rustc_target/loongarch: Fix passing of transparent unions with only one non-ZST member
This ensures that `MaybeUninit<T>` has the same ABI as `T` when passed through an `extern "C"` function.
Fixes https://github.com/rust-lang/rust/issues/115509
r? `@bjorn3`
coverage: Remove debug code from the instrumentor
The coverage instrumentor has an entire module full of complex code that is only used for debugging.
And as I continue to work on coverage, I keep finding that this debug code is constantly causing more trouble than it's worth. It's deeply entangled with current implementation details, such that making any non-trivial change to the instrumentor usually requires major changes to the debug code. And so far I have personally not found any of this debug code to be *useful*.
In light of that situation, I'd like to try just ripping all of it out. If I spend any more time dealing with coverage debug code, I want it to be because I'm writing new and useful tools, not dutifully maintaining a boat-anchor that quite plausibly isn't being used by anyone at all.
---
r? `@ghost`
`@rustbot` label +A-code-coverage
---
[Zulip thread](https://rust-lang.zulipchat.com/#narrow/stream/131828-t-compiler/topic/Removing.20debug.20code.20from.20the.20coverage.20instrumentor)
Correctly deny late-bound lifetimes from parent in anon consts and TAITs
Reuse the `AnonConstBoundary` scope (introduced in #108553, renamed in this PR to `LateBoundary`) to deny late-bound vars of *all* kinds (ty/const/lifetime) in anon consts and TAITs.
Side-note, but I would like to consolidate this with the error reporting for RPITs (E0657):
c4f25777a0/compiler/rustc_hir_analysis/src/collect/resolve_bound_vars.rs (L733-L754) but the semantics about what we're allowed to capture there are slightly different, so I'm leaving that untouched.
Fixes#115474
Custom code classes in docs warning
Fixes https://github.com/rust-lang/rust/issues/115938.
This PR does two things:
1. Unless the `custom_code_classes_in_docs` feature is enabled, it will use the old codeblock tag parser.
2. If there is a codeblock tag that starts with a `.`, it will emit a behaviour change warning.
Hopefully this is the last missing part for this feature until stabilization.
Follow-up of https://github.com/rust-lang/rust/pull/110800.
r? `@notriddle`
Detect cycle errors hidden by opaques during monomorphization
Opaque types may reveal to projections, which themselves normalize to opaques. We don't currently normalize when checking that opaques are cyclical, and we may also not know that the opaque is cyclical until monomorphization (see `tests/ui/type-alias-impl-trait/mututally-recursive-overflow.rs`).
Detect cycle errors in `normalize_projection_ty` and report a fatal overflow (in the old solver). Luckily, this is already detected as a fatal overflow in the new solver.
Fixes#112047
rustc_target/riscv: Fix passing of transparent unions with only one non-ZST member
This ensures that `MaybeUninit<T>` has the same ABI as `T` when passed through an `extern "C"` function.
Fixes https://github.com/rust-lang/rust/issues/115481.
r? `@RalfJung`
Don't modify libstd to dump rustc ICEs
Do a much simpler thing and just dump a `std::backtrace::Backtrace` to file.
r? `@estebank` `@oli-obk`
Fixes#115610
This function is now used to check `#[panic_handler]`, `start` lang item, `main`, `#[start]` and intrinsic functions.
The diagnosis produced are now closer to the ones produced by trait/impl method signature mismatch.
move things out of mir/mod.rs
This moves a bunch of things out of `mir/mod.rs`:
- all const-related stuff to a new file consts.rs
- all statement/place/operand-related stuff to a new file statement.rs
- all pretty-printing related stuff to pretty.rs
`mod.rs` started out with 3100 lines and ends up with 1600. :)
Also there was some pretty-printing stuff in terminator.rs, that also got moved to pretty.rs, and I reordered things in pretty.rs so that it can be grouped by functionality.
Only the commit "use pretty_print_const_value from MIR constant 'extra' printing" has any behavior changes; it resolves the issue of having a fancy and a very crude pretty-printer for `ConstValue`.
r? `@oli-obk`
Don't complain on a single non-exhaustive 1-ZST
r? RalfJung, though you mentioned being busy, so feel free to reassign.
This doesn't actually attempt to make the diagnostic better, so when we have two non-exhaustive 1-ZSTs in a struct, we still just point to one. 🤷Fixes#115922
Explain HRTB + infer limitations of old solver
Add a helpful message when we hit the limitation of the old trait solver where we don't properly normalize GATs with infer vars + bound vars, leading to too-eagerly reporting trait errors that would be later satisfied due to inference.
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 (#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 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#114928
rustdoc-search: add support for type parameters
r? `@GuillaumeGomez`
## Preview
* https://notriddle.com/rustdoc-html-demo-4/advanced-search/rustdoc/read-documentation/search.html
* https://notriddle.com/rustdoc-html-demo-4/advanced-search/std/index.html?search=option%3Coption%3CT%3E%3E%20-%3E%20option%3CT%3E
* https://notriddle.com/rustdoc-html-demo-4/advanced-search/std/index.html?search=option%3CT%3E,%20E%20-%3E%20result%3CT,%20E%3E
* https://notriddle.com/rustdoc-html-demo-4/advanced-search/std/index.html?search=-%3E%20option%3CT%3E
## Description
When writing a type-driven search query in rustdoc, specifically one with more than one query element, non-existent types become generic parameters instead of auto-correcting (which is currently only done for single-element queries) or giving no result. You can also force a generic type parameter by writing `generic:T` (and can force it to not use a generic type parameter with something like `struct:T` or whatever, though if this happens it means the thing you're looking for doesn't exist and will give you no results).
There is no syntax provided for specifying type constraints for generic type parameters.
When you have a generic type parameter in a search query, it will only match up with generic type parameters in the actual function, not concrete types that match, not concrete types that implement a trait. It also strictly matches based on when they're the same or different, so `option<T>, option<U> -> option<U>` matches `Option::and`, but not `Option::or`. Similarly, `option<T>, option<T> -> option<T>` matches `Option::or`, but not `Option::and`.
## Motivation
This feature is motivated by the many "combinitor"-type functions found in generic libraries, such as Option, Future, Iterator, and Entry. These highly-generic functions have names that are almost completely arbitrary, and a type signature that tells you what it actually does.
This PR is a major step towards[^closure] being able to easily search for generic functions by their type signature instead of by name. Some examples of combinators that can be found using this PR (try them out in the preview):
* `option<option<T>> -> option<T>` returns Option::flatten
* `option<T> -> result<T>` returns Option::ok_or
* `option<result<T>> -> result<option<T>>` returns Option::transpose
* `entry<K, V>, FnOnce -> V` returns `Entry::or_insert_with` (and `or_insert_with_key`, since there's no way to specify the generics on FnOnce)
[^closure]:
For this feature to be as useful as it ought to be, you should be able to search for *trait-associated types* and *closures*. This PR does not implement either of these: they are **Future possibilities**.
Trait-associated types would allow queries like `option<T> -> iterator<item=T>` to return `Option::iter`. We should also allow `option<T> -> iterator<T>` to match the associated type version.
Closures would make a good way to query for things like `Option::map`. Closure support needs associated types to be represented in the search index, since `FnOnce() -> i32` desugars to `FnOnce<Output=i32, ()>`, so associated trait types should be implemented first. Also, we'd want to expose an easy way to query closures without specifying which of the three traits you want.
coverage: Fix an unstable-sort inconsistency in coverage spans
This code was calling `sort_unstable_by`, but failed to impose a total order on the initial spans. That resulted in unpredictable handling of closure spans, producing inconsistencies in the coverage maps and in user-visible coverage reports.
This PR fixes the problem by always sorting closure spans before otherwise-identical non-closure spans, and also switches to a stable sort in case the ordering is still not total.
---
In addition to the fix itself, this PR also contains a cleanup to the comparison function that I was working on when I discovered the bug.
Make `TyKind::Adt`'s `Debug` impl be more pretty
Currently `{:?}` on `Ty` for a `TyKind::Adt` would print as `Adt(Foo, [])`. This PR changes it to be `Foo` when there are no generics or `Foo<T>`/`Foo<T, U>` when there _are_ generics. Example from debug log:
`├─0ms DEBUG rustc_hir_analysis::astconv return=Bar<T/#0, U/#1>`
I should have done this in my initial PR for a prettier TyKind: Debug impl but I thought I would need to be accessing generics_of to figure out where in the "path" the generics would have to go??? but no, adts literally only have a single place the generics can go (on the end). Feel a bit silly about this :)
r? `@oli-obk`
Avoid blessing cargo deps's source code in ui tests
Before this PR, the source code of dependencies was included in UI test error messages whenever possible. Unfortunately, "whenever possible" means in some cases the source code wouldn't be injected, resulting in a test failure.
One such case is when `$CARGO_HOME` is remapped to something that is not present on disk [^1]. As the remapped path doesn't exist on disk, the source code wouldn't be showed in `tests/ui/issues/issue-21763.rs`:
```diff
= note: required for `hashbrown::raw::RawTable<(Rc<()>, Rc<()>)>` to implement `Send`
note: required because it appears within the type `HashMap<Rc<()>, Rc<()>, RandomState>`
--> $HASHBROWN_SRC_LOCATION
- |
-LL | pub struct HashMap<K, V, S = DefaultHashBuilder, A: Allocator + Clone = Global> {
- | ^^^^^^^
note: required because it appears within the type `HashMap<Rc<()>, Rc<()>>`
--> $SRC_DIR/std/src/collections/hash/map.rs:LL:COL
note: required by a bound in `foo`
```
This PR fixes the problem by always hiding dependencies source code in the error messages generated during UI tests. This is implemented with a new internal flag, `-Z ignore-directory-in-diagnostics-source-blocks=$path`, which compiletest passes during UI tests. Once this is merged, remapping the Cargo home will be supported.
This PR is best reviewed commit-by-commit.
[^1]: After being puzzled for a bit, I discovered why this never impacted `rust-lang/rust`: we don't remap `$CARGO_HOME` 😅. Instead, we set `$CARGO_HOME` to `/cargo` in CI, which sort-of-but-not-really achieves the same effect.
move required_consts check to general post-mono-check function
This factors some code that is common between the interpreter and the codegen backends into shared helper functions. Also as a side-effect the interpreter now uses the same `eval` functions as everyone else to get the evaluated MIR constants.
Also this is in preparation for another post-mono check that will be needed for (the current hackfix for) https://github.com/rust-lang/rust/issues/115709: ensuring that all locals are dynamically sized.
I didn't expect this to change diagnostics, but it's just cycle errors that change.
r? `@oli-obk`
This code was calling `sort_unstable_by`, but failed to impose a total order on
the initial spans. That resulted in unpredictable handling of closure spans,
producing inconsistencies in the coverage maps and in user-visible coverage
reports.
This patch fixes the problem by always sorting closure spans before
otherwise-identical non-closure spans, and also switches to a stable sort in
case the ordering is still not total.
Turn custom code classes in docs into warning
By habit, since it was a new feature gate, I added a check which emitted an error in case the new syntax was used. However, since rustdoc tags parser was accepting *everything*, using the "new" syntax should never ever emit errors. It now emits a warning.
Follow-up of #110800.
cc `@Manishearth`
r? `@notriddle`
repr(transparent): it's fine if the one non-1-ZST field is a ZST
This code currently gets rejected:
```rust
#[repr(transparent)]
struct MyType([u16; 0])
```
That clearly seems like a bug to me: `repr(transparent)` [got defined ](https://github.com/rust-lang/rust/issues/77841#issuecomment-716575747) as having any number of 1-ZST fields plus optionally one more field; `MyType` clearly satisfies that definition.
This PR changes the `repr(transparent)` logic to actually match that definition.
Rollup of 6 pull requests
Successful merges:
- #114965 (Remove Drop impl of mpsc Receiver and (Sync)Sender)
- #115434 (make `Debug` impl for `ascii::Char` match that of `char`)
- #115477 (Stabilize the `Saturating` type)
- #115611 (add diagnostic for raw identifiers in format string)
- #115654 (improve PassMode docs)
- #115862 (Migrate `compiler/rustc_hir_typeck/src/callee.rs` to translatable diagnostics)
r? `@ghost`
`@rustbot` modify labels: rollup
Make `.rmeta` file in `dep-info` have correct name (`lib` prefix)
Since `filename_for_metadata()` and
`OutputFilenames::path(OutputType::Metadata)` had different logic for the name of the metadata file, the `.d` file contained a file name different from the actual name used. Share the logic to fix the out-of-sync name.
Without this fix, the `.d` file contained
dash-separated_something-extra.rmeta: dash-separated.rs
instead of
libdash_separated_something-extra.rmeta: dash-separated.rs
which is the name of the file that is actually written by the compiler.
Worth noting: It took me several iterations to get all tests to pass, so I am relatively confident that this PR does not break anything.
Closes#68839
add diagnostic for raw identifiers in format string
Format strings don't support raw identifiers (e.g. `format!("{r#type}")`), but they do support keywords in the format string directly (e.g. `format!("{type}")`). This PR improves the error output when attempting to use a raw identifier in a format string and adds a machine-applicable suggestion to remove the `r#`.
fixes https://github.com/rust-lang/rust/issues/115466
`#[diagnostic::on_unimplemented]` without filters
This commit adds support for a `#[diagnostic::on_unimplemented]` attribute with the following options:
* `message` to customize the primary error message
* `note` to add a customized note message to an error message
* `label` to customize the label part of the error message
The relevant behavior is specified in [RFC-3366](https://rust-lang.github.io/rfcs/3366-diagnostic-attribute-namespace.html)
impl Step for IP addresses
ACP: rust-lang/libs-team#235
Note: since this is insta-stable, it requires an FCP.
Separating out from the bit operations PR since it feels logically disjoint, and so their FCPs can be separate.
Accept additional user-defined syntax classes in fenced code blocks
Part of #79483.
This is a re-opening of https://github.com/rust-lang/rust/pull/79454 after a big update/cleanup. I also converted the syntax to pandoc as suggested by `@notriddle:` the idea is to be as compatible as possible with the existing instead of having our own syntax.
## Motivation
From the original issue: https://github.com/rust-lang/rust/issues/78917
> The technique used by `inline-c-rs` can be ported to other languages. It's just super fun to see C code inside Rust documentation that is also tested by `cargo doc`. I'm sure this technique can be used by other languages in the future.
Having custom CSS classes for syntax highlighting will allow tools like `highlight.js` to be used in order to provide highlighting for languages other than Rust while not increasing technical burden on rustdoc.
## What is the feature about?
In short, this PR changes two things, both related to codeblocks in doc comments in Rust documentation:
* Allow to disable generation of `language-*` CSS classes with the `custom` attribute.
* Add your own CSS classes to a code block so that you can use other tools to highlight them.
#### The `custom` attribute
Let's start with the new `custom` attribute: it will disable the generation of the `language-*` CSS class on the generated HTML code block. For example:
```rust
/// ```custom,c
/// int main(void) {
/// return 0;
/// }
/// ```
```
The generated HTML code block will not have `class="language-c"` because the `custom` attribute has been set. The `custom` attribute becomes especially useful with the other thing added by this feature: adding your own CSS classes.
#### Adding your own CSS classes
The second part of this feature is to allow users to add CSS classes themselves so that they can then add a JS library which will do it (like `highlight.js` or `prism.js`), allowing to support highlighting for other languages than Rust without increasing burden on rustdoc. To disable the automatic `language-*` CSS class generation, you need to use the `custom` attribute as well.
This allow users to write the following:
```rust
/// Some code block with `{class=language-c}` as the language string.
///
/// ```custom,{class=language-c}
/// int main(void) {
/// return 0;
/// }
/// ```
fn main() {}
```
This will notably produce the following HTML:
```html
<pre class="language-c">
int main(void) {
return 0;
}</pre>
```
Instead of:
```html
<pre class="rust rust-example-rendered">
<span class="ident">int</span> <span class="ident">main</span>(<span class="ident">void</span>) {
<span class="kw">return</span> <span class="number">0</span>;
}
</pre>
```
To be noted, we could have written `{.language-c}` to achieve the same result. `.` and `class=` have the same effect.
One last syntax point: content between parens (`(like this)`) is now considered as comment and is not taken into account at all.
In addition to this, I added an `unknown` field into `LangString` (the parsed code block "attribute") because of cases like this:
```rust
/// ```custom,class:language-c
/// main;
/// ```
pub fn foo() {}
```
Without this `unknown` field, it would generate in the DOM: `<pre class="language-class:language-c language-c">`, which is quite bad. So instead, it now stores all unknown tags into the `unknown` field and use the first one as "language". So in this case, since there is no unknown tag, it'll simply generate `<pre class="language-c">`. I added tests to cover this.
Finally, I added a parser for the codeblock attributes to make it much easier to maintain. It'll be pretty easy to extend.
As to why this syntax for adding attributes was picked: it's [Pandoc's syntax](https://pandoc.org/MANUAL.html#extension-fenced_code_attributes). Even if it seems clunkier in some cases, it's extensible, and most third-party Markdown renderers are smart enough to ignore Pandoc's brace-delimited attributes (from [this comment](https://github.com/rust-lang/rust/pull/110800#issuecomment-1522044456)).
## Raised concerns
#### It's not obvious when the `language-*` attribute generation will be added or not.
It is added by default. If you want to disable it, you will need to use the `custom` attribute.
#### Why not using HTML in markdown directly then?
Code examples in most languages are likely to contain `<`, `>`, `&` and `"` characters. These characters [require escaping](https://developer.mozilla.org/en-US/docs/Web/HTML/Element/pre) when written inside the `<pre>` element. Using the \`\`\` code blocks allows rustdoc to take care of escaping, which means doc authors can paste code samples directly without manually converting them to HTML.
cc `@poliorcetics`
r? `@notriddle`
closure field capturing: don't depend on alignment of packed fields
This fixes the closure field capture part of https://github.com/rust-lang/rust/issues/115305: field capturing always stops at projections into packed structs, no matter the alignment of the field. This means changing a private field type from `u8` to `u64` can never change how closures capture fields, which is probably what we want.
Here's an example where, before this PR, changing the type of a private field in a repr(Rust) struct can change the output of a program:
```rust
#![allow(dead_code)]
mod m {
// before patch
#[derive(Default)]
pub struct S1(u8);
// after patch
#[derive(Default)]
pub struct S2(u64);
}
struct NoisyDrop;
impl Drop for NoisyDrop {
fn drop(&mut self) {
eprintln!("dropped!");
}
}
#[repr(packed)]
struct MyType {
field: m::S1, // output changes when this becomes S2
other_field: NoisyDrop,
third_field: Vec<()>,
}
fn test(r: MyType) {
let c = || {
let _val = std::ptr::addr_of!(r.field);
let _val = r.third_field;
};
drop(c);
eprintln!("before dropping");
}
fn main() {
test(MyType {
field: Default::default(),
other_field: NoisyDrop,
third_field: Vec::new(),
});
}
```
Of course this is a breaking change for the same reason that doing field capturing in the first place was a breaking change. Packed fields are relatively rare and depending on drop order is relatively rare, so I don't expect this to have much impact, but it's hard to be sure and even a crater run will only tell us so much.
Also see the [nomination comment](https://github.com/rust-lang/rust/pull/115315#issuecomment-1702807825).
Cc `@rust-lang/wg-rfc-2229` `@ehuss`
Make useless_ptr_null_checks smarter about some std functions
This teaches the `useless_ptr_null_checks` lint that some std functions can't ever return null pointers, because they need to point to valid data, get references as input, etc.
This is achieved by introducing an `#[rustc_never_returns_null_ptr]` attribute and adding it to these std functions (gated behind bootstrap `cfg_attr`).
Later on, the attribute could maybe be used to tell LLVM that the returned pointer is never null. I don't expect much impact of that though, as the functions are pretty shallow and usually the input data is already never null.
Follow-up of PR #113657Fixes#114442
Fallback effects even if types also fallback
`||` is short circuiting, so if we do ty/int var fallback, we *don't* do effect fallback 😸
r? `@fee1-dead` or `@oli-obk`
Fixes#115791Fixes#115842
Improve invalid let expression handling
- Move all of the checks for valid let expression positions to parsing.
- Add a field to ExprKind::Let in AST/HIR to mark whether it's in a valid location.
- Suppress some later errors and MIR construction for invalid let expressions.
- Fix a (drop) scope issue that was also responsible for #104172.
Fixes#104172Fixes#104868
Paper over an accidental regression
r? types
cc https://github.com/rust-lang/rust/issues/115781 (do not close issue until beta backport has been performed)
The PR reasons are explained with comments in the source.
In order to keep the diff simple, this PR effectively reverts https://github.com/rust-lang/rust/pull/113661, but only for RPITs. I will submit a follow up PR that fixes this correctly instead of just disabling the newly added check for RPITs. This PR should be significantly easier to review for beta backport
Properly consider binder vars in `HasTypeFlagsVisitor`
Given a PolyTraitRef like `for<'a> Ty: Trait` (where neither `Ty` nor `Trait` mention `'a`), we do *not* return true for `.has_type_flags(TypeFlags::HAS_LATE_BOUND)`, even though binders are supposed to act as if they have late-bound vars even if they don't mention them in their bound value: 31ae3b2bdb. This is because we use `HasTypeFlagsVisitor`, which only computes the type flags for `Ty`, `Const` and `Region` and `Predicates`, and we consequently skip any binders (and setting flags for their vars) that are not contained in one of these types.
This ends up causing a problem, because when we call `TyCtxt::erase_regions` (which both erases regions *and* anonymizes bound vars), we will skip such a PolyTraitRef, not anonymizing it, and therefore not making it structurally equal to other binders. This breaks vtable computations.
This PR computes the flags for all binders we enter in `HasTypeFlagsVisitor` if we're looking for `TypeFlags::HAS_LATE_BOUND` (or `TypeFlags::HAS_{RE,TY,CT}_LATE_BOUND`).
Fixes#115807
Fix the error message for `#![feature(no_coverage)]`
When #114656 was written, the feature flag to replace `no_coverage` was originally spelled `coverage`, but it was eventually changed to `coverage_attribute` instead.
That update happened to miss this error message in `removed.rs`, and unfortunately I only noticed just *after* the original PR was approved and merged.
cc ``@bossmc`` (original author) ``@oli-obk`` (original reviewer)
``@rustbot`` label +A-code-coverage
some ConstValue refactoring
In particular, use AllocId instead of Allocation in ConstValue::ByRef. This helps avoid redundant AllocIds when a `ByRef` constant gets put back into the interpreter.
r? `@oli-obk`
Fixes https://github.com/rust-lang/rust/issues/105536
Rework `no_coverage` to `coverage(off)`
As discussed at the tail of https://github.com/rust-lang/rust/issues/84605 this replaces the `no_coverage` attribute with a `coverage` attribute that takes sub-parameters (currently `off` and `on`) to control the coverage instrumentation.
Allows future-proofing for things like `coverage(off, reason="Tested live", issue="#12345")` or similar.
cleanup leftovers of const_err lint
Some code / comments seem to not have been updated when const_err was turned into a hard error, so we can do a bit of cleanup here.
r? `@oli-obk`
- Add doc comment to new type
- Restore "only supported directly in conditions of `if` and `while` expressions" note
- Rename variant with clearer name
tests: re-enable pretty-std-collections on macOS
Fixes#78665.
I made some small modifications to this test so that it would pass for me locally (though I was only able to test using lldb without built-in Rust support, but that seems to be the mode in which it would fail). I ran it a few hundred times with stage one and stage two to see if I could re-produce the spurious failures that were being reported in #78665 and couldn't. From the discussion in #78665, it seemed like this was related to Xcode versions and could be reproduced locally fairly easily. It's been a couple years since this was disabled so a lot has changed. If this starts failing spuriously again then we can disable it and I can look into that.
r? `@wesleywiser` (discussed in wg-debugging's triage meeting)
This commit adds support for a `#[diagnostic::on_unimplemented]`
attribute with the following options:
* `message` to customize the primary error message
* `note` to add a customized note message to an error message
* `label` to customize the label part of the error message
Co-authored-by: León Orell Valerian Liehr <me@fmease.dev>
Co-authored-by: Michael Goulet <michael@errs.io>
Read from non-scalar constants and statics in dataflow const-prop
DataflowConstProp is designed to handle scalar values. When MIR features an assignment from a non-scalar constant, we need to manually decompose it into the custom state space.
This PR tweaks interpreter callbacks to allow reusing `eval_mir_constant` without having a stack frame to get a span from.
r? `@oli-obk`
cc `@jachris`
test ABI compatibility for some unsized types as well
and test for what `DispatchFromDyn` needs.
Also I ran this on a whole bunch of targets via Miri and added enough `cfg` to make it all work, as documentation for what does and doesn't currently work. (Most of those targets do not have their tests run on CI anyway.)
Here's the shell rune I used for that:
```
for TARGET in x86_64-unknown-linux-gnu x86_64-pc-windows-gnu aarch64-unknown-linux-gnu s390x-unknown-linux-gnu mips64-unknown-linux-gnuabi64 sparc64-unknown-linux-gnu powerpc64-unknown-linux-gnu powerpc64le-unknown-linux-gnu riscv64gc-unknown-linux-gnu loongarch64-unknown-linux-gnu wasm32-unknown-unknown; do
BOOTSTRAP_SKIP_TARGET_SANITY=1 ./x.py run miri --stage 0 --args tests/ui/abi/compatibility.rs --target $TARGET;
done
```
Rollup of 8 pull requests
Successful merges:
- #115548 (Extract parallel operations in `rustc_data_structures::sync` into a new `parallel` submodule)
- #115591 (Add regression test for LLVM 17-rc3 miscompile)
- #115631 (Don't ICE when computing ctype's `repr_nullable_ptr` for possibly-unsized ty)
- #115708 (fix homogeneous_aggregate not ignoring some ZST)
- #115730 (Some more small driver refactors)
- #115749 (Allow loading the SMIR for constants and statics)
- #115757 (Add a test for #108030)
- #115761 (Update books)
r? `@ghost`
`@rustbot` modify labels: rollup
fix homogeneous_aggregate not ignoring some ZST
This is an ABI-breaking change, because it fixes bugs in our ABI code. I'm not sure what that means for this PR, we don't really have a process for such changes, do we? I can only hope nobody relied on the old buggy behavior.
Fixes https://github.com/rust-lang/rust/issues/115664
Don't ICE when computing ctype's `repr_nullable_ptr` for possibly-unsized ty
We may not always be able to compute the layout of a type like `&T` when `T: ?Sized`, even if we're able to estimate its size skeleton.
r? davidtwco
Fixes#115628
Bubble up opaque <eq> opaque operations instead of picking an order
In case we are in `Bubble` mode (meaning every opaque type that is defined in the current crate is treated as if it were in its defining scope), we don't try to register an opaque type as the hidden type of another opaque type, but instead bubble up an obligation to equate them at the query caller site. Usually that means we have a `DefiningAnchor::Bind` and thus can reliably figure out whether an opaque type is in its defining scope. Where we can't, we'll error out, so the default is sound.
With this change we start using `AliasTyEq` predicates in the old solver, too.
fixes https://github.com/rust-lang/rust/issues/108498
But also regresses `tests/ui/impl-trait/anon_scope_creep.rs`. Our use of `Bubble` for `check_opaque_type_well_formed` is going to keep biting us.
r? `@lcnr` `@compiler-errors`
Previously some invalid let expressions would result in both a feature
error and a parsing error. Avoid this and ensure that we only emit the
parsing error when this happens.
There was an incomplete version of the check in parsing and a second
version in AST validation. This meant that some, but not all, invalid
uses were allowed inside macros/disabled cfgs. It also means that later
passes have a hard time knowing when the let expression is in a valid
location, sometimes causing ICEs.
- Add a field to ExprKind::Let in AST/HIR to mark whether it's in a
valid location.
- Suppress later errors and MIR construction for invalid let
expressions.
rustdoc: Add missing "Aliased type" title in the sidebar
Follow-up of https://github.com/rust-lang/rust/pull/115682.
The sections title are supposed to be present in the sidebar, we forgot to put this one so I added it into it. I also added the missing newly created `aliased-type` ID into the `DEFAULT_ID_MAP`.
r? `@notriddle`
Improve diagnostic for generic params from outer items (E0401)
Generalize the wording of E0401 to talk about *outer items* instead of *outer functions* since the current phrasing is outdated. The outer item can be a function, constant, trait, ADT or impl block (see the new UI test for the more exotic examples).
Further, don't suggest introducing generic parameters to constant items unless the feature `generic_const_items` is enabled.
Lastly, make E0401 translatable while we're at it.
Fixes#115720.
Fix incorrect mutable suggestion information for binding in ref pattern like: `let &b = a;`
fixes#114896
I find we have to get pat_span but not local_decl.source_info.span for suggestion. In `let &b = a;` pat_span is &b. I think check `let &b = a` in hir to make sure it is hir::Node::Local(hir::Local {pat: hir::Pat{kind: hir::PatKind::Ref(....... can distinguish it from other situation, but I'm not sure.
If my processing method is not accurate, please guide me to modify it, thank you.
r? `@davidtwco`
Disentangle `Debug` and `Display` for `Ty`.
The `Debug` impl for `Ty` just calls the `Display` impl for `Ty`. This is surprising and annoying. In particular, it means `Debug` doesn't show as much information as `Debug` for `TyKind` does. And `Debug` is used in some user-facing error messages, which seems bad.
This commit changes the `Debug` impl for `Ty` to call the `Debug` impl for `TyKind`. It also does a number of follow-up changes to preserve existing output, many of which involve inserting
`with_no_trimmed_paths!` calls. It also adds `Display` impls for `UserType` and `Canonical`.
Some tests have changes to expected output:
- Those that use the `rustc_abi(debug)` attribute.
- Those that use the `rustc_layout(debug)` attribute.
- Those that use the `EMIT_MIR` annotation.
In each case the output is slightly uglier than before. This isn't ideal, but it's pretty weird (particularly for the attributes) that the output is using `Debug` in the first place. They're fairly obscure attributes (I hadn't heard of them) so I'm not worried by this.
For `async-is-unwindsafe.stderr`, there is one line that now lacks a full path. This is a consistency improvement, because all the other mentions of `Context` in this test lack a path.
The `Debug` impl for `Ty` just calls the `Display` impl for `Ty`. This
is surprising and annoying. In particular, it means `Debug` doesn't show
as much information as `Debug` for `TyKind` does. And `Debug` is used in
some user-facing error messages, which seems bad.
This commit changes the `Debug` impl for `Ty` to call the `Debug` impl
for `TyKind`. It also does a number of follow-up changes to preserve
existing output, many of which involve inserting
`with_no_trimmed_paths!` calls. It also adds `Display` impls for
`UserType` and `Canonical`.
Some tests have changes to expected output:
- Those that use the `rustc_abi(debug)` attribute.
- Those that use the `EMIT_MIR` annotation.
In each case the output is slightly uglier than before. This isn't
ideal, but it's pretty weird (particularly for the attribute) that the
output is using `Debug` in the first place. They're fairly obscure
attributes (I hadn't heard of them) so I'm not worried by this.
For `async-is-unwindsafe.stderr`, there is one line that now lacks a
full path. This is a consistency improvement, because all the other
mentions of `Context` in this test lack a path.
Encode only MIR reachable from other crates
Only reachable items might participate in the code generation in the
downstream crates. Omit redundant optimized MIR of unreachable items
from a crate metadata.
Additionally, include reachable closures in reachable set, so that
unreachable closures can be omitted on the same basis.
Only reachable items might participate in the code generation in the
downstream crates. Omit redundant optimized MIR of unreachable items
from a crate metadata.
Additionally, include reachable closures in reachable set, so that
unreachable closures can be omitted on the same basis.
The lint panicked for an input like 'extern "C" fn(Option<&<T as FooTrait>::FooType>)' because the type T therein cannot be normalized. The normalization failure caused SizeSkeleton::compute() to return an error and trigger a panic in the unwrap().
Fix sanitize/cfg.rs test
* Move needs-sanitizer conditions to specific revisions that require them (otherwise the conditions are mutually exclusive with needs-sanitizer-kcfi and test is always ignored).
* Add missing revisions
Improve "associated type not found" diagnostics
```rs
use core::ops::Deref;
fn foo<T>() where T: Deref<Output = u32> {}
```
Before:
```
error[E0220]: associated type `Output` not found for `Deref`
--> E0220.rs:5:28
|
5 | fn foo<T>() where T: Deref<Output = u32> {}
| ^^^^^^ associated type `Output` not found
```
After:
```
error[E0220]: associated type `Output` not found for `Deref`
--> E0220.rs:5:28
|
5 | fn foo<T>() where T: Deref<Output = u32> {}
| ^^^^^^ help: `Deref` has the following associated type: `Target`
```
---
`@rustbot` label +A-diagnostics +D-papercut
debuginfo: add compiler option to allow compressed debuginfo sections
LLVM already supports emitting compressed debuginfo. In debuginfo=full builds, the debug section is often a large amount of data, and it typically compresses very well (3x is not unreasonable.) We add a new knob to allow debuginfo to be compressed when the matching LLVM functionality is present. Like clang, if a known-but-disabled compression mechanism is requested, we disable compression and emit uncompressed debuginfo sections.
The API is different enough on older LLVMs we just pretend the support
is missing on LLVM older than 16.
Use the same DISubprogram for each instance of the same inlined function within a caller
# Issue Details:
The call to `panic` within a function like `Option::unwrap` is translated to LLVM as a `tail call` (as it will never return), when multiple calls to the same function like this are inlined LLVM will notice the common `tail call` block (i.e., loading the same panic string + location info and then calling `panic`) and merge them together.
When merging these instructions together, LLVM will also attempt to merge the debug locations as well, but this fails (i.e., debug info is dropped) as Rust emits a new `DISubprogram` at each inline site thus LLVM doesn't recognize that these are actually the same function and so thinks that there isn't a common debug location.
As an example of this, consider the following program:
```rust
#[no_mangle]
fn add_numbers(x: &Option<i32>, y: &Option<i32>) -> i32 {
let x1 = x.unwrap();
let y1 = y.unwrap();
x1 + y1
}
```
When building for x86_64 Windows using 1.72 it generates (note the lack of `.cv_loc` before the call to `panic`, thus it will be attributed to the same line at the `addq` instruction):
```llvm
.cv_loc 0 1 3 0 # src\lib.rs:3:0
addq $40, %rsp
retq
leaq .Lalloc_f570dea0a53168780ce9a91e67646421(%rip), %rcx
leaq .Lalloc_629ace53b7e5b76aaa810d549cc84ea3(%rip), %r8
movl $43, %edx
callq _ZN4core9panicking5panic17h12e60b9063f6dee8E
int3
```
# Fix Details:
Cache the `DISubprogram` emitted for each inlined function instance within a caller so that this can be reused if that instance is encountered again.
Ideally, we would also deduplicate child scopes and variables, however my attempt to do that with #114643 resulted in asserts when building for Linux (#115156) which would require some deep changes to Rust to fix (#115455).
Instead, when using an inlined function as a debug scope, we will also create a new child scope such that subsequent child scopes and variables do not collide (from LLVM's perspective).
After this change the above assembly now (with <https://reviews.llvm.org/D159226> as well) shows the `panic!` was inlined from `unwrap` in `option.rs` at line 935 into the current function in `lib.rs` at line 0 (line 0 is emitted since it is ambiguous which line to use as there were two inline sites that lead to this same code):
```llvm
.cv_loc 0 1 3 0 # src\lib.rs:3:0
addq $40, %rsp
retq
.cv_inline_site_id 6 within 0 inlined_at 1 0 0
.cv_loc 6 2 935 0 # library\core\src\option.rs:935:0
leaq .Lalloc_5f55955de67e57c79064b537689facea(%rip), %rcx
leaq .Lalloc_e741d4de8cb5801e1fd7a6c6795c1559(%rip), %r8
movl $43, %edx
callq _ZN4core9panicking5panic17hde1558f32d5b1c04E
int3
```
Using `ld.lld` may have been clever, but that was getting the /system/
ld.lld, not one we may have built as part of building llvm. By using the
warning message coming directly from rustc we now correctly skip the
zlib and zstd tests when the support is missing.
LLVM already supports emitting compressed debuginfo. In debuginfo=full
builds, the debug section is often a large amount of data, and it
typically compresses very well (3x is not unreasonable.) We add a new
knob to allow debuginfo to be compressed when the matching LLVM
functionality is present. Like clang, if a known-but-disabled
compression mechanism is requested, we disable compression and emit
uncompressed debuginfo sections.
The API is different enough on older LLVMs we just pretend the support
is missing on LLVM older than 16.
Rollup of 6 pull requests
Successful merges:
- #104299 (Clarify stability guarantee for lifetimes in enum discriminants)
- #115088 (Fix Step Skipping Caused by Using the `--exclude` Option)
- #115201 (rustdoc: list matching impls on type aliases)
- #115633 (Lint node for `PRIVATE_BOUNDS`/`PRIVATE_INTERFACES` is the item which names the private type)
- #115638 (`-Cllvm-args` usability improvement)
- #115643 (fix: return early when has tainted in mir-lint)
r? `@ghost`
`@rustbot` modify labels: rollup
* Move needs-sanitizer conditions to specific revisions that
require them (otherwise the conditions are mutually exclusive
with needs-sanitizer-kcfi and test is always ignored).
* Add missing revisions
fix: return early when has tainted in mir-lint
Fixes#115203
`a[..]` is of indeterminate size, it had been reported error during borrow check, therefore we skip the mir lint process.
Lint node for `PRIVATE_BOUNDS`/`PRIVATE_INTERFACES` is the item which names the private type
The HIR that the `PRIVATE_BOUNDS` lint should be attached to is the item that has the *bounds*, not the private type. This PR also aligns this behavior with the `EXPORTED_PRIVATE_DEPENDENCIES` lint, which also requires putting the `allow` on the item that names the private type.
Fixes#115475
r? petrochenkov
Add CL and CMD into to pdb debug info
Partial fix for https://github.com/rust-lang/rust/issues/96475
The Arg0 and CommandLineArgs of the MCTargetOptions cpp class are not set within bb548f9645/compiler/rustc_llvm/llvm-wrapper/PassWrapper.cpp (L378)
This causes LLVM to not neither output any compiler path (cl) nor the arguments that were used when invoking it (cmd) in the PDB file.
This fix adds the missing information to the target machine so LLVM can use it.
diagnostics: add test case for trait bounds diagnostic
Closes#82038
It was fixed by https://github.com/rust-lang/rust/pull/89580, a wide-reaching obligation tracking improvement. This commit adds a test case.
Don't suggest dereferencing to unsized type
Rudimentary check that the self type is Sized. I don't really like any of this diagnostics code -- it's really messy and also really prone to false positives and negatives, but oh well.
Fixes#115569
Print the path of a return-position impl trait in trait when `return_type_notation` is enabled
When we're printing a return-position impl trait in trait, we usually just print it like an opaque. This is *usually* fine, but can be confusing when using `return_type_notation`. Print the path of the method from where the RPITIT originates when this feature gate is enabled.
rustdoc: Render private fields in tuple struct as `/* private fields */`
Reopening of https://github.com/rust-lang/rust/pull/110552. All that was missing was a test for the different cases so I added it into the second commit.
Description from the original PR:
> I've gotten some feedback that the current rustdoc rendering of...
>
> ```
> struct HasPrivateFields(_);
> ```
>
> ...is confusing, and I agree with that feedback, especially compared to the field struct case:
>
> ```
> struct HasPrivateFields { /* private fields */ }
> ```
>
> So this PR makes it so that when all of the fields of a tuple variant are private, just render it with the `/* private fields */` comment. We can't *always* render it like that, for example when there's a mix of private and public fields.
cc ````@jsha````
r? ````@notriddle````
Don't require `Drop` for `[PhantomData<T>; N]` where `N` and `T` are generic, if `T` requires `Drop`
fixes https://github.com/rust-lang/rust/issues/115403
fixes https://github.com/rust-lang/rust/issues/115410
This was accidentally regressed in https://github.com/rust-lang/rust/pull/114134, because it was accidentally stabilized in #102204 (cc `@rust-lang/lang,` seems like an innocent stabilization, considering this PR is more of a bugfix than a feature).
While we have a whole month to beta backport this change before the regression hits stable, I'd still prefer not to go through an FCP on this PR (which fixes a regression), if T-lang wants an FCP, I can can open an issue about the change itself.
Stabilize `PATH` option for `--print KIND=PATH`
This PR propose stabilizing the `PATH` option for `--print KIND=PATH`. This option was previously added in https://github.com/rust-lang/rust/pull/113780 (as insta-stable before being un-stablized in https://github.com/rust-lang/rust/pull/114139).
Description of the `PATH` option:
> A filepath may optionally be specified for each requested information kind, in the format `--print KIND=PATH`, just like for `--emit`. When a path is specified, information will be written there instead of to stdout.
------
Description of the original PR [\[link\]](https://github.com/rust-lang/rust/pull/113780#issue-1807080607):
> **Support --print KIND=PATH command line syntax**
>
> As is already done for `--emit KIND=PATH` and `-L KIND=PATH`.
>
> In the discussion of https://github.com/rust-lang/rust/pull/110785, it was pointed out that `--print KIND=PATH` is nicer than trying to apply the single global `-o path` to `--print`'s output, because in general there can be multiple print requests within a single rustc invocation, and anyway `-o` would already be used for a different meaning in the case of `link-args` and `native-static-libs`.
>
> I am interested in using `--print cfg=PATH` in Buck2. Currently Buck2 works around the lack of support for `--print KIND=PATH` by [indirecting through a Python wrapper script](d43cf3a51a/prelude/rust/tools/get_rustc_cfg.py) to redirect rustc's stdout into the location dictated by the build system.
>
> From skimming Cargo's usages of `--print`, it definitely seems like it would benefit from `--print KIND=PATH` too. Currently it is working around the lack of this by inserting `--crate-name=___ --print=crate-name` so that it can look for a line containing `___` as a delimiter between the 2 other `--print` informations it actually cares about. This is commented as a "HACK" and "abuse". 31eda6f7c3/src/cargo/core/compiler/build_context/target_info.rs (L242)
-----
cc `@dtolnay`
r? `@jackh726`
Description of the `PATH` option:
> A filepath may optionally be specified for each requested information
> kind, in the format `--print KIND=PATH`, just like for `--emit`. When
> a path is specified, information will be written there instead of to
> stdout.
Implement refinement lint for RPITIT
Implements a lint that warns against accidentally refining an RPITIT in an implementation. This is not a hard error, and can be suppressed with `#[allow(refining_impl_trait)]`, since this behavior may be desirable -- the lint just serves as an acknowledgement from the impl author that they understand that the types they write in the implementation are an API guarantee.
This compares bounds syntactically, not semantically -- semantic implication is more difficult and essentially relies on adding the ability to keep the RPITIT hidden in the trait system so that things can be proven about the type that shows up in the impl without its own bounds leaking through, either via a new reveal mode or something else. This was experimentally implemented in #111931.
Somewhat opinionated choices:
1. Putting the lint behind `refining_impl_trait` rather than a blanket `refine` lint. This could be changed, but I like keeping the lint specialized to RPITITs so the explanation can be tailored to it.
2. This PR does not include the `#[refine]` attribute or the feature gate, since it's kind of orthogonal and can be added in a separate PR.
r? `@oli-obk`
Lint on invalid usage of `UnsafeCell::raw_get` in reference casting
This PR proposes to take into account `UnsafeCell::raw_get` method call for non-Freeze types for the `invalid_reference_casting` lint.
The goal of this is to catch those kind of invalid reference casting:
```rust
fn as_mut<T>(x: &T) -> &mut T {
unsafe { &mut *std::cell::UnsafeCell::raw_get(x as *const _ as *const _) }
//~^ ERROR casting `&T` to `&mut T` is undefined behavior
}
```
r? `@est31`
fix#115348fix#115348
It looks that:
- In `rustc_mir_build::build`, the body of function will not be built, when the `tcx.check_match(def)` fails due to `non-exhaustive patterns`
- In `rustc_mir_transform::check_unsafety`, the `UnsafetyChecker` collects all `used_unsafe_blocks` in the MIR of a function, and the `UnusedUnsafeVisitor` will visit all `UnsafeBlock`s in the HIR and collect `unused_unsafes`, which are not contained in `used_unsafe_blocks`, and report `unnecessary_unsafe`s
- So the unsafe block in the issue example code will be reported as `unnecessary_unsafe`.
Add explanatory note to 'expected item' error
Fixes#113110
It changes the diagnostic from this:
```
error: expected item, found `5`
--> ../test.rs:1:1
|
1 | 5
| ^ expected item
```
to this:
```
error: expected item, found `5`
--> ../test.rs:1:1
|
1 | 5
| ^ expected item
|
= note: items are things that can appear at the root of a module
= note: for a full list see https://doc.rust-lang.org/reference/items.html
```
