-
Notifications
You must be signed in to change notification settings - Fork 12.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Don't resolve generic impls that may be shadowed by dyn built-in impls #114941
Don't resolve generic impls that may be shadowed by dyn built-in impls #114941
Conversation
Some changes occurred to MIR optimizations cc @rust-lang/wg-mir-opt |
_2 = &(*_1); | ||
- _0 = <T as Any>::type_id(move _2) -> [return: bb1, unwind unreachable]; | ||
+ StorageLive(_3); | ||
+ _3 = std::intrinsics::type_id::<T>() -> [return: bb1, unwind unreachable]; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is the problematic inlined line -- when this function is eventually monomorphized for T = dyn Any
, we're acquiring the type-id of dyn Any
rather than going thru the vtable.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
please move this check to instance resolve instead and add these two tests to make sure that they don't also end up triggering this bug
trait Trait {
fn foo(&self) {}
}
impl<T: ?Sized> Trait for T {
fn foo(&self) {
println!("generic: {}", std::any::type_name::<T>());
}
}
fn bar<T: ?Sized>() -> fn(&T) {
Trait::foo
// If const prop where to propagate the instance
}
fn main() {
bar::<dyn Trait>()(&1);
}
#![feature(inline_const)]
trait Trait {
fn foo(&self) {}
}
impl<T: ?Sized> Trait for T {
fn foo(&self) {
println!("generic: {}", std::any::type_name::<T>());
}
}
fn bar<T: ?Sized>() -> fn(&T) {
const { Trait::foo as fn(&T) }
}
fn main() {
bar::<dyn Trait>()(&1);
}
r? @lcnr |
r=me after a types FCP. While this change is not breaking existing code, it is a new special-case of the types system so I think it's a good idea to notify the rest of the types team of this. |
be4fc7b
to
e8f9d94
Compare
@rfcbot fcp merge See PR description :> |
This comment was marked as outdated.
This comment was marked as outdated.
@rfcbot fcp cancel |
@compiler-errors proposal cancelled. |
@rfcbot fcp merge see PR description :> |
Team member @compiler-errors has proposed to merge this. The next step is review by the rest of the tagged team members: No concerns currently listed. Once a majority of reviewers approve (and at most 2 approvals are outstanding), this will enter its final comment period. If you spot a major issue that hasn't been raised at any point in this process, please speak up! See this document for info about what commands tagged team members can give me. |
This comment has been minimized.
This comment has been minimized.
@bors r- |
4209263
to
a30ad3a
Compare
@bors r=lcnr |
…yn, r=lcnr Don't resolve generic impls that may be shadowed by dyn built-in impls **NOTE:** This is a hack. This is not trying to be a general fix for the issue that we've allowed overlapping built-in trait object impls and user-written impls for quite a long time, and traits like `Any` rely on this (rust-lang#57893) -- this PR specifically aims to mitigate a new unsoundness that is uncovered by the MIR inliner (rust-lang#114928) that interacts with this pre-existing issue. Builtin `dyn Trait` impls may overlap with user-provided blanket impls (`impl<T: ?Sized> Trait for T`) in generic contexts. This leads to bugs when instances are resolved in polymorphic contexts, since we typically prefer object candidates over impl candidates. This PR implements a (hacky) heuristic to `resolve_associated_item` to account for that unfortunate hole in the type system -- we now bail with ambiguity if we try to resolve a non-rigid instance whose self type is not known to be sized. This makes sure we can still inline instances like `impl<T: Sized> Trait for T`, which can never overlap with `dyn Trait`'s built-in impl, but we avoid inlining an impl that may be shadowed by a `dyn Trait`. Fixes rust-lang#114928
☀️ Test successful - checks-actions |
👀 Test was successful, but fast-forwarding failed: 422 Update is not a fast forward |
@bors retry |
Finished benchmarking commit (39efb1c): comparison URL. Overall result: no relevant changes - no action needed@rustbot label: -perf-regression Instruction countThis benchmark run did not return any relevant results for this metric. Max RSS (memory usage)ResultsThis is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.
CyclesResultsThis is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.
Binary sizeResultsThis is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.
Bootstrap: 633.02s -> 632.908s (-0.02%) |
…llaumeGomez Rollup of 6 pull requests Successful merges: - rust-lang#112725 (rustdoc-search: add support for type parameters) - rust-lang#114941 (Don't resolve generic impls that may be shadowed by dyn built-in impls) - rust-lang#115625 (Explain HRTB + infer limitations of old solver) - rust-lang#115839 (Bump libc to 0.2.148) - rust-lang#115924 (Don't complain on a single non-exhaustive 1-ZST) - rust-lang#115946 (panic when encountering an illegal cpumask in thread::available_parallelism) r? `@ghost` `@rustbot` modify labels: rollup
@@ -2945,6 +2945,33 @@ impl<'tcx> Ty<'tcx> { | |||
_ => false, | |||
} | |||
} | |||
|
|||
pub fn is_known_rigid(self) -> bool { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Would be good to have a doc comment explaining what "rigid" means.
For instance, (_, _)
would be considered rigid by this check, right?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah, I can add this.
Rollup merge of rust-lang#114941 - compiler-errors:inline-shadowed-by-dyn, r=lcnr Don't resolve generic impls that may be shadowed by dyn built-in impls **NOTE:** This is a hack. This is not trying to be a general fix for the issue that we've allowed overlapping built-in trait object impls and user-written impls for quite a long time, and traits like `Any` rely on this (rust-lang#57893) -- this PR specifically aims to mitigate a new unsoundness that is uncovered by the MIR inliner (rust-lang#114928) that interacts with this pre-existing issue. Builtin `dyn Trait` impls may overlap with user-provided blanket impls (`impl<T: ?Sized> Trait for T`) in generic contexts. This leads to bugs when instances are resolved in polymorphic contexts, since we typically prefer object candidates over impl candidates. This PR implements a (hacky) heuristic to `resolve_associated_item` to account for that unfortunate hole in the type system -- we now bail with ambiguity if we try to resolve a non-rigid instance whose self type is not known to be sized. This makes sure we can still inline instances like `impl<T: Sized> Trait for T`, which can never overlap with `dyn Trait`'s built-in impl, but we avoid inlining an impl that may be shadowed by a `dyn Trait`. Fixes rust-lang#114928
…Jung Add note to `is_known_rigid` Adds a note requested by `@RalfJung` in rust-lang#114941 (comment) Let me know if there are any other fns that need documentation, I could throw them into this PR too :)
Rollup merge of rust-lang#116041 - compiler-errors:rigid-note, r=RalfJung Add note to `is_known_rigid` Adds a note requested by `@RalfJung` in rust-lang#114941 (comment) Let me know if there are any other fns that need documentation, I could throw them into this PR too :)
Pkgsrc changes: * Remove NetBSD-8 support (embedded LLVm requires newer C++ than what is in -8; it's conceivable that this could still build with an external LLVM) * undo powerpc 9.0 file naming tweak, since we no longer support -8. * Remove patch to LLVM for powerpc now included by upstream. * Minor adjustments, checksum changes etc. Upstream changes: Version 1.74.1 (2023-12-07) =========================== - [Resolved spurious STATUS_ACCESS_VIOLATIONs in LLVM] (rust-lang/rust#118464) - [Clarify guarantees for std::mem::discriminant] (rust-lang/rust#118006) - [Fix some subtyping-related regressions] (rust-lang/rust#116415) Version 1.74.0 (2023-11-16) ========================== Language -------- - [Codify that `std::mem::Discriminant<T>` does not depend on any lifetimes in T] (rust-lang/rust#104299) - [Replace `private_in_public` lint with `private_interfaces` and `private_bounds` per RFC 2145] (rust-lang/rust#113126) Read more in [RFC 2145](https://rust-lang.github.io/rfcs/2145-type-privacy.html). - [Allow explicit `#[repr(Rust)]`] (rust-lang/rust#114201) - [closure field capturing: don't depend on alignment of packed fields] (rust-lang/rust#115315) - [Enable MIR-based drop-tracking for `async` blocks] (rust-lang/rust#107421) Compiler -------- - [stabilize combining +bundle and +whole-archive link modifiers] (rust-lang/rust#113301) - [Stabilize `PATH` option for `--print KIND=PATH`] (rust-lang/rust#114183) - [Enable ASAN/LSAN/TSAN for `*-apple-ios-macabi`] (rust-lang/rust#115644) - [Promote loongarch64-unknown-none* to Tier 2] (rust-lang/rust#115368) - [Add `i686-pc-windows-gnullvm` as a tier 3 target] (rust-lang/rust#115687) Libraries --------- - [Implement `From<OwnedFd/Handle>` for ChildStdin/out/err] (rust-lang/rust#98704) - [Implement `From<{&,&mut} [T; N]>` for `Vec<T>` where `T: Clone`] (rust-lang/rust#111278) - [impl Step for IP addresses] (rust-lang/rust#113748) - [Implement `From<[T; N]>` for `Rc<[T]>` and `Arc<[T]>`] (rust-lang/rust#114041) - [`impl TryFrom<char> for u16`] (rust-lang/rust#114065) - [Stabilize `io_error_other` feature] (rust-lang/rust#115453) - [Stabilize the `Saturating` type] (rust-lang/rust#115477) - [Stabilize const_transmute_copy] (rust-lang/rust#115520) Stabilized APIs --------------- - [`core::num::Saturating`] (https://doc.rust-lang.org/stable/std/num/struct.Saturating.html) - [`impl From<io::Stdout> for std::process::Stdio`] (https://doc.rust-lang.org/stable/std/process/struct.Stdio.html#impl-From%3CStdout%3E-for-Stdio) - [`impl From<io::Stderr> for std::process::Stdio`] (https://doc.rust-lang.org/stable/std/process/struct.Stdio.html#impl-From%3CStderr%3E-for-Stdio) - [`impl From<OwnedHandle> for std::process::Child{Stdin, Stdout, Stderr}`] (https://doc.rust-lang.org/stable/std/process/struct.Stdio.html#impl-From%3CStderr%3E-for-Stdio) - [`impl From<OwnedFd> for std::process::Child{Stdin, Stdout, Stderr}`] (https://doc.rust-lang.org/stable/std/process/struct.Stdio.html#impl-From%3CStderr%3E-for-Stdio) - [`std::ffi::OsString::from_encoded_bytes_unchecked`] (https://doc.rust-lang.org/stable/std/ffi/struct.OsString.html#method.from_encoded_bytes_unchecked) - [`std::ffi::OsString::into_encoded_bytes`] (https://doc.rust-lang.org/stable/std/ffi/struct.OsString.html#method.into_encoded_bytes) - [`std::ffi::OsStr::from_encoded_bytes_unchecked`] (https://doc.rust-lang.org/stable/std/ffi/struct.OsStr.html#method.from_encoded_bytes_unchecked) - [`std::ffi::OsStr::as_encoded_bytes`] (https://doc.rust-lang.org/stable/std/ffi/struct.OsStr.html#method.as_encoded_bytes) - [`std::io::Error::other`] (https://doc.rust-lang.org/stable/std/io/struct.Error.html#method.other) - [`impl TryFrom<char> for u16`] (https://doc.rust-lang.org/stable/std/primitive.u16.html#impl-TryFrom%3Cchar%3E-for-u16) - [`impl<T: Clone, const N: usize> From<&[T; N]> for Vec<T>`] (https://doc.rust-lang.org/stable/std/vec/struct.Vec.html#impl-From%3C%26%5BT;+N%5D%3E-for-Vec%3CT,+Global%3E) - [`impl<T: Clone, const N: usize> From<&mut [T; N]> for Vec<T>`] (https://doc.rust-lang.org/stable/std/vec/struct.Vec.html#impl-From%3C%26mut+%5BT;+N%5D%3E-for-Vec%3CT,+Global%3E) - [`impl<T, const N: usize> From<[T; N]> for Arc<[T]>`] (https://doc.rust-lang.org/stable/std/sync/struct.Arc.html#impl-From%3C%5BT;+N%5D%3E-for-Arc%3C%5BT%5D,+Global%3E) - [`impl<T, const N: usize> From<[T; N]> for Rc<[T]>`] (https://doc.rust-lang.org/stable/std/rc/struct.Rc.html#impl-From%3C%5BT;+N%5D%3E-for-Rc%3C%5BT%5D,+Global%3E) These APIs are now stable in const contexts: - [`core::mem::transmute_copy`] (https://doc.rust-lang.org/beta/std/mem/fn.transmute_copy.html) - [`str::is_ascii`] (https://doc.rust-lang.org/beta/std/primitive.str.html#method.is_ascii) - [`[u8]::is_ascii`] (https://doc.rust-lang.org/beta/std/primitive.slice.html#method.is_ascii) Cargo ----- - [fix: Set MSRV for internal packages] (rust-lang/cargo#12381) - [config: merge lists in precedence order] (rust-lang/cargo#12515) - [fix(update): Clarify meaning of --aggressive as --recursive] (rust-lang/cargo#12544) - [fix(update): Make `-p` more convenient by being positional] (rust-lang/cargo#12545) - [feat(help): Add styling to help output ] (rust-lang/cargo#12578) - [feat(pkgid): Allow incomplete versions when unambigious] (rust-lang/cargo#12614) - [feat: stabilize credential-process and registry-auth] (rust-lang/cargo#12649) - [feat(cli): Add '-n' to dry-run] (rust-lang/cargo#12660) - [Add support for `target.'cfg(..)'.linker`] (rust-lang/cargo#12535) - [Stabilize `--keep-going`] (rust-lang/cargo#12568) - [feat: Stabilize lints] (rust-lang/cargo#12648) Rustdoc ------- - [Add warning block support in rustdoc] (rust-lang/rust#106561) - [Accept additional user-defined syntax classes in fenced code blocks] (rust-lang/rust#110800) - [rustdoc-search: add support for type parameters] (rust-lang/rust#112725) - [rustdoc: show inner enum and struct in type definition for concrete type] (rust-lang/rust#114855) Compatibility Notes ------------------- - [Raise minimum supported Apple OS versions] (rust-lang/rust#104385) - [make Cell::swap panic if the Cells partially overlap] (rust-lang/rust#114795) - [Reject invalid crate names in `--extern`] (rust-lang/rust#116001) - [Don't resolve generic impls that may be shadowed by dyn built-in impls] (rust-lang/rust#114941) Internal Changes ---------------- These changes do not affect any public interfaces of Rust, but they represent significant improvements to the performance or internals of rustc and related tools. None this cycle.
NOTE: This is a hack. This is not trying to be a general fix for the issue that we've allowed overlapping built-in trait object impls and user-written impls for quite a long time, and traits like
Any
rely on this (#57893) -- this PR specifically aims to mitigate a new unsoundness that is uncovered by the MIR inliner (#114928) that interacts with this pre-existing issue.Builtin
dyn Trait
impls may overlap with user-provided blanket impls (impl<T: ?Sized> Trait for T
) in generic contexts. This leads to bugs when instances are resolved in polymorphic contexts, since we typically prefer object candidates over impl candidates.This PR implements a (hacky) heuristic to
resolve_associated_item
to account for that unfortunate hole in the type system -- we now bail with ambiguity if we try to resolve a non-rigid instance whose self type is not known to be sized. This makes sure we can still inline instances likeimpl<T: Sized> Trait for T
, which can never overlap withdyn Trait
's built-in impl, but we avoid inlining an impl that may be shadowed by adyn Trait
.Fixes #114928