Make inductive cycles always ambiguous
This makes inductive cycles always result in ambiguity rather than be treated like a stack-dependent error.
This has some interactions with specialization, and so breaks a few UI tests that I don't agree should've ever worked in the first place, and also breaks a handful of crates in a way that I don't believe is a problem.
On the bright side, it puts us in a better spot when it comes to eventually enabling coinduction everywhere.
## Results
This was cratered in https://github.com/rust-lang/rust/pull/116494#issuecomment-2008657494, which boils down to two regressions:
* `lu_packets` - This code should have never compiled in the first place. More below.
* **ALL** other regressions are due to `commit_verify@0.11.0-beta.1` (edit: and `commit_verify@0.10.x`) - This actually seems to be fixed in version `0.11.0-beta.5`, which is the *most* up to date version, but it's still prerelease on crates.io so I don't think cargo ends up picking `beta.5` when building dependent crates.
### `lu_packets`
Firstly, this crate uses specialization, so I think it's automatically worth breaking. However, I've minimized [the regression](https://crater-reports.s3.amazonaws.com/pr-116494-3/try%23d614ed876e31a5f3ad1d0fbf848fcdab3a29d1d8/gh/lcdr.lu_packets/log.txt) to:
```rust
// Upstream crate
pub trait Serialize {}
impl Serialize for &() {}
impl<S> Serialize for &[S] where for<'a> &'a S: Serialize {}
// ----------------------------------------------------------------------- //
// Downstream crate
#![feature(specialization)]
#![allow(incomplete_features, unused)]
use upstream::Serialize;
trait Replica {
fn serialize();
}
impl<T> Replica for T {
default fn serialize() {}
}
impl<T> Replica for Option<T>
where
for<'a> &'a T: Serialize,
{
fn serialize() {}
}
```
Specifically this fails when computing the specialization graph for the `downstream` crate.
The code ends up cycling on `&[?0]: Serialize` when we equate `&?0 = &[?1]` during impl matching, which ends up needing to prove `&[?1]: Serialize`, which since cycles are treated like ambiguity, ends up in a **fatal overflow**. For some reason this requires two crates, squashing them into one crate doesn't work.
Side-note: This code is subtly order dependent. When minimizing, I ended up having the code start failing on `nightly` very easily after removing and reordering impls. This seems to me all the more reason to remove this behavior altogether.
## Side-note: Item Bounds (edit: this was fixed independently in #121123)
Due to the changes in #120584 where we now consider an alias's item bounds *and* all the item bounds of the alias's nested self type aliases, I've had to add e6b64c61941120f734657106ae2479d05b463197 which is a hack to make sure we're not eagerly normalizing bounds that have nothing to do with the predicate we're trying to solve, and which result in.
This is fixed in a more principled way in #121123.
---
r? lcnr for an initial review
Rollup of 8 pull requests
Successful merges:
- #114009 (compiler: allow transmute of ZST arrays with generics)
- #122195 (Note that the caller chooses a type for type param)
- #122651 (Suggest `_` for missing generic arguments in turbofish)
- #122784 (Add `tag_for_variant` query)
- #122839 (Split out `PredicatePolarity` from `ImplPolarity`)
- #122873 (Merge my contributor emails into one using mailmap)
- #122885 (Adjust better spastorino membership to triagebot's adhoc_groups)
- #122888 (add a couple more tests)
r? `@ghost`
`@rustbot` modify labels: rollup
clean up `Sized` checking
This PR cleans up `sized_constraint` and related functions to make them simpler and faster. This should not make more or less code compile, but it can change error output in some rare cases.
## enums and unions are `Sized`, even if they are not WF
The previous code has some special handling for enums, which made them sized if and only if the last field of each variant is sized. For example given this definition (which is not WF)
```rust
enum E<T1: ?Sized, T2: ?Sized, U1: ?Sized, U2: ?Sized> {
A(T1, T2),
B(U1, U2),
}
```
the enum was sized if and only if `T2` and `U2` are sized, while `T1` and `T2` were ignored for `Sized` checking. After this PR this enum will always be sized.
Unsized enums are not a thing in Rust and removing this special case allows us to return an `Option<Ty>` from `sized_constraint`, rather than a `List<Ty>`.
Similarly, the old code made an union defined like this
```rust
union Union<T: ?Sized, U: ?Sized> {
head: T,
tail: U,
}
```
sized if and only if `U` is sized, completely ignoring `T`. This just makes no sense at all and now this union is always sized.
## apply the "perf hack" to all (non-error) types, instead of just type parameters
This "perf hack" skips evaluating `sized_constraint(adt): Sized` if `sized_constraint(adt): Sized` exactly matches a predicate defined on `adt`, for example:
```rust
// `Foo<T>: Sized` iff `T: Sized`, but we know `T: Sized` from a predicate of `Foo`
struct Foo<T /*: Sized */>(T);
```
Previously this was only applied to type parameters and now it is applied to every type. This means that for example this type is now always sized:
```rust
// Note that this definition is WF, but the type `S<T>` not WF in the global/empty ParamEnv
struct S<T>([T]) where [T]: Sized;
```
I don't anticipate this to affect compile time of any real-world program, but it makes the code a bit nicer and it also makes error messages a bit more consistent if someone does write such a cursed type.
## tuples are sized if the last type is sized
The old solver already has this behavior and this PR also implements it for the new solver and `is_trivially_sized`. This makes it so that tuples work more like a struct defined like this:
```rust
struct TupleN<T1, T2, /* ... */ Tn: ?Sized>(T1, T2, /* ... */ Tn);
```
This might improve the compile time of programs with large tuples a little, but is mostly also a consistency fix.
## `is_trivially_sized` for more types
This function is used post-typeck code (borrowck, const eval, codegen) to skip evaluating `T: Sized` in some cases. It will now return `true` in more cases, most notably `UnsafeCell<T>` and `ManuallyDrop<T>` where `T.is_trivially_sized`.
I'm anticipating that this change will improve compile time for some real world programs.
Without doing so we use the same candidate cache entry
for `?0: Trait<?1>` and `?0: Trait<?0>`. These goals are different
and we must not use the same entry for them.
Overhaul `Diagnostic` and `DiagnosticBuilder`
Implements the first part of https://github.com/rust-lang/compiler-team/issues/722, which moves functionality and use away from `Diagnostic`, onto `DiagnosticBuilder`.
Likely follow-ups:
- Move things around, because this PR was written to minimize diff size, so some things end up in sub-optimal places. E.g. `DiagnosticBuilder` has impls in both `diagnostic.rs` and `diagnostic_builder.rs`.
- Rename `Diagnostic` as `DiagInner` and `DiagnosticBuilder` as `Diag`.
r? `@davidtwco`
There are lots of functions that modify a diagnostic. This can be via a
`&mut Diagnostic` or a `&mut DiagnosticBuilder`, because the latter type
wraps the former and impls `DerefMut`.
This commit converts all the `&mut Diagnostic` occurrences to `&mut
DiagnosticBuilder`. This is a step towards greatly simplifying
`Diagnostic`. Some of the relevant function are made generic, because
they deal with both errors and warnings. No function bodies are changed,
because all the modifier methods are available on both `Diagnostic` and
`DiagnosticBuilder`.
Ignore own item bounds in parent alias types in `for_each_item_bound`
Fixes#120912
I want to get a vibe check on this approach, which is very obviously a hack, but I believe something that is forwards-compatible with a more thorough solution and "good enough for now".
The problem here is that for a really deep rigid associated type, we are now repeatedly considering unrelated item bounds from the parent alias types, meaning we're doing a *lot* of extra work in the MIR inliner for deeply substituted rigid projections.
This feels intimately related to #107614. In that PR, we split *supertrait* bounds (bound which share the same `Self` type as the predicate which is being elaborated) and *implied* bounds (anything that is implied by elaborating the predicate).
The problem here is related to the fact that we don't maintain the split between these two for `item_bounds`. If we did, then when recursing into a parent alias type, we'd want to consider only the bounds that are given by [`PredicateFilter::All`](https://doc.rust-lang.org/nightly/nightly-rustc/rustc_hir_analysis/astconv/enum.PredicateFilter.html#variant.SelfOnly) **except** those given by [`PredicateFilter::SelfOnly`](https://doc.rust-lang.org/nightly/nightly-rustc/rustc_hir_analysis/astconv/enum.PredicateFilter.html#variant.SelfOnly).
Merge `impl_polarity` and `impl_trait_ref` queries
Hopefully this is perf neutral. I want to finish https://github.com/rust-lang/rust/pull/120835 and stop using the HIR in `coherent_trait`, which should then give us a perf improvement.
Dejargonize `subst`
In favor of #110793, replace almost every occurence of `subst` and `substitution` from rustc codes, but they still remains in subtrees under `src/tools/` like clippy and test codes (I'd like to replace them after this)
Remove unused args from functions
`#[instrument]` suppresses the unused arguments from a function, *and* suppresses unused methods too! This PR removes things which are only used via `#[instrument]` calls, and fixes some other errors (privacy?) that I will comment inline.
It's possible that some of these arguments were being passed in for the purposes of being instrumented, but I am unconvinced by most of them.
