Externally implementable items
Supersedes https://github.com/rust-lang/rust/pull/140010
Tracking issue: https://github.com/rust-lang/rust/issues/125418
Getting started:
```rust
#![feature(eii)]
#[eii(eii1)]
pub fn decl1(x: u64)
// body optional (it's the default)
{
println!("default {x}");
}
// in another crate, maybe
#[eii1]
pub fn decl2(x: u64) {
println!("explicit {x}");
}
fn main() {
decl1(4);
}
```
- tiny perf regression, underlying issue makes multiple things in the compiler slow, not just EII, planning to solve those separately.
- No codegen_gcc support, they don't have bindings for weak symbols yet but could
- No windows support yet for weak definitions
This PR merges the implementation of EII for just llvm + not windows, doesn't yet contain like a new panic handler implementation or alloc handler. With this implementation, it would support implementing the panic handler in terms of EII already since it requires no default implementation so no weak symbols
The PR has been open in various forms for about a year now, but I feel that having some implementation merged to build upon
Overhaul filename handling for cross-compiler consistency
This PR overhauls the way we handle filenames in the compiler and `rmeta` in order to achieve achieve cross-compiler consistency (ie. having the same path no matter if the filename was created in the current compiler session or is coming from `rmeta`).
This is required as some parts of the compiler rely on consistent paths for the soundness of generated code (see rust-lang/rust#148328).
In order to achieved consistency multiple steps are being taken by this PR:
- by making `RealFileName` immutable
- by only having `SourceMap::to_real_filename` create `RealFileName`
- currently `RealFileName` can be created from any `Path` and are remapped afterwards, which creates consistency issue
- by also making `RealFileName` holds it's working directory, embeddable name and the remapped scopes
- this removes the need for a `Session`, to know the current(!) scopes and cwd, which is invalid as they may not be equal to the scopes used when creating the filename
In order for `SourceMap::to_real_filename` to know which scopes to apply `FilePathMapping` now takes the current remapping scopes to apply, which makes `FileNameDisplayPreference` and company useless and are removed.
This PR is split-up in multiple commits (unfortunately not atomic), but should help review the changes.
Unblocks https://github.com/rust-lang/rust/pull/147611
Fixes https://github.com/rust-lang/rust/issues/148328
This commit refactors `SourceMap` and most importantly `RealFileName` to
make it self-contained in order to achieve cross-compiler consistency.
This is achieved:
- by making `RealFileName` immutable
- by only having `SourceMap::to_real_filename` create `RealFileName`
- by also making `RealFileName` holds it's working directory,
it's embeddable name and the remapped scopes
- by making most `FileName` and `RealFileName` methods take a scope as
an argument
In order for `SourceMap::to_real_filename` to know which scopes to apply
`FilePathMapping` now takes the current remapping scopes to apply, which
makes `FileNameDisplayPreference` and company useless and are removed.
The scopes type `RemapPathScopeComponents` was moved from
`rustc_session::config` to `rustc_span`.
The previous system for scoping the local/remapped filenames
`RemapFileNameExt::for_scope` is no longer useful as it's replaced by
methods on `FileName` and `RealFileName`.
148725 moved the default to being homogeneous; this adds heterogeneous ones back under an obvious-bikeshed syntax so people can experiment with that as well.
Essentially resolves 149025 by letting them move to this syntax instead.
Move attribute safety checking to attribute parsing
This PR moves attribute safety checking to be done during attribute parsing. The `cfg` and `cfg_attr` attribute no longer need special-cased safety checking, yay!
This PR is a part 1 of 2, in the second part I'd like to define attribute safety in the attribute parsers rather than getting the information from BUILTIN_ATTRIBUTE_MAP, but to keep PRs reviewable lets do that separately.
Fixes https://github.com/rust-lang/rust/issues/148453 by reordering the diagnostics. The "cannot find attribute" diagnostic now appears first, but both diagnostics still appear.
r? `@jdonszelmann`
Stabilize `asm_cfg`
tracking issue: https://github.com/rust-lang/rust/issues/140364
closes https://github.com/rust-lang/rust/issues/140364
Reference PR:
- https://github.com/rust-lang/reference/pull/2063
# Request for Stabilization
## Summary
The `cfg_asm` feature allows `#[cfg(...)]` and `#[cfg_attr(...)]` on the arguments of the assembly macros, for instance:
```rust
asm!( // or global_asm! or naked_asm!
"nop",
#[cfg(target_feature = "sse2")]
"nop",
// ...
#[cfg(target_feature = "sse2")]
a = const 123, // only used on sse2
);
```
## Semantics
Templates, operands, `options` and `clobber_abi` in the assembly macros (`asm!`, `naked_asm!` and `global_asm!`) can be annotated with `#[cfg(...)]` and `#[cfg_attr(...)]`. When the condition evaluates to true, the annotated argument has no effect, and is completely ignored when expanding the assembly macro.
## Documentation
reference PR: https://github.com/rust-lang/reference/pull/2063
## Tests
- [tests/ui/asm/cfg.rs](https://github.com/rust-lang/rust/blob/master/tests/ui/asm/cfg.rs) checks that `cfg`'d arguments where the condition evaluates to false have no effect
- [tests/ui/asm/cfg-parse-error.rs](https://github.com/rust-lang/rust/blob/master/tests/ui/asm/cfg.rs) checks the parsing rules (parsing effectively assumes that the cfg conditions are all true)
## History
- https://github.com/rust-lang/rust/issues/140279
- https://github.com/rust-lang/rust/pull/140367
# Resolved questions
**how are other attributes handled**
Other attributes are parsed, but explicitly rejected.
# unresolved questions
**operand before template**
The current implementation expects at least one template string before any operands. In the example below, if the `cfg` condition evaluates to true, the assembly block is ill-formed. But even when it evaluates to `false` this block is rejected, because the parser still expects just a template (a template is parsed as an expression and then validated to ensure that it is or expands to a string literal).
Changing how this works is difficult.
```rust
// This is rejected because `a = out(reg) x` does not parse as an expresion.
asm!(
#[cfg(false)]
a = out(reg) x, //~ ERROR expected token: `,`
"",
);
```
**lint on positional arguments?**
Adding a lint to warn on the definition or use of positional arguments being `cfg`'d out was discussed in https://github.com/rust-lang/rust/issues/140279#issuecomment-2832237372 and subsequent comments. Such a lint is not currently implemented, but that may not be a blocker based on the comments there.
r? `@traviscross` (I'm assuming you'll reassign as needed)
Port `cfg_select!` to the new attribute parsing system
Best reviewed commit by commit, since it involves some moving around of code
r? `````@jdonszelmann`````
stop specializing on `Copy`
fixes https://github.com/rust-lang/rust/issues/132442
`std` specializes on `Copy` to optimize certain library functions such as `clone_from_slice`. This is unsound, however, as the `Copy` implementation may not be always applicable because of lifetime bounds, which specialization does not take into account; the result being that values are copied even though they are not `Copy`. For instance, this code:
```rust
struct SometimesCopy<'a>(&'a Cell<bool>);
impl<'a> Clone for SometimesCopy<'a> {
fn clone(&self) -> Self {
self.0.set(true);
Self(self.0)
}
}
impl Copy for SometimesCopy<'static> {}
let clone_called = Cell::new(false);
// As SometimesCopy<'clone_called> is not 'static, this must run `clone`,
// setting the value to `true`.
let _ = [SometimesCopy(&clone_called)].clone();
assert!(clone_called.get());
```
should not panic, but does ([playground](https://play.rust-lang.org/?version=stable&mode=debug&edition=2021&gist=6be7a48cad849d8bd064491616fdb43c)).
To solve this, this PR introduces a new `unsafe` trait: `TrivialClone`. This trait may be implemented whenever the `Clone` implementation is equivalent to copying the value (so e.g. `fn clone(&self) -> Self { *self }`). Because of lifetime erasure, there is no way for the `Clone` implementation to observe lifetime bounds, meaning that even if the `TrivialClone` has stricter bounds than the `Clone` implementation, its invariant still holds. Therefore, it is sound to specialize on `TrivialClone`.
I've changed all `Copy` specializations in the standard library to specialize on `TrivialClone` instead. Unfortunately, the unsound `#[rustc_unsafe_specialization_marker]` attribute on `Copy` cannot be removed in this PR as `hashbrown` still depends on it. I'll make a PR updating `hashbrown` once this lands.
With `Copy` no longer being considered for specialization, this change alone would result in the standard library optimizations not being applied for user types unaware of `TrivialClone`. To avoid this and restore the optimizations in most cases, I have changed the expansion of `#[derive(Clone)]`: Currently, whenever both `Clone` and `Copy` are derived, the `clone` method performs a copy of the value. With this PR, the derive macro also adds a `TrivialClone` implementation to make this case observable using specialization. I anticipate that most users will use `#[derive(Clone, Copy)]` whenever both are applicable, so most users will still profit from the library optimizations.
Unfortunately, Hyrum's law applies to this PR: there are some popular crates which rely on the precise specialization behaviour of `core` to implement "specialization at home", e.g. [`libAFL`](https://github.com/AFLplusplus/LibAFL/blob/89cff637025c1652c24e8d97a30a2e3d01f187a4/libafl_bolts/src/tuples.rs#L27-L49). I have no remorse for breaking such horrible code, but perhaps we should open other, better ways to satisfy their needs – for example by dropping the `'static` bound on `TypeId::of`...
rustc_builtin_macros: rename bench parameter to avoid collisions with user-defined function names
Resolvesrust-lang/rust#148275 by preventing name collisions in the `#[bench]` macro.
Previously, a user-defined function named "b" could not be benchmarked because
the macro-generated lambda identity collided with the same name. We now generate
the lambda ident as `__bench_<function_name>`, ensuring it is always distinct
from the user’s function.
Because the prefix is applied recursively (e.g. benchmarking `__bench_b`
produces a lambda ident `__bench___bench_b`), there is no possible function
name that can equal its corresponding lambda ident. This guarantees that
the user can safely bench a function of any valid name without risk of
identifier collision.
mgca: Add ConstArg representation for const items
tracking issue: rust-lang/rust#132980fixesrust-lang/rust#131046fixesrust-lang/rust#134641
As part of implementing `min_generic_const_args`, we need to distinguish const items that can be used in the type system, such as in associated const equality projections, from const items containing arbitrary const code, which must be kept out of the type system. Specifically, all "type consts" must be either concrete (no generics) or generic with a trivial expression like `N` or a path to another type const item.
To syntactically distinguish these cases, we require, for now at least, that users annotate all type consts with the `#[type_const]` attribute. Then, we validate that the const's right-hand side is indeed eligible to be a type const and represent it differently in the HIR.
We accomplish this representation using a new `ConstItemRhs` enum in the HIR, and a similar but simpler enum in the AST. When `#[type_const]` is **not** applied to a const (e.g. on stable), we represent const item right-hand sides (rhs's) as HIR bodies, like before. However, when the attribute is applied, we instead lower to a `hir::ConstArg`. This syntactically distinguishes between trivial const args (paths) and arbitrary expressions, which are represented using `AnonConst`s. Then in `generics_of`, we can take advantage of the existing machinery to bar the `AnonConst` rhs's from using parent generics.
Contract variable declarations
This change adds contract variables that can be declared in the `requires` clause and can be referenced both in `requires` and `ensures`, subject to usual borrow checking rules. This allows any setup common to both the `requires` and `ensures` clauses to only be done once.
In particular, one future use case would be for [Fulminate](https://dl.acm.org/doi/10.1145/3704879)-like ownership assertions in contracts, that are essentially side-effects, and executing them twice would alter the semantics of the contract.
As of this change, `requires` can now be an arbitrary sequence of statements, with the final expression being of type `bool`. They are executed in sequence as expected, before checking if the final `bool` expression holds.
This PR depends on rust-lang/rust#144438 (which has now been merged).
Contracts tracking issue: https://github.com/rust-lang/rust/issues/128044
**Other changes introduced**:
- Contract macros now wrap the content in braces to produce blocks, meaning there's no need to wrap the content in `{}` when using multiple statements. The change is backwards compatible, in that wrapping the content in `{}` still works as before. The macros also now treat `requires` and `ensures` uniformally, meaning the `requires` closure is built inside the parser, as opposed to in the macro.
**Known limiatations**:
- Contracts with variable declarations are subject to the regular borrow checking rules, and the way contracts are currently lowered limits the usefulness of contract variable declarations. Consider the below example:
```rust
#[requires(let init_x = *x; true)]
#[ensures(move |_| *x == 2 * init_x)]
fn double_in_place(x: &mut i32) {
*x *= 2;
}
```
We have used the new variable declarations feature to remember the initial value pointed to by `x`, however, moving `x` into the `ensures` does not pass the borrow checker, meaning the above function contract is illegal. Ideally, something like the above should be expressable in contracts.
Contract variables can be declared in the `requires` clause and
can be referenced both in `requires` and `ensures`, subject to usual
borrow checking rules.
This allows any setup common to both the `requires` and `ensures`
clauses to only be done once.
Guard HIR lowered contracts with `contract_checks`
Refactor contract HIR lowering to ensure no contract code is executed when contract-checks are disabled.
The call to `contract_checks` is moved to inside the lowered fn body, and contract closures are built conditionally, ensuring no side-effects present in contracts occur when those are disabled. This partially addresses rust-lang/rust#139548, i.e. the bad behavior no longer happens with contract checks disabled (`-Zcontract-checks=no`).
The change is made in preparation for adding contract variable declarations - variables declared before the `requires` assertion, and accessible from both `requires` and `ensures`, but not in the function body (PR rust-lang/rust#144444). As those declarations may also have side-effects, it's good to guard them with `contract_checks` - the new lowering approach allows for this to be done easily.
Contracts tracking issue: rust-lang/rust#128044
**Known limiatations**:
- It is still possible to early return from the *function* from within a contract, e.g.
```rust
#[ensures({if x > 0 { return 0 }; |_| true})]
fn foo(x: u32) -> i32 {
42
}
```
When `foo` is called with an argument greater than 0, instead of `42`, `0` will be returned.
As this is not a regression, it is not addressed in this PR. However, it may be worth revisiting later down the line, as users may expect a form of early return from *contract specifications*, and so returning from the entire *function* could cause confusion.
- ~Contracts are still not optimised out when disabled. Currently, even when contracts are disabled, the code generated causes existing optimisations to fail, meaning even disabled contracts could impact runtime performance. This issue is blocking rust-lang/rust#136578, and has not been addressed in this PR, i.e. the `mir-opt` and `codegen` tests that fail in rust-lang/rust#136578 still fail with these new HIR lowering changes.~ Contracts should now be optimised out when disabled, however some regressions tests still need to be added to be sure that is indeed the case.
Fix ICE when using contracts on async functions
Fixesrust-lang/rust#145333
contract is not supported for async functions right now, it's not properly lowered and getting HirId will ICE.
This PR adds checking for async function in expanding AST phase, it's better until we want to fully support async for contracts feature.
The compiler complained about uncecessary parenthesis on contract clauses,
which were insterted by the contract macros. This commit changes the
macro to use braces as the delimiter instead, fixing the issue.
bors