Split-up stability_index query
This PR aims to move deprecation and stability processing away from the monolithic `stability_index` query, and directly implement `lookup_{deprecation,stability,body_stability,const_stability}` queries.
The basic idea is to:
- move per-attribute sanity checks into `check_attr.rs`;
- move attribute compatibility checks into the `MissingStabilityAnnotations` visitor;
- progressively dismantle the `Annotator` visitor and the `stability_index` query.
The first commit contains functional change, and now warns when `#[automatically_derived]` is applied on a non-trait impl block. The other commits should not change visible behaviour.
Perf in https://github.com/rust-lang/rust/pull/143845#issuecomment-3066308630 shows small but consistent improvement, except for unused-warnings case. That case being a stress test, I'm leaning towards accepting the regression.
This PR changes `check_attr`, so has a high conflict rate on that file. This should not cause issues for review.
resolve: Make disambiguators for underscore bindings module-local
Disambiguators attached to underscore name bindings (like `const _: u8 = something;`) do not need to be globally unique, they just need to be unique inside the module in which they live, because the bindings in a module are basically kept as `Map<BindingKey, SomeData>`.
Also, the specific values of the disambiguators are not important, so a glob import of `const _` may have a different disambiguator than the original `const _` itself.
So in this PR the disambiguator is just set to the current number of bindings in the module.
This removes one more piece of global mutable state from resolver and unblocks https://github.com/rust-lang/rust/pull/143884.
Use $crate in macros for rustc_public (aka stable_mir)
This makes `#[macro_use] extern crate rustc_public` unnecessary (which brings all of `rustc_public`'s macros into scope for the entire crate); instead, now you can simply use `rustc_public::run!()`.
Make slice comparisons const
This needed a fix for `derive_const`, too, as it wasn't usable in libcore anymore as trait impls need const stability attributes. I think we can't use the same system as normal trait impls while `const_trait_impl` is still unstable.
r? ```@fee1-dead```
cc rust-lang/rust#143800
Rollup of 11 pull requests
Successful merges:
- rust-lang/rust#142300 (Disable `tests/run-make/mte-ffi` because no CI runners have MTE extensions enabled)
- rust-lang/rust#143271 (Store the type of each GVN value)
- rust-lang/rust#143293 (fix `-Zsanitizer=kcfi` on `#[naked]` functions)
- rust-lang/rust#143719 (Emit warning when there is no space between `-o` and arg)
- rust-lang/rust#143846 (pass --gc-sections if -Zexport-executable-symbols is enabled and improve tests)
- rust-lang/rust#143891 (Port `#[coverage]` to the new attribute system)
- rust-lang/rust#143967 (constify `Option` methods)
- rust-lang/rust#144008 (Fix false positive double negations with macro invocation)
- rust-lang/rust#144010 (Boostrap: add warning on `optimize = false`)
- rust-lang/rust#144049 (rustc-dev-guide subtree update)
- rust-lang/rust#144056 (Copy GCC sources into the build directory even outside CI)
r? `@ghost`
`@rustbot` modify labels: rollup
Fix false positive double negations with macro invocation
This PR fixes false positive double_negations lint when macro expansion has negation and macro caller also has negations.
Fixrust-lang/rust#143980
pass --gc-sections if -Zexport-executable-symbols is enabled and improve tests
Exported symbols are added as GC roots in linking, so `--gc-sections` won't hurt `-Zexport-executable-symbols`.
Fixes the run-make test to work on Linux. Enable the ui test on more targets.
cc rust-lang/rust#84161
Emit warning when there is no space between `-o` and arg
Closesrust-lang/rust#142812
`getopt` doesn't seem to have an API to check this, so we have to check the args manually.
r? compiler
fix `-Zsanitizer=kcfi` on `#[naked]` functions
fixes https://github.com/rust-lang/rust/issues/143266
With `-Zsanitizer=kcfi`, indirect calls happen via generated intermediate shim that forwards the call. The generated shim preserves the attributes of the original, including `#[unsafe(naked)]`. The shim is not a naked function though, and violates its invariants (like having a body that consists of a single `naked_asm!` call).
My fix here is to match on the `InstanceKind`, and only use `codegen_naked_asm` when the instance is not a `ReifyShim`. That does beg the question whether there are other `InstanceKind`s that could come up. As far as I can tell the answer is no: calling via `dyn` seems to work find, and `#[track_caller]` is disallowed in combination with `#[naked]`.
r? codegen
````@rustbot```` label +A-naked
cc ````@maurer```` ````@rcvalle````
Store the type of each GVN value
MIR is fully typed, so type information is an integral part of what defines a value. GVN currently tries to circumvent storing types, which creates all sorts of complexities.
This PR stores the type along with the enum `Value` when defining a value index. This allows to simplify a lot of code.
Fixesrust-lang/rust#128094Fixesrust-lang/rust#135128
r? ``````@ghost`````` for perf
`-Zhigher-ranked-assumptions`: Consider WF of coroutine witness when proving outlives assumptions
### TL;DR
This PR introduces an unstable flag `-Zhigher-ranked-assumptions` which tests out a new algorithm for dealing with some of the higher-ranked outlives problems that come from auto trait bounds on coroutines. See:
* rust-lang/rust#110338
While it doesn't fix all of the issues, it certainly fixed many of them, so I'd like to get this landed so people can test the flag on their own code.
### Background
Consider, for example:
```rust
use std::future::Future;
trait Client {
type Connecting<'a>: Future + Send
where
Self: 'a;
fn connect(&self) -> Self::Connecting<'_>;
}
fn call_connect<C>(c: C) -> impl Future + Send
where
C: Client + Send + Sync,
{
async move { c.connect().await }
}
```
Due to the fact that we erase the lifetimes in a coroutine, we can think of the interior type of the async block as something like: `exists<'r, 's> { C, &'r C, C::Connecting<'s> }`. The first field is the `c` we capture, the second is the auto-ref that we perform on the call to `.connect()`, and the third is the resulting future we're awaiting at the first and only await point. Note that every region is uniquified differently in the interior types.
For the async block to be `Send`, we must prove that both of the interior types are `Send`. First, we have an `exists<'r, 's>` binder, which needs to be instantiated universally since we treat the regions in this binder as *unknown*[^exist]. This gives us two types: `{ &'!r C, C::Connecting<'!s> }`. Proving `&'!r C: Send` is easy due to a [`Send`](https://doc.rust-lang.org/nightly/std/marker/trait.Send.html#impl-Send-for-%26T) impl for references.
Proving `C::Connecting<'!s>: Send` can only be done via the item bound, which then requires `C: '!s` to hold (due to the `where Self: 'a` on the associated type definition). Unfortunately, we don't know that `C: '!s` since we stripped away any relationship between the interior type and the param `C`. This leads to a bogus borrow checker error today!
### Approach
Coroutine interiors are well-formed by virtue of them being borrow-checked, as long as their callers are invoking their parent functions in a well-formed way, then substitutions should also be well-formed. Therefore, in our example above, we should be able to deduce the assumption that `C: '!s` holds from the well-formedness of the interior type `C::Connecting<'!s>`.
This PR introduces the notion of *coroutine assumptions*, which are the outlives assumptions that we can assume hold due to the well-formedness of a coroutine's interior types. These are computed alongside the coroutine types in the `CoroutineWitnessTypes` struct. When we instantiate the binder when proving an auto trait for a coroutine, we instantiate the `CoroutineWitnessTypes` and stash these newly instantiated assumptions in the region storage in the `InferCtxt`. Later on in lexical region resolution or MIR borrowck, we use these registered assumptions to discharge any placeholder outlives obligations that we would otherwise not be able to prove.
### How well does it work?
I've added a ton of tests of different reported situations that users have shared on issues like rust-lang/rust#110338, and an (anecdotally) large number of those examples end up working straight out of the box! Some limitations are described below.
### How badly does it not work?
The behavior today is quite rudimentary, since we currently discharge the placeholder assumptions pretty early in region resolution. This manifests itself as some limitations on the code that we accept.
For example, `tests/ui/async-await/higher-ranked-auto-trait-11.rs` continues to fail. In that test, we must prove that a placeholder is equal to a universal for a param-env candidate to hold when proving an auto trait, e.g. `'!1 = 'a` is required to prove `T: Trait<'!1>` in a param-env that has `T: Trait<'a>`. Unfortunately, at that point in the MIR body, we only know that the placeholder is equal to some body-local existential NLL var `'?2`, which only gets equated to the universal `'a` when being stored into the return local later on in MIR borrowck.
This could be fixed by integrating these assumptions into the type outlives machinery in a more first-class way, and delaying things to the end of MIR typeck when we know the full relationship between existential and universal NLL vars. Doing this integration today is quite difficult today.
`tests/ui/async-await/higher-ranked-auto-trait-11.rs` fails because we don't compute the full transitive outlives relations between placeholders. In that test, we have in our region assumptions that some `'!1 = '!2` and `'!2 = '!3`, but we must prove `'!1 = '!3`.
This can be fixed by computing the set of coroutine outlives assumptions in a more transitive way, or as I mentioned above, integrating these assumptions into the type outlives machinery in a more first-class way, since it's already responsible for the transitive outlives assumptions of universals.
### Moving forward
I'm still quite happy with this implementation, and I'd like to land it for testing. I may work on overhauling both the way we compute these coroutine assumptions and also how we deal with the assumptions during (lexical/nll) region checking. But for now, I'd like to give users a chance to try out this new `-Zhigher-ranked-assumptions` flag to uncover more shortcomings.
[^exist]: Instantiating this binder with infer regions would be incomplete, since we'd be asking for *some* instantiation of the interior types, not proving something for *all* instantiations of the interior types.
Unify `CoroutineWitness` sooner in typeck, and stall coroutine obligations based off of `TypingEnv`
* Stall coroutine obligations based off of `TypingMode` in the old solver.
* Eagerly assign `TyKind::CoroutineWitness` to the witness arg of coroutines during typeck, rather than deferring them to the end of typeck.
r? lcnr
This is part of https://github.com/rust-lang/rust/issues/143017.
Rollup of 11 pull requests
Successful merges:
- rust-lang/rust#143326 (Remove deprecated `Error::description` impl from `c_str::FromBytesWithNulError`)
- rust-lang/rust#143431 (Use relative visibility when noting sealed trait to reduce false positive)
- rust-lang/rust#143550 (resolve: Use interior mutability for extern module map)
- rust-lang/rust#143631 (update to literal-escaper-0.0.5)
- rust-lang/rust#143793 (Opaque type collection: Guard against endlessly recursing free alias types)
- rust-lang/rust#143880 (tests: Test line debuginfo for linebreaked function parameters)
- rust-lang/rust#143914 (Reword mismatched-lifetime-syntaxes text based on feedback)
- rust-lang/rust#143926 (Remove deprecated fields in bootstrap)
- rust-lang/rust#143955 (Make frame spans appear on a separate trace line)
- rust-lang/rust#143975 (type_id_eq: check that the hash fully matches the type)
- rust-lang/rust#143984 (Fix ice for feature-gated `cfg` attributes applied to the crate)
r? `@ghost`
`@rustbot` modify labels: rollup
Fix ice for feature-gated `cfg` attributes applied to the crate
This PR fixes two fixes:
1. When a feature gated option of the `cfg` attribute is applied to the crate, an ICE would occur because features are not yet available at that stage. This is fixed by ignoring the feature gate at that point, the attribute will later be re-checked (this was already done) when the feature gate is available. Fixes https://github.com/rust-lang/rust/issues/143977
2. Errors and lints on the `cfg` attribute applied to the crate would be produced twice, because of the re-checking. This is fixed by not producing any errors and lints during the first run.
The added regression test checks both problems.
r? ``@jdonszelmann``
type_id_eq: check that the hash fully matches the type
The previous logic wouldn't always detect when the hash mismatches the provenance. Fix that by adding a new helper, `read_type_id`, that reads a single type ID while fully checking it for validity and consistency.
r? ``@oli-obk``
Make frame spans appear on a separate trace line
This PR changes tracing_chrome's `tracing::Layer` so that if a span has the "tracing_separate_line" field as one of the span arguments, that span is put on a separate trace line. See https://github.com/rust-lang/miri/pull/4451 for an earlier attempt and for screenshots explaining better what I mean by "separate trace line".
This PR also makes the "frame" span use this feature (so it appears on a separate trace line, see https://github.com/rust-lang/miri/pull/4451 for motivation), but passes `tracing::field::Empty` as the span parameter value so it is ignored by other tracing layers (e.g. the logger):
```rust
info_span!("frame", tracing_separate_line = Empty, "{}", instance);
```
<details><summary>Also see the following discussion I had with ``@RalfJung</summary>``
> Is there no way to attach metadata we could use instead?
[These](https://docs.rs/tracing-core/0.1.34/src/tracing_core/metadata.rs.html#57) are the **static** metadata items we can control about a span. We can't add more metadata outside of them. The most relevant are:
- `name` (for the frame span it's currently "frame")
- `target` which acts as the category (for the frame span it's currently "rustc_const_eval::interpret::stack" by default)
- `fields` which contains a list of the *names* of each of the arguments passed to the `span!` macro (for the frame span it's currently ["message"], where "message" is the default identifier for data passed in the `format!` syntax)
When the tracing code is called at runtime, the **dynamic** values of the arguments are collected into a [`ValueSet`](https://docs.rs/tracing-core/0.1.34/src/tracing_core/field.rs.html#166). Each argument value stored there corresponds with one of the static names stored in `fields` (see above).
---
We have already determined that filtering out spans by `name` is not a good idea, and I would say the same goes for `target`. Both the `name` and the `target` fields are printed to stderr when `MIRI_LOG=` is enabled, so changing them to contain an identifier (e.g. "frame:tracing_separate_root" instead of "frame" as the name) would uselessly clutter the text logs (unless we add one more filter [there](https://github.com/rust-lang/rust/blob/master/compiler/rustc_log/src/lib.rs#L137), but then it gets even more complicated).
```rust
// examples of how the above (problematic) solutions would look like
info_span!("frame:tracing_separate_root", "{}", instance);
info_span!(target: "tracing_separate_root", "frame", "{}", instance);
```
---
So that leaves us with `fields` and their runtime values. Now, my initial thought (inspired by [this comment](https://github.com/rust-lang/miri/pull/4451#issuecomment-3068072303)) was to use a field with the static name "tracing_separate_root" and with a dynamic boolean value of "true". In `tracing_chrome.rs` we can easily check if this field is true and act accordingly. This would work but then again this field would also be picked up by the logger when `MIRI_LOG=` is enabled, and would uselessly clutter the text logs.
```rust
// example of how the above (problematic) solution would look like
info_span!("frame", tracing_separate_root = true, "{}", instance);
```
---
To avoid cluttering the text logs, we can instead set "tracing_separate_root" to the dynamic value of `tracing::field::Empty`. Citing from [here](https://docs.rs/tracing/0.1.41/tracing/field/struct.Empty.html), "when a field’s value is `Empty`, it will not be recorded". "not being recorded" means that the field and its value won't be printed to stderr text logs, nor will it be printed by any other tracing layers that might be attached in the future. In `tracing_chrome.rs` we would still be able to check if "tracing_separate_root" is in the list of static `fields`, and act accordingly. So I believe this solution would effectively allow us to attach metadata to a span in a way that does not clutter logs and still allows being read in `tracing_chrome.rs`.
If we ever wanted to pass arbitrary metadata (i.e. not just a present/not present flag), it would be possible with a custom `Empty` that also holds data and implement `Value` without doing anything ([like `Empty` does](https://docs.rs/tracing-core/0.1.34/src/tracing_core/field.rs.html#775)).
```rust
// example of how the above solution would look like
info_span!("frame", tracing_separate_root = tracing::field::Empty, "{}", instance);
```
</details>
