Propagate coercion cause into `try_coerce`
Currently, `coerce_inner` discards its `ObligationCause`
when calling `try_coerce`. This interfers with other
diagnostc improvements I'm working on, since we will lose
the original span by the time the actual coercion occurs.
Additionally, we now use the span of the trailing expression
(rather than the span of the entire function) when performing
a coercion in `check_return_expr`. This currently has no visible
effect on any of the unit tests, but will unblock future
diagnostic improvements.
Start block is not allowed to have basic block predecessors
* The MIR validator is extended to detect potential violations.
* The start block has no predecessors after building MIR, so no changes are required there.
* The SimplifyCfg could previously violate this requirement when collapsing goto chains, so transformation is disabled for the start block, which also substantially simplifies the implementation.
* The LLVM function entry block also must not have basic block predecessors. Previously, to ensure that code generation had to perform necessary adjustments. This now became unnecessary.
The motivation behind the change is to align with analogous requirement in LLVM, and to avoid potential latent bugs like the one reported in #88043.
Disable the evaluation cache when in intercrate mode
It's possible to use the same `InferCtxt` with both
an intercrate and non-intercrate `SelectionContext`. However,
the local (inferctxt) evaluation cache is not aware of this
distinction, so this kind of `InferCtxt` re-use will pollute
the cache wth bad results.
This commit avoids the issue by disabling the evaluation cache
entirely during intercrate mode.
Simplify lazy DefPathHash decoding by using an on-disk hash table.
This PR simplifies the logic around mapping `DefPathHash` values encountered during incremental compilation to valid `DefId`s in the current session. It is able to do so by using an on-disk hash table encoding that allows for looking up values directly, i.e. without deserializing the entire table.
The main simplification comes from not having to keep track of `DefPathHashes` being used during the compilation session.
Skip single use lifetime lint for generated opaque types
Fix: #77175
The opaque type generated by the desugaring process of an async function uses the lifetimes defined by the originating function. The DefId for the lifetimes in the opaque type are different from the ones in the originating async function - as they should be, as far as I understand, and could therefore be considered a single use lifetimes, this causes the single_use_lifetimes lint to fail compilation if explicitly denied. This fix skips the lint for lifetimes used only once in generated opaque types for an async function that are declared in the parent async function definition.
More info in the comments on the original issue: 1 and 2
Use explicit log level in tracing instrument macro
Specify a log level in tracing instrument macro explicitly.
Additionally reduce the used log level from a default info level to a
debug level (all of those appear to be developer oriented logs, so there
should be no need to include them in release builds).
Move the Lock into symbol::Interner
This makes it easier to make the symbol interner (near) lock free in case of concurrent accesses in the future.
With https://github.com/rust-lang/rust/pull/87867 landed this shouldn't affect performance anymore.
Fast reject for NeedsNonConstDrop
Hopefully fixes the regression in #88558.
I've always wanted to help with the performance of rustc, but it doesn't feel the same when you are fixing a regression caused by your own PR...
r? `@oli-obk`
If any block on a goto chain has more than one predecessor, then the new
start block would have basic block predecessors.
Skip the transformation for the start block altogether, to avoid
violating the new invariant that the start block does not have any basic
block predecessors.
inline(always) on check_recursion_limit
r? `@oli-obk`
#88558 caused a regression, this PR adds `#[inline(always)]` to `check_recursion_limit`, a possible suspect of that regression.
Extend the `DepthFirstSearch` iterator so that it can be re-used and
extended with add'l start nodes. Then replace the FxHashSets of nodes
we were using in the fallback analysis with a single iterator. This
way we won't re-walk portions of the graph that are reached more than
once, and we also do less allocation etc.
Instead, we now record those type variables that are the target of a
`NeverToAny` adjustment and consider those to be the "diverging" type
variables. This allows us to remove the special case logic that
creates a type variable for `!` in coercion.
The comment seems incorrect. Testing revealed that the examples in
question still work (as well as some variants) even without the
special casing here.
We now fallback type variables using the following rules:
* Construct a coercion graph `A -> B` where `A` and `B` are unresolved
type variables or the `!` type.
* Let D be those variables that are reachable from `!`.
* Let N be those variables that are reachable from a variable not in
D.
* All variables in (D \ N) fallback to `!`.
* All variables in (D & N) fallback to `()`.
Suggest removing `#![feature]` for library features that have been stabilized
Issue: https://github.com/rust-lang/rust/issues/88802
Delayed the check if #![feature] has been used to enable lib features in a non-nightly build to occur after TyCtxt has been constructed.
Allow `panic!("{}", computed_str)` in const fn.
Special-case `panic!("{}", arg)` and translate it to `panic_display(&arg)`. `panic_display` will behave like `panic_any` in cosnt eval and behave like `panic!(format_args!("{}", arg))` in runtime.
This should bring Rust 2015 and 2021 to feature parity in terms of `const_panic`; and hopefully would unblock the stabilisation of #51999.
`@rustbot` modify labels: +T-compiler +T-libs +A-const-eval +A-const-fn
r? `@oli-obk`
Fix handling of `hir::GenericArg::Infer` in `wrong_number_of_generic_args.rs`
Fixes#87563. More precisely, I have fixed the "index out of bounds" error, which is what #87563 is about. The example given there still ICEs due to running into this `todo!()`, but I'd say that this is a separate issue:
c3c0f80d60/compiler/rustc_typeck/src/astconv/mod.rs (L460-L463)
Do not issue E0071 if a type error has already been reported
Fixes#88844. A suggested fix is already included in the error message for E0412, so with my changes, E0071 is simply not emitted anymore if the type in question is a "type error". This makes sense, I think, because we cannot confidently state that something is "not a struct" if we couldn't resolve it properly; and it's unnecessary to pollute the output with this additional error message, as it is a direct consequence of the former error.
I have also addressed the issue mentioned in https://github.com/rust-lang/rust/issues/88844#issuecomment-917324856 by changing the fixed example in the documentation to more closely match the erroneous code example.
Point to closure when emitting 'cannot move out' for captured variable
Attempts to fix#87456. The error message now points to the capturing closure, but I was not able to explain _why_ the closure implements `Fn` or `FnMut` (`TypeckResults::closure_kind_origins` did not contain anything for the closure in question).
cc `@Aaron1011`
Remove concept of 'completion' from the projection cache
Fixes#88910
When we initially store a `NormalizedTy` in the projection cache,
we discard all obligations that we can (while ensuring that we
don't cause any issues with incremental compilation).
Marking a projection cache entry as 'completed' discards all
obligations associated with it. This can only cause problems,
since any obligations stored in the cache are there for a reason
(e.g. they evaluate to `EvaluatedToOkModuloRegions`).
This commit removes `complete` and `complete_normalized` entirely.
PassWrapper: handle separate Module*SanitizerPass
Change ab41eef9aca3 in LLVM split MemorySanitizerPass into
MemorySanitizerPass for functions and ModuleMemorySanitizerPass for
modules. There's a related change for ThreadSanitizerPass, and in here
since we're using a ModulePassManager I only add the module flavor of
the pass on LLVM 14.
r? `@nikic` cc `@nagisa`
Improve error message for type mismatch in generator arguments
Fixes#88653. The code example given there is invalid because the `Generator` trait (unlike the `Fn` traits) does not take the generator arguments in tupled-up form (because there can only be one argument, from my understanding). Hence, the type error in the example in #88653 is correct, because the given generator takes a `bool` argument, whereas the function's return type talks about a generator with a `(bool,)` argument.
The error message is both confusing and wrong, though: It is wrong because it displays the wrong "expected signature", and it is confusing because both the "expected" and "found" notes point at the same span. With my changes, I get the following, more helpful output:
```
error[E0631]: type mismatch in generator arguments
--> test.rs:5:22
|
5 | fn foo(bar: bool) -> impl Generator<(bool,)> {
| ^^^^^^^^^^^^^^^^^^^^^^^ expected signature of `fn((bool,)) -> _`
6 | |bar| {
| ----- found signature of `fn(bool) -> _`
```
Couple of changes to FileSearch and SearchPath
* Turn a couple of regular comments into doc comments
* Move `get_tools_search_paths` from `FileSearch` to `Session`
* Use Lrc instead of Option to avoid duplication of a `SearchPath`
Fix ICE in `improper_ctypes_definitions` lint with all-ZST transparent types
Fixes#87496. There is also another function in the same file that looks fishy, but I haven't been able to produce an ICE there, and in any case, it's not related to #87496:
fd853c00e2/compiler/rustc_lint/src/types.rs (L720-L734)
r? ```@JohnTitor```
Introduce a fast path that avoids the `debug_tuple` abstraction when deriving Debug for unit-like enum variants.
The intent here is to allow LLVM to remove the switch entirely in favor of an
indexed load from a table of constant strings, which is likely what the
programmer would write in C. Unfortunately, LLVM currently doesn't perform this
optimization due to a bug, but there is [a
patch](https://reviews.llvm.org/D109565) that fixes this issue. I've verified
that, with that patch applied on top of this commit, Debug for unit-like tuple
variants becomes a load, reducing the O(n) code bloat to O(1).
Note that inlining `DebugTuple::finish()` wasn't enough to allow LLVM to
optimize the code properly; I had to avoid the abstraction entirely. Not using
the abstraction is likely better for compile time anyway.
Part of #88793.
r? `@oli-obk`
Point at argument instead of call for their obligations
When an obligation is introduced by a specific `fn` argument, point at
the argument instead of the `fn` call if the obligation fails to be
fulfilled.
Move the information about pointing at the call argument expression in
an unmet obligation span from the `FulfillmentError` to a new
`ObligationCauseCode`.
When giving an error about an obligation introduced by a function call
that an argument doesn't fulfill, and that argument is a block, add a
span_label pointing at the innermost tail expression.
Current output:
```
error[E0425]: cannot find value `x` in this scope
--> f10.rs:4:14
|
4 | Some(x * 2)
| ^ not found in this scope
error[E0277]: expected a `FnOnce<({integer},)>` closure, found `Option<_>`
--> f10.rs:2:31
|
2 | let p = Some(45).and_then({
| ______________________--------_^
| | |
| | required by a bound introduced by this call
3 | | |x| println!("doubling {}", x);
4 | | Some(x * 2)
| | -----------
5 | | });
| |_____^ expected an `FnOnce<({integer},)>` closure, found `Option<_>`
|
= help: the trait `FnOnce<({integer},)>` is not implemented for `Option<_>`
```
Previous output:
```
error[E0425]: cannot find value `x` in this scope
--> f10.rs:4:14
|
4 | Some(x * 2)
| ^ not found in this scope
error[E0277]: expected a `FnOnce<({integer},)>` closure, found `Option<_>`
--> f10.rs:2:22
|
2 | let p = Some(45).and_then({
| ^^^^^^^^ expected an `FnOnce<({integer},)>` closure, found `Option<_>`
|
= help: the trait `FnOnce<({integer},)>` is not implemented for `Option<_>`
```
Partially address #27300. Will require rebasing on top of #88546.
Currently, `coerce_inner` discards its `ObligationCause`
when calling `try_coerce`. This interfers with other
diagnostc improvements I'm working on, since we will lose
the original span by the time the actual coercion occurs.
Additionally, we now use the span of the trailing expression
(rather than the span of the entire function) when performing
a coercion in `check_return_expr`. This currently has no visible
effect on any of the unit tests, but will unblock future
diagnostic improvements.
Rollup of 10 pull requests
Successful merges:
- #88292 (Enable --generate-link-to-definition for rustc's docs)
- #88729 (Recover from `Foo(a: 1, b: 2)`)
- #88875 (cleanup(rustc_trait_selection): remove vestigial code from rustc_on_unimplemented)
- #88892 (Move object safety suggestions to the end of the error)
- #88928 (Document the closure arguments for `reduce`.)
- #88976 (Clean up and add doc comments for CStr)
- #88983 (Allow calling `get_body_with_borrowck_facts` without `-Z polonius`)
- #88985 (Update clobber_abi list to include k[1-7] regs)
- #88986 (Update the backtrace crate)
- #89009 (Fix typo in `break` docs)
Failed merges:
r? `@ghost`
`@rustbot` modify labels: rollup
Previously "trait bounds other than `Sized` on const fn parameters are unstable"
error was used for both trait bounds (<T: Trait>) and trait objects (dyn Trait).
This was pretty confusing.
This patch adds a separeta error for trait objects: "trait objects in const fn
are unstable". The error for trait bounds is otherwise intact.
If #![feature] is used outside the nightly channel for only lib
features, the check will be delayed to the stability pass after
parsing. This is done so that appropriate help messages can be shown if
the #![feature] has been used needlessly
Allow calling `get_body_with_borrowck_facts` without `-Z polonius`
For my [static analysis tool](https://github.com/willcrichton/flowistry), I need to access the set of outlives-constraints. Recently, #86977 merged a way to access these facts via Polonius. However, the merged implementation requires `-Z polonius` to be provided to use this feature. This uses Polonius for borrow checking on the entire crate, which as described [here](https://rust-lang.zulipchat.com/#narrow/stream/186049-t-compiler.2Fwg-polonius/topic/Polonius.20performance.20in.20a.20rustc.20plugin/near/251301631), is very slow.
This PR allows `get_body_with_borrowck_facts` to be called without `-Z polonius`. This is essential for my tool to run in a sensible length of time. This is a temporary patch as the Polonius-related APIs develop -- I can update my code as future changes happen.
Additionally, this PR also makes public two APIs that were previously public but then became private after `rustc_mir` got broken up: `rustc_mir_dataflow::framework::graphviz` and `rustc_mir_transform::MirPass`. I need both of these for my analysis tool. (I can break this change into a separate PR if necessary.)
cleanup(rustc_trait_selection): remove vestigial code from rustc_on_unimplemented
This isn't allowed by the validator, and seems to be unused.
When it was added in ed10a3faae,
it was used on `Sized`, and that usage is gone.
Recover from `Foo(a: 1, b: 2)`
Detect likely `struct` literal using parentheses as delimiters and emit
targeted suggestion instead of type ascription parse error.
Fix#61326.
In some cases, we emit borrowcheck diagnostics pointing
at a particular field expression in a struct expression
(e.g. `MyStruct { field: my_expr }`). However, this
behavior currently relies on us choosing the
`ConstraintCategory::Boring` with the 'correct' span.
When adding additional variants to `ConstraintCategory`,
(or changing existing usages away from `ConstraintCategory::Boring`),
the current behavior can easily get broken, since a non-boring
constraint will get chosen over a boring one.
To make the diagnostic output less fragile, this commit
adds a `ConstraintCategory::Usage` variant. We use this variant
for the temporary assignments created for each field of
an aggregate we are constructing.
Using this new variant, we can emit a message mentioning
"this usage", emphasizing the fact that the error message
is related to the specific use site (in the struct expression).
This is preparation for additional work on improving NLL error messages
(see #57374)
Disable RemoveZsts in generators to avoid query cycles
Querying layout of a generator requires its optimized MIR. Thus
computing layout during MIR optimization of a generator might create a
query cycle. Disable RemoveZsts in generators to avoid the issue
(similar approach is used in ConstProp transform already).
Fixes#88972.
Change ab41eef9aca3 in LLVM split MemorySanitizerPass into
MemorySanitizerPass for functions and ModuleMemorySanitizerPass for
modules. There's a related change for ThreadSanitizerPass, and in here
since we're using a ModulePassManager I only add the module flavor of
the pass on LLVM 14.
r? @nikic cc @nagisa
When evaluating an `ExprKind::Call`, we first have to `check_expr` on it's
callee. When this one is a `ExprKind::Path`, we had to evaluate the bounds
introduced for its arguments, but by the time we evaluated them we no
longer had access to the argument spans. Now we special case this so
that we can point at the right place on unsatisfied bounds. This also
allows the E0277 deduplication to kick in correctly, so we now emit
fewer errors.
When giving an error about an obligation introduced by a function call
that an argument doesn't fulfill, and that argument is a block, add a
span_label pointing at the innermost tail expression.
Move the information about pointing at the call argument expression in
an unmet obligation span from the `FulfillmentError` to a new
`ObligationCauseCode`.
Add non_exhaustive_omitted_patterns lint related to rfc-2008-non_exhaustive
Fixes: #84332
This PR adds `non_exhaustive_omitted_patterns`, an allow by default lint that is triggered when a `non_exhaustive` type is missing explicit patterns. The warning or deny attribute can be put above the wildcard `_` pattern on enums or on the expression for enums or structs. The lint is capable of warning about multiple types within the same pattern. This lint will not be triggered for `if let ..` patterns.
```rust
// crate A
#[non_exhaustive]
pub struct Foo {
a: u8,
b: usize,
}
#[non_exhaustive]
pub enum Bar {
A(Foo),
B,
}
// crate B
#[deny(non_exhaustive_omitted_patterns)] // here
match Bar::B {
Bar::B => {}
#[deny(non_exhaustive_omitted_patterns)] // or here
_ => {}
}
#[warn(non_exhaustive_omitted_patterns)] // only here
let Foo { a, .. } = Foo::default();
#[deny(non_exhaustive_omitted_patterns)]
match Bar::B {
// triggers for Bar::B, and Foo.b
Bar::A(Foo { a, .. }) => {}
// if the attribute was here only Bar::B would cause a warning
_ => {}
}
```
When a macro is used in the trailing expression position of a block
(e.g. `fn foo() { my_macro!() }`), we currently parse it as an
expression, rather than a statement. As a result, we ended up
using the `NodeId` of the containing statement as our `lint_node_id`,
even though we don't normally do this for macro calls.
If such a macro expands to an expression with a `#[cfg]` attribute,
then the trailing statement can get removed entirely. This lead to
an ICE, since we were usng the `NodeId` of the expression to emit
a lint.
Ths commit makes us skip updating `lint_node_id` when handling
a macro in trailing expression position. This will cause us to
lint at the closest parent of the macro call.
It's possible to use the same `InferCtxt` with both
an intercrate and non-intercrate `SelectionContext`. However,
the local (inferctxt) evaluation cache is not aware of this
distinction, so this kind of `InferCtxt` re-use will pollute
the cache wth bad results.
This commit avoids the issue by disabling the evaluation cache
entirely during intercrate mode.
Highlight the `const fn` if error happened because of a bound on the impl block
Currently, for the following code, the compiler produces the errors like the
following:
```rust
struct Type<T>(T);
impl<T: Clone> Type<T> {
const fn f() {}
}
```
```text
error[E0658]: trait bounds other than `Sized` on const fn parameters are unstable
--> ./test.rs:3:6
|
3 | impl<T: Clone> Type<T> {
| ^
|
= note: see issue #57563 <https://github.com/rust-lang/rust/issues/57563> for more information
= help: add `#![feature(const_fn_trait_bound)]` to the crate attributes to enable
```
This can be confusing (especially to newcomers) since the error mentions "const fn parameters", but highlights only the impl.
This PR adds function highlighting, changing the error to the following:
```text
error[E0658]: trait bounds other than `Sized` on const fn parameters are unstable
--> ./test.rs:3:6
|
3 | impl<T: Clone> Type<T> {
| ^
4 | pub const fn f() {}
| ---------------- function declared as const here
|
= note: see issue #57563 <https://github.com/rust-lang/rust/issues/57563> for more information
= help: add `#![feature(const_fn_trait_bound)]` to the crate attributes to enable
```
---
I've originally wanted to point directly to `const` token, but couldn't find a way to get it's span. It seems like this span is lost during the AST -> HIR lowering.
Also, since the errors for object casts in `const fn`s (`&T` -> `&dyn Trait`) seem to trigger the same error, this PR accidentally changes these errors too. Not sure if it's desired or how to fix this.
P.S. it's my first time contributing to diagnostics, so feedback is very appreciated!
---
r? ```@estebank```
```@rustbot``` label: +A-diagnostics
Revert anon union parsing
Revert PR #84571 and #85515, which implemented anonymous union parsing in a manner that broke the context-sensitivity for the `union` keyword and thus broke stable Rust code.
Fix#88583.
Accept `m!{ .. }.method()` and `m!{ .. }?` statements.
This PR fixes something that I keep running into when using `quote!{}.into()` in a proc macro to convert the `proc_macro2::TokenStream` to a `proc_macro::TokenStream`:
Before:
```
error: expected expression, found `.`
--> src/lib.rs:6:6
|
4 | quote! {
5 | ...
6 | }.into()
| ^ expected expression
```
After:
```
```
(No output, compiles fine.)
---
Context:
For expressions like `{ 1 }` and `if true { 1 } else { 2 }`, we accept them as full statements without a trailing `;`, which means the following is not accepted:
```rust
{ 1 } - 1 // error
```
since that is parsed as two statements: `{ 1 }` and `-1`. Syntactically correct, but the type of `{ 1 }` should be `()` as there is no `;`.
However, for specifically `.` and `?` after the `}`, we do [continue parsing it as an expression](13db8440bb/compiler/rustc_parse/src/parser/expr.rs (L864-L876)):
```rust
{ "abc" }.len(); // ok
```
For braced macro invocations, we do not do this:
```rust
vec![1, 2, 3].len(); // ok
vec!{1, 2, 3}.len(); // error
```
(It parses `vec!{1, 2, 3}` as a full statement, and then complains about `.len()` not being a valid expression.)
This PR changes this to also look for a `.` and `?` after a braced macro invocation. We can be sure the macro is an expression and not a full statement in those cases, since no statement can start with a `.` or `?`.
Introduce -Z remap-cwd-prefix switch
This switch remaps any absolute paths rooted under the current
working directory to a new value. This includes remapping the
debug info in `DW_AT_comp_dir` and `DW_AT_decl_file`.
Importantly, this flag does not require passing the current working
directory to the compiler, such that the command line can be
run on any machine (with the same input files) and produce the
same results. This is critical property for debugging compiler
issues that crop up on remote machines.
This is based on adetaylor's dbc4ae7cba
Major Change Proposal: https://github.com/rust-lang/compiler-team/issues/450
Discussed on #38322. Would resolve issue #87325.
Specify a log level in tracing instrument macro explicitly.
Additionally reduce the used log level from a default info level to a
debug level (all of those appear to be developer oriented logs, so there
should be no need to include them in release builds).
Querying layout of a generator requires its optimized MIR. Thus
computing layout during MIR optimization of a generator might create a
query cycle. Disable RemoveZsts in generators to avoid the issue
(similar approach is used in ConstProp transform already).
Use a separate interner type for UniqueTypeId
Using symbol::Interner makes it very easy to mixup UniqueTypeId symbols
with the global interner. In fact the Debug implementation of
UniqueTypeId did exactly this.
Using a separate interner type also avoids prefilling the interner with
unused symbols and allow for optimizing the symbol interner for parallel
access without negatively affecting the single threaded module codegen.
Const drop
The changes are pretty primitive at this point. But at least it works. ^-^
Problems with the current change that I can think of now:
- [x] `~const Drop` shouldn't change anything in the non-const world.
- [x] types that do not have drop glues shouldn't fail to satisfy `~const Drop` in const contexts. `struct S { a: u8, b: u16 }` This might not fail for `needs_non_const_drop`, but it will fail in `rustc_trait_selection`.
- [x] The current change accepts types that have `const Drop` impls but have non-const `Drop` glue.
Fixes#88424.
Significant Changes:
- `~const Drop` is no longer treated as a normal trait bound. In non-const contexts, this bound has no effect, but in const contexts, this restricts the input type and all of its transitive fields to either a) have a `const Drop` impl or b) can be trivially dropped (i.e. no drop glue)
- `T: ~const Drop` will not be linted like `T: Drop`.
- Instead of recursing and iterating through the type in `rustc_mir::transform::check_consts`, we use the trait system to special case `~const Drop`. See [`rustc_trait_selection::...::candidate_assembly#assemble_const_drop_candidates`](https://github.com/fee1-dead/rust/blob/const-drop/compiler/rustc_trait_selection/src/traits/select/candidate_assembly.rs#L817) and others.
Changes not related to `const Drop`ping and/or changes that are insignificant:
- `Node.constness_for_typeck` no longer returns `hir::Constness::Const` for type aliases in traits. This was previously used to hack how we determine default bound constness for items. But because we now use an explicit opt-in, it is no longer needed.
- Removed `is_const_impl_raw` query. We have `impl_constness`, and the only existing use of that query uses `HirId`, which means we can just operate it with hir.
- `ty::Destructor` now has a field `constness`, which represents the constness of the destructor.
r? `@oli-obk`
Fixes#88910
When we initially store a `NormalizedTy` in the projection cache,
we discard all obligations that we can (while ensuring that we
don't cause any issues with incremental compilation).
Marking a projection cache entry as 'completed' discards all
obligations associated with it. This can only cause problems,
since any obligations stored in the cache are there for a reason
(e.g. they evaluate to `EvaluatedToOkModuloRegions`).
This commit removes `complete` and `complete_normalized` entirely.
Add linting on non_exhaustive structs and enum variants
Add ui tests for non_exhaustive reachable lint
Rename to non_exhaustive_omitted_patterns and avoid triggering on if let
This encoding allows for random access without an expensive upfront decoding
state which in turn allows simplifying the DefPathIndex lookup logic without
regressing performance.
Enum should prefer discriminant zero for niche
Given an enum with unassigned zero-discriminant, rust should prefer it for niche selection.
Zero as discriminant for `Option<Enum>` makes it possible for LLVM to optimize resulting asm.
- Eliminate branch when expected value coincides.
- Use smaller instruction `test eax, eax` instead of `cmp eax, ?`
- Possible interaction with zeroed memory?
Example:
```rust
pub enum Size {
One = 1,
Two = 2,
Three = 3,
}
pub fn handle(x: Option<Size>) -> u8 {
match x {
None => {0}
Some(size) => {size as u8}
}
}
```
In this case discriminant zero is available as a niche.
Above example on nightly:
```asm
mov eax, edi
cmp al, 4
jne .LBB0_2
xor eax, eax
.LBB0_2:
ret
```
PR:
```asm
mov eax, edi
ret
```
I created this PR because I had a performance regression when I tried to use an enum to represent legal grapheme byte-length for utf8.
Using an enum instead of `NonZeroU8` [here](d683304f5d/src/internal/decoder_incomplete.rs (L90))
resulted in a performance regression of about 5%.
I consider this to be a somewhat realistic benchmark.
Thanks to `@ogoffart` for pointing me in the right direction!
Edit: Updated description
Improve error message for missing trait in trait impl
Fixes#88818. For the following example:
```rust
struct S { }
impl for S { }
```
the current output is:
```
error: missing trait in a trait impl
--> t1.rs:2:5
|
2 | impl for S { }
| ^
```
With my changes, I get:
```
error: missing trait in a trait impl
--> t1.rs:2:5
|
2 | impl for S { }
| ^
|
help: add a trait here
|
2 | impl Trait for S { }
| +++++
help: for an inherent impl, drop this `for`
|
2 - impl for S { }
2 + impl S { }
|
```
Fix duplicate bounds for const_trait_impl
Fixes#88383.
Compare the constness of the candidates before winnowing and removing a `~const` `BoundCandidate`.
Use smaller spans for some structured suggestions
Use more accurate suggestion spans for
* argument parse error
* fully qualified path
* missing code block type
* numeric casts
Mark all of these as locals so the debugger does not try to interpret
them as being a pointer to the value. This extends the approach used in
PR #81898.
Currently, for the following code, the compiler produces the errors like the
following error:
```rust
struct Type<T>
impl<T: Clone> Type<T> {
fn const f() {}
}
```
```text
error[E0658]: trait bounds other than `Sized` on const fn parameters are unstable
--> ./test.rs:3:6
|
3 | impl<T: Clone> Type<T> {
| ^
|
= note: see issue #57563 <https://github.com/rust-lang/rust/issues/57563> for more information
= help: add `#![feature(const_fn_trait_bound)]` to the crate attributes to enable
```
This can be confusing (especially to newcomers) since the error mentions
"const fn parameters", but highlights only the impl.
This commits adds function highlighting, changing the error to the following:
```text
error[E0658]: trait bounds other than `Sized` on const fn parameters are unstable
--> ./test.rs:3:6
|
3 | impl<T: Clone> Type<T> {
| ^
4 | pub const fn f() {}
| ---------------- function declared as const here
|
= note: see issue #57563 <https://github.com/rust-lang/rust/issues/57563> for more information
= help: add `#![feature(const_fn_trait_bound)]` to the crate attributes to enable
```
Using symbol::Interner makes it very easy to mixup UniqueTypeId symbols
with the global interner. In fact the Debug implementation of
UniqueTypeId did exactly this.
Using a separate interner type also avoids prefilling the interner with
unused symbols and allow for optimizing the symbol interner for parallel
access without negatively affecting the single threaded module codegen.
This PR adds a suggestion to replace an inexisting field for an
unmentioned field. Given the following code:
```rust
enum Foo {
Bar { alpha: u8, bravo: u8, charlie: u8 },
}
fn foo(foo: Foo) {
match foo {
Foo::Bar {
alpha,
beta, // `bravo` miswritten as `beta` here.
charlie,
} => todo!(),
}
}
```
the compiler now emits the error messages below.
```text
error[E0026]: variant `Foo::Bar` does not have a field named `beta`
--> src/lib.rs:9:13
|
9 | beta, // `bravo` miswritten as `beta` here.
| ^^^^
| |
| variant `Foo::Bar` does not have this field
| help: `Foo::Bar` has a field named `bravo`: `bravo`
```
Note that this suggestion is available iff the number of inexisting
fields and unmentioned fields are both 1.
ARMv6K Nintendo 3DS Tier 3 target added
Addition of the target specifications to build .elf files for Nintendo 3DS (ARMv6K, Horizon). Requires devkitARM 3DS toolkit for system libraries and arm-none-eabi-gcc linker.
Introduce NullOp::AlignOf
This PR introduces `Rvalue::NullaryOp(NullOp::AlignOf, ty)`, which will be lowered from `align_of`, similar to `size_of` lowering to `Rvalue::NullaryOp(NullOp::SizeOf, ty)`.
The changes are originally part of #88700 but since it's not dependent on other changes and could have performance impact on its own, it's separated into its own PR.
Add -Z panic-in-drop={unwind,abort} command-line option
This PR changes `Drop` to abort if an unwinding panic attempts to escape it, making the process abort instead. This has several benefits:
- The current behavior when unwinding out of `Drop` is very unintuitive and easy to miss: unwinding continues, but the remaining drops in scope are simply leaked.
- A lot of unsafe code doesn't expect drops to unwind, which can lead to unsoundness:
- https://github.com/servo/rust-smallvec/issues/14
- https://github.com/bluss/arrayvec/issues/3
- There is a code size and compilation time cost to this: LLVM needs to generate extra landing pads out of all calls in a drop implementation. This can compound when functions are inlined since unwinding will then continue on to process drops in the callee, which can itself unwind, etc.
- Initial measurements show a 3% size reduction and up to 10% compilation time reduction on some crates (`syn`).
One thing to note about `-Z panic-in-drop=abort` is that *all* crates must be built with this option for it to be sound since it makes the compiler assume that dropping `Box<dyn Any>` will never unwind.
cc https://github.com/rust-lang/lang-team/issues/97
Rework DepthFirstSearch API
This expands the API to be more flexible, allowing for more visitation patterns
on graphs. This will be useful to avoid extra datasets (and allocations) in
cases where the expanded DFS API is sufficient.
This also fixes a bug with the previous DFS constructor, which left the start
node not marked as visited (even though it was immediately returned).
Commit written by ```@nikomatsakis``` originally, cherry picked from several commits in work on never type stabilization, but stands alone.
generic_const_exprs: use thir for abstract consts instead of mir
Changes `AbstractConst` building to use `thir` instead of `mir` so that there's less chance of consts unifying when they shouldn't because lowering to mir dropped information (see `abstract-consts-as-cast-5.rs` test)
r? `@lcnr`
Detect stricter constraints on gats where clauses in impls vs trait
I might try to see if I can do a bit more to improve these diagnostics, but any initial feedback is appreciated. I can also do any additional work in a followup PR.
r? `@estebank`
Encode spans relative to the enclosing item
The aim of this PR is to avoid recomputing queries when code is moved without modification.
MCP at https://github.com/rust-lang/compiler-team/issues/443
This is achieved by :
1. storing the HIR owner LocalDefId information inside the span;
2. encoding and decoding spans relative to the enclosing item in the incremental on-disk cache;
3. marking a dependency to the `source_span(LocalDefId)` query when we translate a span from the short (`Span`) representation to its explicit (`SpanData`) representation.
Since all client code uses `Span`, step 3 ensures that all manipulations
of span byte positions actually create the dependency edge between
the caller and the `source_span(LocalDefId)`.
This query return the actual absolute span of the parent item.
As a consequence, any source code motion that changes the absolute byte position of a node will either:
- modify the distance to the parent's beginning, so change the relative span's hash;
- dirty `source_span`, and trigger the incremental recomputation of all code that
depends on the span's absolute byte position.
With this scheme, I believe the dependency tracking to be accurate.
For the moment, the spans are marked during lowering.
I'd rather do this during def-collection,
but the AST MutVisitor is not practical enough just yet.
The only difference is that we attach macro-expanded spans
to their expansion point instead of the macro itself.
Refactor query forcing
The control flow in those functions was very complex, with several layers of continuations.
I tried to simplify the implementation, while keeping essentially the same logic.
Now, all code paths go through `try_execute_query` for the actual query execution.
Communication with the `dep_graph` and the live caches are the only difference between query getting/ensuring/forcing.
Fix ICE for functions with more than 65535 arguments
This pull request fixes#88577 by changing the `param_idx` field in the `Param` variant of `WellFormedLoc` from `u16` to `u32`, thus allowing for more than 65,535 arguments in a function. Note that I also added a regression test, but needed to add `// ignore-tidy-filelength` because the test is more than 8000 lines long.
Change more x64 size checks to not apply to x32.
Commit 95e096d6 changed a bunch of size checks already, but more have
been added, so this fixes the new ones the same way: the various size
checks that are conditional on target_arch = "x86_64" were not intended
to apply to x86_64-unknown-linux-gnux32, so add
target_pointer_width = "64" to the conditions.
Fix non-capturing closure return type coercion
Fixes#88097. For the example given there:
```rust
fn peculiar() -> impl Fn(u8) -> u8 {
return |x| x + 1
}
```
which incorrectly reports an error, I noticed something weird in the debug log:
```
DEBUG rustc_typeck::check::coercion coercion::try_find_coercion_lub([closure@test.rs:2:12: 2:21], [closure@test.rs:2:12: 2:21], exprs=1 exprs)
```
Apparently, `try_find_coercion_lub()` thinks that the LUB for two closure types always has to be a function pointer (which explains the `expected closure, found fn pointer` error in #88097). There is one corner case where that isn't true, though — namely, when the two closure types are equal, in which case the trivial LUB is the type itself. This PR fixes this by inserting an explicit check for type equality in `try_find_coercion_lub()`.
rustc: use more correct span data in for loop desugaring
Fixes#82462
Before:
help: consider adding semicolon after the expression so its temporaries are dropped sooner, before the local variables declared by the block are dropped
|
LL | for x in DroppingSlice(&*v).iter(); {
| +
After:
help: consider adding semicolon after the expression so its temporaries are dropped sooner, before the local variables declared by the block are dropped
|
LL | };
| +
This seems like a reasonable fix: since the desugared "expr_drop_temps_mut" contains the entire desugared loop construct, its span should contain the entire loop construct as well.
Debug for unit-like enum variants.
The intent here is to allow LLVM to remove the switch entirely in favor of an
indexed load from a table of constant strings, which is likely what the
programmer would write in C. Unfortunately, LLVM currently doesn't perform this
optimization due to a bug, but there is [a
patch](https://reviews.llvm.org/D109565) that fixes this issue. I've verified
that, with that patch applied on top of this commit, Debug for unit-like tuple
variants becomes a load, reducing the O(n) code bloat to O(1).
Note that inlining `DebugTuple::finish()` wasn't enough to allow LLVM to
optimize the code properly; I had to avoid the abstraction entirely. Not using
the abstraction is likely better for compile time anyway.
Part of #88793.
Now that we encode spans relative to the items, the item's own span is
never actually hashed as part of the HIR.
In consequence, we explicitly include it in the crate hash to avoid
missing cross-crate invalidations.
RustWrapper: avoid deleted unclear attribute methods
These were deleted in https://reviews.llvm.org/D108614, and in C++ I
definitely see the argument for their removal. I didn't try and
propagate the changes up into higher layers of rustc in this change
because my initial goal was to get rustc working against LLVM HEAD
promptly, but I'm happy to follow up with some refactoring to make the
API on the Rust side match the LLVM API more directly (though the way
the enum works in Rust makes the API less scary IMO).
r? ``@nagisa`` cc ``@nikic``
Emit proper errors when on missing closure braces
This commit focuses on emitting clean errors for the following syntax
error:
```
Some(42).map(|a|
dbg!(a);
a
);
```
Previous implementation tried to recover after parsing the closure body
(the `dbg` expression) by replacing the next `;` with a `,`, which made
the next expression belong to the next function argument. As such, the
following errors were emitted (among others):
- the semicolon token was not expected,
- a is not in scope,
- Option::map is supposed to take one argument, not two.
This commit allows us to gracefully handle this situation by adding
giving the parser the ability to remember when it has just parsed a
closure body inside a function call. When this happens, we can treat the
unexpected `;` specifically and try to parse as much statements as
possible in order to eat the whole block. When we can't parse statements
anymore, we generate a clean error indicating that the braces are
missing, and return an ExprKind::Err.
Closes#88065.
r? `@estebank`
Fix stray notes when the source code is not available
Fixes#87060. To reproduce it with a local build of rustc, you have to copy the compiler (e.g. `build/x86_64-unknown-linux-gnu/stage1/`) somewhere and then rename the compiler source directory (maybe there is a smarter way as well). Then, rustc won't find the standard library sources and report stray notes such as
```
note: deref defined here
```
with no location for "here". Another example I've found is this:
```rust
use std::ops::Add;
fn foo<T: Add<Output=()>>(x: T) {
x + x;
}
fn main() {}
```
```
error[E0382]: use of moved value: `x`
--> binop.rs:4:9
|
3 | fn foo<T: Add<Output=()>>(x: T) {
| - move occurs because `x` has type `T`, which does not implement the `Copy` trait
4 | x + x;
| ----^
| | |
| | value used here after move
| `x` moved due to usage in operator
|
note: calling this operator moves the left-hand side
help: consider further restricting this bound
|
3 | fn foo<T: Add<Output=()> + Copy>(x: T) {
| ^^^^^^
error: aborting due to previous error
```
where, again, the note is supposed to point somewhere but doesn't. I have fixed this by checking whether the corresponding source code is actually available before emitting the note.
Add proc_macro::Span::{before, after}.
This adds `proc_macro::Span::before()` and `proc_macro::Span::after()` to get a zero width span at the start or end of the span.
These are equivalent to rustc's `Span::shrink_to_lo()` and `Span::shrink_to_hi()` but with a less cryptic name. They are useful when generating diagnostlics like "missing \<thing\> after \<thing\>".
E.g.
```rust
syn::Error::new(ident.span().after(), "missing `:` after field name").into_compile_error()
```
This commit focuses on emitting clean errors for the following syntax
error:
```
Some(42).map(|a|
dbg!(a);
a
);
```
Previous implementation tried to recover after parsing the closure body
(the `dbg` expression) by replacing the next `;` with a `,`, which made
the next expression belong to the next function argument. As such, the
following errors were emitted (among others):
- the semicolon token was not expected,
- a is not in scope,
- Option::map is supposed to take one argument, not two.
This commit allows us to gracefully handle this situation by adding
giving the parser the ability to remember when it has just parsed a
closure body inside a function call. When this happens, we can treat the
unexpected `;` specifically and try to parse as much statements as
possible in order to eat the whole block. When we can't parse statements
anymore, we generate a clean error indicating that the braces are
missing, and return an ExprKind::Err.
This reverts commit 059b68dd67.
Note that this was manually adjusted to retain some of the refactoring
introduced by commit 059b68dd67, so that it could
likewise retain the correction introduced in commit
5b4bc05fa5
Move *_max methods back to util
change to inline instead of inline(always)
Remove valid_range_exclusive from scalar
Use WrappingRange instead
implement always_valid_for in a safer way
Fix accidental edit
Only follow backwards edges during get_moved_indexes if the move path is
definitely initialized at loop entry. Otherwise, the error occurred prior to the
loop, so we ignore the backwards edges to avoid generating misleading "value
moved here, in previous iteration of loop" errors.
This patch also slightly improves the analysis of inits, including
NonPanicPathOnly initializations (which are ignored by
drop_flag_effects::for_location_inits). This is required for the definite
initialization analysis, but may also help find certain skipped reinits in rare
cases.
Patch passes all non-ignored src/test/ui testcases.
Split rustc_mir
The `rustc_mir` crate is the second largest in the compiler.
This PR splits it up into 5 crates:
- rustc_borrowck;
- rustc_const_eval;
- rustc_mir_dataflow;
- rustc_mir_transform;
- rustc_monomorphize.
Improve diagnostics for unary plus operators (#88276)
This pull request improves the diagnostics emitted on parsing a unary plus operator. See #88276.
Before:
```
error: expected expression, found `+`
--> src/main.rs:2:13
|
2 | let x = +1;
| ^ expected expression
```
After:
```
error: leading `+` is not supported
--> main.rs:2:13
|
2 | let x = +1;
| ^
| |
| unexpected `+`
| help: try removing the `+`
```
This expands the API to be more flexible, allowing for more visitation patterns
on graphs. This will be useful to avoid extra datasets (and allocations) in
cases where the expanded DFS API is sufficient.
This also fixes a bug with the previous DFS constructor, which left the start
node not marked as visited (even though it was immediately returned).
Suggest deriving traits if possible
This only applies to builtin derives as I don't think there is a
clean way to get the available derives in typeck.
Closes#85851
Remove `hir::GenericBound::Unsized`
Rather than "moving" the `?Sized` bounds to the param bounds, just also check where clauses in `astconv`. I also did some related cleanup here, but that's not strictly neccesary. Also going to do a perf run here.
r? `@estebank`
