miri: improve errors for type validity assertion failures
Miri has pretty nice errors for type validity violations, printing which field in the type the problem occurs at and so on.
However, we don't see these errors when using e.g. `mem::zeroed` as that uses `assert_zero_valid` to bail out before Miri can detect the UB.
Similar to what we did with `@saethlin's` UB checks, I think we should disable such language UB checks in Miri so that we can get better error messages. If we go for this we should probably say this in the intrinsic docs as well so that people don't think they can rely on these intrinsics catching anything.
Furthermore, I slightly changed `MaybeUninit::assume_init` so that the `.value` field does not show up in error messages any more.
`@rust-lang/miri` what do you think?
interpret: move the native call preparation logic into Miri
`@nia-e` has to do a bunch of changes to this logic for her native call ptrace work, and it's getting annoying that the logic is split between Miri and rustc. So this moves the logic to Miri, keeping just the generic traversal part in rustc. It is unfortunate that this means we have to expose `get_alloc_raw`/`get_alloc_raw_mut`... I hope the function name is scary enough to reduce the risk of misuse.
r? `@oli-obk`
Rollup of 11 pull requests
Successful merges:
- rust-lang/rust#131923 (Derive `Copy` and `Hash` for `IntErrorKind`)
- rust-lang/rust#138340 (Remove some unsized tuple impls now that we don't support unsizing tuples anymore)
- rust-lang/rust#141219 (Change `{Box,Arc,Rc,Weak}::into_raw` to only work with `A = Global`)
- rust-lang/rust#142212 (bootstrap: validate `rust.codegen-backends` & `target.<triple>.codegen-backends`)
- rust-lang/rust#142237 (Detect more cases of unused_parens around types)
- rust-lang/rust#142964 (Attribute rework: a parser for single attributes without arguments)
- rust-lang/rust#143070 (Rewrite `macro_rules!` parser to not use the MBE engine itself)
- rust-lang/rust#143235 (Assemble const bounds via normal item bounds in old solver too)
- rust-lang/rust#143261 (Feed `explicit_predicates_of` instead of `predicates_of`)
- rust-lang/rust#143276 (loop match: handle opaque patterns)
- rust-lang/rust#143306 (Add `track_caller` attributes to trace origin of Clippy lints)
r? `@ghost`
`@rustbot` modify labels: rollup
try-job: aarch64-apple
try-job: x86_64-msvc-1
try-job: x86_64-gnu
try-job: dist-i586-gnu-i586-i686-musl
try-job: test-various
Rollup of 9 pull requests
Successful merges:
- rust-lang/rust#141847 (Explain `TOCTOU` on the top of `std::fs`, and reference it in functions)
- rust-lang/rust#142138 (Add `Vec::into_chunks`)
- rust-lang/rust#142321 (Expose elf abi on ppc64 targets)
- rust-lang/rust#142886 (ci: aarch64-gnu: Stop skipping `panic_abort_doc_tests`)
- rust-lang/rust#143194 (fix bitcast of single-element SIMD vectors)
- rust-lang/rust#143231 (Suggest use another lifetime specifier instead of underscore lifetime)
- rust-lang/rust#143232 ([COMPILETEST-UNTANGLE 3/N] Use "directives" consistently within compiletest)
- rust-lang/rust#143258 (Don't recompute `DisambiguatorState` for every RPITIT in trait definition)
- rust-lang/rust#143274 (ci: support optional jobs)
r? `@ghost`
`@rustbot` modify labels: rollup
Add `track_caller` attributes to trace origin of Clippy lints
This allows the use of `-Z track-diagnostics` to see the origin of Clippy lints emission, as is already the case for lints coming from rustc.
[COMPILETEST-UNTANGLE 3/N] Use "directives" consistently within compiletest
Instead of using *both* "headers" and "directives" within compiletest to refer to the same thing. This of course induces some churn, but it's been bugging me for a while, and I rather do the self-consistency changes now than later.
The first commit tries to be mostly move-only to help per-file git history.
I intend to revisit rustc-dev-guide's testing docs, but I don't want to do it on rust-lang/rust side because it would need syncing and might conflict.
Remove let_chains unstable feature
Per https://github.com/rust-lang/rust/issues/53667#issuecomment-3016742982 (but then I also noticed rust-lang/rust#140722)
This replaces the feature gate with a parser error that says let chains require 2024.
A lot of tests were using the unstable feature. I either added edition:2024 to the test or split out the parts that require 2024.
compiletest has confusingly two terminology to refer to the same concept
-- "headers" and "directives". To make this more self-consistent and
less confusing, stick with "directives" only.
This commit **intentionally** tries to be limited to move-only (modulo
some key usage reference renames) to help git history.
Remove support for `dyn*` from the compiler
This PR removes support for `dyn*` (https://github.com/rust-lang/rust/issues/102425), which are a currently un-RFC'd experiment that was opened a few years ago to explore a component that we thought was necessary for AFIDT (async fn in dyn trait).
It doesn't seem like we are going to need `dyn*` types -- even in an not-exposed-to-the-user way[^1] -- for us to implement AFIDT. Given that AFIDT was the original motivating purpose of `dyn*` types, I don't really see a compelling reason to have to maintain their implementation in the compiler.
[^1]: Compared to, e.g., generators whih are an unstable building block we use to implement stable syntax like `async {}`.
We've learned quite a lot from `dyn*`, but I think at this point its current behavior leads to more questions than answers. For example, `dyn*` support today remains somewhat fragile; it ICEs in many cases where the current "normal" `dyn Trait` types rely on their unsizedness for their vtable-based implementation to be sound I wouldn't be surprised if it's unsound in other ways, though I didn't play around with it too much. See the examples below.
```rust
#![feature(dyn_star)]
trait Foo {
fn hello(self);
}
impl Foo for usize {
fn hello(self) {
println!("hello, world");
}
}
fn main() {
let x: dyn* Foo = 1usize;
x.hello();
}
```
And:
```rust
#![feature(dyn_star)]
trait Trait {
type Out where Self: Sized;
}
fn main() {
let x: <dyn* Trait as Trait>::Out;
}
```
...and probably many more problems having to do with the intersection of dyn-compatibility and `Self: Sized` bounds that I was too lazy to look into like:
* GATs
* Methods with invalid signatures
* Associated consts
Generally, `dyn*` types also end up getting in the way of working with [normal `dyn` types](https://github.com/rust-lang/rust/issues/102425#issuecomment-1712604409) to an extent that IMO outweighs the benefit of experimentation.
I recognize that there are probably other, more creative usages of `dyn*` that are orthogonal to AFIDT. However, I think any work along those lines should first have to think through some of the more fundamental interactions between `dyn*` and dyn-compatibility before we think about reimplementing them in the type system.
---
I'm planning on removing the `DynKind` enum and the `PointerLike` built-in trait from the compiler after this PR lands.
Closesrust-lang/rust#102425.
cc `@eholk` `@rust-lang/lang` `@rust-lang/types`
Closesrust-lang/rust#116979.
Closesrust-lang/rust#119694.
Closesrust-lang/rust#134591.
Closesrust-lang/rust#104800.
Remove `ItemKind::descr` method
Follow-up of rust-lang/rust#143234.
After this PR is merged, it will remain two `descr` methods:
* `hir::GenericArg::descr`
* `hir::AssocItemConstraintKind::descr`
For both these enums, I don't think there is the right equivalent in `hir::DefKind` so unless I missed something, we can't remove these two methods because we can't convert these enums into `hir::DefKind`.
r? `@oli-obk`
[COMPILETEST-UNTANGLE 2/N] Make some compiletest errors/warnings/help more visually obvious
This PR makes some compiletest errors/warnings/help more visually obvious. Note that this is only intended to help visually -- the error handling in compiletest is still a mess.
r? ghost
Upgrade dependencies in run-make-support
The main purpose of this is to upgrade `object` and `gimli`, which will allow us to drop outdated versions once backtrace also updates. The only semver breakage in `object`'s is in `elf::R_RISCV_GNU_*` and `pe::IMAGE_WEAK_EXTERN_*` constants, as well as Mach-O dyld, which don't appear to be used here. `gimli` is similar, there is only minor breakage related to dyld.
These version upgrades were also done in the library.
`bstr`, `similar`, and `regex` are also upgraded to the latest minor version here to match what the lockfile already uses. The `regex` comment about `memchr` version hasn't been relevant to this lockfile since e95d15a115 ("Pin memchr to 2.5.0 in the library rather than rustc_ast") and is no longer relevant in the library lockfile either.
Object Changelog: https://github.com/gimli-rs/object/blob/master/CHANGELOG.md#0370
Gimli changelog: https://github.com/gimli-rs/gimli/blob/master/CHANGELOG.md#0320
Introduce `ByteSymbol`
It's like `Symbol` but for byte strings. The interner is now used for both `Symbol` and `ByteSymbol`. E.g. if you intern `"dog"` and `b"dog"` you'll get a `Symbol` and a `ByteSymbol` with the same index and the characters will only be stored once.
The motivation for this is to eliminate the `Arc`s in `ast::LitKind`, to make `ast::LitKind` impl `Copy`, and to avoid the need to arena-allocate `ast::LitKind` in HIR. The latter change reduces peak memory by a non-trivial amount on literal-heavy benchmarks such as `deep-vector` and `tuple-stress`.
`Encoder`, `Decoder`, `SpanEncoder`, and `SpanDecoder` all get some changes so that they can handle normal strings and byte strings.
The main purpose of this is to upgrade `object` and `gimli`, which will
allow us to drop outdated versions once backtrace also updates.
The only semver breakage in `object`'s is in `elf::R_RISCV_GNU_*` and
`pe::IMAGE_WEAK_EXTERN_*` constants, as well as Mach-O dyld, which don't
appear to be used here. `gimli` is similar, there is only minor breakage
related to dyld.
These version upgrades were also done in the library.
`bstr`, `similar`, and `regex` are also upgraded to the latest minor
version here to match what the lockfile already uses. The `regex`
comment about `memchr` version hasn't been relevant to this lockfile
since e95d15a115 ("Pin memchr to 2.5.0 in the library rather than
rustc_ast") and is no longer relevant in the library lockfile either.
Object Changelog: https://github.com/gimli-rs/object/blob/master/CHANGELOG.md#0370
Gimli changelog: https://github.com/gimli-rs/gimli/blob/master/CHANGELOG.md#0320
[COMPILETEST-UNTANGLE 1/N] Move some some early config checks to the lib and move the compiletest binary
This is part of a patch series to untangle `compiletest` to hopefully nudge it towards being more maintainable.
This PR:
- Moves some early config checks (some warnings) to the compiletest library.
- Moves `src/main.rs` to `src/bin/main.rs` to make the separation (as in, compiletest's library component vs the tool binary component) more obvious.
r? ``@Kobzol`` (or reroll)
Make combining LLD with external LLVM config a hard error
Younger me made this only a warning in https://github.com/rust-lang/rust/pull/139853, because our post-dist tests were relying on this. But that was not a good idea, because there are a bunch of places in bootstrap that outright try to build LLD/copy LLD to sysroot when `lld_enabled` is true (rightfully so), which is causing issues (https://github.com/rust-lang/rust/issues/143076). Instead of piling more hacks, I'd like to just disallow this, and if we need to use a hack, do it only for our CI.
If this breaks the CI post-dist tests, I'll either add some special environment variable for it, or, as an alternative, make the error back into a warning, but also disable `lld_enabled` when this situation happens.
try-job: dist-x86_64-linux
Fixes: https://github.com/rust-lang/rust/pull/143175
give Pointer::into_parts a more scary name and offer a safer alternative
`into_parts` is a bit too innocent of a name for a somewhat subtle operation.
r? `@oli-obk`
Miri: handling of SNaN inputs in `f*::pow` operations
fixes [miri/#4286](https://github.com/rust-lang/miri/issues/4286) and related to rust-lang/rust#138062 and [miri/#4208](https://github.com/rust-lang/miri/issues/4208#issue-2879058184).
For the following cases of the powf or powi operations, Miri returns either `1.0` or an arbitrary `NaN`:
- `powf(SNaN, 0.0)`
- `powf(1.0, SNaN)`
- `powi(SNaN, 0)`
Also added a macro in `miri/tests/pass/float.rs` which conveniently checks if both are indeed returned from such an operation.
Made these changes in the rust repo so I could test against stdlib, since these were impacted some time ago and were fixed in rust-lang/rust#138062. Tested with:
```fish
env MIRIFLAGS=-Zmiri-many-seeds ./x miri --no-fail-fast std core coretests -- f32 f64
```
This was successful. This does take a while, so I recommend using `--no-doc` and separate use of `f32` or `f64`
The pr is somewhat split up into 3 main commits, which implement the cases described above. The first commit also introduces the macro, and the last commit is just a global refactor of some things.
r? `@RalfJung`
`tests/ui`: A New Order [13/N]
Some `tests/ui/` housekeeping, to trim down number of tests directly under `tests/ui/`. Part of rust-lang/rust#133895.
r? ```@jieyouxu```
It's like `Symbol` but for byte strings. The interner is now used for
both `Symbol` and `ByteSymbol`. E.g. if you intern `"dog"` and `b"dog"`
you'll get a `Symbol` and a `ByteSymbol` with the same index and the
characters will only be stored once.
The motivation for this is to eliminate the `Arc`s in `ast::LitKind`, to
make `ast::LitKind` impl `Copy`, and to avoid the need to arena-allocate
`ast::LitKind` in HIR. The latter change reduces peak memory by a
non-trivial amount on literal-heavy benchmarks such as `deep-vector` and
`tuple-stress`.
`Encoder`, `Decoder`, `SpanEncoder`, and `SpanDecoder` all get some
changes so that they can handle normal strings and byte strings.
This change does slow down compilation of programs that use
`include_bytes!` on large files, because the contents of those files are
now interned (hashed). This makes `include_bytes!` more similar to
`include_str!`, though `include_bytes!` contents still aren't escaped,
and hashing is still much cheaper than escaping.
To make it obvious `compiletest`-the-tool has two components:
1. The core compiletest library, and
2. The tool binary, which will be executed by bootstrap.
Guillaume Gomez