Multiple people have asked for them, in
https://github.com/rust-lang/rust/issues/49137.
Given that the unsigned ones already exist,
they are very easy to add and not an additional maintenance burden.
Turns out that the equality check for regions is rather expensive, and
the current early exit check works in such a way, that the comparison is
even done twice. As we only really care about the case of equal scopes,
we can perform a faster, more specialized check and move it up one
level, so we can eventually skip the additional full comparison as well.
Prepare beta 1.33.0
This PR includes the usual changes for a new beta, and suppresses a few lints on libcore: those lints are false positives caused by an internal attribute (`rustc_layout_scalar_valid_range_start`) and only happen on stage0.
r? @Mark-Simulacrum
It's a level of indirection that hurts far more than it helps. The code
is simpler without it. (This commit cuts more than 120 lines of code.)
In particular, this commit removes some unnecessary `Span`s within
`DeclKind` that were always identical to those in the enclosing `Stmt`,
and some unnecessary allocations via `P`.
Always calculate glob map but only for glob uses
Previously calculating glob map was *opt-in*, however it did record node id -> ident use for every use directive. This aims to see if we can unconditionally calculate the glob map and not regress performance.
Main motivation is to get rid of some of the moving pieces and simplify the compilation interface - this would allow us to entirely remove `CrateAnalysis`. Later, we could easily expose a relevant query, similar to the likes of `maybe_unused_trait_import` (so using precomputed data from the resolver, but which could be rewritten to be on-demand).
r? @nikomatsakis
Local perf run showed mostly noise (except `ctfe-stress-*`) but I'd appreciate if we could do a perf run run here and double-check that this won't regress performance.
Benefits:
- It lets us move the `NodeId` field out of every `hir::StmtKind`
variant `NodeId` to a more sensible spot.
- It eliminates sadness in `Stmt::fmt`.
- It makes `hir::Stmt` match `ast::Stmt`.
Use `clear_if_dirty` on std for backend changes, just as we do for
changes to rustc itself, so new codegen is correctly applied to all
later compiler stages.
Fixes#48298.
Implement basic input validation for built-in attributes
Correct top-level shape (`#[attr]` vs `#[attr(...)]` vs `#[attr = ...]`) is enforced for built-in attributes, built-in attributes must also fit into the "meta-item" syntax (aka the "classic attribute syntax").
For some subset of attributes (found by crater run), errors are lowered to deprecation warnings.
NOTE: This PR previously included https://github.com/rust-lang/rust/pull/57367 as well.
In particular, I can use the following in my `config.toml`:
```
[build]
host = ["i686-unknown-linux-gnu", "x86_64-unknown-linux-gnu"]
target = ["i686-unknown-linux-gnu", "x86_64-unknown-linux-gnu"]
```
Before this change, my attempt to run the test suite would fail
because the error output differs depending on what your host and
targets are.
----
To be concrete, here are the actual messages one can observe:
```
% ./build/x86_64-unknown-linux-gnu/stage1/bin/rustc ../src/test/ui/huge-enum.rs -Aunused --target=x86_64-unknown-linux-gnu
error: the type `std::option::Option<[u32; 35184372088831]>` is too big for the current architecture
error: aborting due to previous error
% ./build/x86_64-unknown-linux-gnu/stage1/bin/rustc ../src/test/ui/huge-enum.rs -Aunused --target=i686-unknown-linux-gnu
error: the type `std::option::Option<[u32; 536870911]>` is too big for the current architecture
error: aborting due to previous error
% ./build/i686-unknown-linux-gnu/stage1/bin/rustc ../src/test/ui/huge-enum.rs -Aunused --target=i686-unknown-linux-gnu
error: the type `std::option::Option<[u32; 536870911]>` is too big for the current architecture
error: aborting due to previous error
% ./build/i686-unknown-linux-gnu/stage1/bin/rustc ../src/test/ui/huge-enum.rs -Aunused --target=x86_64-unknown-linux-gnu
error: the type `[u32; 35184372088831]` is too big for the current architecture
error: aborting due to previous error
```
To address these variations, I changed the test to be more aggressive
in its normalization strategy. We cannot (and IMO should not)
guarantee that `Option` will appear in the error output here. So I
normalized both types `Option<[u32; N]>` and `[u32; N]` to just `TYPE`
rustc: Remove platform intrinsics crate
This was originally attempted in #57048 but it was realized that we
could fully remove the crate via the `"unadjusted"` ABI on intrinsics.
This means that all intrinsics in stdsimd are implemented directly
against LLVM rather than using the abstraction layer provided here. That
ends up meaning that this crate is no longer used at all.
This crate developed long ago to implement the SIMD intrinsics, but we
didn't end up using it in the long run. In that case let's remove it!
It's simpler and makes some benchmark run up to 1% faster. It also makes
`hir::ExprKind` more like `ast::ExprKind` (which underwent the
equivalent change in #55777).
