x86 Windows, LLVM uses 'L' as the prefix for any private global symbols (`PrivateGlobalPrefix`), so when the `raw-dylib` feature creates an undecorated function symbol that begins with an 'L' LLVM misinterprets that as a private global symbol that it created and so fails the compilation at a later stage since such a symbol must have a definition.
## Fix Details
Mark the function we are creating for `raw-dylib` with `DllImport` storage class (this was already being done for MSVC at a later point for `callee::get_fn` but not for GNU (due to "backwards compatibility")): this will cause LLVM to prefix the name with `__imp_` and so it won't mistake it for a private global symbol.
Matthias Krüger [Mon, 21 Nov 2022 13:11:10 +0000 (14:11 +0100)]
Rollup merge of #104500 - WaffleLapkin:deref-the-compiler, r=wesleywiser
`rustc_ast`: remove `ref` patterns
Or in other words use match ergonomics in `rustc_ast`. I do plan to do the same with other crates, but to keep the diff sane, let's do them one at a time.
Matthias Krüger [Mon, 21 Nov 2022 13:11:09 +0000 (14:11 +0100)]
Rollup merge of #104420 - TethysSvensson:master, r=JohnTitor
Fix doc example for `wrapping_abs`
The `max` variable is unused. This change introduces the `min_plus` variable, to make the example similar to the one from `saturating_abs`. An alternative would be to remove the unused variable.
bors [Mon, 21 Nov 2022 12:17:03 +0000 (12:17 +0000)]
Auto merge of #103491 - cjgillot:self-rpit, r=oli-obk
Support using `Self` or projections inside an RPIT/async fn
I reuse the same idea as https://github.com/rust-lang/rust/pull/103449 to use variances to encode whether a lifetime parameter is captured by impl-trait.
The current implementation of async and RPIT replace all lifetimes from the parent generics by `'static`. This PR changes the scheme
```rust
impl<'a> Foo<'a> {
fn foo<'b, T>() -> impl Into<Self> + 'b { ... }
}
// NEW
fn foo<'b, T>() -> Foo::<'a>::foo::<'b, T>::opaque::<'b> { ... }
^^ the `Self` stays `Foo<'a>`
}
```
There is the same issue with projections. In the example, substitute `Self` by `<T as Trait<'b>>::Assoc` in the sugared version, and `Foo<'_a>` by `<T as Trait<'_b>>::Assoc` in the desugared one.
This allows to support `Self` in impl-trait, since we do not replace lifetimes by `'static` any more. The same trick allows to use projections like `T::Assoc` where `Self` is allowed. The feature is gated behind a `impl_trait_projections` feature gate.
The implementation relies on 2 tweaking rules for opaques in 2 places:
- we only relate substs that correspond to captured lifetimes during TypeRelation;
- we only list captured lifetimes in choice region computation.
For simplicity, I encoded the "capturedness" of lifetimes as a variance, `Bivariant` vs `Invariant` for unused vs captured lifetimes. The `variances_of` query used to ICE for opaques.
Impl-trait that do not reference `Self` or projections will have their variances as:
- `o` (invariant) for each parent type or const;
- `*` (bivariant) for each parent lifetime --> will not participate in borrowck;
- `o` (invariant) for each own lifetime.
Impl-trait that does reference `Self` and/or projections will have some parent lifetimes marked as `o` (as the example above), and participate in type relation and borrowck. In the example above, `variances_of(opaque) = ['_a: o, '_b: *, T: o, 'b: o]`.
r? types
cc `@compiler-errors` , as you asked about the issue with `Self` and projections.
The current approach to field accesses in derived code:
- Normal case: `&self.0`
- In a packed struct that derives `Copy`: `&{self.0}`
- In a packed struct that doesn't derive `Copy`: `let Self(ref x) = *self`
The `let` pattern used in the third case is equivalent to the simpler
field access in the first case. This commit changes the third case to
use a field access.
The commit also combines two boolean arguments (`is_packed` and
`always_copy`) into a single field (`copy_fields`) earlier, to save
passing both around.
bors [Mon, 21 Nov 2022 01:44:12 +0000 (01:44 +0000)]
Auto merge of #102717 - beetrees:repr128-c-style-debuginfo, r=nagisa
Pass 128-bit C-style enum enumerator values to LLVM
Pass the full 128 bits of C-style enum enumerators through to LLVM. This means that debuginfo for C-style repr128 enums is now emitted correctly for DWARF platforms (as compared to not being correctly emitted on any platform).
Matthias Krüger [Sun, 20 Nov 2022 22:50:30 +0000 (23:50 +0100)]
Rollup merge of #104643 - pnkfelix:examples-for-chunks-remainder, r=scottmcm
add examples to chunks remainder methods.
add examples to chunks remainder methods.
my motivation for adding the examples was to make it very clear that the state of the iterator (in terms of where its cursor lies) has no effect on what remainder returns.
Also fixed some links to rchunk remainder methods.
Matthias Krüger [Sun, 20 Nov 2022 22:50:28 +0000 (23:50 +0100)]
Rollup merge of #104581 - notriddle:notriddle/js-iife-2, r=GuillaumeGomez
rustdoc: remove unused JS IIFE from main.js
This [IIFE] made sense when it was added in deaf5e200e79a75ac57d3f0952f6758a38168e52 and there was a local variable scoped to it, but now it calls a function, but declares nothing.
[IIFE]: https://developer.mozilla.org/en-US/docs/Glossary/IIFE "immediately invoked function expression"
Matthias Krüger [Sun, 20 Nov 2022 22:50:27 +0000 (23:50 +0100)]
Rollup merge of #104504 - compiler-errors:fru-syntax-note, r=estebank
Add a detailed note for missing comma typo w/ FRU syntax
Thanks to `@pierwill` for working on this with me!
Fixes #104373, perhaps `@alice-i-cecile` can comment on the new error for the example provided on that issue -- feedback is welcome.
```
error[E0063]: missing field `defaulted` in initializer of `Outer`
--> $DIR/multi-line-fru-suggestion.rs:14:5
|
LL | Outer {
| ^^^^^ missing `defaulted`
|
note: this expression may have been misinterpreted as a `..` range expression
--> $DIR/multi-line-fru-suggestion.rs:16:16
|
LL | inner: Inner {
| ________________^
LL | | a: 1,
LL | | b: 2,
LL | | }
| |_________^ this expression does not end in a comma...
LL | ..Default::default()
| ^^^^^^^^^^^^^^^^^^^^ ... so this is interpreted as a `..` range expression, instead of functional record update syntax
help: to set the remaining fields from `Default::default()`, separate the last named field with a comma
|
LL | },
| +
error: aborting due to previous error
For more information about this error, try `rustc --explain E0063`.
```
Matthias Krüger [Sun, 20 Nov 2022 22:50:27 +0000 (23:50 +0100)]
Rollup merge of #104487 - klensy:ntapi, r=Mark-Simulacrum
update ntapi dep to remove future-incompat warning
This fixes warning https://github.com/rust-lang-ci/rust/actions/runs/3477235400/jobs/5813202075#step:25:217
`warning: the following packages contain code that will be rejected by a future version of Rust: ntapi v0.3.7`
by upgrading `sysinfo` version (https://github.com/GuillaumeGomez/sysinfo/blob/master/CHANGELOG.md#0267)
There was some breaking changes in `sysinfo`:
* 0.25.0 (System::refresh_cpu behaviour changed: it only computes CPU usage and doesn't retrieve CPU frequency.) not affected?
* 0.26.0 (Switch memory unit from kilobytes to bytes) fixed.
Matthias Krüger [Sun, 20 Nov 2022 22:50:26 +0000 (23:50 +0100)]
Rollup merge of #101310 - zachs18:rc_get_unchecked_mut_docs_soundness, r=Mark-Simulacrum
Clarify and restrict when `{Arc,Rc}::get_unchecked_mut` is allowed.
(Tracking issue for `{Arc,Rc}::get_unchecked_mut`: #63292)
(I'm using `Rc` in this comment, but it applies for `Arc` all the same).
As currently documented, `Rc::get_unchecked_mut` can lead to unsoundness when multiple `Rc`/`Weak` pointers to the same allocation exist. The current documentation only requires that other `Rc`/`Weak` pointers to the same allocation "must not be dereferenced for the duration of the returned borrow". This can lead to unsoundness in (at least) two ways: variance, and `Rc<str>`/`Rc<[u8]>` aliasing. ([playground link](https://play.rust-lang.org/?version=nightly&mode=debug&edition=2021&gist=d7e2d091c389f463d121630ab0a37320)).
This PR changes the documentation of `Rc::get_unchecked_mut` to restrict usage to when all `Rc<T>`/`Weak<T>` have the exact same `T` (including lifetimes). I believe this is sufficient to prevent unsoundness, while still allowing `get_unchecked_mut` to be called on an aliased `Rc` as long as the safety contract is upheld by the caller.
## Alternatives
* A less strict, but still sound alternative would be to say that the caller must only write values which are valid for all aliased `Rc`/`Weak` inner types. (This was [mentioned](https://github.com/rust-lang/rust/issues/63292#issuecomment-568284090) in the tracking issue). This may be too complicated to clearly express in the documentation.
* A more strict alternative would be to say that there must not be any aliased `Rc`/`Weak` pointers, i.e. it is required that get_mut would return `Some(_)`. (This was also mentioned in the tracking issue). There is at least one codebase that this would cause to become unsound ([here](https://github.com/kaimast/lsm-rs/blob/be5a164d770d850d905e510e2966ad4b1cc9aa5e/src/memtable.rs#L166), where additional locking is used to ensure unique access to an aliased `Rc<T>`; I saw this because it was linked on the tracking issue).
bors [Sun, 20 Nov 2022 19:14:14 +0000 (19:14 +0000)]
Auto merge of #104646 - matthiaskrgr:rollup-7xnhzf0, r=matthiaskrgr
Rollup of 6 pull requests
Successful merges:
- #104537 (fix std::thread docs are unclear regarding stack sizes)
- #104558 (Don't assume `FILE_ID_BOTH_DIR_INFO` will be aligned)
- #104564 (interpret: use Either over Result when it is not representing an error condition)
- #104568 (clarify that realloc refreshes pointer provenance even when the allocation remains in-place)
- #104611 (rustdoc: use real buttons for scrape examples controls)
- #104640 (Migrate kdb style to CSS variables)
bors [Sun, 20 Nov 2022 10:09:39 +0000 (10:09 +0000)]
Auto merge of #103390 - compiler-errors:metadata-mod-regions, r=eholk
Check fat pointer metadata compatibility modulo regions
Regions don't really mean anything anyways during hir typeck.
If this `erase_regions` makes anyone nervous, it's probably equally valid to just equate the types using a type relation, but regardless we should _not_ be using strict type equality while region variables are present.
bors [Sun, 20 Nov 2022 04:28:21 +0000 (04:28 +0000)]
Auto merge of #104629 - JohnTitor:rollup-vp3m98i, r=JohnTitor
Rollup of 6 pull requests
Successful merges:
- #103901 (Add tracking issue for `const_arguments_as_str`)
- #104112 (rustdoc: Add copy to the description of repeat)
- #104435 (`VecDeque::resize` should re-use the buffer in the passed-in element)
- #104467 (Fix substraction with overflow in `wrong_number_of_generic_args.rs`)
- #104608 (Cleanup macro matching recovery)
- #104626 (Fix doctest errors related to rustc_middle)
Yuki Okushi [Sun, 20 Nov 2022 04:15:59 +0000 (13:15 +0900)]
Rollup merge of #104608 - Nilstrieb:fixmed, r=compiler-errors
Cleanup macro matching recovery
The retry has been implemented already in #104335. Also removes a `HACK` comment that's not really needed anymore because the "don't recover during macro matching" isn't really a hack but correct behavior.
Yuki Okushi [Sun, 20 Nov 2022 04:15:59 +0000 (13:15 +0900)]
Rollup merge of #104467 - fuzzypixelz:fix/attempt-to-substract-with-overflow, r=compiler-errors
Fix substraction with overflow in `wrong_number_of_generic_args.rs`
Fixes #104287
This issue happens in the `suggest_moving_args_from_assoc_fn_to_trait_for_qualified_path` function, which seems to run before the error checking facilities can catch an invalid use of generic arguments. Thus we get a subtraction with overflow because the code implicitly assumes that the source program makes sense (or is this assumption not true even if the program is correct?).
Yuki Okushi [Sun, 20 Nov 2022 04:15:59 +0000 (13:15 +0900)]
Rollup merge of #104435 - scottmcm:iter-repeat-n, r=thomcc
`VecDeque::resize` should re-use the buffer in the passed-in element
Today it always copies it for *every* appended element, but one of those clones is avoidable.
This adds `iter::repeat_n` (https://github.com/rust-lang/rust/issues/104434) as the primitive needed to do this. If this PR is acceptable, I'll also use this in `Vec` rather than its custom `ExtendElement` type & infrastructure that is harder to share between multiple different containers:
bors [Sun, 20 Nov 2022 01:16:52 +0000 (01:16 +0000)]
Auto merge of #104522 - RalfJung:try_normalize_after_erasing_regions, r=oli-obk
try_normalize_after_erasing_regions: promote an assertion to always run
In https://github.com/rust-lang/miri/issues/2433 this assertion has been seen to trigger, so it might be worth actually checking this? Regressing debug assertions are very easy to miss until much later, and then they become quite hard to debug.
bors [Sat, 19 Nov 2022 22:25:18 +0000 (22:25 +0000)]
Auto merge of #104470 - ehuss:cdb-dupe-last-command, r=jyn514
Don't duplicate last cdb debuginfo test command
cdb scripts interpret a blank line to mean "repeat the last command", similar to what happens when running the debugger from a console. The code for compiletest that constructs the debugger script was inserting a blank line between the last command and the "quit" command. This caused the last command to be executed twice. This can cause some confusion since the `-check` lines are expecting the output in a certain order. But printing the last command twice causes that order-assumption to fail, and that can cause confusion.
This fixes it by removing the blank line.
AFAICT, gdb and lldb scripts don't have the same behavior with blank lines (and the gdb code doesn't add any blank lines anyways).
* Fix doc examples for Platforms with underaligned integer primitives.
* Mutable pointer doc examples use mutable pointers.
* Fill out tracking issue.
* Minor formatting changes.