Overhaul `AssocItem`
`AssocItem` has multiple fields that only make sense some of the time. E.g. the `name` can be empty if it's an RPITIT associated type. It's clearer and less error prone if these fields are moved to the relevant `kind` variants.
r? ``@fee1-dead``
compiletest: Add an experimental new executor to replace libtest
This PR adds a new "executor" to compiletest for running the list of collected tests, to eventually replace the current dependency on unstable libtest internals.
The new executor is currently inactive by default. It must be activated explicitly by passing `-n` or `--new-executor` to compiletest, e.g. `./x test ui -- -n`.
(After some amount of wider manual testing, the new executor will hopefully be made the default, and the libtest dependency can be removed. Contributors should not notice any change.)
The new executor is a stripped-down rewrite of the subset of libtest needed by compiletest.
# Supported functionality
- Specifying the number of concurrent tests with `RUST_TEST_THREADS`
- Filtering and skipping tests by name (substring or exact-match)
- Forcibly running ignored tests with `--ignored`
- Optional fail-fast with `--fail-fast`
- JSON output, compatible with bootstrap's parser for libtest output
- Running each test in its own thread
- Short backtraces that ignore the executor itself
- Slow test detection, with a hard-coded timeout of 60 seconds
- Capturing stdout/stderr, via `#![feature(internal_output_capture)]`
- Suppressing output capture with `--no-capture`
# Unsupported functionality
- Non-JSON output, as this is handled by bootstrap instead
- Separate code path for concurrency=1, as the concurrent path should handle this case naturally
- Fallback to running tests synchronously if new threads can't be spawned
- Special handling of hosts that don't support basic functionality like threads or timers
- Our ability to test *targets* should be unaffected
- Graceful handling of some edge cases that could occur in arbitrary user-written unit tests, but would represent bugs in compiletest
- Due to the current need for output capture, the new executor is still not entirely written in stable Rust
---
r? jieyouxu
std: sys: process: uefi: Use NULL stdin by default
According to the docs in `Command::output`:
> By default, stdout and stderr are captured (and used to provide the
resulting output). Stdin is not inherited from the parent and any attempt by the child process to read from the stdin stream will result in the stream immediately closing.
This was being violated by UEFI which was inheriting stdin by default.
While the docs don't explicitly state that the default should be NULL, the behaviour seems like reading from NULL.
UEFI however, has a bit of a problem. The `EFI_SIMPLE_TEXT_INPUT_PROTOCOL` only provides support for reading 1 key press. This means that you either get an error, or it is assumed that the keypress was read successfully. So there is no way to have a successful read of length 0. Currently, I am returning UNSUPPORTED error when trying to read from NULL stdin. On linux however, you will get a read of length 0 for Null stdin.
One possible way to get around this is to translate one of the UEFI errors to a read 0 (Maybe unsupported?). It is also possible to have a non-standard error code, but well, not sure if we go that route.
Alternatively, if meaning of Stdio::Null is platform dependent, it should be fine to keep the current behaviour of returning an error.
cc ```@nicholasbishop``` ```@dvdhrm```
Allow const patterns of matches to contain pattern types
Trying to pattern match on a type containing a pattern type will currently fail with an ICE
```rust
error: internal compiler error: compiler/rustc_mir_build/src/builder/matches/test.rs:459:18: invalid type for non-scalar compare: (u32) is 1..
--> src/main.rs:22:5
|
22 | TWO => {}
| ^^^
```
because the compiler tries to generate a MIR `BinOp(Eq)` operation on a pattern type, which is not supported. While we could support that, there are side effects of allowing this (none that would compile, but the compiler would simultaneously think it could `==` pattern types and that it could not because `PartialEq` is not implemented. So instead I change the logic for pattern matching to transmute pattern types to their base type before comparing.
r? ```@BoxyUwU```
cc #123646 ```@scottmcm``` ```@joshtriplett```
ci: add runners for vanilla LLVM 20
Ubuntu 25.04 has `llvm-20` packages that we can start testing with.
The `Dockerfile` is otherwise the same as the `llvm-18`/`19` runners.
try-job: x86_64-gnu-llvm-20-1
try-job: x86_64-gnu-llvm-20-2
try-job: x86_64-gnu-llvm-20-3
Enable contracts for const functions
Use `const_eval_select!()` macro to enable contract checking only at runtime. The existing contract logic relies on closures, which are not supported in constant functions.
This commit also removes one level of indirection for ensures clauses since we no longer build a closure around the ensures predicate.
Resolves#136925
**Call-out:** This is still a draft PR since CI is broken due to a new warning message for unreachable code when the bottom of the function is indeed unreachable. It's not clear to me why the warning wasn't triggered before.
r? ```@compiler-errors```
To accurately reflect that RPITIT assoc items don't have a name. This
avoids the use of `kw::Empty` to mean "no name", which is error prone.
Helps with #137978.
re-use `Sized` fast-path
There's an existing fast path for the `type_op_prove_predicate` predicate, checking for trivially `Sized` types, which can be re-used when evaluating obligations within queries. This should improve performance and was found to be beneficial in #137944.
r? types
Rollup of 8 pull requests
Successful merges:
- #139127 (Fix up partial res of segment in primitive resolution hack)
- #139392 (Detect and provide suggestion for `&raw EXPR`)
- #139767 (Visit place in `BackwardIncompatibleDropHint` statement)
- #139777 (Remove `define_debug_via_print` for `ExistentialProjection`, use regular structural debug impl)
- #139796 (ptr docs: add missing backtics around 'usize')
- #139801 (Add myself to mailmap)
- #139804 (use `realpath` in `bootstrap.py` when creating build-dir)
- #139807 (Improve wording of post-merge report)
r? `@ghost`
`@rustbot` modify labels: rollup
use `realpath` in `bootstrap.py` when creating build-dir
Fixes#139800
r? `@jieyouxu`
My use case for `./build` being a symlink is this: my "default" ~~partition~~ btrfs subvolume is snapshotted/backed up. I don't want to backup target-likes, so I move them to a special subvolume which isn't backed up. `./build` is a symlink into that subvolume. (`build.build-dir` configuration is not fully sufficient, it is still nice to be able to check build files with `ls ./build` or call tools from there)
ptr docs: add missing backtics around 'usize'
We almost always have the backticks, except here... so let's just fix that. Barely worth a PR but it's user-visible docs so here we go.
Remove `define_debug_via_print` for `ExistentialProjection`, use regular structural debug impl
The pretty print impl for `ExistentialProjection` always prints `AssocItem = Ty`:
6e83046233/compiler/rustc_middle/src/ty/print/pretty.rs (L3293-L3299)
We can't change this, b/c it's used for both pretty printing dyn types and for legacy symbol mangling.
Unfortunately, we also use this printing procedure for `Debug` impls. That means that it leaves out the *trait name* and *trait args* when debug printing an `ExistentialProjection` (or an `ExistentialPredicate` which has a variant for `ExistentialProjection`). This leads to awkward situations, like the two seemingly identical existential projection predicates present in a `dyn Trait` type using the definition below:
```rust
trait Super { type Assoc; }
trait Foo: Super<A, Assoc = i32> + Super<B, Assoc = i32> {}
```
Namely, they both just render as `Projection(Assoc = i32)`! This makes debugging `dyn Trait` type system bugs really hard, so let's use the *regular* debug impl for `ExistentialProjection`.
Visit place in `BackwardIncompatibleDropHint` statement
Remove a weird hack from the `LocalUpdater` where we were manually visiting the place stored in a `StatementKind::BackwardIncompatibleDropHint` because the MIR visitor impls weren't doing so.
Also, clean up `BackwardIncompatibleDropHint`s in `CleanupPostBorrowck`, since they're not needed for runtime MIR.
Detect and provide suggestion for `&raw EXPR`
When emitting an error in the parser, and we detect that the previous token was `raw` and we *could* have consumed `const`/`mut`, suggest that this may have been a mistyped raw ref expr. To do this, we add `const`/`mut` to the expected token set when parsing `&raw` as an expression (which does not affect the "good path" of parsing, for the record).
This is kind of a rudimentary error improvement, since it doesn't actually attempt to recover anything, leading to some other knock-on errors b/c we still treat `&raw` as the expression that was parsed... but at least we add the suggestion! I don't think the parser grammar means we can faithfully recover `&raw EXPR` early, i.e. during `parse_expr_borrow`.
Fixes#133231
Fix up partial res of segment in primitive resolution hack
There is a hack in the resolver:
```
// In `a(::assoc_item)*` `a` cannot be a module. If `a` does resolve to a module we
// don't report an error right away, but try to fallback to a primitive type.
```
This fixes up the resolution for primitives which would otherwise resolve to a module, but we weren't also updating the res of the path segment, leading to weird diagnostics.
We explicitly call `self.r.partial_res_map.insert` instead of `record_partial_res` b/c we have recorded a partial res already, and we specifically want to override it.
cc https://github.com/rust-lang/rust/issues/139095#issuecomment-2764371934
Rollup of 9 pull requests
Successful merges:
- #138336 (Improve `-Z crate-attr` diagnostics)
- #139636 (Encode dep node edge count as u32 instead of usize)
- #139666 (cleanup `mir_borrowck`)
- #139695 (compiletest: consistently use `camino::{Utf8Path,Utf8PathBuf}` throughout)
- #139699 (Proactively update coroutine drop shim's phase to account for later passes applied during shim query)
- #139718 (enforce unsafe attributes in pre-2024 editions by default)
- #139722 (Move some things to rustc_type_ir)
- #139760 (UI tests: migrate remaining compile time `error-pattern`s to line annotations when possible)
- #139776 (Switch attrs to `diagnostic::on_unimplemented`)
r? `@ghost`
`@rustbot` modify labels: rollup
`hir::AssocItem` currently has a boolean `fn_has_self_parameter` field,
which is misplaced, because it's only relevant for associated fns, not
for associated consts or types. This commit moves it (and renames it) to
the `AssocKind::Fn` variant, where it belongs.
This requires introducing a new C-style enum, `AssocTag`, which is like
`AssocKind` but without the fields. This is because `AssocKind` values
are passed to various functions like `find_by_ident_and_kind` to
indicate what kind of associated item should be searched for, and having
to specify `has_self` isn't relevant there.
New methods:
- Predicates `AssocItem::is_fn` and `AssocItem::is_method`.
- `AssocItem::as_tag` which converts `AssocItem::kind` to `AssocTag`.
Removed `find_by_name_and_kinds`, which is unused.
`AssocItem::descr` can now distinguish between methods and associated
functions, which slightly improves some error messages.
UI tests: migrate remaining compile time `error-pattern`s to line annotations when possible
There's a number of cases in which `error-pattern` is still necessary even for compile time checking.
- It checks something that compiler writes directly into stderr as text, and not to the structured json output. This includes some stuff reported during compiler panics, and also diagnostics that happen very early, for example when parsing the command line.
- It checks something that exists only in the full rendered diagnostic test, but not in its structured components, for example code fragments or output of `-Ztrack-diagnostics`. (The latter can probably be converted to structured form though.)
This is continuation of https://github.com/rust-lang/rust/pull/139137.
r? `@jieyouxu`
Move some things to rustc_type_ir
This moves
- `PatternKind`
- `FlagComputation`
- `TypeWalker`
into rustc_type_ir.
Not strictly required for rust-analyzer next-solve integration, but helps with code duplication.
r? types
enforce unsafe attributes in pre-2024 editions by default
New unsafe attributes should emit an error when used without the `unsafe(...)` in all editions.
The `no_mangle`, `link_section` and `export_name` attributes are exceptions, and can still be used without an unsafe in earlier editions. The only attributes for which this change is relevant right now are `#[ffi_const]` and `#[ffi_pure]`.
This change is required for making `#[unsafe(naked)]` sound in pre-2024 editions.
Improve `-Z crate-attr` diagnostics
- Show the `#![ ... ]` in the span (to make it clear that it should not
be included in the CLI argument)
- Show more detailed errors when the crate has valid token trees but
invalid syntax.
Previously, `crate-attr=feature(foo),feature(bar)` would just say
"invalid crate attribute" and point at the comma. Now, it explicitly
says that the comma was unexpected, which is useful when using
`--error-format=short`. It also fixes the column to show the correct
span.
- Recover from parse errors. Previously we would abort immediately on
syntax errors; now we go on to try and type-check the rest of the
crate.
The new diagnostic code also happens to be slightly shorter.
r? diagnostics
