| Age | Commit message (Collapse) | Author | Lines |
|---|
|
compiler-errors:associated-type-bounds-in-bad-position, r=cjgillot
Ban associated type bounds in bad positions
We should not try to lower associated type bounds into TAITs in positions where `impl Trait` is not allowed (except for in `where` clauses, like `where T: Trait<Assoc: Bound>`).
This is achieved by using the same `rustc_ast_lowering` machinery as impl-trait does to characterize positions as universal/existential/disallowed.
Fixes #106077
Split out the first commit into #108066, since it's not really related.
|
|
r=compiler-errors,davidtwco,estebank,oli-obk
diagnostics: if AssocFn has self argument, describe as method
Discussed in https://rust-lang.zulipchat.com/#narrow/stream/147480-t-compiler.2Fwg-diagnostics/topic/.22associated.20function.22.20vs.20.22method.22/near/329265515
This commit also changes the tooltips on rustdoc intra-doc links targeting methods.
For anyone not sure why this is being done, see the Reference definitions of these terms in <https://doc.rust-lang.org/1.67.1/reference/items/associated-items.html#methods>
> Associated functions whose first parameter is named `self` are called methods and may be invoked using the [method call operator](https://doc.rust-lang.org/1.67.1/reference/expressions/method-call-expr.html), for example, `x.foo()`, as well as the usual function call notation.
In particular, while this means it's technically correct for rustc to refer to a method as an associated function (and there are a few cases where it'll still do so), rustc *must never* use the term "method" to refer to an associated function that does not have a `self` parameter.
|
|
|
|
|
|
Rollup of 6 pull requests
Successful merges:
- #107736 ( Rename atomic 'as_mut_ptr' to 'as_ptr' to match Cell (ref #66893) )
- #108176 (Don't delay `ReError` bug during lexical region resolve)
- #108315 (Lint dead code in closures and generators)
- #108342 (apply query response: actually define opaque types)
- #108344 (Fix test filename for #105700)
- #108353 (resolve: Remove `ImportResolver`)
Failed merges:
- #107911 (Add check for invalid #[macro_export] arguments)
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
This makes the check for when associated type bounds more accurate
|
|
Fix test filename for #105700
The test is for #105700 rather than #21102
|
|
Lint dead code in closures and generators
Fixes #108296
I think this might be a potentially breaking change, but restores the behaviour of pre-1.64.
`@rustbot` label +A-lint
|
|
r=oli-obk
Don't delay `ReError` bug during lexical region resolve
Lexical region resolution returns a list of `RegionResolutionError` which don't necessarily correspond to diagnostics being emitted. The compiler may, validly, throw away these resolution errors and do something else. Therefore it's not valid to use `ReError` during lifetime resolution, since we may actually be on a totally fine compilation path.
For example, the `implied_bounds_entailment` lint runs region resolution twice, and only emits an error if it fails both times. If we delay a bug and create a `ReError` during this first run, then we will ICE.
Fixes #108170
----
Side-note: this is conceptually equivalent to how we can't necessarily delay bugs or create `ty::Error` during trait solving/fulfillment, since the compiler is allowed to throw away these fulfillment errors to do other things. It's only once we actually emit an error (`report_region_errors` / `report_fulfillment_errors`)
|
|
|
|
|
|
|
|
|
|
|
|
Fix overlapping spans in removing extra arguments
Fixes #108225
Each span is already extended to include the previous comma, so extending to the *next* comma is unecessary and causes an ICE with assertions on.
``@rustbot`` label +A-diagnostics
|
|
Convert a hard-warning about named static lifetimes into lint "unused_lifetimes"
Fixes https://github.com/rust-lang/rust/issues/96956.
Some changes are ported from https://github.com/rust-lang/rust/pull/98079, thanks to jeremydavis519.
r? `@estebank` `@petrochenkov`
Any feedback is appreciated!
## Actions
- [x] resolve conflicts
- [x] fix build
- [x] address review comments in last pr
- [x] update tests
|
|
Fix ICE on type alias in recursion
Fixes #108160
|
|
Define the `named_static_lifetimes` lint
This lint will replace the existing hard-warning.
Replace the named static lifetime hard-warning with the new lint
Update the UI tests for the `named_static_lifetimes` lint
Remove the direct dependency on `rustc_lint_defs`
fix build
Signed-off-by: Zhi Qi <qizhi@pingcap.com>
use "UNUSED_LIFETIMES" instead
Signed-off-by: Zhi Qi <qizhi@pingcap.com>
update 1 test and fix typo
Signed-off-by: Zhi Qi <qizhi@pingcap.com>
update tests
Signed-off-by: Zhi Qi <qizhi@pingcap.com>
fix tests: add extra blank line
Signed-off-by: Zhi Qi <qizhi@pingcap.com>
|
|
Use DefKind to give more item kind information during BindingObligation note
The current label says "required by a bound in this". When I see that label, my immediate impression is "this... **what**?". It feels like it was cut short.
Alternative to this would be saying "in this item", but adding the item kind is strictly more informational and adds very little overhead to the existing error presentation.
|
|
compiler-errors:non_lifetime_binders-type-match-ice, r=davidtwco
Make sure `test_type_match` doesn't ICE with late-bound types
Fixes #108190 (in a kind of hacky way, anyways doesn't really matter)
|
|
Make hidden type registration opt-in, so that each site can be reviewed on its own and we have the right defaults for trait solvers
r? `@lcnr`
pulled out of https://github.com/rust-lang/rust/pull/107891 as it is the uncontroversial part
|
|
its own and we have the right defaults for trait solvers
|
|
|
|
Move `Fn*` traits malformedness protections to typeck
I found it strange that we were doing a custom well-formedness check just for the `Fn*` traits' `call_*` fn items. My understanding from the git history is that this is just to avoid ICEs later on in typeck.
Well, that well-formedness check isn't even implemented correctly for `FnOnce::call_once`, or `FnMut::call_mut` for that matter. Instead, this PR just makes the typeck checks more robust, and leaves it up to the call-site to report errors when lang items are implemented in funny ways.
This coincidentally fixes another ICE where a the `Add` lang item is implemented with a `add` item that's a const instead of a method.
|
|
Rollup of 7 pull requests
Successful merges:
- #108000 (lint: don't suggest MaybeUninit::assume_init for uninhabited types)
- #108105 (Explain the default panic hook better)
- #108141 (Add rpitit queries)
- #108272 (docs: wrong naming convention in struct keyword doc)
- #108285 (remove unstable `pick_stable_methods_before_any_unstable` flag)
- #108289 (Name placeholder in some region errors)
- #108290 (Add a test for default trait method with RPITITs)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
r=oli-obk
Add a test for default trait method with RPITITs
This didn't work in #107013, but now that #108203 has landed, let's make sure we don't regress it.
r? types
|
|
Name placeholder in some region errors
Also don't print `ReVar` or `ReLateBound` as debug... these error messages are super uncommon anyways, but in the case they do trigger, let's be slightly more helpful.
|
|
lint: don't suggest MaybeUninit::assume_init for uninhabited types
Creating a zeroed uninhabited type such as `!` or an empty enum with `mem::zeroed()` (or transmuting `()` to `!`) currently triggers this lint:
```rs
warning: the type `!` does not permit zero-initialization
--> test.rs:5:23
|
5 | let _val: ! = mem::zeroed();
| ^^^^^^^^^^^^^
| |
| this code causes undefined behavior when executed
| help: use `MaybeUninit<T>` instead, and only call `assume_init` after initialization is done
|
= note: the `!` type has no valid value
```
The `MaybeUninit` suggestion in the help message seems confusing/useless for uninhabited types, as such a type cannot be fully initialized in the first place (as the note implies).
This PR limits this help message to inhabited types which can be initialized
|
|
Use `ThinVec` more in the AST
r? `@ghost`
|
|
|
|
Miri: basic dyn* support
As usual I am very unsure about the dynamic dispatch stuff, but it passes even the `Pin<&mut dyn* Trait>` test so that is something.
TBH I think it was a mistake to make `dyn Trait` and `dyn* Trait` part of the same `TyKind` variant. Almost everywhere in Miri this lead to the wrong default behavior, resulting in strange ICEs instead of nice "unimplemented" messages. The two types describe pretty different runtime data layout after all.
Strangely I did not need to do the equivalent of [this diff](https://github.com/rust-lang/rust/pull/106532#discussion_r1087095963) in Miri. Maybe that is because the unsizing logic matches on `ty::Dynamic(.., ty::Dyn)` already? In `unsized_info` I don't think the `target_dyn_kind` can be `DynStar`, since then it wouldn't be unsized!
r? `@oli-obk` Cc `@eholk` (dyn-star) https://github.com/rust-lang/rust/issues/102425
|
|
|
|
give the resolver access to TyCtxt
The resolver is now created after TyCtxt is created. Then macro expansion and name resolution are run and the results fed into queries just like before this PR.
Since the resolver had (before this PR) mutable access to the `CStore` and the source span table, these two datastructures are now behind a `RwLock`. To ensure that these are not mutated anymore after the resolver is done, a read lock to them is leaked right after the resolver finishes.
### PRs split out of this one and leading up to it:
* https://github.com/rust-lang/rust/pull/105423
* https://github.com/rust-lang/rust/pull/105357
* https://github.com/rust-lang/rust/pull/105603
* https://github.com/rust-lang/rust/pull/106776
* https://github.com/rust-lang/rust/pull/106810
* https://github.com/rust-lang/rust/pull/106812
* https://github.com/rust-lang/rust/pull/108032
|
|
|
|
|
|
|
|
This commit changes the sequence parsers to produce `ThinVec`, which
triggers numerous conversions.
|
|
|
|
|
|
|
|
r=WaffleLapkin
Refine error span for trait error into borrowed expression
Extends the error span refinement in #106477 to drill into borrowed expressions just like tuples/struct/enum literals. For example,
```rs
trait Fancy {}
trait Good {}
impl <'a, T> Fancy for &'a T where T: Good {}
impl <S> Good for Option<S> where S: Iterator {}
fn want_fancy<F>(f: F) where F: Fancy {}
fn example() {
want_fancy(&Some(5));
// (BEFORE) ^^^^^^^^ `{integer}` is not an iterator
// (AFTER) ^ `{integer}` is not an iterator
}
```
Existing heuristics try to find the right part of the expression to "point at"; current heuristics look at e.g. struct constructors and tuples. This PR adds a new check for borrowed expressions when looking into a borrowed type.
|
|
Rollup of 5 pull requests
Successful merges:
- #108124 (Document that CStr::as_ptr returns a type alias)
- #108171 (Improve building compiler artifacts output)
- #108200 (Use restricted Damerau-Levenshtein distance for diagnostics)
- #108259 (remove FIXME that doesn't require fixing)
- #108265 ("`const` generic" -> "const parameter")
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
|
|
|
|
|
|
|
|
"`const` generic" -> "const parameter"
|
|
r=tmiasko
Use restricted Damerau-Levenshtein distance for diagnostics
This replaces the existing Levenshtein algorithm with the Damerau-Levenshtein algorithm. This means that "ab" to "ba" is one change (a transposition) instead of two (a deletion and insertion). More specifically, this is a _restricted_ implementation, in that "ca" to "abc" cannot be performed as "ca" → "ac" → "abc", as there is an insertion in the middle of a transposition. I believe that errors like that are sufficiently rare that it's not worth taking into account.
This was first brought up [on IRLO](https://internals.rust-lang.org/t/18227) when it was noticed that the diagnostic for `prinltn!` (transposed L and T) was `print!` and not `println!`. Only a single existing UI test was effected, with the result being an objective improvement.
~~I have left the method name and various other references to the Levenshtein algorithm untouched, as the exact manner in which the edit distance is calculated should not be relevant to the caller.~~
r? ``@estebank``
``@rustbot`` label +A-diagnostics +C-enhancement
|
|
create dummy placeholder crate to prevent compiler from panicing
This PR is to address the panic found in https://github.com/rust-lang/rust/issues/105700.
There are 2 separate things going on with this panic.
First the code could not generate a dummy response for crate fragment types when it hits the recursion limit.
This PR adds the method to the trait implementation for `DymmyResult` to be able to create a dummy crate node.
This stops the panic from happening.
The second thing that is not addressed (and maybe does not need addressing? 🤷🏻)
is that when you have multiple attributes it ends up treating attributes that follow another as being the result of expanding the former (maybe there is a better way to say that). So you end up hitting the recursion limit. Even though you would think there is no expansion happening here.
If you did not hit the recursion limit the compiler would output that `invalid_attribute` does not exists. But it currently exits before the resolution step when the recursion limit is reached here.
|
|
|
|
Use covariance on type relations of field projection types if possible
It's fine to use covariance here unless we're in a mutating context.
Fixes https://github.com/rust-lang/rust/issues/96514
Supersedes https://github.com/rust-lang/rust/pull/105958
r? `@lcnr`
|
